Professional Documents
Culture Documents
and Properties of
Petroleum Fractions
Characterization
and Properties of
Petroleum Fractions
First Edition
M. R. Riazi
Professor of Chemical Engineering
Kuwait University
ASTM
100 Barr Harbor
West Conshohocken, PA 19428-2959
Riazi, M.-R.
Characterization and properties of petroleum fractions / M.-R. Riazi--1 st ed.
p. cm.--(ASTM manual series: MNL50)
ASTM stock number: MNL50
Includes bibliographical references and index.
ISBN 0-8031-3361-8
1. Characterization. 2. Physical property estimation. 3. Petroleum fractions--crude oils.
TP691.R64 2005
666.5---dc22
2004059586
Copyright 9 2005 AMERICAN SOCIETY FOR TESTING AND MATERIALS, West Conshohocken,
PA. All rights reserved. This material may not be reproduced or copied, in whole or in part, in any
printed, mechanical, electronic, film, or other distribution and storage media, without the written
consent of the publisher.
Photocopy Rights
NOTE: This publication does not purport to address all of the safety problems associated with
its use. It is the responsibility of the user of this publication to establish appropriate safety and
health practices and determine the applicability of regulatory limitations prior to use.
Printed in Philadelphia, PA
January 2005
To
Shiva, Touraj, and Nazly
Contents
Foreword xvii
Preface xix
Chapter 1--Introduction 1
Nomenclature 1
1.1 Nature of Petroleum Fluids 1
1.1.1 Hydrocarbons 3
1.1.2 Reservoir Fluids and Crude Oil 5
1.1.3 Petroleum Fractions and Products 7
1.2 Types and Importance of Physical Properties 10
1.3 Importance of Petroleum Fluids Characterization 12
1.4 Organization of the Book 15
1.5 Specific Features of this Manual 15
1.5.1 Introduction of Some Existing Books 15
1.5.2 Special Features of the Book 16
1.6 Applications of the Book 16
1.6.1 Applications in Petroleum Processing
(Downstream) 17
1.6.2 Applications in Petroleum Production
(Upstream) 17
1.6.3 Applications in Academia 17
1.6.4 Other Applications 17
1.7 Definition of Units and the Conversion Factors 17
1.7.1 Importance and Types of Units 17
1.7.2 Fundamental Units and Prefixes 18
1.7.3 Units of Mass 18
1.7.4 Units of Length 18
1.7.5 Units of Time 18
1.7.6 Units of Force 19
1.7.7 Units of Moles 19
1.7.8 Units of Molecular Weight 19
1.7.9 Units of Pressure 19
1.7.10 Units of Temperature 19
1.7.11 Units of Volume, Specific Volume, and
Molar Volume---The Standard Conditions 20
1.7.12 Units of Volumetric and Mass Flow Rates 20
1.7.13 Units of Density and Molar Density 20
1.7.14 Units of Specific Gravity 21
1.7.15 Units of Composition 21
1.7.16 Units of Energy and Specific Energy 22
1.7.17 Units of Specific Energy per Degrees 22
1.7.18 Units of Viscosity and Kinematic Viscosity 23
1.7.19 Units of Thermal Conductivity 23
1.7.20 Units of Diffusion Coefficients 23
1.7.21 Units of Surface Tension 24
1.7.22 Units of Solubility Parameter 24
1.7.23 Units of Gas-to-Oil Ratio 24
vii
viii CONTENTS
1
Introduction
gas), liquids (crude oil), semisolids (bitumen), or solids (wax organic compounds into petroleum, and (3) a sealed reservoir
or asphaltite). Liquid fuels are normally produced from liq- space to store the hydrocarbons produced. The conditions re-
uid hydrocarbons, although conversion of nonliquid hydro- quired for the process of conversion of organic compounds
carbons such as coal, oil shale, and natural gas to liquid fuels into petroleum (as shown through Eq. (1.2) are (1) geologic
is being investigated. In this book, only petroleum hydrocar- time of about 1 million years, (2) m a x i m u m pressure of
bons in the form of gas or liquid, simply called petroleum flu- about 17 MPa (2500 psi), and (3) temperature not exceed-
ids, are considered. Liquid petroleum is also simply called oil. ing 100-120~ (~210-250~ If a leak occurred sometime
Hydrocarbon gases in a reservoir are called a natural gas or in the past, the exploration well will encounter only small
simply a gas. An underground reservoir that contains hydro- amounts of residual hydrocarbons. In some cases bacteria
carbons is called petroleum reservoir and its hydrocarbon con- may have biodegraded the oil, destroying light hydrocarbons.
tents that can be recovered through a producing well is called An example of such a case would be the large heavy oil accu-
reservoir fluid. Reservoir fluids in the reservoirs are usually in mulations in Venezuela. The hydrocarbons generated grad-
contact with water in porous media conditions and because ually migrate from the original beds to more porous rocks,
they are lighter than water, they stay above the water level such as sandstone, and form a petroleum reservoir. A series
under natural conditions. of reservoirs within a c o m m o n rock is called an oil field.
Although petroleum has been known for many centuries, Petroleum is a mixture of hundreds of different identifiable
the first oil-producing well was discovered in 1859 by E.L. hydrocarbons, which are discussed in the next section. Once
Drake in the state of Pennsylvania and that marked the petroleum is accumulated in a reservoir or in various sedi-
birth of modern petroleum technology and refining. The ments, hydrocarbon compounds may be converted from one
main elements of petroleum are carbon (C) and hydrogen form to another with time and varying geological conditions.
(H) and some small quantities of sulfur (S), nitrogen (N), This process is called in-situ alteration, and examples of chem-
and oxygen (O). There are several theories on the formation ical alteration are thermal maturation and microbial degra-
of petroleum. It is generally believed that petroleum is de- dation of the reservoir oil. Examples of physical alteration of
rived from aquatic plants and animals through conversion of petroleum are the preferential loss of low-boiling constituents
organic compounds into hydrocarbons. These animals and by the diffusion or addition of new materials to the oil in
plants under aquatic conditions have converted inorganic place from a source outside the reservoir [1]. The main dif-
compounds dissolved in water (such as carbon dioxide) to ference between various oils from different fields around the
organic compounds through the energy provided by the sun. world is the difference in their composition of hydrocarbon
An example of such reactions is shown below: compounds. Two oils with exactly the same composition have
identical physical properties under the same conditions [2].
(1.1) 6CO2 + 6H20 d- energy --~ 602 + C6H1206
A good review of statistical data on the amount of oil and
in which C6H1206 is an organic compound called carbohy- gas reservoirs, their production, processing, and consump-
drate. In some cases organic compounds exist in an aquatic tion is usually reported yearly by the Oil and Gas Journal
environment. For example, the Nile river in Egypt and the (OGJ). An annual refinery survey by OGJ is usually published
Uruguay river contain considerable amounts of organic ma- in December of each year. OGJ also publishes a forecast and
terials. This might be the reason that most oil reservoirs are review report in January and a midyear forecast report in
located near the sea. The organic compounds formed may be July of each year. In 2000 it was reported that total proven oil
decomposed into hydrocarbons under certain conditions. reserves is estimated at 1016 billion bbl (1.016 x 10 tz bbl),
which for a typical oil is equivalent to approximately 1.39 x
(1.2) (CHEO)n --~ xCO2 d-yCH4
1011 tons. The rate of oil production was about 64.6 million
in which n, x, y, and z are integer numbers and yCHz is the bbl/d (~3.23 billion ton/year) through more than 900 000 pro-
closed formula for the produced hydrocarbon compound. ducing wells and some 750 refineries [3, 4]. These numbers
Another theory suggests that the inorganic compound cal- vary from one source to another. For example, Energy Infor-
cium carbonate (CaCO3) with alkali metal can be converted to mation Administration of US Department of Energy reports
calcium carbide (CaC2), and then calcium carbide with water world oil reserves as of January 1, 2003 as 1213.112 billion
(H20) can be converted to acetylene (C2H2). Finally, acetylene bbl according to OGJ and 1034.673 billion bbl according to
can be converted to petroleum [ 1]. Conversion of organic mat- World Oil (www.eia.doe.gov/emeu/iea). According to the OGJ
ters into petroleum is called maturation. The most important worldwide production reports (Oil and Gas Journal, Dec. 22,
factors in the conversion of organic compounds to petroleum 2003, p. 44), world oil reserves estimates changed from 999.78
hydrocarbons are (1) heat and pressure, (2) radioactive rays, in 1995 to 1265.811 billion bbl on January 1, 2004. For the
such as g a m m a rays, and (3) catalytic reactions. Vanadium- same period world gas reserves estimates changed from 4.98 x
and nickel-type catalysts are the most effective catalysts in 1015 scf to 6.0683 x 1015 scf. In 2003 oil consumption was
the formation of petroleum. For this reason some of these about 75 billion bbl/day, and it is expected that it will in-
metals may be found in small quantities in petroleum fluids. crease to more than 110 million bbl/day by the year 2020.
The role of radioactive materials in the formation of hydro- This means that with existing production rates and reserves,
carbons can be best observed through radioactive bombard- it will take nearly 40 years for the world's oil to end. Oil
ing of fatty acids (RCOOH) that form paraffin hydrocarbons. reserves life (reserves-to-production ratio) in some selected
Occasionally traces of radioactive materials such as uranium countries is given by OGJ (Dec. 22, 2004, p. 45). According
and potassium can also be found in petroleum. In summary, to 2003 production rates, reserves life is 6.1 years in UK,
the following steps are required for the formation of hydrocar- 10.9 years in US, 20 years in Russia, 5.5 years in Canada,
bons: (1) a source of organic material, (2) a process to convert 84 years in Saudi Arabia, 143 years in Kuwait, and 247 years
1. INTRODUCTION 3
in Iraq. As in January l, 2002, the total number of world oil interconversion schemes may occur during the formation,
wells was 830 689, excluding shut or service wells (OGJ, Dec. maturation, and in-situ alteration of petroleum.
22, 2004). Estimates of world oil reserves in 1967 were at Paraffins are also called alkanes and have the general for-
418 billion and in 1987 were at 896 billion bbl, which shows mula of C, Han+a, where n is the number of carbon atoms.
an increase of 114% in this period [5]. Two-thirds of these Paraffins are divided into two groups of normal and isoparaf-
reserves are in the Middle East, although this portion de- fins. Normal paraffins or normal alkanes are simply written
pends on the type of oil considered. Although some people as n-paraffins or n-alkanes and they are open, straight-chain
believe the Middle East has a little more than half of world saturated hydrocarbons. Paraffins are the largest series of hy-
oil reserves, it is believed that many undiscovered oil reser- drocarbons and begin with methane (CH4), which is also rep-
voirs exist offshore under the sea, and with increase in use resented by C1. Three n-alkanes, methane (C1), ethane (C2),
of the other sources of energy, such as natural gas or coal, and n-butane (C4), are shown below:
and through energy conservation, oil production may well
continue to the end of the century. January 2000, the total H H H H H H H
amount of gas reserves was about 5.15 • 1015 scf, and I I I I I I I
H--C--H H--C--C--H H--C--C--C--C--H
its production in 1999 was about 200 x 109 scf/d (5.66 x
109 sm3/d) through some 1500 gas plants [3]. In January I I I I I I I
H H H H H H H
2004, according to OGJ (Dec. 22, 2004, p. 44), world natu-
ral gas reserves stood at 6.068 • 1015 scf (6068.302 trillion
scf). This shows that existing gas reserves may last for some Methane Ethane n-Butane
70 years. Estimated natural gas reserves in 1975 were at
2.5 x 1015 scf (7.08 x 1013 sm3), that is, about 50% of current (CH4) (C2H6) (C4H1~
reserves [6]. In the United States, consumption of oil and gas
The open formula for n-C4 can also be shown as CH3--
in 1998 was about 65% of total energy consumption. Crude
CH2--CH2--CH3 and for simplicity in drawing, usually the
oil demand in the United State in 1998 was about 15 million
CH3 and CH2 groups are not written and only the carbon-
bbl/d, that is, about 23% of total world crude production [3].
Worldwide consumption of natural gas as a clean fuel is on carbon bonds are drawn. For example, for a n-alkane com-
pound of n-heptadecane with the formula of C17H36, the
the rise, and attempts are underway to expand the trans-
structure can also be shown as follows:
fer of natural gas through pipelines as well as its conver-
sion to liquid fuels such as gasoline. The world energy con-
sumption is distributed as 35% through oil, 31% through
coal, and 23% through natural gas. Nearly 11% of total
world energy is produced through nuclear and hydroelectric n-Heptadecane (C17H36)
sources [ 1].
The second group of paraffins is called isoparaffins; these
are branched-type hydrocarbons and begin with isobutane
(methylpropane), which has the same closed formula as n-
1.1.1 Hydrocarbons butane (Call10). Compounds of different structures but the
In early days of chemistry science, chemical compounds were same closed formula are called isomers. Three branched or
divided into two groups: inorganic and organic, depending isoparaffin compounds are shown below:
on their original source. Inorganic compounds were obtained
from minerals, while organic compounds were obtained from CH3 CH3 CH3
living organisms and contained carbon. However, now or- CH3--CH--CH3 CH3--CH~CH2--CH3 CH3--CH--CH2--CH2--CH2--CH2--CH3
ganic compounds can be produced in the laboratory. Those
organic compounds that contain only elements of carbon (C) isobutane isopen~ane
(methylbutane) isooctane(2-methylheptane)
and hydrogen (H) are called hydrocarbons, and they form
the largest group of organic compounds. There might be as (C4HIo) (C5H12) (C8HI8)
many as several thousand different hydrocarbon compounds
in petroleum reservoir fluids. Hydrocarbon compounds have In the case of isooctane, if the methyl group (CH3) is at-
a general closed formula of CxHy, where x and y are integer tached to another carbon, then we have another compound
numbers. The lightest hydrocarbon is methane (CH4), which (i.e., 3-methylheptane). It is also possible to have more than
is the main component in a natural gas. Methane is from a one branch of CH3 group, for example, 2,3-dimethylhexane
group of hydrocarbons called paraffins. Generally, hydrocar- and 2-methylheptane, which are simply shown as following:
bons are divided into four groups: (1) paraffins, (2) olefins,
(3) naphthenes, and (4) aromatics. Paraffins, olefins, and
naphthenes are sometime called aliphatic versus aromatic
compounds. The International Union of Pure and Applied
Chemistry (IUPAC) is a nongovernment organization that
provides standard names, nomenclature, and symbols for dif- 2-Methylheptane (CsHls) 2,3-Dimethylhexane (C8H18)
ferent chemical compounds that are widely used [7]. The
relationship between the various hydrocarbon constituents Numbers refer to carbon numbers where the methyl group
of crude oils is hydrogen addition or hydrogen loss. Such is attached. For example, 1 refers to the first carbon either
4 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
1.0E+O0
0 10 20 30 40 50
Number of CarbonAtoms
FIG. 1.1reNumber of possible alkane isomers.
Cyclopentane Methylcyclopentane Ethylcyclohexane
from the right or from the left. There are 2 isomers for bu- (CsHIo) (C6HI2) (C8H~6)
tane and 3 for pentane, but there are 5 isomers for hexane, 9
for heptane, 18 for octane (C8H18), and 35 for nonane. Sim-
If there is only one alkyl group from n-paraffins (i.e., methyl,
ilarly, dodecane (C12H26) has 355, while octadecane (C18H38)
ethyl, propyl, n-butyl .... ) attached to a cyclopentane hydro-
has 60523 and C40 has 62 x 1012 isomers [1, 8, 9]. The num-
carbon, the series is called n-alkylcyclopentanes, such as the
ber of isomers rapidly increases with the number of carbon
two hydrocarbons shown above where on each junction of the
atoms in a molecule because of the rapidly rising number of
ring there is a CH2 group except on the alkyl group juncture
their possible structural arrangements as shown in Fig. 1.1.
where there is only a CH group. For simplicity in drawing,
For the paraffins in the range of Cs-C12, the number of iso-
these groups are not shown. Similarly there is a homologous
mers is more than 600 although only about 200-400 of them
have been identified in petroleum mixtures [ 10]. Isomers have napthenic series of n-alkylcyclohexanes with only one satu-
different physical properties. The same increase in number rated ring of cyclohexane, such as ethylcyclohexane shown
of isomers with molecular weight applies to other hydro- above. Napthenic hydrocarbons with only one ring are also
carbon series. As an example, the total number of hydrocar- called monocycloparaffins or mononaphthenes. In heavier oils,
bons (from different groups) having 20 carbon atoms is more saturated multirings attached to each other called polycy-
than 300000 [10]! cloparaffins orpolynaphthenes may also be available. Thermo-
Under standard conditions (SC) of 20~ and 1 atm, the dynamic studies show that naphthene rings with five and six
carbon atoms are the most stable naphthenic hydrocarbons.
first four members of the alkane series (methane, ethane,
The content of cycloparaffins in petroleum may vary up to
propane, and butane) are in gaseous form, while from C5Hl1
(pentane) to n-hexadecane (C16H36)they are liquids, and from 60%. Generally, any petroleum mixture that has hydrocarbon
n-heptadecane (C17H38) the compounds exist as waxlike solids compounds with five carbon atoms also contains naphthenic
at this standard temperature and pressure. Paraffins from C1 compounds.
A r o m a t i c s are an important series of hydrocarbons found
to C40 usually appear in crude oil and represent up to 20% of
crude by volume. Since paraffins are fully saturated (no dou- in almost every petroleum mixture from any part of the world.
ble bond), they are stable and remain unchanged over long Aromatics are cyclic but unsaturated hydrocarbons that begin
periods of geological time. with benzene molecule (C6H6) and contain carbon-carbon
Olefms are another series of noncyclic hydrocarbons but double bonds. The name aromatic refers to the fact that such
they are unsaturated and have at least one double bond hydrocarbons commonly have fragrant odors. Four different
between carbon-carbon atoms. Compounds with one dou- aromatic compounds are shown below:
ble bond are called monoolefins or alkenes, such as ethene
\
(also named ethylene: CH2=CH2) and propene or propylene
(CH2=CH--CH3). Besides structural isomerism connected
with the location of double bond, there is another type of iso-
merism called geometric isomerism, which indicates the way
9
atoms are oriented in space. The configurations are differen-
tiated in their names by the prefixes cis- and trans- such as (C6H6) (C7H8) (C8H1o) (C1o~8)
cis- and trans-2-butene. Monoolefins have a general formula
Benzene Toluene O-xylene Naphthalene
of CnH2n. If there are two double bonds, the olefin is called
diolefin (or diene), such as butadiene (CH2=CH--CH=CH2). (Methylbenzene) ( 1,2-Dimethylbenzene)
1. INTRODUCTION 5
p r o d u c e d at SC i n s t a n d a r d cubic feet (scf) to the a m o u n t of properties outside the corresponding limits m e n t i o n e d ear-
liquid oil produced at the SC i n stock t a n k barrel (stb). Other lier. D e t e r m i n a t i o n of a type of reservoir fluid by the above
u n i t s of GOR are discussed i n Section 1.7.23 a n d its calcula- rule of t h u m b based on the GOR, API gravity of stock t a n k
tion is discussed i n Chapter 9. Generally, reservoir fluids are oil, or its color is not possible for all fluids. A more accu-
categorized into four or five types (their characteristics are rate m e t h o d of d e t e r m i n i n g the type of a reservoir fluid is
given in Table 1.1). These five fluids i n the direction of in- based o n the phase behavior calculations, its critical point,
creasing GOR are black oil, volatile oil, gas condensate, wet a n d shape of the phase d i a g r a m which will be discussed i n
gas, and dry gas. Chapters 5 a n d 9. I n general, oils produced from wet gas,
If a gas after surface separator, u n d e r SC, does n o t pro- gas condensate, volatile oil, a n d black oil increase i n spe-
duce any liquid oil, it is called dry gas. A n a t u r a l gas that after cific gravity (decrease in API gravity a n d quality) in the same
p r o d u c t i o n at the surface facilities can produce a little liquid order. Here quality of oil indicates lower carbon, sulfur, nitro-
oil is called wet gas. The word wet does n o t m e a n that the gen, a n d metal contents which correspond to higher heating
gas is wet with water, b u t refers to the h y d r o c a r b o n liquids value. Liquids from black oils are viscous a n d black i n color,
that condense at surface conditions. I n dry gases n o liquid while the liquids from gas condensates or wet gases are clear
h y d r o c a r b o n is formed at the surface conditions. However, a n d colorless. Volatile oils produce fluids b r o w n with some
both dry a n d wet gases are i n the category of n a t u r a l gases. red/green color liquid. Wet gas contains less m e t h a n e t h a n a
Volatile oils have also been called high-shrinkage crude oil a n d dry gas does, b u t a larger fraction of C2-C 6 c o m p o n e n t s . Ob-
near-critical oils, since the reservoir t e m p e r a t u r e a n d pressure viously the m a i n difference between these reservoir fluids is
are very close to the critical p o i n t of such oils, b u t the critical their respective composition. An example of c o m p o s i t i o n of
t e m p e r a t u r e is always greater t h a n the reservoir t e m p e r a t u r e different reservoir fluids is given i n Table 1.2.
[i 1]. Gases a n d gas c o n d e n s a t e fluids have critical tempera- I n Table 1.2, C7+ refers to all h y d r o c a r b o n s having seven
tures less t h a n the reservoir temperature. Black oils c o n t a i n or higher c a r b o n atoms a n d is called heptane-plus fraction,
heavier c o m p o u n d s a n d therefore the API gravity of stock while C6 refers to a group of all h y d r o c a r b o n s with six car-
t a n k oil is generally lower t h a n 40 a n d the GOR is less t h a n b o n atoms (hexanes) that exist in the fluid. MT+ a n d SG7+ are
1000 scf/stb. The specifications given i n Table 1.1 for various the molecular weight a n d specific gravity at 15.5~ (60~ for
reservoir fluids, especially at the b o u n d a r i e s b e t w e e n differ- the C7+ fraction of the mixture, respectively. It should be re-
ent types, are arbitrary a n d vary from one source to a n o t h e r alized that molecular weight a n d specific gravity of the whole
[9, 11]. It is possible to have a reservoir fluid type that has reservoir fluid are less t h a n the corresponding values for the
TABLE 1.2---Composition (mol%) and properties of various reservoir fluids and a crude oil
Component Dry gas~ Wet gasb Gas condensateC Volatileoild Blackoile Crudeoilf
CO2 3.70 0.00 0.18 1.19 0.09 0.00
N2 0.30 0.00 0.13 0.51 2.09 0.00
H2S 0.00 0.00 0.00 0.00 1.89 0.00
C1 96.00 82.28 61.92 45.21 29.18 0.00
C2 0.00 9.52 14.08 7.09 13.60 0.19
C3 0.00 4.64 8.35 4.61 9.20 1.88
iC4 0.00 0.64 0.97 1.69 0.95 0.62
nC4 0.00 0.96 3.41 2.81 4.30 3.92
iC5 0.00 0.35 0.84 1.55 1.38 2.11
nC5 0.00 0.29 1.48 2.01 2.60 4.46
C6 0.00 0.29 1.79 4.42 4.32 8.59
C7+ 0.00 1.01 6.85 28.91 30.40 78.23
Total 100.00 100.00 100.00 100.00 100.00 100.00
GOR (scf/stb) ... 69917 4428 1011 855
M7+ ... 113 143 190 209.8 266
SG7+ (at 15.5~ ... 0.794 0.795 0.8142 0.844 0.895
API7+ 46.7 46.5 42.1 36.1 26.6
"Gas sample from Salt Lake, Utah [12].
bWetgas data from McCaln[11].
CGas condensate sample from Samson County,Texas (M. B. Standing, personal notes, Department of Petroleum
Engineering,NorwegianInstitute of Technology,Trondheim,Norway, 1974).
dVolatile oil sample from Raleigh Field, Smith County,Mississipi(M. B. Standing,personal notes, Departmentof
Petroleum Engineering,NorwegianInstitute of Technology,Trondheim,Norway, 1974).
eBlack oil sample from M. Ghuraiba,M.Sc. Thesis, Kuwait University,Kuwait, 2000.
fA crude oil sampleproduced at stock tank conditions.
1. I N T R O D U C T I O N 7
heptane-plus fraction. For example, for the crude oil sample content of a crude is less than 0.5 wt% it is called a sweet
in Table 1.2, the specific gravity of the whole crude oil is 0.871 oil. It should be realized that these ranges for the gravity and
or API gravity of 31. Details of such calculations are discussed sulfur content are relative and may vary from one source to
in Chapter 4. These compositions have been determined from another. For example, Favennec [15] classifies heavy crude as
recombination of the compositions of corresponding sepa- those with API less than 22 and light crude having API above
rator gas and stock tank liquid, which have been measured 33. Further classification of crude oils will be discussed in
through analytical tools (i.e., gas chromatography, mass spec- Chapter 4.
trometry, etc.). Composition of reservoir fluids varies with the
reservoir pressure and reservoir depth. Generally in a produc- 1.1.3 P e t r o l e u m Fractions and Products
ing oil field, the sulfur and amount of heavy compounds in-
crease versus production time [10]. However, it is important A crude oil produced after necessary field processing and
to note that within an oil field, the concentration of light hy- surface operations is transferred to a refinery where it is
drocarbons and the API gravity of the reservoir fluid increase processed and converted into various useful products. The
with the reservoir depth, while its sulfur and C7+ contents de- refining process has evolved from simple batch distillation
crease with the depth [ 1]. The lumped C7+ fraction in fact is in the late nineteenth century to today's complex processes
a mixture of a very large number of hydrocarbons, up to C40 through modern refineries. Refining processes can be gener-
or higher. As an example the number of pure hydrocarbons ally divided into three major types: (1) separation, (2) con-
from C5 to C9 detected by chromatography tools in a crude oil version, and (3) finishing. Separation is a physical process
from North Sea reservoir fluids was 70 compounds. Detailed where compounds are separated by different techniques. The
composition of various reservoir fluids from the North Sea most important separation process is distillation that occurs
fields is provided by Pedersen et al. [13]. As shown in Chapter in a distillation column; compounds are separated based on
9, using the knowledge of the composition of a reservoir fluid, the difference in their boiling points. Other major physical
one can determine a pressure-temperature (PT) diagram of separation processes are absorption, stripping, and extrac-
the fluid. And on the basis of the temperature and pressure tion. In a gas plant of a refinery that produces light gases,
of the reservoir, the exact type of the reservoir fluid can be the heavy hydrocarbons (Cs and heavier) in the gas mixture
determined from the PT diagram. are separated through their absorption by a liquid oil sol-
Reservoir fluids from a producing well are conducted to vent. The solvent is then regenerated in a stripping unit. The
two- or three-stage separators which reduce the pressure and conversion process consists of chemical changes that occur
temperature of the stream to atmospheric pressure and tem- with hydrocarbons in reactors. The purpose of such reactions
perature. The liquid leaving the last stage is called stock tank is to convert hydrocarbon compounds from one type to an-
oil (STO) and the gas released in various stages is called as- other. The most important reaction in m o d e m refineries is
sociated gas. The liquid oil after necessary field processing is the cracking in which heavy hydrocarbons are converted to
called crude oil. The main factor in operation and design of an lighter and more valuable hydrocarbons. Catalytic cracking
oil-gas separator is to find the optimum operating conditions and thermal cracking are commonly used for this purpose.
of temperature and pressure so that the amount of produced Other types of reactions such as isomerization or alkylation
liquid (oil) is maximized. Such conditions can be determined are used to produce high octane number gasoline. Finishing is
through phase behavior calculations, which are discussed in the purification of various product streams by processes such
detail in Chapter 9. Reservoir fluids from producing wells are as desulfurization or acid treatment of petroleum fractions to
mixed with free water. The water is separated through gravi- remove impurities from the product or to stabilize it.
tational separators based on the difference between densities After the desalting process in a refinery, the crude oil en-
of water and oil. Remaining water from the crude can be re- ters the atmospheric distillation column, where compounds
moved through dehydration processes. Another surface oper- are separated according to their boiling points. Hydrocarbons
ation is the desalting process that is necessary to remove the in a crude have boiling points ranging from -160~ (boil-
salt content of crude oils. Separation of oil, gas, and water ing point of methane) to more than 600~ (ll00~ which
from each other and removal of water and salt from oil and is the boiling point of heavy compounds in the crude oil.
any other process that occurs at the surface are called surface However, the carbon-carbon bond in hydrocarbons breaks
production operations [14]. down at temperatures around 350~ (660~ This process is
The crude oil produced from the atmospheric separator has called cracking and it is undesirable during the distillation
a composition different from the reservoir fluid from a pro- process since it changes the structure of hydrocarbons. For
ducing well. The light gases are separated and usually crude this reason, compounds having boiling points above 350~
oils have almost no methane and a small C2-C3 content while (660+~ called residuum are removed from the bottom of
the C7+ content is higher than the original reservoir fluid. As atmospheric distillation column and sent to a vacuum dis-
an example, the composition of a crude oil produced through tillation column. The pressure in a vacuum distillation col-
a three-stage separator from a reservoir fluid is also given in umn is about 50-100 m m Hg, where hydrocarbons are boiled
Table 1.2. Actually this crude is produced from a black oil at much lower temperatures. Since distillation cannot com-
reservoir fluid (composition given in Table 1.2). Two impor- pletely separate the compounds, there is no pure hydrocarbon
tant characterisitcs of a crude that determine its quality are as a product of a distillation column. A group of hydrocarbons
the API gravity (specific gravity) and the sulfur content. Gen- can be separated through distillation according to the boiling
erally, a crude with the API gravity of less than 20 (SG > 0.934) point of the lightest and heaviest compounds in the mixtures,
is called heavy crude and with API gravity of greater than 40 The lightest product of an atmospheric column is a mixture of
(SG < 0.825) is called light crude [1, 9]. Similarly, if the sulfur methane and ethane (but mainly ethane) that has the boiling
8 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
range of -180 to -80~ ( - 2 6 0 to -40~ which corresponds molecular weight multiring aromatics. The vacuum residuum
to the boiling point of methane and ethane. This mixture, may be mixed with lighter products to produce a more valu-
which is in the form of gas and is known as fuel gas, is actu- able blend.
ally a petroleum fraction. In fact, during distillation a crude Distillation of a crude oil can also be performed in the lab-
is converted into a series of petroleum fractions where each oratory to divide the mixture into many narrow boiling point
one is a mixture of a limited number of hydrocarbons with a range fractions with a boiling range of about 10~ Such nar-
specific range of boiling point. Fractions with a wider range row range fractions are sometimes referred to as petroleum
of boiling points contain greater numbers of hydrocarbons. cuts. When boiling points of all the cuts in a crude are known,
All fractions from a distillation column have a known boiling then the boiling point distribution (distillation curve) of the
range, except the residuum for which the upper boiling point
is usually not known. The boiling point of the heaviest com-
ponent in a crude oil is not really known, but it is quite high. 60
The problem of the nature and properties of the heaviest com-
pounds in crude oils and petroleum residuum is still under
investigation by researchers [i 6, 17]. Theoretically, it can be
assumed that the boiling point of the heaviest component in a
Residuum
crude oil is infinity. Atmospheric residue has compounds with 50
.655
carbon number greater than 25, while vacuum residue has
compounds with carbon number greater than 50 (M > 800).
Some of the petroleum fractions produced from distillation
Vacuum
columns with their boiling point ranges and applications are 40 Gas Oil
given in Table 1.3. The boiling point and equivalent carbon
number ranges given in this table are approximate and they
may vary according to the desired specific product. For ex- 9
ample, the light gases fraction is mainly a mixture of ethane, E t-
propane, and butane; however, some heavier compounds zt- 30 fl_
(C5+) may exist in this fraction. The fraction is further frac- o Heavy t-
Gas Oil
tionated to obtain ethane (a fuel gas) and propane and butane A t m o s p h e r i c Distillation 46.1% o
(petroleum gases). The petroleum gases are liquefied to get
liquefied petroleum gas (LPG) used for home cooking pur- 345
Light
poses. In addition the isobutane may be separated for the gas 20
mixture to be used for improving vapor pressure characteris- ,205
Kerosene Gas Oil
tics (volatility) of gasoline in cold weathers. These fractions
may go through further processes to produce desired prod-
ucts. For example, gas oil may go through a cracking process 10-
to obtain more gasoline. Since distillation is not a perfect sep-
aration process, the initial and final boiling points for each V a c u u m Distillation 53.9o'/o
fraction are not exact and especially the end points are ap- ,q
proximate values. Fractions may be classified as narrow or t 90
wide depending on their boiling point range. As an example, 0 i i
is given in Table 1.4 and is graphically shown in Fig. 1.3. Volume Percent
The weight and volume percentages for the products are
near each other. More than 50% of the crude is processed LightGases ~ LightGaserine
in vacuum distillation unit. The vacuum residuum is mainly
resin and asphaltenes-type compounds composed of high FIG. 1.3--Products and composition of Alaska crude oil.
1. INTRODUCTION 9
whole crude can be obtained. Such distillation data and their will be discussed in Chapter 3. The U.S. gasoline demand
uses will be discussed in Chapters 3 and 4. In a petroleum in 1964 was 4,4 million bbl/d and has increased from 7.2 to
cut, hydrocarbons of various types are lumped together in 8.0 million bbl/d in a period of 7 years from 1991 to 1998
four groups of paraffins (P), olefins (O), naphthenes (N), and [6, 20]. In 1990, gasoline was about a third of refinery prod-
aromatics (A). For olefin-free petroleum cuts the composi- ucts in the United States.
tion is represented by the PNA content. If the composition 3. Kerosene and jet fuel are mainly used for lighting and jet
of a hydrocarbon mixture is known the mixture is called a engines, respectively. The main characteristics are sulfur
defined mixture, while a petroleum fraction that has an un- content, cold resistance (for jet fuel), density, and ignition
known composition is called an undefined fraction. quality,
As mentioned earlier, the petroleum fractions presented 4. Diesel and heating oil are used for motor fuel and domestic
in Table 1.3 are not the final products of a refinery. They purposes. The main characteristics are ignition (for diesel
go through further physicochemical and finishing processes oil), volatility, viscosity, cold resistance, density, sulfur con-
to get the characteristics set by the market and government tent (corrosion effects), and flash point (safety factor).
regulations. After these processes, the petroleum fractions 5. Residual fuel oil is used for industrial fuel, for thermal pro-
presented in Table 1.3 are converted to petroleum products. duction of electricity, and as motor fuel (low speed diesel
The terms petroleum fraction, petroleum cut, and petroleum engines). Its main characteristics are viscosity (good at-
product are usually used incorrectly, while one should re- omization for burners), sulfur content (corrosion), stabil-
alize that petroleum fractions are products of distillation ity (no decantation separation), cold resistance, and flash
columns in a refinery before being converted to final prod- point for safety.
ucts. Petroleum cuts may have very narrow boiling range
which may be produced in a laboratory during distillation The major nonfuel petroleum products are [18] as follows:
of a crude. In general the petroleum products can be divided
into two groups: (1) fuel products and (2) nonfuel products. i. Solvents are light petroleum cuts in the C4-C14 range and
The major fuel petroleum products are as follows: have numerous applications in industry and agriculture.
As an example of solvents, white spirits which have boiling
1. Liquefied petroleum gases (LPG) that are mainly used for points between 135 and 205~ are used as paint thinners.
domestic heating and cooking (50%), industrial fuel (clean The main characteristics of solvents are volatility, purity,
fuel requirement) (15%), steam cracking feed stock (25%), odor, and toxicity. Benzene, toluene, and xylenes are used
and as a motor fuel for spark ignition engines (10%). The as solvents for glues and adhesives and as a chemical for
world production in 1995 was 160 million ton per year petrochemical industries.
(---5 million bbl/d) [20]. LPG is basically a mixture of 2. Naphthas constitute a special category of petroleum sol-
propane and butane. vents whose boiling points correspond to the class of white
2. Gasoline is perhaps one of the most important products of spirits. They can be classified beside solvents since they are
a refinery. It contains hydrocarbons from C4 to Cll (molec- mainly used as raw materials for petrochemicals and as
ular weight of about 100-110). It is used as a fuel for cars. the feeds to steam crackers. Naphthas are thus industrial
Its main characteristics are antiknock (octane number), intermediates and not consumer products. Consequently,
volatility (distillation data and vapor pressure), stability, naphthas are not subject to government specifications but
and density. The main evolution in gasoline production has only to commercial specifications.
been the use of unleaded gasoline in the world and the use 3. Lubricants are composed of a main base stock and addi-
of reformulated gasoline (RFG) in the United States. The tives to give proper characteristics. One of the most im-
RFG has less butane, less aromatics, and more oxygenates. portant characteristics of lubricants is their viscosity and
The sulfur content of gasoline should not exceed 0.03% by viscosity index (change of viscosity with temperature). Usu-
weight. Further properties and characteristics of gasoline ally aromatics are eliminated from lubricants to improve
10 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
their viscosity index. Lubricants have structure similar In general, more than 2000 petroleum products within some
to isoparaffinic compounds. Additives used for lubricants 20 categories are produced in refineries in the United States
are viscosity index additives such as polyacrylates and [ 1, 19]. Blending techniques are used to produce some of these
olefin polymers, antiwear additives (i.e., fatty esters), an- products or to improve their quality. The product specifica-
tioxidants (i.e., alkylated aromatic amines), corrosion in- tions must satisfy customers' requirements for good perfor-
hibitors (i.e., fatty acids), and antifoaming agents (i.e., poly- mance and government regulations for safety and environ-
dimethylsiloxanes). Lubricating greases are another class ment protection. To be able to plan refinery operations, the
of lubricants that are semisolid. The properties of lubri- availability of a set of product quality prediction methods is
cants that should be known are viscosity index, aniline therefore very important.
point (indication of aromatic content), volatility, and car- There are a number of international organizations that are
bon residue. known as standard organizations that recommend specific
4. Petroleum waxes are of two types: the paraffin waxes in characteristics or standard measuring techniques for various
petroleum distillates and the microcrystalline waxes in pe- petroleum products through their regular publications. Some
troleum residua. In some countries such as France, paraf- of these organizations in different countries that are known
fin waxes are simply called paraffins. Paraffin waxes are with their abbreviations are as follows:
high melting point materials used to improve the oil's pour
point and are produced during dewaxing of vacuum dis- 1. ASTM (American Society for Testing and Materials) in the
tillates. Paraffin waxes are mainly straight chain alkanes United States
2. ISO (International Organization for Standardization),
(C18-C36) with a very small proportion of isoalkanes and
cycloalkanes. Their freezing point is between 30 and 70~ which is at the international level
and the average molecular weight is around 350. When 3. IP (Institute of Petroleum) in the United Kingdom
present, aromatics appear only in trace quantities. Waxes 4. API (American Petroleum Institute) in the United States
from petroleum residua (microcrystalline form) are less 5. AFNOR (Association Francaise de Normalisation), an offi-
defined aliphatic mixtures of n-alkanes, isoalkanes, and cy- cial standard organization in France
cloalkanes in various proportions. Their average molecular 6. Deutsche Institut fur Norrnung (DIN) in Germany
weights are between 600 and 800, carbon number range is 7. Japan Institute of Standards (J-IS) in Japan
a l k a n e s C30-C60 , and the freezing point range is 60-90~ ASTM is composed of several committees in which the D-02
[ 13]. Paraffin waxes (when completely dearomatized) have committee is responsible for petroleum products and lubri-
applications in the food industry and food packaging. They cants, and for this reason its test methods for petroleum ma-
are also used in the production of candles, polishes, cos- terials are designated by the prefix D. For example, the test
metics, and coatings [ 18]. Waxes at ordinary temperature of method ASTM D 2267 provides a standard procedure to de-
25~ are in solid states although they contain some hydro- termine the benzene content of gasoline [22]. In France this
carbons in liquid form. When melted they have relatively test method is designated by EN 238, which are documented
low viscosity. in AFNOR information document M 15-023. Most standard
5. Asphalt is another major petroleum product that is pro- test methods in different countries are very similar in prac-
duced from vacuum distillation residues. Asphalts contain tice and follow ASTM methods but they are designated by
nonvolatile high molecular weight polar aromatic com- different codes. For example the international standard ISO
pounds, such as asphaltenes (molecular weights of several 6743/0, accepted as the French standard NF T 60-162, treats
thousands) and cannot be distilled even under very high all the petroleum lubricants, industrial oils, and related prod-
vacuum conditions. In some countries asphalt is called ucts. The abbreviation NF is used for the French standard,
bitumen, although some suggest these two are different while EN is used for European standard methods [ 18].
petroleum products. Liquid asphaltic materials are in- Government regulations to protect the environment or to
tended for easy applications to roads. Asphalt and bitu- save energy, in m a n y cases, rely on the recommendations
men are from a category of products called hydrocarbon of official standard organizations. For example, in France,
binders. Major properties to determine the quality of as- AFNOR gives specifications and requirements for various
phalt are flash point (for safety), composition (wax con- petroleum products. For diesel fuels it recommends (after
tent), viscosity and softening point, weathering, density or 1996) that the sulfur content should not exceed 0.05 wt% and
specific gravity, and stability or chemical resistance. the flash point should not be less than 55~ [18].
6. There are some other products such as white oils (used in
pharmaceuticals or in the food industry), aromatic extracts
(used in the paint industry or the manufacture of plastics), 1.2 T Y P E S A N D IMPORTANCE
and coke (as a fuel or to produce carbon elecrodes for alu- OF PHYSICAL P R O P E R T I E S
m i n u m refining). Petroleum cokes generally have boiling
points above 1100+~ (~2000+~ molecular weight of On the basis of the production and refining processes de-
above 2500+, and carbon number of above 200+. Aromatic scribed above it may be said that the petroleum industry
extracts are black materials, composed essentially of con- is involved with m a n y types of equipment for production,
densed polynuclear aromatics and of heterocyclic nitrogen transportation, and storage of intermediate or final petroleum
and/or sulfur compounds. Because of this highly aromatic products. Some of the most important units are listed below.
structure, the extracts have good solvent power.
i. Gravity decanter (to separate oil and water)
Further information on technology, properties, and test- 2. Separators to separate oil and gas
ing methods of fuels and lubricants is given in Ref. [21]. 3. Pumps, compressors, pipes, and valves
1. I N T R O D U C T I O N 11
3. Heat capacity (Cp) of a fluid is needed in design and oper- such as density, boiling point, molecular weight, and refrac-
ation of heat transfer units such as heat exchangers. tive index are called physical properties. Properties such as en-
4. Enthalpy (H) of a fluid is needed in energy balance cal- thalpy, heat capacity, heat of vaporization, equilibrium ratios,
culations, heat requirements needed in design and oper- and fugacity are called thermodynamic properties. Viscosity,
ation of distillation, absorption, stripping columns, and thermal conductivity, diffusion coefficient, and surface ten-
reactors. sion are in the category of physical properties but they are also
5. Heat of vaporization (AHvap) is needed in calculation of called transport properties. In general all the thermodynamic
heat requirements in design and operation of reboilers or and physical properties are called thermophysical properties.
condensers. But they are commonly referred to as physical properties or
6. Heats of formation (hHf), combustion (AHc), and reaction simply properties, which is used in the title of this book.
(AHr) are used in calculation of heating values of fuels A property of a system depends on the thermodynamic state
and the heat required/generated in reactors and furnaces of the system that is determined by its temperature, pressure,
in refineries. Such information is essential in design and and composition. A process to experimentally determine var-
operations of burners, furnaces, and chemical reactors. ious properties for all the industrially important materials,
These properties together with the Gibbs free energy are especially complex mixtures such as crude oils or petroleum
used in calculation of equilibrium constants in chemical products, would be prohibitive in both cost and time, indeed
reactions to determine the optimum operating conditions it could probably never be completed. For these reasons ac-
in reactors for best conversion of feed stocks into the prod- curate methods for the estimation of these properties are be-
ucts. coming increasingly important. In some references the term
7. Viscosity (t*) is another useful property in petroleum pro- property prediction is used instead of property estimation;
duction, refining, and transportation. It is used in reser- however, in this book as generally adopted by most scientists
voir simulators to estimate the rate of oil or gas flow both terms are used for the same purpose.
and their production. It is needed in calculation of power
required in mixers or to transfer a fluid, the amount of
pressure drop in a pipe or column, flow measurement de- 1.3 I M P O R T A N C E OF P E T R O L E U M F L U I D S
vices, and design and operation of oil/water separators CHARACTERIZATION
[37, 38].
8. Thermal conductivity (k) is needed for design and opera- In the previous section, various basic characteristic para-
tion of heat transfer units such as condensers, heat ex- meters for petroleum fractions and crude oils were intro-
changers, as well as chemical reactors [39]. duced. These properties are important in design and oper-
9. Diffusivity or diffusion coefficient (D) is used in calcula- ation of almost every piece of equipment in the petroleum in-
tion of mass transfer rates and it is a useful property in dustry. Thermodynamic and physical properties of fluids are
design and operation of reactors in refineries where feed generally calculated through standard methods such as cor-
and products diffuse in catalyst pores. In petroleum pro- responding state correlations or equations of state and other
duction, a gas injection technique is used in improved oil pressure-volume-temperature (PVT) relations. These corre-
recovery where a gas diffuses into oil under reservoir con- lations and methods have a generally acceptable degree of ac-
ditions; therefore, diffusion coefficient is also required in curacy provided accurate input parameters are used. When
reservoir simulation and modeling [37, 40-42]. using cubic equation of state to estimate a thermodynamic
10. Surface tension (a) or interfacial tension (IFT) is used property such as absolute density for a fluid at a known tem-
mainly by the reservoir engineers in calculation of cap- perature and pressure, the critical temperature (Tc), critical
illary pressure and rate of oil production and is needed pressure (Pc), acentric factor (~0), and molecular weight (M)
in reservoir simulators [37]. In refineries, IFT is a use- of the system are required. For most pure compounds and hy-
ful parameter to determine foaming characteristics of oils drocarbons these properties are known and reported in var-
and the possibility of having such problems in distillation, ious handbooks [36, 47-50]. If the system is a mixture such
absorption, or stripping columns [43]. It is also needed as a crude oil or a petroleum fraction then the pseudocritical
in calculation of the rate of oil dispersion on seawater properties are needed for the calculation of physical proper-
surface polluted by an oil spill [44]. ties. The pseudocritical properties cannot be measured but
11. Equilibrium ratios (Ki) and fugacity coefficients (~Pi) are have to be calculated through the composition of the mix-
the most important thermodynamic properties in all ture. Laboratory reports usually contain certain measured
phase behavior calculations. These calculations include properties such as distillation curve (i.e., ASTM D 2887) and
vapor-liquid equilibria, bubble and dew point pressure, the API gravity or specific gravity of the fraction. However,
pressure-temperature phase diagram, and GOR. Such cal- in some cases viscosity at a certain temperature, the per-
culations are important in design and operation of distilla- cent of paraffin, olefin, naphthene, and aromatic hydrocar-
tion, absorption and stripping units, gas-processing units, bon groups, and sulfur content of the fraction are measured
gas-oil separators at production fields, and to determine and reported. Petroleum fractions are mixtures of many com-
the type of a reservoir fluid [45, 46]. pounds in which the specific gravity can be directly measured
for the mixture, but the average boiling point cannot be mea-
Generally, the first set of properties introduced above sured. Calculation of average boiling point from distillation
(temperature-independent) are the basic parameters that are data, conversion of various distillation curves from one type
used to estimate physical and thermodynamic properties to another, estimation of molecular weight, and the PNA com-
given in the second set (temperature-dependent). Properties position of fractions are the initial steps in characterization of
1. INTRODUCTION 13
properties of petroleum fractions can save a large portion of in Chapter 5 will be used in the prediction of physical
such huge additional investment and operating costs. properties discussed in the follow-up chapters. Fundamental
thermodynamic relations for calculation of thermodynamic
properties are presented in Chapter 6. The last three chapters
1.4 O R G A N I Z A T I O N O F T H E B O O K of the book show applications of methods presented in Chap-
ters 2-6 for calculation of various physical, thermodynamic,
As the title of the book portrays and was discussed in Sec- and transport properties. Methods of calculation and esti-
tions 1.2 and 1.3, the book presents methods of characteriza- mation of density and vapor pressure are given in Chapter 7.
tion and estimation of thermophysical properties of hydrocar- Thermal properties such as heat capacity, enthalpy, heat
bons, defined mixtures, undefined petroleum fractions, crude of vaporization, heats of combustion and reaction, and
oils, and reservoir fluids. The entire book is written in nine the heating value of fuels are also discussed in Chapter 7.
chapters in a way such that in general every chapter requires Predictive methods for transport properties namely viscosity,
materials presented in previous chapters. In addition there is thermal conductivity, diffusixdty, and surface tension are
an appendix and an index. Chapter 1 gives a general intro- given in Chapter 8 [30,31,42,43,69,70]. Finally, phase
duction to the subject from basic definition of various terms, equilibrium calculations, estimation of equilibrium ratios,
the nature of petroleum, its formation and composition, types GOR, calculation of pressure-temperature (PT) diagrams,
of petroleum mixtures, and the importance of characteriza- solid formations, the conditions at which asphaltene, wax,
tion and property prediction to specific features of the book and hydrate are formed, as well as their preventive methods
and its application in the petroleum industry and academia. are discussed in Chapter 9.
Because of the importance of units in property calculations, The book is written according to the standards set by ASTM
the last section of Chapter 1 deals with unit conversion fac- for its publication. Every chapter begins with a general intro-
tors especially between SI and English units for the parame- duction to the chapter. Since in the following chapters for
ters used in the book. Chapter 2 is devoted to properties and most properties several predictive methods are presented, a
characterization of pure hydrocarbons from C1 to C22 from section on conclusion and recommendations is added at the
different hydrocarbon groups, especially from homologous end of the chapter. Practical problems as examples are pre-
groups commonly found in petroleum fluids. Properties of sented and solved for each property discussed in each chap-
some nonhydrocarbons found with petroleum fluids such as ter. Finally, the chapter ends by a set of exercise problems
H20, H2S, CO2, and Nz are also given. Basic parameters are followed by a citation section for the references used in the
defined at the beginning of the chapter, followed by charac- chapter.
terization of pure hydrocarbons. Predictive methods for vari- The Appendix gives a summary of definitions of terms and
ous properties of pure hydrocarbons are presented and com- properties used in this manual according to the ASTM dictio-
pared with each other. A discussion is given on the state-of- nary as well as the Greek letters used in this manual. Finally
the-arts development of predictive methods. The procedures the book ends with an index to provide a quick guide to find
presented in this chapter are essential for characterization of specific subjects.
petroleum fractions and crude oils discussed in Chapters 3
and 4.
Chapter 3 discusses various characterization methods for 1.5 S P E C I F I C F E A T U R E S O F
petroleum fractions and petroleum products. Characteriza- THIS MANUAL
tion parameters are introduced and analytical instruments in
laboratory are discussed. In this chapter one can use min- In this part several existing books in the area of character-
i m u m laboratory data to characterize petroleum fractions ization and physical properties of petroleum fractions are
and to determine the quality of petroleum products. Esti- introduced and their differences with the current book are
mation of some basic properties such as molecular weight, discussed. Then some special features of this book are pre-
molecular-type composition, sulfur content, flash, pour point sented.
and freezing points, critical constants, and acentric factor for
petroleum fractions are presented in this chapter. A theoret-
1.5.1 Introduction of S o m e Existing Books
ical discussion on development of characterization methods
and generation of predictive correlations from experimental There are several books available that deal with physical prop-
data is also presented. Methods of Chapter 3 are extended to erties of petroleum fractions and hydrocarbon systems. The
Chapter 4 for the characterization of various reservoir fluids most comprehensive and widely used book is the API Tech-
and crude oils. Chapters 2-4 are perhaps the most impor- nical Data Book--Petroleum Refining [47]. It is a book with
tant chapters in this book, as the methods presented in these 15 chapters in three volumes, and the first edition appeared
chapters influence the entire field of physical properties in the in mid 1960s. Every 5 years since, some chapters of the book
remaining chapters. have been revised and updated. The project has been con-
In Chapter 5, PVT relations, equations of state, and ducted at the Pennsylvania State University and the sixth
corresponding state correlations are presented [31,63-65]. edition was published in 1997. It contains a data bank on
The use of the velocity of light and sound in developing properties of pure hydrocarbons, chapters on characteriza-
equations of state is also presented [31, 66-68]. Equations of tion of petroleum fractions, thermodynamic and transport
state and corresponding state correlations are powerful tools properties of liquid and gaseous hydrocarbons, their mix-
in the estimation of volumetric, physical, transport, and tures, and undefined petroleum fractions. For each property,
thermodynamic properties [64, 65, 69]. Procedures outlined one predictive method that has been approved and selected
16 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
by the API-TDB committees as the best available method is m a n y methods available in a process simulator. However, the
presented. This book will be referred as API-TDB throughout book has been written in a language that is understandable
this book. to undergraduate and graduate students in all areas of engi-
Another important book in this area is The Properties of neering and science. It contains practical solved problems as
Gases and Liquids that was originally written by Reid and well as exercise problems so that the book would be suitable
Sherwood in 1950s and it has been revised and updated nearly as a text for educational purposes.
every decade. The fifth and latest edition was published in Special features of this book are Chapters 2, 3, and 4 that
November 2000 [36] by three authors different from the orig- deal with the characterization of hydrocarbons, petroleum
inal two authors. The book has been an excellent reference for fractions, and crude oils and their impact on the entire field
students and practical engineers in the industry over the past of property prediction methods. It discusses both light as
five decades. It discusses various methods for prediction of well as heavy fractions and presents methods of prediction
properties of pure hydrocarbons as well as nonhydrocarbons of the important characteristics of petroleum products from
and their defined mixtures. However, it does not treat un- m i n i m u m laboratory data and easily measurable parame-
defined petroleum fractions, crude oils, and reservoir fluids. ters. It presents several characterization methods developed
Most of the methods for properties of pure compounds re- in recent years and not documented in existing references.
quire the chemical structure of compounds (i.e., group con- The book also presents various predictive methods, including
tribution techniques). The book compares various methods the most accurate and widely used method for each property
and gives its recommendations for each method. and discusses points of strength, weaknesses, and limitations.
There are several other books in the area of properties of Recommended methods are based on the generality, simplic-
oils that document empirically developed predictive methods, ity, accuracy, and availability of input parameters. This is
among them is the book Properties of Oils and Natural Gases, another special feature of the book. In Chapters 5 and 6 it
by Pedersen et al. [ 13]. The book mainly treats reservoir fluids, discusses equations of state based on the velocity of sound
especially gas condensates from North Sea, and it is mainly and light and how these two measurable properties can be
a useful reference for reservoir engineers. Books by McCain used to predict thermodynamic and volumetric properties of
[11], Ahmed [71], Whitson [45], and Danesh [72] are all writ- fluids, especially heavy compounds and their mixtures [31,63,
ten by reservoir engineers and contain information mainly for 66-68]. Significant attention is given throughout the book on
phase behavior calculations needed in petroleum production how to estimate properties of heavy hydrocarbons, petroleum
and reservoir simulators. However, they contain some useful fractions, crude oils, and reservoir fluids. Most of the methods
information on methods of prediction of some physical prop- developed by Riazi and coworkers [23, 24, 26-33, 51-56, 63,
erties of petroleum fractions. Another good reference book 65-70], which have been in use by the petroleum industry
was written by Tsonopoulos et al. [73] on thermodynamic [47, 75-82], are documented in this book. In addition, a new
and transport properties of coal liquids in the mid 1980s. experimental technique to measure diffusion coefficients in
Although there are m a n y similarities between coal liquids reservoir fluids under reservoir conditions is presented in
and petroleum fractions, the book does not consider crude Chapter 8 [42]. In Chapter 9 some new methods for determi-
oils and reservoir fluids. But it provides some useful correla- nation of onset of solid formation are introduced. Reported
tions for properties of coal liquids. The book by Wauquier [ 18] experimental data on characteristics and properties of var-
on petroleum refining has several useful chapters on charac- ious oils from different parts of the world are included in
terization and physical properties of petroleum fractions and various chapters for direct evaluations and testing of meth-
finished products. It also provides the test methods accord- ods. Although both gases and liquids are treated in the book,
ing to European standards. Some organizations' Web sites emphasis is on the liquid fractions. Generally, the methods
also provide information on fluid physical properties. A good of estimation of properties of gases are more accurate than
example of such online information is provided by National those for liquid systems. Most of the methods presented in the
Institute of Standards (http://webbook.nist.gov) which gives book are supported by some scientific basis and they are not
molecular weight, names, formulas, structure, and some data simply empirical correlations derived from a certain group of
on various compounds [74]. data. This widens the application of the methods presented
in the book to different types of oils. However, all basic pa-
rameters and necessary engineering concepts are defined in
1.5.2 Special Features of the Book
a way that is understandable for those nonengineer scientists
This book has objectives and aims that are different from who are working in the petroleum or related industry. Nearly
the books mentioned in Section 1.5.1. The main objective all methods are expressed through mathematical relations so
of this book has been to provide a quick reference in the they are convenient for computer applications; however, most
area of petroleum characterization and properties of various of them are simple such that the properties can be calculated
petroleum fluids for the people who work in the petroleum by hand calculators for a quick estimate whenever applica-
industry and research centers, especially in petroleum pro- ble special methods are given for coal liquid fractions. This is
cessing (downstream), petroleum production (upstream), and another unique feature of this book.
related industries. One special characteristic of the book is its
discussion on development of various methods which would
help the users of process/reservoir simulators to become fa- 1.6 A P P L I C A T I O N S OF T H E B O O K
miliar with the nature of characterization and property esti-
mation methods for petroleum fractions. This would in turn The information that is presented in the book may be applied
help them to choose the proper predictive method among the and used in all areas of the petroleum industries: production,
1. INTRODUCTION 17
processing, and transportation. It can also be used as a organizations that set the standards for such materials. As
textbook for educational purposes. Some of the applica- an example, the amount of sulfur or aromatic contents of a
tions of the materials covered in the book were discussed in fuel can be estimated through minimum laboratory data to
Sections 1.2 and 1.3. The applications and uses of the book check if they meet the market demand or government regu-
may he summarized as follows. lations for environmental protection. This book can be used
to determine properties of crude oil, its products, and natural
gases that are needed for transportation and storage. Exam-
1.6.1 A p p l i c a t i o n s in Petroleum Processing
ples of such properties are density, boiling point, flash and
(Downstream)
pour points, sulfur content, vapor pressure, and viscosity.
Engineers, scientists, and operators working in various sec- The book can also be used to determine the properties of
tors of petroleum processing and refining or related industries oils for clean-up operations where there is an oil spill on sea-
can use the entire material discussed in the book. It helps water. To simulate the fate of an oil spill and the rate of its
laboratory people in refineries to measure useful properties disappearance at least the following properties are needed in
and to test the reliability of their measurements. The book order to use appropriate simulators [44, 83-85]:
should be useful for engineers and researchers to analyze ex-
9 Characterization of petroleum fractions (Chapter 3)
perimental data and develop their own predictive methods.
9 Pour point (Chapter 3)
It is also intended to help people who are involved with de-
9 Characterization of crude oil (Chapter 4)
velopment of computer softwares and process simulators for
9 Solubility parameter (Chapters 4, 6, and 9)
design and operation of units and equipments in petroleum
9 Density (Chapters 5 and 7)
refineries. Another objective was to help users of such simu-
9 Vapor pressure (Chapter 7)
lators to be able to select an appropriate predictive method
9 Viscosity, diffusion coefficient, and surface tension
for a particular application based on available data on the
(Chapter 8)
fraction.
Accurate prediction of the fate of a crude oil spill depends on
the characterization technique used to estimate the physical
1.6.2 Applications in Petroleum Production
properties. For example, to estimate how much of the ini-
(Upstream)
tial oil would be vaporized after a certain time, accurate val-
Reservoir, chemical, and mechanical engineers may use the ues of the diffusion coefficient, vapor pressure, and molecular
book in reservoir simulators, design and operation of surface weight are needed in addition to an appropriate characteriza-
separators in production fields, and feasibility studies for en- tion method to split the crude into several pseudocomponents
hanced oil recovery projects, such as gas injection projects. E833.
Another application of the book by reservoir engineers is to
simulate laboratory data on PVT experiments for the reser-
voir fluids, determination of the nature and type of reservoir 1.7 DEFINITION OF UNITS AND
fluids, and calculation of the initial amounts of oil and gas in THE CONVERSION FACTORS
the reservoir. Reservoir engineers may also use Chapter 9 to
determine the conditions that a solid may form, amount of An estimated physical property is valuable only if it is ex-
solid formation, and method of its prevention during produc- pressed in an appropriate unit. The most advanced process
tion. Practically all chapters of the book should be useful for simulators and the most sophisticated design approaches
reservoir engineers. fail to perform properly if appropriate units are not used.
This is particularly important for the case of estimation
of physical properties through various correlations or re-
1.6.3 Applications in Academia
porting the experimental data. Much of the confusion with
Although the original goal and aim in writing this book was reported experimental data arises from ambiguity in their
to prepare a reference manual for the industry, laboratories, units. If a density is reported without indicating the tem-
and research institutions in the area of petroleum, it has been perature at which the density has been measured, this value
written in a way such that it can also be used as a textbook has no use. In this part basic units for properties used in
for educational purposes. It can be used as a text for an elec- the book are defined and conversion factors between the
tive course for either undergraduate (senior level) or graduate most commonly used units are given for each property.
level. Students from chemical, petroleum, and mechanical en- Finally some units specifically used in the petroleum indus-
gineering fields as well as from chemistry and physics can take try are introduced. Interested readers may also find other
the course and understand the contents of the book. However, information on units from online sources (for example,
it should not be hard for students from other fields of engi- http://physics.nist.gov/cuu/contents/index.html).
neering and science to use this book. The book may also be
used to conduct short courses in the petroleum industry.
1.7.1 Importance and Types of U n i t s
The petroleum industry and its research began and grew
1.6.4 Other Applications
mainly in the United States during the last century. The rela-
There are several other areas in which the book can be used. tions developed in the 1930s, 1940s, and 1950s were mainly
One may use this book to determine the quality of crude oils, graphical. The best example of such methods is the Winn
petroleum fuels, and products for marketing and government nomogram developed in the late 1950s [86]. However, with the
18 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
b i r t h of the c o m p u t e r a n d its expansion, m o r e analytical As a n e x a m p l e 1 000 000 Pa can be expressed as 1 MPa. These
m e t h o d s in the f o r m of e q u a t i o n s were developed in the 1960s prefixes are n o t used in c o n j u n c t i o n with the English units.
a n d m a i n l y in the 1970s a n d 1980s. Nearly all correlations However, in the English system of units w h e n v o l u m e t r i c
a n d g r a p h i c a l m e t h o d s that were developed until the early quantities of gases are p r e s e n t e d in large n u m b e r s , usually
1980s are in English units. However, starting from the 1980s every 1000 units is expressed by one prefix of M. F o r ex-
m a n y b o o k s a n d h a n d b o o k s a p p e a r e d in the SI units (from Le ample, 2000 scf of gas is expressed as 2 Mscf a n d similarly
Systeme I n t e r n a t i o n a l d'Unites). The general t r e n d is to unify 2 000 000 scf is w r i t t e n as 2 MMscf. Other s y m b o l s usu-
all engineering b o o k s a n d d o c u m e n t s in SI units to be used ally used to express large quantities are b for billion
b y the i n t e r n a t i o n a l community. However, m a n y books, re- (1000 million or 109) a n d tr for trillion (one m i l l i o n millions
ports, h a n d b o o k s , a n d equations a n d figures in various publi- or 1012).
cations are still in English units. The United States a n d United
K i n g d o m b o t h officially use the English system of units.
Therefore, it is essential t h a t engineers be f a m i l i a r with b o t h 1.7.3 Units o f Mass
unit systems of English a n d SI. The o t h e r unit system that
is s o m e t i m e s u s e d for s o m e p r o p e r t i e s is the cgs (centimeter, The m a s s is s h o w n b y m a n d its u n i t in SI is kg (kilogram), in
gram, second) unit, w h i c h is derived f r o m the SI unit. cgs is g (gram), a n d in the English unit system is Ibm (pound-
Since the b o o k is p r e p a r e d for an i n t e r n a t i o n a l audience, mass). On m a n y occasions the s u b s c r i p t m is d r o p p e d for lb
the p r i m a r y unit system u s e d for equations, tables, a n d fig- w h e n it is referred to mass. In the English u n i t system, units
ures is the SI; however, it has b e e n tried to p r e s e n t equivalent of o u n c e (oz) a n d grains are also used for m a s s units s m a l l e r
of n u m b e r s a n d values of p r o p e r t i e s in b o t h SI a n d English t h a n a p o u n d . F o r larger values of mass, unit of ton is used,
units. There are s o m e figures t h a t are t a k e n f r o m o t h e r ref- w h i c h is defined in three forms of long, short, a n d metric.
erences in the literature a n d are in English units a n d t h e y Generally the t e r m t o n is a p p l i e d to the m e t r i c t o n (1000 kg).
have been p r e s e n t e d in t h e i r original form. There are s o m e The conversion factors are as follows:
special units that are c o m m o n l y u s e d to express s o m e spe-
cial properties. F o r example, viscosity is usually expressed in 1 kg = 1000g = 2.204634 lb --- 35.27392 oz
centipoise (cp), k i n e m a t i c viscosity in centistoke (cSt), den-
sity in g/cm 3, specific gravity (SG) at s t a n d a r d t e m p e r a t u r e 11b = 0.45359kg = 453.59g = 16oz = 7000 g r a i n
of 60~ o r the GOR in scf/stb. F o r such properties, these pri- 1 g = 0.001 kg = 0.002204634 lb = 15.4324 g r a i n
m a r y units have b e e n u s e d t h r o u g h o u t the book, while t h e i r i ton (metric) = 1000kg = 2204.6341b
respective equivalent values in SI are also presented.
1 t o n (short) = 2000 lb = 907.18 kg
I ton (long) = 22401b = 1016kg = 1.12ton (short)
1.7.2 F u n d a m e n t a l Units and Prefixes = 1.1016 t o n (metric)
appropriately. The conversion factors a m o n g these units are unit for the m o l e c u l a r weight is not m e n t i o n e d ; however, one
as follows: m u s t realize that it is not a dimensionless parameter. Most
1 year = 365 d = 8 7 6 0 h = 5.256 • 105 m i n = 3.1536 x 107 s recent c o m p i l a t i o n s of m o l a r masses are p r o v i d e d by Coplen
[87].
1 d = 2.743973 x 10 3 year = 24 h = 1440 m i n = 8.64 x 104 s
l h = 1.14155 • 10-4year = 4.16667 • 10-2d
= 60 m i n = 3600 s 1.7.9 U n i t s o f P r e s s u r e
i m i n = 1.89934 • 10-6year = 6.94444 • 10 -4 d Pressure is the force exerted by a fluid per unit area; therefore,
= 1.66667 x 10-2h = 6 0 s in the SI system it has the unit of N / m 2, w h i c h is called Pascal
1 s = 3.17098 x 10 -8 year = 1.157407 • 10 -5 d (Pa), and in the English system has the unit of lbf/ft2 (psf) or
= 2.77777 x 10 -4 h = 1.66667 x 10 -2 m i n lbf/in. 2 (psi). Other units c o m m o n l y used for the pressure are
the bar (bar) an d standard atmosphere (atm). Pressure m a y
also be expressed in t er m s of m m Hg. In this b o o k units of
1.7.6 U n i t s o f F o r c e MPa, kPa, bar, atm, or psi are c o m m o n l y used for pressure.
As m e n t i o n e d above, the u n i t of force in the SI system is The c o n v e r s i o n factors are given as follows:
N e w t o n (N) an d in the English u n i t system is p o u n d - f o r c e 1 a t m = 1.01325 b ar = 101 3 2 5 P a = 101.325kPa
(lbf). 1 lbf is equivalent to the weight of a mass of 1 lbm at = 0.101325 MPa = 14.696psi
the sea level w h e r e the acceleration of gravity is 32.174 ft/s 2
i a t m = 1.0322 k g f / c m 2 = 760 m m Hg (torr) = 29.921 in. Hg
(9.807 m/s2). In the cgs system, the unit of force is dyne (dyn).
A n o t h e r unit for the force in the m e t r i c system is kgf, w h i c h = 10.333 m H 2 0 (4~
is equivalent to the weight of a mass of 1 kg at the sea level. 1 b a r = 0.98692 a t m = 1 • 105 Pa = 100kPa
The conversion factors are as follows: = 0.1 MPa = 14.5038psi
1 N = 1 kg. m / s 2 = 105 dyn = 0.2248 lbf = 1.01968 x 10 -1 kgf 1 Pa = i x 10 - 3 k P a = 1 x 10 - 6 M P a = 9.8692 x 10 - 6 a t m
= 1 • 1 0 - 5 b ar = 1.45037 x 10-4psi
1 lbf = 4.4482 N = 0.45359 kgf
1 psi = 6.804573 • 10 -2 at m = 6.89474 x 1 0 - 2 b a r
1 kgf = 9.807 N = 2.204634 lbf
= 6.89474 x 10 -3 MPa
1 dyn = 10 5 N = 2.248 x 10 -6 lbf
1 psf = 144 psi = 9.79858 a t m = 9.92843 b a r = 0.99285 MPa
1 k g f / c m 2 = 0.96784 a t m = 0.98067 b ar = 14.223 psi
1.7.7 U n i t s o f M o l e s
The actual pressure of a fluid is the absolute pressure, w h i c h
A not h er unit to present a m o u n t of m a t t e r especially in en-
is m e a s u r e d relative to v a c u u m . However, s o m e pressure
gineering calculations is mole (mol), w h i c h is defined as the
m e a s u r e m e n t devices are calibrated to read zero in the at-
ratio of mass (m) to m o l e c u l a r weight (M).
m o s p h e r e an d they s h o w the difference b e t w e e n the abso-
m lute and a t m o s p h e r i c pressure. This difference is called gage
(1.6) n = --
M pressure. N o r m a l l y "a" is used to indicate the absolute value
In SI system the unit of m o l e is kmol, w h e r e m in the above (i.e., psia, bara) an d "g" is used to s h o w the gage pressure
e q u a t i o n is in kg. In the English system, the unit of m o l is (i.e., psig). However, for absolute pressure very often "a" is
lbmol. In the cgs system, the unit of m o l is gmol, w h i c h is d r o p p e d f r o m the unit (i.e., psi, atm, bar). A n o t h e r unit for
usually wr i t t en as mol. F o r example, for m e t h a n e (molecular the pressure is v a c u u m pressure that is defined for pressure
weight 16.04) i m o l of the gas has mass of 16.04 g. One m o l e b el o w a t m o s p h e r i c pressure. Relations b e t w e e n these units
of any substance contains 6.02 x 1023 n u m b e r of molecules are as follows:
(Avogadro's number). The c o n v e r s i o n factors b e t w e e n vari-
ous units of m o l es are the same as given for the mass in (1.7) Pgage = Pabs -- Patm
Section 1.7.3. (1.8) Pabs = Patna -- Pvac
1 krnol = l O 0 0 m o l = 2.2046341bmol
Generally gage pressure unit is used to express pressures
1 l b m o l = 0.45359 k m o l = 453.59 m o l above the a t m o s p h e r i c pressures an d v a c u u m pressure unit is
1 tool = 0.001 krnol = 0.002204634 l b m o l used for pressures b el o w a t m o s p h e r i c and m a y be expressed
in various units (i.e., m m Hg, psi).
1.7.8 U n i t s o f M o l e c u l a r W e i g h t
1.7.10 U n i t s o f T e m p e r a t u r e
M olecu l ar weight or m o l a r mass s h o w n by M is a n u m b e r
that 1 m o l of any substance has equivalent mass of M g. In T e m p e r a t u r e (T) is the m o s t i m p o r t a n t p a r a m e t e r affecting
the SI system the unit of M is kg/kmol a n d in the English properties of fluids an d it is r ep r esen t ed in Centigrade (~
system the unit is lb/lbmol, while in the cgs system the unit an d Kelvin (K) in the SI system an d in F a h r e n h e i t (~ and
of M is g/mol. Molecular weight is r e p r e s e n t e d by the s a m e degrees Ran k i n e (~ in the English unit system. T e m p e r a t u r e
n u m b e r in all unit systems regardless of the system used. As in m o s t eq u at i o n s is in absolute degrees of Kelvin or Rank-
an example, m e t h a n e has the m o l e c u l a r w e i g h t of 16 g/mol, ine. However, acco r d i n g to the definition of Kelvin a nd de-
16 lb/lbmol, an d 16 kg/krnol in the unit systems of cgs, SI, grees Ran k i n e w h e r e there is a t e m p e r a t u r e difference (AT),
a nd English, respectively. F o r this reason, in m a n y cases the unit of ~ is the s a m e as K an d ~ is the s a m e as ~ These
20 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
t e m p e r a t u r e u n i t s are related t h r o u g h the following relations: this reason, to express a m o u n t of gases i n terms of volume,
n o r m a l l y some SC are defined. The SC i n the metric SI u n i t s
(1.9) T(K) = T(~ + 273.15 are 0~ a n d 1 a t m a n d i n the English system are 60~ a n d
1 atm. U n d e r these conditions m o l a r volume of a n y gas is
(1.10) T(~ = T(~ + 459.67
equivalent to 22.4 L/mol (in SI) a n d 379 scf/lbmol (in English
(1.11) AT(K) = AT(~ units). I n reservoir engineering calculations a n d p e t r o l e u m
i n d u s t r y in general, the SC in the SI units are also set at 60~
(1.12) AT(~ = AT(~
(15.5~ or 289 K) a n d 1 atm. The choice of s t a n d a r d temper-
(1.13) T(~ = 1.8T(K) ature a n d pressure (STP) varies from one source to another.
I n this book w h e n the s t a n d a r d T a n d P are n o t specified the
(1.14) T(~ = 1.8T(~ + 32
STP refers to 289 K a n d I atm, which is equivalent to the STP
in English u n i t system rather t h a n SI system (273 K a n d 1
As a n example, absolute t e m p e r a t u r e of 100 K is equivalent to
atm). However, for liquid systems the volume is less affected
1.8 x 100 or 180~ Therefore, the conversion factors between
by pressure a n d for this reason specification of t e m p e r a t u r e
K a n d ~ are as follows:
alone is sufficient.
1 K = 1.8~ (for absolute t e m p e r a t u r e T a n d the t e m p e r a t u r e
difference, AT) 1.7.12 Units of Volumetric and Mass Flow Rates
1~C -- 1.8 ~F (only for the t e m p e r a t u r e difference, A T) Most processes i n the petroleum i n d u s t r y are c o n t i n u o u s a n d
usually the volume or mass quantities are expressed i n the
1.7.11 Units of Volume, Specific Volume, form of rate defined as volume or mass per u n i t time. One
and Molar V o l u m e - - T h e Standard Conditions particular volumetric flow rate used for liquids i n the English
system is gallon (U.S.) per m i n u t e a n d is k n o w n as GPM.
Volume (V) has the d i m e n s i o n of cubic length (L 3) a n d thus Some of the conversion factors for these quantities are
i n SI has the u n i t of m 3 a n d i n English its u n i t is cubic feet
(cf or ft3). Some units particularly used for liquids in the SI l m 3 / s = 1 • 103 L/s = 1.5851 x 104 GPM
system are liter (L), cm 3 (cc), or milliliter (mL) a n d i n English = 5.4345 x I0 s b b l / d = 1.27133 x 105 fta/h
u n i t s are gallon (in U.S. or Imperial) a n d barrel (bbl). Volume 1 fta/h = 7.86558 x 10-4m3/s = 0.12468GPM
of one u n i t mass of a fluid is called specific volume a n d the = 4.27466 b b l / d
volume of 1 mol of a fluid is called m o l a r volume. Some of 1 GPM = 2.228 x 10 -3 ft3/s = 8.0205 m a / h = 34.285 b b l / d
the conversion factors are as follows.
I b b l / d = 2.9167 x 10 -2 GPM = 1.8401 • 10 -4 ma/s
1 m 3 = 106 cm 3 = 1000 L = 35.315 f t 3 ~- 264.18 gallon (U.S,) = 0.23394 ft3/h
= 35.316 ft 3 = 6.29bbl The conversion factors for the mass rates are as follows:
I ft 3 = 2.8316 x 1 0 - 2 m 3 = 28.316L = 7.4805 gallon(U.S.) 1 kg/s -- 7.93656 x 103 lb/h --- 3.5136 x 107 ton/year
1 bbl = 42 gallon(U.S.) = 158.98 L = 34.973 gallon (Imperial) 1 lb/s = 1.63295 x 103 kg/h = 39.1908ton/d
1 gallon (U.S.) = 0.8327 gallon (Imperial)
= 0.023809 bbl = 3.7853 L The same conversion factors apply to m o l a r rates.
1 mL = i c m 3 = 10-3L = 1 0 - 6 m 3 ---0.061024in. 3
1.7.13 Units of Density and Molar Density
For the m o l a r volumes some of the conversion factors are Density s h o w n by d or p is defined as mass per u n i t volume
given as follows: a n d it is reciprocal of specific volume. The conversion factors
can be o b t a i n e d from reversing those of specific volume in
1 m 3 / k m o l = 1 L / m o l = 0.001 m 3 / m o l = 1000 c m 3 / m o l Section 1.7.11.
= 16.019 ft3/lbmol 1 kg/m3 = 6.24259 x 10-21b/ft3 = 1 x 10-3 g/cm 3
1 ft3/lbmol = 6.24259 x 10 -2 m 3 / k m o l = 8.3455 x 10 -3 lb/gal
= 6.24259 x 10 -5 m a / m o l = 62.4259 c m 3 / m o l 1 lb/ft 3 = 16.019kg/m 3 = 1.6019 • 10-2 g/cm 3
1 c m a / m o l = i m L / m o l = 1 L / k m o l = 0.001 m a / k m o l = 0.13368 lb/gal
= 1.6019 x 10 -2 fta/lbmol 1 g/cm 3 = 1 kg/L = 103 kg/m 3 = 62.42591b/ft 3
It should be noted that the same conversion factors apply -= 8.3455 lb/gal
to specific volumes. For example, 1 lb/gal = 1.19825 x 102 kg/m 3 = 7.48031b/ft 3
= 0.119825 g/cm 3
1 ft3/lb = 6.24259 • 10 -2 m3/kg = 62.4259cm3/g
Density m a y also be presented i n terms of n u m b e r of moles
Since volume a n d specific or m o l a r volumes depend o n tem- per u n i t volume, which is called molar density a n d is recipro-
perature a n d pressure of the system, values of volume in a n y cal of m o l a r volume. It can be o b t a i n e d by dividing absolute
u n i t system are meaningless if the conditions are n o t spec- density to molecular weight.The conversion factors for m o l a r
ified. This is particularly i m p o r t a n t for gases in which both density are exactly the same as those for the absolute density
t e m p e r a t u r e a n d pressure strongly influence the volume. For (i.e., I mol/cm 3 = 62.4259 lbmol/ft3). I n practical calculations
1. I N T R O D U C T I O N 21
the conversion factors may be simplified without major volume fraction can be obtained from the following equation:
error in the calculations. For example, 62.4 instead of 62.4259
or 7.48 instead of 7.4803 are used in practical calculations. Xwi/ SGi
In expressing values of densities, similar to specific volumes, (1.16) xvi -- ~N=I Xwi/SGi
the SC must be specified. Generally densities of liquid hydro-
carbons are reported either in the form of specific gravity at
15.5~ (60~ or the absolute density at 20~ and 1 atm in in which x~ is the volume fraction and SGi is the specific grav-
g/cm 3. ity of component i. In Eq. (1.16) density (d) can also be used
instead of specific gravity. If mole and weight fractions are
multiplied by 100, then composition is calculated on the per-
1.7.14 Units of Specific Gravity
centage basis. In a similar way the conversion of composition
For liquid systems, the specific gravity (SG) is defined as the from volume to weight and then to mole fraction can be ob-
ratio of density of a liquid to that of water, and therefore, it is tained by reversing the above equations. The composition of
a dimensionless quantity. However, the temperature at which a component in a liquid mixture may also be presented by its
specific gravity is reported should be specified. The specific molar density, units of which were discussed in Section 1.7.13.
gravity is also called relative density versus absolute density. Generally, a solution with solute molarity of 1 has 1 mol of
For liquid petroleum fractions and crude oils, densities of solute per 1 L of solution (1 mol/L). Through use of both
both the oil and water are expressed at the SC of 60~ (15.5~ molecular weight of solute and density of solution one can
and 1 atm, and they are usually indicated as SG at 60~176 obtain weight fraction from molarity. Another unit to express
or simply SG at 60~ Another unit for the specific gravity of concentration of a solute in a liquid solution is molality. A so-
liquid hydrocarbons is defined by the American Petroleum lution with molality of 1 has 1 mol of solute per i kg of liquid
Institute (API) and is called API degree and is defined in terms solvent.
of SG at 60~ (API = 141.5/SG-131.5). For gases, the spe- Another unit for the composition in small quantities is the
cific gravity is defined as the ratio of density of the gas to ppm (part per million), which is defined as the ratio of unit
that of the air at the SC, which is equivalent to the ratio of weight (or volume) of a component to 106 units of weight or
molecular weights. Further discussion on specific gravity, def- volume for the whole mixture. Therefore, ppm can be pre-
initions, and methods of calculation are given in Chapter 2 sented in terms of both volume or weight. Usually in gases
(Section 2.1.3). the ppm is presented in terms of volume and in liquids it is
expressed in terms of weight. When ppm is presented in terms
of weight, its relation with wt% is 1 ppm = 10.4 wt%. For ex-
1.7.15 Units of Composition
ample, the maximum allowable concentration of H2 S in air
Composition is the most important characteristic of homoge- for prolonged exposure is 10 ppm or 0.001 wt%. There is an-
nous mixtures in which two or more components are uni- other smaller unit definedas part per billion known as ppb
formly mixed in a single phase. Because of the nature of (1 ppm = 1000 ppb). In the United States a gas is considered
petroleum fluids, accurate knowledge of composition is im- "sweet" if the amount of its H2S content is no more than one
portant. Generally composition is expressed as percent- quarter grain per i00 scf of gas. This is almost equivalent to
age (%) or as fraction (percent/100) in terms of weight, mole, 4 x 10.4 mol fraction [88]. This is in turn equivalent to 4 ppm
and volume. Density of the components (or pseudocompo- on the gas volume basis. Gas composition may also be rep-
nents) constituting a mixture is required to convert composi- resented in terms of partial pressure where sum of all partial
tion from weight basis to volume basis or vice versa. Similarly pressures is equivalent to the total pressure.
conversion of composition from mole basis to weight basis In general, the composition of gases is presented in volume
or vice versa requires molecular weight of the constituting or mole fractions, while the liquid composition may be pre-
components (or pseudocomponents). Mole, weight, and vol- sented in any form of weight, mole, or volume. For gases at
ume fractions are shown by Xm, X~, and xv, respectively. Mole, low pressures (< 1 atm where a gas may be considered an ideal
weight, and volume percentages are shown by mol%, wt%, gas) mole fraction and volume fractions are the same. How-
and vol%, respectively. Some references use mol/mol, wt/wt, ever, generally under any conditions, volume and mole frac-
and vol/vol to express fractional compositions. For normal- tions are considered the same for gases and vapor mixtures.
ized compositions, the sum of fractions for all components For narrow boiling range petroleum fractions with composi-
in a mixture is 1 (Y~xi = 1) and the sum of all percentages is tions presented in terms of PNA percentages, it is assumed
100. If the molecular weights of all components in a mixture that densities and molecular weights for all three representa-
are the same, then the mole fraction and weight fraction are tive pseudocompoents are nearly the same. Therefore, with
identical. Similarly, if the density (or specific gravity) of all a good degree of approximation, it is assumed that the PNA
components is the same, the weight and volume fractions are composition in all three unit systems are the same and for
identical. The formula to calculate weight fraction from mole this reason on many occasions the PNA composition is repre-
fraction is given as sented only in terms of percentage (%) or fraction without in-
XmiMi dicating their weight or volume basis. However, this is not the
(1.15) Xwi - - ~N 1 xmiMi case for the crude or reservoir fluid compositions where the
composition is presented in terms of boiling point (or carbon
where N is the total number of components, Mi is the molec- number) and not in the form of molecular type. The following
ular weight, and Xwi and Xmi are the weight and mole fractions example shows conversion of composition from one type to
of component i, respectively. The conversion from weight to another for a crude sample.
22 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 1.5--Conversion of composition of a crude oil sample from mole to weight and volume percent.
Component tool% Molecularweight (M) Specific
gravity(SG) wt% vol%
C2 0.19 30.07 0.356 0.03 0.06
C3 1.88 44.10 0.508 0.37 0.64
iC4 0.62 58.12 0.563 0.16 0.25
nC4 3.92 58.12 0.584 1.02 1.52
iC5 2.11 72.15 0.625 0.68 0.95
nC5 4.46 72.15 0.631 1.44 1.98
C6 (fraction) 8.59 82.00a 0.690 3.15 3.97
C7+ (fraction) 78.23 266.00 0.895 93.15 90.63
Sum 100.00 100.00 100.00
~This is m o l e c u l a r w e i g h t of C6 h y d r o c a r b o n g r o u p a n d s h o u l d not be m i s t a k e n with M of nC6 w h i c h is 86.2.
Example / . / - - T h e composition of a Middle East crude 1 kW. h = 3600kJ = 3412.2 Btu = 2.655 • 106 ft-lbf
oil is given in Table 1.5 i n terms of mol% with k n o w n
molecular weight a n d specific gravity for each c o m p o n e n t / Energy per u n i t mass is called specific energy that m a y be
p s e u d o c o m p o n e n t . Calculate the c o m p o s i t i o n of the crude i n used to present properties such as specific enthalpy, specific
both wt% a n d vo1%. i n t e r n a l energy, specific heats of reaction, a n d c o m b u s t i o n or
the heating values of fuels. Some of the conversion factors are
Solution--In this table values of molecular weight a n d spe- given below.
cific gravity for pure c o m p o u n d s are o b t a i n e d from Chapter 2
(Table 2.1), while for the C6 group, values are taken from 1 J/g = 103 J/kg = 1 kJ/kg -- 0.42993 Btu/lb
Chapter 4 a n d for the C7 ~ fraction, values are given by the 1 Btu/lb = 2.326 J/g = 0.55556 cal/g
laboratory. Conversion calculations are based o n Eqs. (1.15)
a n d (1.16) o n the percentage basis a n d the results are also The same conversion factors apply to the units of m o l a r en-
given in Table 1.5. I n this calculation it is seen that i n terms ergy such as m o l a r enthalpy.
of wt% a n d vo1%, heavier c o m p o u n d s (i.e., C7+) have higher
values t h a n in terms of mol%. t
1.7.17 Units of Specific Energy per Degrees
1.7.16 Units of Energy and Specific Energy Properties such as heat capacity have the u n i t of specific en-
ergy per degrees. The conversion factors are as follows:
Energy in various forms (i.e., heat, work) has the u n i t of Joule
(1 J -- 1 N- m) i n the SI a n d ft-lbf in the English system. Val- J J
ues of heat are also presented in terms of calorie (in SI) a n d 1~ = 1 x 10 .3 kgOC = 1 = 0.23885 Btu
BTU (British T h e r m a l Unit) i n the English system. There are
two types of joules: absolute joules a n d i n t e r n a t i o n a l joules, cal Btu J
where i Joule (int.) =1.0002 Joule (abs.). In this book only ab- 1~ = 1~ = 4.1867 go--~
solute joules is used a n d it is designated by J. There are also
two types of calories: t h e r m o c h e m i c a l a n d I n t e r n a t i o n a t i o n a l As m e n t i o n e d i n Section 1.7.13, for the difference i n tem-
Steam Tables, where I cal (international steam tables) -- perature (AT), units of ~ a n d K are the same. There-
1.0007 cal (thermochemical) as defined i n the API-TDB [47]. fore, the units of heat capacity m a y also be represented
I n this book cal refers to the i n t e r n a t i o n a l steam tables unless i n terms of specific energy per Kelvin or degrees R a n k i n e
otherwise is specified. I n the cgs system the u n i t of energy is (i.e., 1 ~ = 1 lb.~
Bm = 1 ~ g.~ = 1 ~ ) " The same conversion fac-
~-.~
dyn-cm, which is also called erg. The u n i t of power in the SI tors apply to u n i t s of m o l a r energy per degrees such as m o l a r
system is J/s or watt (W). Therefore, kW.h equivalent to 3600 heat capacity.
kJ is also a u n i t for the energy. The p r o d u c t of pressure a n d Another p a r a m e t e r which has the u n i t of m o l a r energy per
volume (PV) m a y also present the u n i t of energy. Some of the degrees is the universal gas constant (R) used in thermody-
conversion factors for the u n i t s of energy are given as follows: n a m i c relations a n d equations of state. However, the u n i t of
t e m p e r a t u r e for this p a r a m e t e r is the absolute t e m p e r a t u r e
1 J = 1 N. m -- 10 -3 kJ = 107erg -- 0.23885 cal (K or ~ a n d ~ or ~ m a y never be used in this case. Similar
= 9.4783 x 10-4 Btu = 2.778 x 1 0 - 7 k W . h conversion factors as those used for the heat capacity given
1 J = 3.725 x 10 -7 h p . h ---=0.73756 ft.lbf = 9.869 L.atm above also apply to the u n i t s of gas constants in terms of mo-
I cal ( I n t e r n a t i o n a l Tables) -- 3.9683 x 10 -3 Btu = 4.187 J lar energy per absolute degrees.
= 3.088 ft.lbf = 1.1630 x 1 0 - 6 k W . h
Btu cal cal(thermochemical)
1 cal (thermochemical) = 1 cal = 3.9657 x 10 .3 Btu 1 - 1 - 1.0007
lbmol 9~ mol. K mol. K
= 4.184J = 3.086 ft-lbf = 1.1622 x 1 0 - 6 k W . h
J
1Btu = 1055 J = 251.99 cal = 778.16 ft. lbf = 4.1867 x 1 0 3 _
kmol. K
= 2.9307 x 1 0 - 4 k W . h
1 ft-lbf = 1.3558J = 0.32384cal = 1.2851 x 10-3 Btu Numerical values of the gas c o n s t a n t are given i n Sec-
= 3.766 x 1 0 - 7 k W . h tion 1.7.24.
1. I N T R O D U C T I O N 23
1.7.18 Units of Viscosity and Kinematic Viscosity simplified to the following form at the t e m p e r a t u r e s of 311
(100~ a n d 372 K, (210~ respectively [1].
Viscosity (absolute viscosity) s h o w n b y / ~ is a p r o p e r t y t h a t
c h a r a c t e r i z e s the fluidity of fluids a n d it has the d i m e n s i o n (1.19) SUS100F = 4.632v100F vl00F > 7 5 c S t
of m a s s p e r length p e r t i m e (M/L. t). If the r e l a t i o n b e t w e e n
d i m e n s i o n s of force ( F ) a n d m a s s (M) is u s e d (F = M.L.t-2),
(1.20) SUS210F = 4.664VZ10F V210F> 75 cSt
t h e n absolute viscosity finds the d i m e n s i o n of F.t.L -2 w h i c h
is the s a m e as d i m e n s i o n for the p r o d u c t of p r e s s u r e a n d w h e r e Vl00F is the k i n e m a t i c viscosity at 100~ (311 K) in cSt.
time. Therefore, in the SI system the unit of viscosity is Pa- s As a n example, a p e t r o l e u m fraction with k i n e m a t i c viscosity
(N. m-2.s). In the cgs system the unit of viscosity is in g/cm- s of 5 cSt at 311 K has a n equivalent S a y b o h Universal Viscosity
that is called poise (p) a n d its h u n d r e d t h is called centipoise of 42.4 SUS as calculated f r o m Eq. (1.17).
(cp), w h i c h is equivalent to mi]li-Pa, s ( m P a . s). The conver- A n o t h e r unit for the viscosity is S F S (Saybolt foural sec-
sion factors in various units are given below. onds) expressed for S a y b o l t foural viscosity, w h i c h is m e a -
1 cp = 1.02 x 10 -4 kgf. s/m 2 : 1 • 10 -3 Pa- s = 1 m P a . s s u r e d in a way s i m i l a r to S a y b o l t universal viscosity b u t
m e a s u r e d b y a larger orifice (ASTM D 88). The conversion
= 10-2p = 2.089 x 10 -5 lbf. s/ft 2 = 2.4191b/h- ft
f r o m cSt to S F S is expressed t h r o u g h the following equations
= 3.6 kg/h. m
at two reference t e m p e r a t u r e s of 122~ (323 K) a n d 210~
1 Pa.s = 1 k g / m . s = 1000 cp = 0.67194 lb/ft- s (372 K) [47].
1 lb/h.ft = 8.634 x 10 -6 lbf. s/ft 2 = 0.4134 cp = 1.488 kg/h. m
13924
i kgf. s/m 2 = 9.804 x 103cp = 9.804 Pa. s = 0.20476 lbf 9s/ft 2 (1.21) SFS122F = 0.4717Vm22F +
U 1F2 22 -- 72.59VI22F + 6816
1 lbf. s/ft 2 = 4.788 • 104 cp = 4.884 kgf. s/m 2
5610
The ratio of viscosity to density is k n o w n as kinematic vis- (1.22) SFS210F = 0.4792v2mF + v2210F+ 2130
cosity (v) a n d has the d i m e n s i o n of L/t 2. In the cgs system,
the unit of k i n e m a t i c vsicosity is cm2/s also called stoke (St)
F o r conversion of Saybolt foural viscosity (SFS) to k i n e m a t i c
a n d its h u n d r e d t h is centistoke (cSt). The conversion factors
viscosity (cSt.), the above equations should be u s e d in reverse
are given below.
o r to use t a b u l a t e d values given b y API-TDB [47]. As an exam-
1 ft2/h = 2.778 • 10-4 ft2/s --- 0.0929m2/h = 25.81 cSt ple, an oil w i t h Saybolt foural viscosity of 450 S F S at 210~
1 ft2/s = 9.29 • 104 cSt = 334.5 m2/h has a k i n e m a t i c viscosity of 940 cSt. Generally, viscosity of
highly viscous oils is p r e s e n t e d by SUS or S F S units.
1 cSt = 10 -2 S t - - 10-6m2/s = 1 mmZ/s = 3.875 x 10 -2 ftZ/h
= 1.076 • 10 -s ft2/s
1 m2/s = 104 St = 106 cSt = 3.875 x 104 ft2/h
1.7.19 Units of Thermal Conductivity
T h e r m a l conductivity (k) as discussed in C h a p t e r 8 r e p r e s e n t s
A n o t h e r u n i t to express k i n e m a t i c viscosity of liquids is
a m o u n t of h e a t passing t h r o u g h a u n i t a r e a of a m e d i u m for
Saybolt universal seconds (SUS), w h i c h is the unit for the
one unit of t e m p e r a t u r e g r a d i e n t ( t e m p e r a t u r e difference p e r
S a y b o l t universal viscosity (ASTM D 88). Definition of viscos-
unit length). Therefore, it has the d i m e n s i o n of energy p e r
ity gravity c o n s t a n t (VGC) is b a s e d on SUS unit for the viscos-
t i m e p e r a r e a p e r t e m p e r a t u r e gradient. In the SI units it is
ity at two reference t e m p e r a t u r e s of 100 a n d 210~ (37.8 a n d
expressed in J/s- m . K. Since t h e r m a l conductivity is defined
98.9 ~ The VGC is u s e d in C h a p t e r 3 to e s t i m a t e the com-
b a s e d on a t e m p e r a t u r e difference (AT), the u n i t of ~ m a y
p o s i t i o n of heavy p e t r o l e u m fractions. The relation b e t w e e n
also be used i n s t e a d of K. Because J/s is defined as w a t t (W),
SUS a n d cSt is a function of t e m p e r a t u r e a n d it is given in the
the u n i t of t h e r m a l conductivity in the SI system is u s u a l l y
API TDB [47]. The analytical relations to convert cSt to SUS
w r i t t e n as W/re. K. I n the English system, the unit of ther-
are given b e l o w [47].
m a l conductivity is Btu a n d in s o m e references is w r i t t e n
SUSeq = 4.6324VT as ~9' h .o~ t/" w h i c h is the. ratio of h e a t flux to the. t e m p e r. a t u r e
[1.0 + 0.03264VT] gradient. The conversion factors b e t w e e n various u m t s are
+ given below.
[(3930.2 + 262.7vr + 23.97v~- + 1.646v 3) x 10 -5]
(1.17) 1 W/re. K (J/s. m.~ = 0.5778 Btu/fl. h.~
= 1.605 • 10 -4 Btu/ft. s.~
w h e r e vr is the k i n e m a t i c viscosity at t e m p e r a t u r e T in cSt. = 0.8593 kcal/h- m-~
The SUSeq calculated from this relation is converted to the
S U S r at the d e s i r e d t e m p e r a t u r e of T t h r o u g h the following 1 Btu/ft-h-~ = 1.7307 W / m . K
relation. 1 cal/cm 9s.~ = 242.07 Btu/ft 9h-~ = 418.95 W / m . K
squared length per time (L2]t). Usually it is expressed i n u n i t s of scf/stft 3. This u n i t is exactly the same as sm3/stm 3
cm2/s. i n the SI unit.
i cm2/s = 10 -4 m2/s : 9.29 x 10 -6 f t 2 / s : 3.3445 x 10 -4 f t 2 / h The conversion factors b e t w e e n these three u n i t s for the GOR
(GLR) are given as follows:
Saudi Arabian crude of 35.5 API (SG = 0.847), the follow- factors with ratios of unity as follows:
ing conversion factors apply between weight and volume of
crudes and the rates: 802~ol =(802~ol) x m
X ~l 6 g ~1 0 ~ J x
Example 1.2--The molar heating value of m e t h a n e is 802 kJ/ [0.07 x 251.99 x 4.187 x 1000 x 1.8 m_~.~
mol. Calculate the heating value of methane in the units of
cal/g and Btu/lb. The molecular weight of methane is 16.0,
= 121.18 m W / m K
Solution--In this calculation a practicing engineer has to re- Examples 1.2 and 1.3 show that with the knowledge of only
m e m b e r the following basic conversion factors: 1 lb = 453.6 g, very few conversion factors and basic definitions of funda-
1 cal = 4.187 J, and 1 Btu = 252 cal. The value of molecular mental units, one can obtain the conversion factor between
weight indicates that 1 tool = 16 g. In the conversion process any two unit systems for any property without use of a refer-
the initial unit is multiplied by a series of k n o w n conversion ence conversion table.
26 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
Chemistry Symposia, American Chemical Society (ACS), [57] Watson, IC M., Nelson, E. E, and Murphy, G. B.,
Vol. 40, No. 8, 2001, pp. 1975-1980. "Characterization of Petroleum Fractions," Industrial and
[36] Poling, B. E., Prausnitz, J. M., and O'Connell, J. E, Properties of Engineering Chemistry, Vol. 27, 1935, pp. 1460-1464.
Gases and Liquids, 5th ed., McGraw-Hill, New York, 2000. [58] Brule, M. R., Kumar, K. H., and Watanasiri, S.,
[37] Mattax, C. C. and Dalton, R. L., Reservoir Simulation, SPE "Characterization Methods Improve Phase Behavior
Nomograph 13, Society of Petroleum Engineers, Richardson, Predictions," Oil & Gas Journal, February 11, 1985, pp. 87-93.
TX, 1990. [59] Riazi, M. R., A1-Sahhaf, T. A., and Al-Shammari, M. A.,
[38] Riazi, M. R. and Faghri, A., "Effect of the Interfacial Drag on "A Generalized Method for Estimation of Critical Constants,"
Gas Absorption with Chemical Reaction in a Vertical Tube," Fluid Phase Equilibria, Vol. 147, 1998, pp. 1-6.
American Institute of Chemical Engineers Journal, [60] Dohrn, R. and Pfohl, O. "Thermophysical
Vol. 32, No. 4, 1986, pp. 696-699. Properties--Industrial Directions," Paper presented at the
[39] Dabir, B., Riazi, M. R., and Davoudirad, H. R., "Modeling of Ninth International Conference on Properties and Phase
Falling Film Reactors," Chemical Engineering Science, Vol. 51, Equilibria for Product and Process Design (PPEPPD 2001),
No. 11, 1996, pp. 2553-2558. Kurashiki, Japan, May 20-25, 2001.
[40] Riazi, M. R., Whitson, C. H., and da Silva, E, "Modelling of [61] McCabe, W. L., Smith, J. C., and Harriot, P., Unit Operations of
Diffusional Mass Transfer in Naturally Fractured Reservoirs," Chemical Engineering, 4th ed., McGraw-Hill, New York, 1985.
Journal of Petroleum Science and Engineering, Vol. 10, 1994, [62] Peridis, S., Magoulas, K., and Tassios, D., "Sensitivity of
pp. 239-253. Distillation Column Design to Uncertainties in Vapor-Liquid
[41] Riazi, M, R. and Whitson, C. H., "Estimating Diffusion Equilibrium Information," Separation Science and Technology,
Coefficients of Dense Fluids," Industrial and Engineering Vol. 28, No. 9, 1993, pp. 1753-1767.
Chemistry Research, Vol. 32, No. 12, 1993, pp. 3081-3088. [63] Riazi, M. R. and Mansoori, G. A., "Simple Equation of State
[42] Riazi, M. R., "A New Method for Experimental Measurement of Accurately Predicts Hydrocarbon Densities," Oil and Gas
Diffusion Coefficient in Reservoir Fluids," Journal of Petroleum Journal, July 12, 1993, pp. 108-111.
Science and Engineering, Vol. 14, 1996, pp. 235-250. [64] Lee, B. I. and Kesler, M. G., "A Generalized Thermodynamic
[43] Kister, H. Z., Distillation Operations, McGraw-Hill, New York, Correlation Based on Three- Parameter Corresponding States,"
1990. American Institute of Chemical Engineers Journal, Vol. 21, No. 5,
[44] Villoria, C. M., Anselmi, A. E., Intevep, S. A., and Garcia, E R., 1975, pp. 510-527.
"An Oil Spill Fate Model," SPE23371, Society of Petroleum [65] Riazi, M. R. and Daubert, T. E., "Application of Corresponding
Engineers, 1991, pp. 445-454. States Principles for Prediction of Self-Diffusion Coefficients in
[45] Whitson, C. H. and Brule, M. R., Phase Behavior, Monograph Liquids," American Institute of Chemical Engineers Journal,
Volume 20, Society of Petroleum Engineers, Richardson, TX, Vol. 26, No. 3, 1980, pp. 386-391.
2000. [66] Riazi, M. R. and Mansoori, G. A., "Use of the Velocity of Sound
[46] Firoozabadi, A., Thermodynamics of Hydrocarbon Reservoirs, in Predicting the PVT Relations," Fluid Phase Equilibria, Vol. 90,
McGraw-Hill, New York, 1999. 1993, pp. 251-264.
[47] Daubert, T. E. and Danner, R. E, Eds., API Technical Data [67] Shabani, M. R., Riazi, M. R., and Shabau, H. I., "Use of Velocity
Book--Petroleum Refining, 6th ed., American Petroleum of Sound in Predicting Thermodynamic Properties from Cubic
Institute (API), Washington, DC, 1997. Equations of State," Canadian Journal of Chemical Engineering,
[48] AIChE DIPPR| Design Institute for Physical Vol. 76, 1998, pp. 281-289.
Property Data (DIPPR), EPCON International, Houston, TX, [68] Riazi, M. R. and Roomi, Y. A., "Use of Velocity of Sound in
1996. Estimating Thermodynamic Properties of Petroleum
[49] Rowley, R. L., Wilding, W. V., Oscarson, J. L., Zundel, N. A., Fractions,"Preprints of Division of Petroleum Chemistry
Marshall, T. L., Daubert, T. E., and Danner, R. P., DIPPR Data Symposia, American Chemical Society (ACS), August 2000,
Compilation of Pure Compound Properties, Design Institute for Vol. 45, No. 4, pp. 661-664.
Physical Properties (DIPPR), Taylor & Francis, New York, 2002 [69] Riazi, M. R. and Faghri, A., "Prediction of Thermal
(http://dippr.byu.edu). Conductivity of Gases at High Pressures," American Institute of
[50] Thermodynamic Research Center, National Institute of Chemical Engineers Journal, Vol.31, No.l, 1985, pp. 164-167.
Standards and Technology (NIST), Boulder, CO, URL: [70] Riazi, M. R. and Faghri, A., "Thermal Conductivity of Liquid
http://www.nist.gov/. and Vapor Hydrocarbon Systems: Pentanes and Heavier at Low
[51] Riazi, M. R. and Daubert, T. E., "Analytical Correlations Pressures," Industrial and Engineering Chemistry, Process
Interconvert Distillation Curve Types," Oil and Gas Journal, Design and Development, Vol. 24, No. 2, 1985, pp. 398-401.
August 25, 1986, pp. 50-57. [71] Ahmed, T., Hydrocarbon Phase Behavior, Gulf Publishing,
[52] Riazi, M. R. and Daubert, T. E., "Improved Characterization of Houston, TX, 1989.
Wide Boiling Range Undefined Petroleum Fractions," Industrial [72] Danesh, A., PVT and Phase Behavior of Petroleum Reservoir
and Engineering Chemistry Research, Vol. 26, 1987, pp. 629-632. Fluids, Elsevier, Amsterdam, 1998.
[53] Riazi, M. R. and Daubert, T. E., "Predicting Flash Points and [73] Tsonopoulos, C., Heidman, J. L., and Hwang, S.-C.,
Pour Points of Petroleum Fractions," Hydrocarbon Processing, Thermodynamic and Transport Properties of Coal Liquids, An
September 1987, pp. 81-84. Exxon Monograph, Wiley, New York, 1986.
[54] Riazi, M. R., "A Distribution Model for C7+ Fractions [74] National Institute of Standards and Technology (NIST),
Characterization of Petroleum Fluids," Industrial and Boulder, CO, 2003 (http://webbook.nist.gov/chemistry/).
Engineering Chemistry Research, Vol. 36, 1997, pp. 4299-4307. [75] Ahmed, T., Reservoir Engineering Handbook, Gulf Publishing,
[55] Riazi, M. R., "Distribution Model for Properties of Houston, TX, 2000.
Hydrocarbon-Plus Fractions," Industrial and Engineering [76] Edmister, W. C. and Lee, B. I., Applied Hydrocarbon
Chemistry Research, Vol. 28, 1989, pp. 1731-1735. Thermodynamics, 2nd ed., Gulf Publishing, Houston, TX, 1985.
[56] Riazi, M. R. and Al-Sahhaf, T. A., "Physical Properties of Heavy [77] HYSYS, "Reference Volume 1., Version 1.1", HYSYS Reference
Petroleum Fractions and Crude Oils," Fluid Phase Equilibria, Manual for Computer Software, HYSYS Conceptual Design,
Vol. 117, 1996, pp. 217-224. Hyprotech Ltd., Calgary, Alberta, Canada, 1996.
1. I N T R O D U C T I O N 29
[78] EPCON, "API Tech Database Software," EPCON International, [84] Riazi, M. R. and Edalat, M., "Prediction of the Rate of Oil
Houston, TX, 2000 (www.epcon.com). Removal from Seawater by Evaporation and Dissolution,"
[79] Stange, E. and Johannesen, S. O., "HYPO*S, A Program for Journal of Petroleum Science and Engineering, Vol. 16, 1996,
Heavy End Characterization," An Internal Program from the pp. 291-300.
Norsk Hydro Oil and Gas Group (Norway), Paper presented at [85] Fingas, M. E, "A Literature Review of the Physics
the 65th Annual Convection, Gas Processors Association, San and Predictive Modeling of Oil Spill Evaporation,"
Antonio, TX, March 10, 1986. Journal of Hazardous Materials, Vol. 42, 1995, pp. 157-
[80] Aspen Plus, "Introductory Manual Software Version," Aspen 175.
Technology, Inc., Cambridge, MA, December 1986. [86] Winn, E W., "Physical Properties by Nomogram," Petroleum
[81] PRO/II, "Keyword Manual," Simulation Sciences Inc., Refiners, Vol. 36, No. 21, 1957, pp. 157.
Fullerton, CA, October 1992. [87] Coplen, T. B., "Atomic Weights of the Elements 1999," Pure and
[82] Riazi, M. R., "Estimation of Physical Properties and Applied Chemistry, Vol. 73, No. 4, 2001, pp. 667-683.
Composition of Hydrocarbon Mixures," Analytical Advartces [88] Maddox, R. N., Gas Conditioning and Processing, Volume 4,
for Hydrocarbon Research, edited by Hsu, C. Samuel, Kluwer Campbell Petroleum Series, John M. Campbell, Norman, OK,
Academic/Plenum Publishers, New York, NY, 2003, pp. 1-26. 1985.
[83] Riazi, M. R. and A1-Enezi, G., "A Mathematical Model for the [89] Riazi, M. R., "Estimation of Rates and Enhancement Factors in
Rate of Oil Spill Disappearance from Seawater for Kuwaiti Gas Absorption with Zero-Order Reaction and Gas Phase Mass
Crude and Its Products," Chemical Engineering Journal, Vol. 73, Transfer Resistances," Chemical Engineering Science, Vol. 41,
1999, pp. 161-172. No. 11, 1986, pp. 2925-2929.
MNL50-EB/Jan. 2005
Characterization and
Properties of Pure
Hydrocarbons
30
As DISCUSSEDIN CHAPTER1, the characterization of petroleum pressure to normal boiling point requires a vapor pressure re-
fractions and crude oils depends on the characterization and lation and methods for its calculation for petroleum fractions
properties of pure hydrocarbons. Calculation of the prop- are discussed in Chapter 3. The boiling point, when available,
erties of a mixture depends on the properties of its con- is one of the most important characterization parameters for
stituents. In this chapter, first basic parameters and properties hydrocarbons and is frequently used in property estimation
of pure compounds are defined. These properties are either methods.
temperature-independent or values of some basic properties
at a fixed temperature. These parameters are the basis for
2.1.3 Density, Specific Gravity, and API Gravity
calculation of various physical properties discussed in this
book. Reported values of these parameters for more than Density is defined as mass per unit volume of a fluid. Density
100 selected pure compounds are given in Section 2.2. These is a state function and for a pure compound depends on both
values will be used extensively in the following chapters, es- temperature and pressure and is shown by p. Liquid densities
pecially in Chapters 3 to determine the quality and properties decrease as temperature increases but the effect of pressure
of petroleum fractions. In Section 2.3, the characterization of on liquid densities at moderate pressures is usually negligible.
hydrocarbons is introduced, followed by the development of At low and moderate pressures (less than a few bars), satu-
a generalized correlation for property estimation that is a rated liquid density is nearly the same as actual density at the
unique feature of this chapter. Various correlations and meth- same temperature. Methods of the estimation of densities of
ods for the estimation of these basic parameters for pure hy- fluids at various conditions are discussed in Chapters 5 and 7.
drocarbons and narrow boiling range petroleum fractions are However, liquid density at the reference conditions of 20~
presented in different sections. Finally, necessary discussion (293 K) and 1 atm is shown b y d and it is used as a characteri-
and recommendations for the selection of appropriate pre- zation parameter in this chapter as well as Chapter 3. Parame-
dictive methods for various properties are presented. ter d is also called absolute density to distinguish from relative
density. Other parameters that represent density are specific
volume (l/d), molar volume (M/d), and molar density (d/M).
2.1 DEFINITION O F BASIC PROPERTIES Generally, absolute density is used in this book as the charac-
teristic parameter to classify properties of hydrocarbons.
In this section, all properties of pure hydrocarbons presented Liquid density for hydrocarbons is usually reported in
in Section 2.2 are defined. Some specific characteristics of terms of specific gravity (SG) or relative density defined as
petroleum products, such as cetane index and pour point, are
defined in Chapter 3. Definitions of general physical proper- density of liquid at temperature T
(2.1) SG =
ties such as thermal and transport properties are discussed in density of water at temperature T
corresponding chapters where their estimation methods are Since the standard conditions adopted by the petroleum in-
presented. dustry are 60 ~F (15.5 ~C) and 1 atm, specific gravities of liquid
hydrocarbons are normally reported at these conditions. At a
2.1.1 Molecular Weight reference temperature of 60~ (15.5~ the density of liquid
water is 0.999 g/cm 3 (999 kg/m 3) or 8.337 lb/gal(U.S.). There-
The units and definition of molecular weight or molar mass, fore, for a hydrocarbon or a petroleum fraction, the specific
M, was discussed in Section 1.7.8. The molecular weight of a
gravity is defined as
pure compound is determined from its chemical formula and
the atomic weights of its elements. The atomic weights of the (2.2) SG (60~176 -- density of liquid at 60~ in g/cm 3
elements found in a petroleum fluid are C = 12.011, H = 1.008, 0.999 g/cm 3
S = 32.065, O = 16.0, and N = 14.01, as given by the IUPAC
standard [ 1]. As an example, the molecular weight of methane Water density at 60~ is 0.999 or almost 1 g/cm3; therefore,
(CH4) is calculated as 12.011 + 4 x 1.008 = 16.043 kg/kmol or values of specific gravities are nearly the same as the density
16.043 g/tool (0.01604 kg/mol) or 16.043 lb/lbmol. Molecular of liquid at 15.5~ (289 K) in g/cm 3. The Society of Petroleum
weight is one of the characterization parameters for hydro- Engineers usually uses y for the specific gravity and in some
carbons. references it is designated by S. However, in this book SG de-
notes the specific gravity. Since most of hydrocarbons found
in reservoir fluids have densities less than that of water, spe-
2.1.2 Boiling Point cific gravities of hydrocarbons are generally less than 1. Spe-
The boiling point of a pure compound at a given pressure cific gravity defined by Eq. (2.2) is slightly different from the
is the temperature at which vapor and liquid exist together specific gravity defined in the SI system as the ratio of the den-
at equilibrium. If the pressure is 1 atm, the boiling point is sity of hydrocarbon at 15~ to that of water at 4~ designated
called the normal boiling point. However, usually the term b y d 15. Note that density of water at 4~ is exactly I g/cm 3 and
boiling point, Tb, is used instead of normal boiling point and therefore d15 is equal to the density of hydrocarbon at 15~
for other pressures the term saturation temperature is used. in g/cm 3. The relation between these two specific gravities is
In some cases, especially for heavy hydrocarbons in which approximately given as follows:
thermal cracking may occur at high temperatures, boiling
(2.3) SG = 1.001 d4is
points at pressures other than atmospheric is specified. Boil-
ing points of heavy hydrocarbons are usually measured at 1, In this book specific gravity refers to SG at 60~176 (15.5~
10, or 50 m m Hg. The conversion of boiling point from low In the early years of the petroleum industry, the American
32 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
Petroleum Institute (API) defined the API gravity (degrees at 20~ and 1 atm. The instrument to measure the refractive
API) to quantify the quality of petroleum products and crude index is called a refractometer and is discussed in Chapter 3.
oils. The API gravity is defined as [2] In some references the values of refractive index are reported
at 25~ however, in this book the refractive index at 20~ and
141.5
(2.4) API gravity - SG (at 60~ - 131.5 1 arm is used as a characterization parameter for hydrocar-
bons and petroleum fractions. As is shown in this chapter and
Degrees API was derived from the degrees Baum6 in which Chapter 3, refractive index is a very useful characterization
it is defined in terms of specific gravity similar to Eq. (2.4) parameter for pure hydrocarbons and petroleum fractions,
except numerical values of 140 and 130 were used instead especially in relation with molecular type composition. Val-
of 141.5 and 131.5, respectively. Liquid hydrocarbons with ues of n vary from about 1.3 for propane to 1.6 for some
lower specific gravities have higher API gravity. Aromatic hy- aromatics. Aromatic hydrocarbons have generally higher
drocarbons have higher specific gravity (lower API gravity) n values than paraffinic compounds as shown in Table 2.1.
than do paraffinic hydrocarbons. For example, benzene has
SG of 0.8832 (API of 28.72) while n-hexane with the same car-
bon number has SG of 0.6651 (API gravity of 81.25). A liquid 2.1.5 Critical Constants (T~, Pc, V~, Zc)
with SG of i has API gravity of 10. Once Eq. (2.4) is reversed it The critical point is a point on the pressure-volume-
can be used to calculate specific gravity from the API gravity. temperature diagram where the saturated liquid and satu-
141.5 rated vapor are identical and indistinguishable. The temper-
(2.5) SG = ature, pressure, and volume of a pure substance at the critical
API gravity + 131.5
point are called critical temperature (To), critical pressure (Pc),
The definition of specific gravity for gases is somewhat dif- and critical volume (Vc), respectively. In other words, the crit-
ferent. It is defined as relative density of gas to density of air at ical temperature and pressure for a pure compound are the
standard conditions. In addition, density of gases is a strong highest temperature and pressure at which the vapor and liq-
function of pressure. Since at the standard conditions (15.5 ~C uid phase can coexist at equilibrium. In fact, for a pure com-
and 1 arm) the density of gases are estimated from the ideal pound at temperatures above the critical temperature, it is
gas law (see Chapter 5), the specific gravity of a gas is pro- impossible to liquefy a vapor no matter how high the pres-
portional to the ratio of molecular weight of gas (Mg) to the sure is. A fluid whose temperature and pressure are above
molecular weight of air (28.97). the critical point is called supercritical fluid. For pure com-
pounds, critical temperature and pressure are also called true
(2.6) SGg-- Mg critical temperature and true critical pressure. However, as
28.97
will be discussed in Chapter 3, pseudocritical properties are
Therefore, to obtain the specific gravity of a gas, only its defined for mixtures and petroleum fractions, which are dif-
molecular weight is needed. For a mixture, Mg can be deter- ferent from true critical properties. Pseudocritical properties
mined from the gas composition, as discussed in Chapter 3. are important in process calculations for the estimation of
thermophysical properties of mixtures.
2.1.4 R e f r a c t i v e Index The critical compressibility factor, Z~, is defined from To, Pc,
and Vc according to the general definition of compressibility
Refractive index or refractivity for a substance is defined as factor.
the ratio of velocity of light in a vacuum to the velocity of
light in the substance (fluid) and is a dimensionless quantity PcVc
(2.8) Zc =
shown by n: RTc
velocity of light in the vacuum where R is the universal gas constant. According to Eq. (2.8),
(2.7) n= Zc is dimensionless and Vc must be in terms of molar vol-
velocity of light in the substance
ume (i.e., cm3/mol) to be consistent with R values given in
In other words, when a light beam passes from one substance Section 1.7.24. Critical temperature, pressure, and volume
(air) to another (a liquid), it is bent or refracted because of ( Tc, Pc, Vc) are called the critical constants or critical properties.
the difference in speed between the two substances. In fact, Critical constants are important characteristics of pure com-
refractive index indicates the degree of this refraction. Refrac- pounds and mixtures and are used in corresponding states
tive index is a state function and depends on the temperature correlations and equations of state (EOS) to calculate PVT and
and pressure of a fluid. Since the velocity of light in a fluid is many other thermodynamic, physical, and transport proper-
less than the velocity of light in a vacuum, its value for a fluid ties. Further discussion on the critical point of a substance
is greater than unity. Liquids have higher values of refractive is given in Chapter 5. As was discussed in Section 1.3, the re-
index than that of gases. For gases the values of refractive sults of EOS calculations very much depend on the values of
index are very close to unity. critical properties used. Critical volume may be expressed in
All frequencies of electromagnetic radiation (light) travel terms of specific critical volume (i.e., ma/kg), molar critical
at the same speed in vacuum (2.998 x l0 s m/s); however, in a volume (i.e., ma/kinol), or critical density dc (i.e., kg/m 3) or
substance the velocity of light depends on the nature of the critical molar density (i.e., kmol/m3). Critical density is re-
substance (molecular structure) as well as the frequency of lated to the critical molar volume as
the light. For this reason, standard values of refractive index
must be measured at a standard frequency. Usually the refrac- M
(2.9) d~ = - -
tive index of hydrocarbons is measured by the sodium D line vo
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 33
Experimental values of critical properties have been re- the saturation temperature becomes the normal boiling point.
ported for a large number of pure substances. However, for Vapor pressure increases with temperature and the highest
hydrocarbon compounds, because of thermal cracking that value of vapor pressure for a substance is its critical pressure
occurs at higher temperatures, critical properties have been (Pc) in which the corresponding temperature is the critical
measured up to C18 [2]. Recently some data on critical proper- temperature (To). When a liquid is open to the atmosphere at
ties of n-alkanes from C19 to C36 have been reported [3]. How- a temperature T in which the vapor pressure of liquid is pvap,
ever, such data have not yet been universally confirmed and vol% of the compound vapors in the air is
they are not included in major data sources. Reported data
on critical properties of such heavy compounds are generally (2.11) vol% = 100 • \ Pa ]
predicted values and vary from one source to another. For ex-
ample, the API-TDB [2] reports values of 768 K and 11.6 bar where Pa is the atmospheric pressure. Derivation of Eq. (2.11)
for the critical temperature and pressure of n-eicosane, while is based on the fact that vapor pressure is equivalent to partial
these values are reported as 767 K and 11.1 bar by Poling pressure (mole fraction • total pressure) and in gases under
et al. [4]. Generally, as boiling point increases (toward heav- low-pressure conditions, mole fraction and volume fraction
ier compounds), critical temperature increases while critical are the same. At sea level, where P~ = 1 atm, calculation of
pressure decreases. As shown in Section 2.2, aromatics have vol% of hydrocarbon vapor in the air from Eq. (2.11) is simply
higher Tc and Pc relative to those of paraffinic compounds 100 pvap, if pwp is in atm.
with the same carbon atoms. Vapor pressure is a very important thermodynamic prop-
erty of any substance and it is a measure of the volatility of
2.1.6 Acentric Factor a fluid. Compounds with a higher tendency to vaporize have
higher vapor pressures. More volatile compounds are those
Acentric factor is a parameter that was originally defined by that have lower boiling points and are called light compounds.
Pitzer to improve accuracy of corresponding state correla- For example, propane (Ca) has boiling point less than that of
tions for heavier and more complex compounds [5, 6]. Acen- n-butane (nCa) and as a result it is more volatile. At a fixed
tric factor is a defined parameter and not a measurable quan- temperature, vapor pressure of propane is higher than that
tity. It is a dimensionless parameter represented by w and is of butane. In this case, propane is called the light compound
defined as (more volatile) and butane the heavy compound. Generally,
(2.10) w = - log10 (PIvap) -- 1.0 more volatile compounds have higher critical pressure and
lower critical temperature, and lower density and lower boil-
where ing point than those of less volatile (heavier) compounds, al-
pvap = reduced vapor pressure, pvap/Pc, dimensionless
though this is not true for the case of some isomeric com-
pvap = vapor pressure at T = 0.7 Tc (reduced temperature
pounds. Vapor pressure is a useful parameter in calculations
of 0.7), bar related to hydrocarbon losses and flammability of hydrocar-
Pc -- critical pressure, bar bon vapor in the air (through Eq. 2.11). More volatile com-
T = absolute temperature, K pounds are more ignitable than heavier compounds. For ex-
Tc = critical temperature, K ample, n-butane is added to gasoline to improve its ignition
Acentric factor is defined in a way that for simple fluids such characteristics. Low-vapor-pressure compounds reduce evap-
as argon and xenon it is zero and its value increases as the oration losses and chance of vapor lock. Therefore, for a fuel
size and shape of molecule changes. For methane w -- 0.001 there should be a compromise between low and high vapor
and for decane it is 0.489. Values reported for acentric fac- pressure. However, as will be seen in Chapter 6, one of the
tor of pure compounds are calculated based on Eq. (2.10), major applications of vapor pressure is in calculation of equi-
which depends on the values of vapor pressure. For this rea- librium ratios (Ki values) for phase equilibrium calculations.
son values reported for the acentric factor of a compound may Methods of calculation of vapor pressure are given in detail in
slightly vary from one source to another depending on the re- Chapter 7. For pure hydrocarbons, values of vapor pressure at
lation used to estimate the vapor pressure. In addition, since the reference temperature of 100~ (38~ are provided by the
calculation of the acentric factor requires values of critical API [2] and are given in Section 2.2. For petroleum fractions,
temperature and pressure, reported values for w also depend as will be discussed in Chapter 3, method of Reid is used to
on the values of Tc and Pc used. measure vapor pressure at 100~ Reid vapor pressure (RVP)
is measured by the ASTM test method D 323 and it is approx-
2.1.7 Vapor Pressure imately equivalent to vapor pressure at 100~ (38~ RVP is
a major characteristic of gasoline fuel and its prediction is
In a closed container, the vapor pressure of a pure compound discussed in Chapter 3.
is the force exerted per unit area of walls by the vaporized
portion of the liquid. Vapor pressure, pvap, can also be de-
fined as a pressure at which vapor and liquid phases of a 2.1.8 Kinematic Viscosity
pure substance are in equilibrium with each other. The vapor
Kinematic viscosity is defined as the ratio of absolute (dy-
pressure is also called saturation pressure, psat, and the cor-
namic) viscosity/z to absolute density p at the same temper-
responding temperature is called saturation temperature. In
ature in the following form:
an open air under atmospheric pressure, a liquid at any tem-
perature below its boiling point has its own vapor pressure (2.12) v = ~-
that is less than 1 atm. When vapor pressure reaches 1 atm, P
34 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
As discussed in Section 1.7.18, kinematic viscosity is ex- flash point greater than 80~ (ASTM D 92 or ISO 2592 test
pressed in cSt, SUS, and SFS units. Values of kinematic vis- methods). Flash point should not be mistaken with fire point,
cosity for pure liquid hydrocarbons are usually measured and which is defined as the minimum temperature at which the
reported at two reference temperatures of 38~ (100~ and hydrocarbon will continue to burn for at least 5 s after being
99~ (210~ in cSt. However, other reference temperatures ignited by a flame.
of 40~ (104~ 50~ (122~ and 60~ (140~ are also used
to report kinematic viscosities of petroleum fractions. Liq- 2.1.11 Autoignition Temperature
uid viscosity decreases with an increase in temperature (see
Section 2.7). Kinematic viscosity, as it is shown in Chapter 3, This is the minimum temperature at which hydrocarbon va-
is a useful characterization parameter, especially for heavy por when mixed with air can spontaneously ignite without
fractions in which the boiling point may not be available. the presence of any external source. Values of autoignition
temperature are generally higher than flash point, as given
in Table 2.2 for some pure hydrocarbons. Values of autoigni-
2.1.9 Freezing and Melting Points tion temperature for oils obtained from mineral sources are
Petroleum and most petroleum products are in the form of a in the range of 150-320~ (300-500~ for gasoline it is about
liquid or gas at ambient temperatures. However, for oils con- 350~ (660~ and for alcohol is about 500~ (930~ [7].
taining heavy compounds such as waxes or asphaltinic oils, With an increase in pressure the autoignition temperature
problems may arise from solidification, which cause the oil decreases. This is particularly important from a safety point
to lose its fluidity characteristics. For this reason knowledge of view when hydrocarbons are compressed.
of the freezing point is important and it is one of the ma-
jor specifications of jet fuels and kerosenes. For a pure com-
2.1.12 F l a m m a b i l i t y R a n g e
pound the freezing point is the temperature at which liquid
solidifies at 1 atm pressure. Similarly the melting point, TM, To have a combustion, three elements are required: fuel (hy-
is the temperature that a solid substance liquefies at 1 atm. drocarbon vapor), oxygen (i.e., air), and a spark to initiate the
A pure substance has the same freezing and melting points; combustion. One important parameter to have a good com-
however, for petroleum mixtures, there are ranges of melting bustion is the ratio of air to hydrocarbon fuel. The combustion
and freezing points versus percent of the mixture melted or does not occur if there is too much air (little fuel) or too lit-
frozen. For a mixture, the initial melting point is close to the tle air (too much fuel). This suggests that combustion occurs
melting point of the lightest compound in the mixture, while when hydrocarbon concentration in the air is within a certain
the initial freezing point is close to the freezing point (or melt- range. This range is called flammability range and is usually
ing point) of the heaviest compound in the mixture. Since the expressed in terms of lower and upper volume percent in the
melting point increases with molecular weight, for petroleum mixture of hydrocarbon vapor and air. The actual volume per-
mixtures the initial freezing point is greater than the initial cent of hydrocarbon vapor in the air may be calculated from
melting point. For petroleum mixtures an equivalent term of Eq. (2.1 i) using vapor pressure of the hydrocarbon. If the cal-
pour point instead of initial melting point is defined, which culated vol% of hydrocarbon in the air is within the flamma-
will be discussed in Chapter 3. Melting point is an important bility range then the mixture is flammable by a spark or flame.
characteristic parameter for petroleum and paraffinic waxes.
2.1.13 Octane Number
2.1.10 Flash Point Octane number is a parameter defined to characterize an-
Flash point, TF, for a hydrocarbon or a fuel is the minimum tiknock characteristic of a fuel (gasoline) for spark ignition
temperature at which vapor pressure of the hydrocarbon is engines. Octane number is a measure of fuel's ability to re-
sufficient to produce the vapor needed for spontaneous igni- sist auto-ignition during compression and prior to ignition.
tion of the hydrocarbon with the air with the presence of an Higher octane number fuels have better engine performance.
external source, i.e., spark or flame. From this definition, it is The octane number of a fuel is measured based on two ref-
clear that hydrocarbons with higher vapor pressures (lighter erence hydrocarbons of n-heptane with an assigned octane
compounds) have lower flash points. Generally flash point number of zero and isooctane (2,2,4-trimethylpentane) with
increases with an increase in boiling point. Flash point is assigned octane number of 100. A mixture of 70 vol% isooc-
an important parameter for safety considerations, especially tane and 30 vol% n-heptane has an octane number of 70.
during storage and transportation of volatile petroleum prod- There are two methods of measuring octane number of a fuel
ucts (i.e., LPG, light naphtha, gasoline) in a high-temperature in the laboratory. The methods are known as motor octane
environment. The surrounding temperature around a storage number (MON) and research octane number (RON). The MON
tank should always be less than the flash point of the fuel is indicative of high-speed performance (900 rpm) and is mea-
to avoid possibility of ignition. Flash point is used as an in- sured under heavy road conditions (ASTM D 357). The RON
dication of the fire and explosion potential of a petroleum is indicative of normal road performance under low engine
product. Estimation of the flash point of petroleum fractions speed (600 rpm) city driving conditions (ASTM D 908). The
is discussed in Chapter 3, and data for flash points of some third type of octane number is defined as posted octane num-
pure hydrocarbons are given in Table 2.2. These data were ob- ber (PON), which is the arithmetic average of the MON and
tained using the closed cup apparatus as described in ASTM RON [PON = (MON + RON)/2]. Generally isoparaffins have
D 93 (ISO 2719) test method. There is another method of higher octane number than do normal paraffins. Naphthenes
measuring flash point known as open cup for those oils with have relatively higher octane number than do corresponding
2. C H A R A C T E R I Z A T I O N AND PROPERTIES OF PURE HYDROCARBONS 35
paraffins and aromatics have very high octane numbers. The Tb is the mean average boiling point (also see Chapter 3). The
octane number of a fuel can be improved by adding tetra- purpose of definition of this factor was to classify the type
ethyl-lead (TEL) or methyl-tertiary-butyl-ether (MTBE). Use of hydrocarbons in petroleum mixtures. The naphthenic hy-
of lead (Pb) to improve octane number of fuels is limited in drocarbons have Kw values between paraffinic and aromatic
many industrial countries. In these countries MTBE is used compounds. In general, aromatics have low Kw values while
for octane number improvement. However, there are prob- paraffins have high values. However, as will be discussed in
lems of groundwater contamination with MTBE. MTBE has Chapter 3 there is an overlap between values of Kw from dif-
MON and RON of 99 and 115, respectively [8]. Lead gener- ferent hydrocarbon groups. The Watson K was developed in
ally improves octane number of fuels better than MTBE. The 1930s by using data for the crude and products available in
addition of 0.15 g Pb/L to a fuel of RON around 92 can im- that time. Now the base petroleum stocks in general vary sig-
prove its octane number by 2-3 points. With 0.6 g Pb/L one nificantly from those of 1930s [10, 11]. However, because it
may improve the octane number by 10 points [8]. However, combines two characterization parameters of boiling point
as mentioned above, because of environmental hazards use and specific gravity it has been used extensively in the devel-
of lead is restricted in many North American and West Euro- opment of many physical properties for hydrocarbons and
pean countries. Values of the octane number measured with- petroleum fractions [2, 11, 12].
out any additives are called clear octane number. For pure
hydrocarbons values of clear MON and RON are given in Sec-
2.1.16 Refractivity Intercept
tion 2.2. Estimation of the octane number of fuels is discussed
in Chapter 3. Kurtz and Ward [13] showed that a plot of refractive index
against density for any homologous hydrocarbon group is
2.1.14 Aniline Point linear. For example, plot of refractive index of n-paraffins ver-
sus density (d20) in the carbon number range of C5-C45 is a
The aniline point for a hydrocarbon or a petroleum fraction is
straight line represented by equation n = 1.0335 + 0.516d20,
defined as the minimum temperature at which equal volumes
with R E value of 0.9998 (R2 = I, for an exact linear relation).
of liquid hydrocarbon and aniline are miscible. Aniline is an
Other hydrocarbon groups show similar performance with
aromatic compound with a structure of a benzene molecule
an exact linear relation between n and d. However, the inter-
where one atom of hydrogen is replaced by the -NH2 group cept for various groups varies and based on this observation
(C6Hs-NH2). The aniline point is important in characteriza-
they defined a characterization parameter called refractivity
tion of petroleum fractions and analysis of molecular type.
intercept, R~, in the following form:
As discussed in Chapter 3, the aniline point is also used as a
characterization parameter for the ignition quality of diesel d
(2.14) Ri = n - -
fuels. It is measured by the ASTM D 611 test method. Within 2
a hydrocarbon group, aniline point increases with molecu- where n and d are refractive index and density of liquid hy-
lar weight or carbon number, but for the same carbon num- drocarbon at the reference state of 20~ and I atm in which
ber it increases from aromatics to paraffinic hydrocarbons. density must be in g/cm 3. Ri is high for aromatics and low for
Aromatics have very low aniline points in comparison with naphthenic compounds, while paraffins have intermediate/~
paraffins, since aniline itself is an aromatic compound and it values.
has better miscibility with aromatic hydrocarbons. Generally,
oils with higher aniline points have lower aromatic content.
Values of the aniline point for pure hydrocarbons are given 2.1.17 Viscosity Gravity Constant
in Table 2.2, and its prediction for petroleum fractions is dis- Another parameter defined in the early years of petroleum
cussed in Chapter 3. characterization is the viscosity gravity c o n s t a n t (VGC). This
parameter is defined based on an empirical relation developed
2.1.15 Watson K between Saybolt viscosity (SUS) and specific gravity through
Since the early years of the petroleum industry it was desired a constant. VGC is defined at two reference temperatures of
to define a characterization parameter based on measurable 38~ (100~ and 99~ (210~ as [14]
parameters to classify petroleum and identify hydrocarbon 10SG - 1.0752 log10(V38 - 38)
molecular types. The Watson characterization factor denoted (2.15) VGC =
10 - logl0(V3s - 38)
by Kw is one of the oldest characterization factors originally
defined by Watson et al. of the Universal Oil Products (UOP) SG - 0.24 - 0.022 log10(V99 - 35.5)
in mid 1930s [9]. For this reason the parameter is sometimes (2.16) VGC =
0.755
called UOP characterization factor and is defined as
where
(1.8Tb) 1/3
(2.13) Kw = V38 = viscosity at 38~ (100~ in SUS (Saybolt Universal
SG Seconds)
where V99 = Saylbolt viscosity (SUS) at 99~ (210~
Tb = normal boiling point K Conversion factors between cSt and SUS are given in Sec-
SG = specific gravity at 15.5~ tion 1.7.18. Equations (2.15) and (2.16) do not give identical
In the original definition of Kw, boiling point is in degrees values for a given compound but calculated values are close to
Rankine and for this reason the conversion factor of 1.8 each other, except for very low viscosity oils. Equation (2.16)
is used to have Tb in the SI unit. For petroleum fractions is recommended only when viscosity at 38~ (100~ is not
36 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
available. VGC varies for paraffinic hydrocarbons from 0.74 to general formula of CnHan+2, alkylcyclopentanes or alkylcy-
0.75, for naphthenic from 0.89 to 0.94, and for aromatics from clohexanes (naphthenes) have formula of C, H2,, and alkyl-
0.95 to 1.13 [15]. In Chapter 3, VGC along with other param- benzenes (aromatics) have formula of CnH2n-6 (n ~ 6). This
eters has been used to estimate the composition of petroleum shows that at the same carbon number, the atomic ratio of
fractions. Values of VGC for some hydrocarbons are given in number of carbon (C) atoms to number of hydrogen (H)
Table 2.3. The main limitation in use of VGC is that it cannot atoms increases from paraffins to naphthenes and aromat-
be defined for compounds or fractions with viscosities less ics. For example, n-hexane (C6H14), cyclohexane (C6H12), and
than 38 SUS (--~3.6 cSt) at 38~ A graphical method to esti- benzene (C6H6)from three different hydrocarbon groups all
mate VGC of petroleum fractions is presented in Chapter 3. have six carbon atoms, but have different CH atomic ratios
ASTM D 2501 suggests calculation of VGC using specific of 6/14, 6/12, and 6/6, respectively. If CH atomic ratio is mul-
gravity and viscosity in mm2/s (cSt) at 40~ v40 in the follow- tiplied by the ratio of atomic weights of carbon (12.011) to
ing form: hydrogen (1.008), then CH weight ratio is obtained. For ex-
ample, for n-hexane, the CH weight ratio is calculated as
SG - 0.0664 - 0.1154 logl0(v40 - 5.5) (6/14)x(12.011/1.008) = 5.107, This number for benzene is
(2.17) VGC--
0.94 - 0.109 logl0(v40 - 5.5) 11.92. Therefore, CH weight ratio is a parameter that is ca-
Values of VGC calculated from Eq. (2.17) are usually very pable of characterizing the hydrocarbon type. In addition,
close to values obtained from Eq. (2.15). If viscosity at 40~ within the same hydrocarbon group, the CH value changes
is available, use of Eq. (2.17) is recommended for calculation from low to high carbon number. For example, methane has
of VGC. Another relation to calculate VGC in metric units was CH value of 2.98, while pentane has CH value of 4.96. For
proposed by Kurtz et al. [16] in terms of kinematic viscosity extremely large molecules (M ~ o0), the CH value of all hy-
and density at 20~ which is also reported in other sources drocarbons regardless of their molecular type approaches the
[17]. limiting value of 5.96. This parameter is used in Section 2.3
to estimate hydrocarbon properties, and in Chapter 3 it is
(2.18) VGC= d - 0.1384 log10(v20-20) +0.0579 used to estimate the composition of petroleum fractions. In
0.1526[7.14 - logl0(V20 - 20)] some references HC atomic ratio is used as the character-
izing parameter. According to the definition, the CH weight
in which d is density at 20~ and 1 atm in g/cm 3 and v20 is the
ratio and HC atomic ratio are inversely proportional. The
kinematic viscosity at 20~ in cSt. In this method viscosity
limiting value of HC atomic ratio for all hydrocarbon types
of oil at 20~ must be greater than 20 cSt. However, when
is 2.
there is a choice Eq. (2.15) should be used for the procedures
Another use of CH weight ratio is to determine the quality of
described in Chapter 3.
a fossil-type fuel. Quality and the value of a fuel is determined
from its heat of combustion and heating value. Heating value
Example 2.1--API RP-42 [ 18] reports viscosity of some heavy of a fuel is the amount of heat generated by complete com-
hydrocarbons, 1,1-Di-(alphadecalyl)hendecane (C31H56) is a bustion of 1 unit mass of the fuel. For example, n-hexane has
naphthenic compound with molecular weight of 428.8 and the heating value of 44734 kJ/kg (19232 Btu/lb) and benzene
specific gravity of 0.9451. The kinematic viscosity at 38~ has the heating value of 40142 kJ/kg (17258 Btu/lb). Calcula-
(100~ is 20.25 cSt. Calculate the viscosity gravity constant tion of heating values are discussed in Chapter 7. From this
for this compound. analysis it is clear that as CH value increases the heating value
decreases. Hydrogen (H2), which has a CH value of zero, has a
Solution--Using Eq. (1.17), the viscosity is converted from heating value more than that of methane (CH4) and methane
cSt to SUS: V3s = 99.5 SUS. Substituting values of V3s and has a heating value more than that of any other hydrocarbon.
SG = 0.9451 into Eq. (2.15) gives VGC = 0.917. The VGC may Heavy aromatic hydrocarbons that have high CH values have
be calculated from Eq. (2.17) with direct substitution of vis- lower heating values. In general, by moving toward lower CH
cosity in the cSt unit. Assuming there is a slight change in value fuel, not only do we have better heating value but also
viscosity from 38 to 40~ the same value of viscosity at 38~ better and cleaner combustion of the fuel. It is for this reason
is used for v40. Thus v40 ~ 20.25 cSt (mm2/s) and Eq. (2.17) that the use of natural gas is preferable to any other type of
gives VGC = 0.915. The small difference between calculated fuel, and hydrogen is an example of a perfect fuel with zero
values of VGC because in Eq. (2.17) viscosity at 40~ must CH weight ratio (CH = 0), while black carbon is an example
be used, which is less than the viscosity at 38~ Calculated of the worst fuel with a CH value of infinity. Values of CH for
VGC is within the range of 0.89-0.94 and thus the hydrocar- pure hydrocarbons are given in Section 2.2 and its estimation
bon must be a naphthenic compound (also see Fig. 3.22 in methods are given in Section 2.6.3.
Chapter 3). #
1. API: Technical Data Book~Petroleum Refining [2]. The first sources for physical property data (http://tigger.uic.edu/
chapter of API-TDB compiles basic properties of more than ~mansoori/Thermodynamic.Data.and.Property-html/).
400 pure hydrocarbons and some nonhydrocarbons that
are important in petroleum refining. For some compounds In m a n y occasions different sources provide different values
where experimental data are not available, predicted values for a particular property depending on the original source
from the methods recommended by the API are given. of data. Calculated properties such as critical constants and
2. DIPPR: Design Institute for Physical and Property Data acentric factor for compounds heavier than C18 should be
[ 19]. The project initially supported by the AIChE began in taken with care as in different sources different methods have
early 1980s and gives various physical properties for both been used to predict these parameters.
hydrocarbon and nonhydrocarbon compounds important
in the industry. A computerized version of this data bank is 2.2.2 Properties o f Selected Pure Compounds
provided by EPCON [20].
3. TRC Thermodynamic Tables--Hydrocarbons [21]. The The basic properties of pure hydrocarbons from different
Thermodynamic Research Center (formerly at the Texas groups that will be used in the predictive methods presented
A&M University) currently at the National Institute of Stan- in the following chapters are tabulated in Tables 2.1 and 2.2.
dards and Technology (NIST) at Boulder, CO, (http://www. The basic properties of M, TM, Tb, SG, d20, n20, To, Pc, Vc, Zc,
trc.nist.gov/) in conjunction with the API Research Project and ~0are presented in Table 2.1. Secondary properties of kine-
44 [22] has regularly published physical and basic proper- matic viscosity, API gravity, Kw, vapor pressure, aniline point,
ties of large number of pure hydrocarbons. flash and autoignition points, flammability range, and oc-
4. API Research Project 44 [22]. This project sponsored by the tane number are given in Table 2.2. Compounds selected are
API was conducted at Texas A&M University and provides mainly hydrocarbons from paraffins, naphthenes, and aro-
physical properties of selected hydrocarbons. matics that constitute crude oil and its products. However,
5. API Research Project 42 [18]. This data compilation com- some olefinic and nonhydrocarbons found with petroleum
pleted in the 1960s provides experimental data on den- fluids are also included. Most of the compounds are from ho-
sity, refractive index, viscosity, and vapor pressure for mologous hydrocarbon groups that are used as model com-
more than 300 hydrocarbons with carbon number greater pounds for characterization of petroleum fractions discussed
than Cu. in Chapter 3. The properties tabulated are the basic prop-
6. Dortmund Data Bank (DDB) [23]. This project on physical erties needed in characterization techniques and thermody-
properties has been conducted at the University of Olden- namic correlations for physical properties of petroleum frac-
burg in Germany. DDB contains experimental data from tions. Although there are separate chapters for estimation of
open literature on various thermodynamic properties of density, viscosity, or vapor pressure, these properties at some
pure compounds and some defined mixtures. Data have reference temperatures are provided because of their use in
been programmed in a computer software convenient for the characterization methods given in Chapter 3 and 4. Other
extracting data. Majority of data are on thermodynamic physical properties such as heat capacity or transport prop-
properties, such as vapor-liquid equilibrium (VLE), activ- erties are given in corresponding chapters where the predic-
ity coefficients, and excess properties. However, data on tive methods are discussed. Data for more than 100 selected
viscosity, density, vapor pressure, thermal conductivity, and compounds are presented in this section and are limited to
surface tension have also been complied as mentioned in C22 mainly due to the lack of sufficient experimental data for
their Web site. Unfortunately they have not compiled char- heavier compounds. Data presented in Tables 2.1 and 2.2 are
acteristic data on hydrocarbons and petroleum fractions taken from the API-TDB [2, 22]. Standard methods of mea-
important in the petroleum industry. Also the data on trans- surement of these properties are presented in Chapter 3.
port properties are mainly for pure compounds at atmo-
spheric pressures. Example 2.2--Assume large amount of toluene is poured on
7. The fourth and fifth editions of The Properties of Gases the ground in an open environment at which the temperature
and Liquids [4] also provide various properties for more is 38~ (100~ Determine if the area surrounding the liquid
than 400 pure compounds (hydrocarbons and nonhydro- surface is within the flammability range.
carbons). However, data in this book have been mainly
taken from the TRC Tables [21 ]. Solution--From Table 2.2, the flammability range is 1.2-7.1
8. There are also some free online sources that one may use to vol% of toluene vapor. From this table, the vapor pressure of
obtain some physical property data. The best example is the toluene at 38~ is 0.071 bar (0.07 atm). Substituting this vapor
one provided by NIST (http://webbook.nist.gov). Various pressure value in Eq. (2.11) gives the value of vol% = 100 x
universities and researchers have also developed special on- 0.07/1.0 = 7% of toluene in the air mixture. This number is
line sources for free access to some data on physical proper- within the flammability range (1.2 < 7 < 7.1) and therefore
ties. For example the Center for Research in Computational the surrounding air is combustible. r
Thermochemistry (CRCT) of Ecole Ploytechnique Mon-
treal provides online calculational software at http://www.
2.2.3 Additional Data on Properties
crct.polymfl.caJfact/index.php/. The Center for Applied
o f H e a v y Hydrocarbons
Thermodynamic Studies (CATS) at the University of Idaho
also provides softwares for property calculations at its web- Some data on density, refractive index, and viscosity of some
site (http://www.webpages.uidaho.edu/~cats/). G. A. Man- heavy hydrocarbons are given in Table 2.3. These data are
soori in his personal Web site also provides some online taken from API RP 42 [18]. Values of P~ and VGC in the table
TABLE 2.1--Basic physical properties of some selected compounds.
TM, Tb, SG, d2o, TC, Pc, VC,
Compound Formula Nc M ~ ~ at 60~ g/ore3 n20 ~ bar cm3/mol Zc
Paraffins
Methane CH4 1 16.0 -182.5 -161.5 0.2999 ... -82.59 45.99 98.65 0.2864 0.0115
Ethane C2H6 2 30.1 -182.8 -88.6 0.3554 0.33"86 ... 32.17 48.72 145.48 0.2792 0.0995
Propane C3H8 3 44,1 -187.7 -42.0 0.5063 0.4989 ... 96.68 42.48 200.14 0.2765 0.1523
n-Butane C4H10 4 58.1 -138.3 -0.5 0.5849 0.5791 1.3326 151.97 37.96 255.09 0.2740 0.2002
n-Pentane C5H12 5 72.2 -129.7 36.1 0.6317 0.6260 1.3575 196.55 33.70 313.05 0.2702 0.2515
n-Hexane C6H14 6 86.2 -95.3 68.7 0.6651 0.6605 1.3749 234.45 30.25 371.22 0.2661 0.3013
n-Heptane C7H16 7 100.2 -90.6 98.4 0.6902 0.6857 1.3876 267.05 27.40 427.88 0.2610 0.3495
n-Octane C8H18 8 114.2 -56.8 125.7 0.7073 0.7031 1.3974 295.55 24.90 486.35 0.2561 0.3996
n-Nonane C9H20 9 128.3 -53.5 150.8 0.7220 0.7180 1.4054 321.45 22.90 543.67 0,2519 0.4435
n-Decane C10H22 10 142,3 -29.6 174.2 0.7342 0.7302 1.4119 344.55 21.10 599.58 0.2463 0.4923
n-Undecane Cl1H24 11 156.3 -25.6 195.9 0.7439 0.7400 1.4151 365.85 19.50 658.69 0,2418 0.5303
n-Dodecane C12H26 12 170,3 -9.6 216.3 0.7524 0.7485 1.4195 384.85 18.20 715.67 0.2381 0.5764
n-Tridecane C13H28 13 184.4 -5.4 235.5 0.7611 0.7571 1.4235 401.85 16.80 774.60 0.2319 0.6174
n-Tetradecane C14H30 14 198.4 5.9 253.6 0.7665 0.7627 1.4269 419.85 15.70 829.82 0.2261 0.6430
n-Pentadecane C15H32 15 212.4 9.9 270.7 0.7717 0.7680 1.4298 434.85 14.80 888.49 0.2234 0.6863
t~ n-Hexadecane C16H34 16 226.4 18.2 286.9 0.7730 0.7729 1.4325 449.85 14.00 944.33 0.2199 0.7174
n-Heptadecane C17H36 17 240.5 22.0 302.2 0.7752 0.7765 1.4348 462.85 13.40 999.83 0.2189 0.7697
n-Octadecane C18H38 18 254.5 28.2 316.7 0.7841 0.7805 1.4369 473.85 12.70 1059.74 0,2167 0.8114
n-Nonadecane C19H40 19 268.5 31.9 329.9 0.7880 0.7844 1.4388 484.85 12.10 1119.82 0.2150 0.8522
n-Eicosane C20H42 20 282.6 36.4 343.8 0.7890 0.7871 1.4405 494.85 11.60 1169.50 0.2125 0.9069
n-Heneicosane C21H44 21 296.6 40.2 356.5 0.7954 0.7906 1.4440 504.85 11.10 1229.41 0.2110 0.9420
n-Docosane C22H46 22 310.6 44.0 368.6 0.7981 0.7929 1.4454 513.85 10,60 1289.49 0.2089 0.9722
isobutane (2-Methylpropane) C4H10 4 58.1 -159.6 -11.7 0.5644 0.5584 134.99 36.48 262.71 0.2824 0.1808
isopentane (2-Methylbutane) CsHI2 5 72.2 -159.9 27.8 0,6265 0.6213 1.3"5"37 187.28 33.81 305.84 0.2701 0.2275
2-Methylpentane C6H14 6 86.2 -95.3 60.3 0.6577 0.6529 1.3715 224.35 30.05 371.22 0.2661 0.2774
2-Methylhexane C7H16 7 100.2 -118.3 90.1 0.6822 0.6778 1,3849 257.22 27.34 421.00 0.2610 0.3277
2-Methylheptane C8H18 8 114.2 -109.0 117.7 0.7029 0.6987 1.3949 286.49 24.84 487.78 0.2604 0.3772
2,2,4-Trimethylpentane (isooctane) C8H18 8 114.2 -107.4 99.2 0.6988 0.6945 1.3915 270.81 25.68 467.81 0.2656 0.3022
2-Methyloctane C9H20 9 128.3 -80.4 143.3 0.7176 0.7134 1.4031 313.60 22.90 541.27 0.2541 0.4212
2-Methylnonane C10H22 10 142.3 -74.7 167.0 0.7307 0.7266 1.4100 336.85 21.20 582.70 0.2436 0.4723
Olefins
Ethene (Ethylene) C2H4 2 28.1 -169.2 -103.7 0.1388 ... 9.19 50.40 131.00 0.2813 0.0865
Propene (Propylene) C3H6 3 42.1 -185.3 -47.7 0.5192 0.5'1"11 ... 92.42 46.65 188.36 0,2891 0.1398
1-Butene C4H8 4 56.1 -185.4 -6.3 0.6001 0.5938 146.80 40.43 239.24 0.2770 0.1905
1-Pentene CsHt0 5 70,1 -165.2 30.0 0.6456 0,6402 1.3715 191.63 35.13 295.10 0.2683 0.2312
1-Hexene C6H12 6 84.2 -139.8 63.5 0.6790 0.6740 1.3879 230.88 31.40 354.13 0.2654 0.2804
1-Heptene C7H14 7 98.2 -118.9 93.6 0,7015 0.6971 1.3998 264.14 28.30 413.15 0.2617 0.3310
1-Octene C8H16 8 112.2 -101.7 121.3 0.7181 0.7140 1.4087 293.50 25.68 460.26 0.2509 0.3764
1-Nonene C9H18 9 126.2 -81.4 146.9 0.7330 0.7290 ... 320.10 23.30 528.04 0.2494 0.4171
1-Decene C10H20 10 140.3 -66.3 170.6 0.7450 0.7410 ... 343.25 22.18 584.08 0.2528 0.4800
Acetylene C2H2 2 26.0 -80.8 -83.8 ... 0.4001 ... 35.17 61.39 112.97 0.2706 0.1873
1,3-Butadiene C4H6 4 54.1 -108.9 -4.4 ... 0.6219 ... 152.02 42.77 220.49 0.2668
Naphthenes
Cyclopentane CsH10 5 70.1 -93.8 49.3 0.7502 0.7456 1.4065 238.61 45.02 257.89 0.2729 0.1959
Methylcyclopentane C6H12 6 84.2 -142.4 71.8 0.7540 0.7491 1.4097 259.64 37.85 318.92 0.2725 0.2302
Ethylcyclopentane C7H14 7 98.2 -138.4 103.5 0.7712 0.7667 1.4198 296.37 33.98 375.14 0.2692 0.2716
Propylcyclopentane C8H16 8 112.2 -117.3 131.0 0.7811 0.7768 1.4263 322.85 30.20 428.03 0.2609 0.3266
n-Butylcyclopemane C9H18 9 126.2 -108.0 156.6 0.7893 0.7851 1.4316 347.85 27.20 483.11 0.2545 0.3719
n-Pentylcyclopentane C10H20 10 140.3 -83.0 180.5 0.7954 ... 1.4352 370.65 24.50 536.79 0.2457 0,4184
n-Hexylcyclopentane CllH22 11 154.3 -73.0 202.9 0.8006 ... 1.4386 390.95 22.20 592.40 0.2382 0.4646
n-Heptylcyclopentane C12H24 12 168.3 -53.0 223.9 0.8051 ... 1.4416 409.45 20.10 647.30 0.2293 0,5100
n-Octylcyclopentane C13H26 13 182.4 -44.0 243.5 0.8088 ... 1.4446 426.35 18.30 702.38 0.2210 0.5525
n-Nonylcyclopentane C14H28 14 196.4 -29.0 262.0 0.8121 ... 1.4467 441.75 16.70 757.64 0.2129 0,5956
n-Decylcyclopentane C15H30 15 210.4 -22.1 279.4 0.8149 ... 1.4486 455.95 15.30 811.76 0.2049 0,6314
n -Undecylcyclopentane C16H32 16 224.4 -10.0 295.8 0.8175 ... 1.4503 469.05 14.00 867,27 0.1968 0.6741
n-Dodecylcyclopentane C17H34 17 238.5 -5.0 311.2 0.8197 ... 1.4518 481.25 12.80 921.48 0.1881 0.7163
n-Tridecylcyclopentane C18H36 18 252.5 5.0 325.9 0.8217 ... 1.4531 492.45 11.80 977.26 0.1812 0.7582
n -Tetradecylcyclopentane CIgHa8 19 266.5 9.0 340.0 0.8235 ... 1.4543 502.85 10.90 1031,55 0.1743 0,7949
n-Pemadecylcyclopentane C20H40 20 280.5 17.0 353.0 0.8252 ... 1.4554 512.55 10.00 1087.59 0.1665 0.8395
n -Hexadecylcyclopentane C21H42 21 294,6 21.0 366.0 0.8267 ... 1.4564 521.55 9.20 1141,97 0.1590 0.8755
n-Heptadecylcyclopentane C22H44 22 308.6 27.0 377.0 0.8280 1.4573 538.01 11.91 1198.28 0.2115 0.9060
Cyclohexane C6H12 6 84.2 6.5 80.7 0.7823 0.8021 1.4262 280.43 40.73 307.89 0.2725 0.2096
Methylcyclohexane C7H14 7 98.2 -126.6 100.9 0,7748 0.7702 1.4231 299.04 34.71 367,79 0.2684 0,2350
Ethylcyclohexane C8H16 8 112.2 -111.3 131.8 0.7926 0.7884 1.4330 336.00 30.40 430.13 0.2582 0.2455
Propylcyclohexane C9H18 9 126.2 -94.9 156.8 0.7981 0.7940 1.4371 366.00 28.07 476,81 0.2519 0.2595
n-Butylcyclohexane C10H20 10 140.3 -74.7 181.0 0.8033 0.7993 1.4408 393.85 25.70 534.16 0.2476 0.2743
t~ Aromatics
Benzene C6H 6 6 78.1 5.5 80.1 0.8832 0.8780 1.5011 289.01 48.98 258.94 0.2714 0,2100
Methylbenzene (Toluene) C7H 8 7 92.1 -95.0 110.6 0.8741 0.8685 1.4969 318.65 41.06 315.80 0.2635 0.2621
Ethylbenzene C8H10 8 106.2 -95.0 136.2 0.8737 0.8678 1.4959 344.05 36.06 373.81 0.2627 0.3026
Propylbenzene C9H12 9 120.2 -99.6 159.2 0.8683 0.8630 1.4920 365.23 32.00 439.71 0.2651 0.3447
n-Butylbenzene C10H14 10 134.2 -87.9 183.3 0.8660 0.8610 1.4898 387.40 28.87 496.89 0.2612 0.3938
n-Pentylbenzene CllH16 11 148.2 -75.0 205.5 0.8624 0.8583 1.4878 406.75 26.04 549.74 0.2532 0,4378
n-Hexylbenzene ClaHl8 12 162.3 --61.2 226.1 0.8622 0.8581 1.4864 424.85 23.80 592.64 0.2431 0.4790
n-Heptylbenzene C13H20 13 176.3 -48.0 246.1 0.8617 0.8576 1.4854 440.85 21.80 648,27 0.2381 0.5272
n-Octylbenzene C14H22 14 190.3 -36.0 264.4 0.8602 0.8562 1.4845 455.85 20.20 703.41 0.2344 0.5670
n-Nonytbenzene C15H24 15 204.4 -24.2 282.1 0.8596 0.8557 1.4838 467.85 18.95 752,70 0.2315 0.6331
n-Decylbenzene C16H26 16 218.4 -14.4 297.9 0.8590 0.8551 1.4832 479.85 17.70 812.55 0.2297 0.6797
n-Undecylbenzene C17H28 17 232.4 -5.2 313.3 0.8587 0.8548 1.4828 490.85 16.72 867.64 0.2284 0.7333
n-Dodecylbenzene C18H30 18 246.4 2.8 327.6 0.8595 0.8556 1.4824 506.85 15.60 923.08 0.2221 0.7333
n -Tridecylbenzene C19H32 19 260.5 10.0 341.3 0.8584 0.8545 1.4821 516.85 14.80 977.25 0.2202 0.7799
n-Tetradecylbenzene C20H34 20 274.5 16.0 354.0 0.8587 0.8553 1.4818 526.85 14.00 1029,88 0.2168 0.8130
n-Pentadecylbenzene C21H36 21 288.5 22.0 366.0 0.8587 0.8540 1.4815 535.85 13.30 1089.71 0.2155 0.8567
n-Hexadecylbenzene C22H38 22 302.5 27.0 378.0 0.8586 0.8541 1.4813 544.85 12.70 1140,80 0.2130 0.8996
o-xylene (o-dimethylbenzene) C8H10 8 106.2 -25.2 144.4 0.8849 0,8799 .,, 357.18 37.34 369.17 0.2630 0.3104
m-xylene CsH10 8 106.2 -47.9 139.1 0.8691 0.8643 1.4972 343.90 35.36 375,80 0.2590 0.3259
(Continued)
TABLE 2.1--(Continued)
TM, Tb, SG, d20, Tc, Pc, Vct
Compound Formula Nc M ~ ~ at 60~ g/cm 3 n20 ~ bar cm3/mol Zc
p-xylene C8H10 8 106.2 13.3 138.4 0.8654 0.8608 1.4958 343.08 35.11 379.11 0.2598 0.3215
isopropylbenzene C9H12 9 120.2 -96.0 152.4 0.8682 0.8632 1.4915 357.95 32.09 426.95 0.2611 0.3258
1,2,3-Trimethylbenezene C9H12 9 120.2 -25.4 176.1 0.8985 0.8942 1.5130 391.38 34.54 414.20 0.2590 0.3664
isobutylbenzene C10H14 10 134.2 -51.5 172.8 0.8577 0.8532 1.4865 377.00 30.40 455.83 0.2564 0.3797
o-Cymene C10H14 10 134.2 -71.5 178.2 0.8812 0.8765 1.5005 388.85 29.30 489.35 0.2605 0.3372
Cyclohexylbenzene C12H16 12 160.3 7.0 240.1 0.8617 0.9430 ... 470.85 28.80 529.25 0.2464 0.3783
Naphthalene C10H8 10 128.2 80.3 218.0 1.0281 1.0238 ... 475.20 40.51 412.89 0.2688 0.3022
1-Methylnaphthalene CllH10 11 142.2 -30.5 244.7 1.0242 1.1020 ... 498.89 36.60 465.17 0.2652 0.3478
Styrene C8H8 8 104.2 -30.6 145.2 0.9097 0.9049 ... 362.85 38.40 351.76 0.2555 0.2971
Diphenyl methane C13H12 13 168.2 25.2 264.3 1.0101 1.0244 ... 494.85 29.20 547.20 0.2502 0.4615
Nonhydrocarbons
O Water H20 ... 18.0 0.0 100.0 1.0000 0.9963 ... 373.98 220.55 55.91 0.2292 0.3449
C a r b o n dioxide CO2 1 44.0 -56.6 -78.5 0.8172 0.7745 ... 31.06 73.83 93.96 0.2743 0.2236
H y d r o g e n Sulfide H2S ... 34.1 -85.5 -60.4 0.8012 0.7901 ... 100.38 89.63 98.51 0.2843 0.0942
Nitrogen N2 ... 28.0 -210.0 -195.8 0.8094 . . . . . . -146.95 34.00 89.21 0.2891 0.0377
Oxygen 02 ... 32.0 -218.8 -183.0 1.1420 . . . . . . -118.57 50.43 73.32 0.2877 0.0222
Ammonia NH3 ... 17.0 -77.7 -33.4 0.6165 0.6093 ... 132.50 112.80 72.51 0.2425 0.2526
Carbon monoxide CO 1 28.0 -205.0 -191.5 . . . . . . . . . -140.23 34.99 94.43 0.2990 0.0482
Hydrogen H2 ... 2.0 -259.2 -252.8 . . . . . . . . . -239.96 13.13 64.28 0.3059 -0.2160
Air N2+O2 ... 29.0 -214.0 -194.5 -140.70 37.74 91.45 0.3134 0.0074
Methanol CHaOH 1 32.04 -97.7 64.7 0.7993 0.7"944 1.3265" 239.49 80.97 118.00 0.2240 0.5640
Ethanol C2HsOH 2 46.07 -114.1 78.29 0.7950 0.7904 1.3594" 240.77 61.48 167.00 0.2400 0.6452
Isobutanol C4H9OH 4 74.12 -108.0 107.66 0.8063 0.8015 1.3938" 274.63 43.02 272.97 0.2580 0.5848
Nc: carbon number; d20: liquid density at 20~ nz0: refractive index at 20~ (at 25~ TM: freezing point; Tb: normal boiling point; Tc: critical temperature;
Pc: critical pressure; Vc: critical volume; Zc: critical compressibility factor; to: acentric factor. Data source: API-TDB [2].
*SG of light gases are at low temperatures where the gas is in liquid form.
TABLE 2.2---Secondary properties o f some selected compounds.
p,sat Flammability
API at 100~ Viscosity, cSt TF AI, AP, range, vol.% RON
Compound Formula Arc gravity Kw bar v380oo) V99(210) CH ~ ~ ~ Min Max clear
Paraffins
Methane CH4 1 340,3 19.53 344.737 . . . . . . 2.98 . . . . . . 536.9 5.00 1.00
Ethane C2H6 2 266.6 19.49 55.1579 ... 3.97 . . . . . . 471.9 2.90 13.00
Propane C3H8 3 148.0 14.74 12.9621 0.1775 . . 4.47
. . . ... . 449.9 2.10 9.50 1.8
n-Butane C4H10 4 110.4 13.49 3,5608 0.2533 0.1686 4.77 -.. 83.2 287.9 1.80 8.40 93.8
n-Pentane C5H12 5 92.5 13.02 1.0745 0.3394 0.2643 4.96 --40.0 70.8 242.9 1.40 8.30 61.7
n-Hexane C6H14 6 81.2 12.79 0.3435 0.4152 ... 5.11 --21.7 68.7 224.9 1.20 7.70 24.8
n-Heptane C7H16 7 73.5 12.67 0.1112 0.5046 0.3515 5.21 --4.1 ... 203.9 1.00 7.00 0.0
n-Octane C8H18 8 68.6 12.66 0.0370 0.6364 0.3997 5.30 12.9 ... 205.9 0.96
n-Nonane C9H20 9 64.5 12.66 0.0125 0.8078 0.4694 5.36 30.9 ... 204.9 0.87 2.90
n-Decane CloH22 10 61.2 12.67 0.0042 1.0154 0.5537 5.42 45.9 ... 200.9 0.78 2.60
n-Undecane CllH24 11 58.7 12.71 0.0014 1.2588 0.6397 5.46 65.0 ... 201.9
n-Dodecane C12H26 12 56.6 12.74 0.0005 1.5452 0.7469 5.50 73.9 ... 202.9
n-Tridecane C13H28 13 54.4 12.76 0.0002 1.8634 0.8624 5.53 78.9 ... 201.9
4~ n-Tetradecane C14H30 14 53.1 12.82 0.0001 2.2294 0.9885 5.56 100.0 ... 199.9
n-Pentadecane C15H32 15 51.9 12.87 0.0000 2.6415 1.1328 5.59 113.9 ... 201.9
n-Hexadecane C16H34 16 51.6 12.97 0.0000 3.1229 1.2859 5.61 135.1 ... 201.9
n-Heptadecane C17H36 17 51.0 13.05 0.0000 3.6045 1.4413 5.63 147.8 ... 201.9
n-Octadecane C18H38 18 49.0 13.01 0.0000 4.1620 1.5815 5.64 165.1 ... 201.9
n-Nonadecane C 19H40 19 48.1 13.04 0.0000 4.6090 1.7940 5.66 167.9 ... 201.9
n-Eicosane C20H42 20 47.8 13.12 0.0000 5.3165 1.9889 5.67 166.9 ... 201.9
n-Heneicosane C21H44 21 46.4 13.11 ... ... 2.1703 5.69 176.9 ... 201.9
n-Docosane C22H46 22 45.8 13.15 ... 2.4099 5.70 184.9 201.9
isobutane (2-Methylpropane) C4H10 4 119.2 13.78 5.0199 0.2"773 0.1873 4.77 - - H 107".7 460.1 1.80 8.40 8;1
i s o p e n t a n e (2-Methylbutane) C5H12 5 94.4 13.01 1.4110 0.3066 ... 4.96 -57.2 70.8 420.1 1.40 8.30 92.3
2-Methylpentane C6H14 6 83.6 12.82 0.4666 0.3862 .,. 5.11 -35.2 73,8 ... 1.20 7.70 73.4
2-Methylhexane C7H16 7 75.9 12.72 0.1562 0.4730 ... 5.21 -23.2 74.1 ... 1.00 7.00 42.4
2-Methylheptane C8H18 8 69.8 12.65 0.0528 0.5908 0.3635 5.30 4.1 73.9 ,.. 0.98 ... 20.6
2,2,4-Trimethylpentane (isooctane) C8H18 8 71.0 12.52 0.1181 0.6077 0.3738 5.30 -12.2 79.5 ... 1.00 ... 100.0
2-Methyloctane C9H20 9 65.7 12.66 0.0177 0.7382 0.4329 5.36 22.9 . . . . . .
2-Methylnonane C10H22 10 62.1 12.66 0.0058 0.9636 0.5401 5.42 40.9 . . . . . .
(Continued)
TABLE 2.2--(Continued)
p,sat Flammability
API at 100~ Viscosity, cSt TF AI, AP, range, vol. % RON
Compound Formula Arc gravity Kw bar v38(100) v99(210) CH ~ ~ ~ Min Max clear
Olefins
Ethene (Ethylene) C2H4 2 888.0 48.49 . . . . . . 5.96 2.30 32.30
Propene (Propylene) C3H 6 3 141.0 14.26 15.7812 0.1"801 ... 5.96 -108.2 455.1 2.00 11.00 100.2
1-Butene C4H8 4 104.3 13.05 4.2953 . . . . . . 5.96 383.9 1.60 9.30 97.4
1-Pentene CsH10 5 87.7 12.66 1.3203 ... 5.96 -i8".2 19.1 272.9 1.50 8.70 90.9
1-Hexene C6H12 6 76.9 12.46 0.4143 0.3415 ... 5.96 -31.2 22.8 253.1 1.00 7.50 76.4
1-Heptene C7H14 7 70.2 12.41 0.1353 0.4317 0.3043 5.96 0.1 27.3 263.1 0.80 6.90 54.5
1-Octene C8H16 8 65.5 12.42 0.0453 0.5657 0.3590 5.96 20.9 32.6 230.1 0.80 6.80 28.7
1-Nonene C9H18 9 61.5 12.43 0.0152 0.7004 0.4294 5.96 26.9 38.1 236,9 0.60 6.00
1-Decene CmH20 10 58.4 12.45 0.0051 0.8848 0.5027 5.96 47.1 44,2 235.1 0.55 5.70
Acetylene C2H2 2 . . . . . . . . . . . . 11.92 -18.2 305.1 2.50 80.00
1,3 Butadiene C4H6 4 . . . . . . 4.0906 0.2030 0.2248 7.94 428.9
Naphthenes
Cyclopentane CsH10 5 57.1 11.12 0.6839 0.4973 .,. 5.96 -39.2 16.8 361.1 1.40 9.40 100.1
Methylcyclopentane C6H12 6 56.2 11.31 0.3106 0.5646 ... 5,96 -27.2 33.1 328.9 1.20 8.35 91.4
Ethylcyclopentane C7H14 7 52.0 11.39 0.0970 0.6199 .., 5.96 -4.1 36.8 260.1 1.10 6.70 67.2
Propylcyclopentane C8H~6 8 49.7 11.51 0.0325 0.7257 0.4613 5.96 15.9 44.5 269.1 0.95 6.40 31.2
n-Butylcyclopentane C9H18 9 47.8 11.63 0.0108 0.9118 0.5148 5.96 31.9 48.8 250.1 0.80 5.90 97.0
n-Pentylcyclopentane CloH20 10 46.4 11.75 ... 1,1280 0.6200 5.96 0.74 5.47
4~ n-Hexylcyclopentane CllH22 11 45.2 11.86 ... 1.4150 0.7300 5.96 0.68 5.20
t~
n-Heptylcyclopentane C12H24 12 44.3 11.97 .., 1.7480 0.8500 5.96 0.62 5.06
n-Octylcyclopentane C13H26 13 43.5 12.07 ... 2.1300 0.9800 5.96 0.57 5.01
n-Nonylcyclopentane C14H28 14 42.7 12.16 ... 2.5700 1.1200 5.96 0.53 5.07
n-Decylcyclopentane C15H30 I5 42.1 12.25 ... 3.0500 1.2700 5.96 0.50 5.24
n-Undecylcyclopentane C16H32 16 41.6 12.33 ... 3.6300 1.4400 5.96 0.47 5.53
n-Dodecylcyclopentane C17H34 17 41.1 12.41 ... 4.2500 1.6100 5.96 0.44 5.95
n-Tridecylcyclopentane C18H36 18 40.7 12.48 ... 4.9500 1.7800 5.96 0.41 6.53
n-Tetradecylcyclopentane C19H38 19 40.3 12.55 ... 5.7100 1.9800 5.96 0.39 7.33
n-Pentadecylcyclopentane C20H40 20 40.0 12.61 ... 6.5600 2.1900 5.96 0.37 8.38
n-Hexadecylcyclopentane C21H42 21 39.7 12.67 ... 7.4900 2.4000 5.96 0.35 9.79
n-Heptadecylcyclopentane C22H44 22 39.4 12.73 . . . . . . 5.96 0.34 11.67
Cyclohexane C6H12 6 49,4 11.00 0,2"274 0.9419 ... 5.96 -ik0 31.1 26;.1 1.30 8.00 83.o
Methylcyclohexane C7H14 7 51.1 11.31 0.1106 0.7640 0.4757 5.96 -5.9 41.1 285.1 1.15 7.20 74.8
Ethylcyclohexane C8H16 8 47.0 11.35 0.0333 0.8629 0.5122 5.96 22.1 43.8 261.9 0.90 6.60 45.6
Propylcyclohexane C9H18 9 45.8 11.50 0.0117 1.0010 0.5759 5.96 30.9 49.8 248.1 0.95 5.90 17.8
n-Butylcydohexane C10H20 10 44.6 11.64 0.0040 1.2539 0.6100 5.96 47.9 54.4 246.1 0.85 5.50
Aromatics
Benzene C6H6 6 28,7 9.74 0.2216 0.5927 0.3306 11.92 -11.2 -30.0 560.1 1.40 7.10
Methylbenzene (Toluene) C7H8 7 30.4 10.11 0.0710 0.5604 0.3433 10.43 4.9 -30.0 480.1 1.20 7.10 lOS'.8
Ethylbenzene C8H1o 8 30.5 10.34 0.0257 0.6540 0.3970 9.53 15.1 -30.0 430.1 1.00 6.70 100.8
Propylbenzene C9H12 9 31.5 10.59 0.0100 0.7977 0.4534 8.94 30.1 -30.0 456,1 0.88 6.00 101.5
n-Butylbenzene CloH14 10 31.9 10.82 0.0033 0.9483 0.5186 8.51 50.1 -30.0 410.1 0.80 5.80 100.4
n-Pentylbenzene CllH16 11 32.6 11,03 0.0011 1.1824 0.6295 8.19 65.1 . . . . . . 0.80 5.50 ...
n-Hexylbenzene C12H18 12 32.6 11.19 0.0004 1.4419 0.7334 7.94 80.1 0.70 5.30
n-Heptylbenzene C13H20 13 32.7 11.35 0.0001 1.7546 0.8546 7.75 95.1 . . . . . . 0.70 5.10 ...
n-Octylbenzene C14H22 14 33.0 11.50 0.0000 2.0974 0.9763 7.58 107.1 . . . . . . 0.70 4.90 ...
n-Nonylbenzene C15H24 15 33.1 11.63 0.0000 2.5329 1.1089 7.45 98.9 . . . . . . 0.60 4.70 ...
n-Decylbenzene C16H26 16 33.2 11.75 0.0000 3.0119 1.2490 7.33 106.9 . . . . . . 0.60 4.60 ...
n-Undecylbenzene C17Hz8 17 33.3 11.86 0.0000 3.5672 1.3983 7.23 143.9 . . . . . . 0.60 4.40 ...
n-Dodecylbenzene C18H30 18 33.1 11.94 0.0000 4.1857 1.5509 7.15 140.9 . . . . . . 0.60 4.30 ...
n-Tridecylbenzene C19H32 19 33.3 12.05 0.0000 4.9566 1.7258 7.07 164.9 . . . . . . 0.50 4.20 ...
n-Tetradecylbenzene C20H34 20 33.3 12.13 0.0000 5.7107 1.8853 7.01 173.9 . . . . . . 0.50 4.10 ...
n-Pentadecylbenzene C21H36 21 33.3 12.20 0.0000 6.5716 2.0862 6.95 182.9 . . . . . . 0.50 4.00 ...
n-Hexadecylbenzene C22H38 22 33,3 12.28 0.0000 7.4138 2.2939 6.90 192.9 ... 0.50 3.90 ...
o-xylene ( o - d i m e t h y l b e n z e n e ) C8H10 8 28.4 10.27 0.0182 0.4238 9.53 16.9 463". 1 1.00 6.00
m-xylene CsHIo 8 31.3 10.42 0.0226 0.5936 0.3644 9.53 25.1 -3'0.0 465.1 1.10 7.00 104.0
p-xylene CsH10 8 32.0 10.46 0.0237 0.6167 0.3704 9.53 25.1 -30.0 528.1 1.10 7.00 103.4
isopropylbenzene C9H12 9 31.5 10.54 0.0129 0.7474 0.4229 8.94 43.9 -15.0 ... 0.88 6.50 102.1
1,2,3-Trimethylbenezene C9H12 9 26.0 10.37 0.0049 0.8317 0.3979 8.94 51.2 . . . . . . 0.88 5.20 101.4
isobutylbenzene CIoH14 10 33.5 10.84 0.0057 0.9895 0.5166 8.51 55.1 . . . . . . 0.80 6.00 101.6
o-Cymene CloH14 10 29.1 10.59 0.0045 0.9739 0.5979 8.51 53.1 . . . . . . 0,80 5.20 100.6
Cyclohexylbenzene C12H16 12 32.7 11.30 0.0002 2.0086 0.7936 8.94 98.9 . . . . . . 0.70 5.40 ...
Naphthalene C10H8 10 6.1 9.34 . . . . . . 0.7747 14.89 80.1 ... 526.1 0.88 5.90 ...
1-Methylnaphthalene CllH10 11 6.7 9.54 0.0002 2.1837 0.9204 13.11 82.1 ... 0.80 5.30
Styrene C8H8 8 24.0 10.00 0.0170 0.6637 0.3783 11.92 31.9 ... 490.1 1.10 6.10 10"3'.0
DiphenyI m e t h a n e C13H12 13 8.6 9.79 0.0001 2.1883 0.9834 12.91 130.1 . . . . . . 0.70 5.20 ...
Nonhydrocarbons
Water H20 . 10.0 8.76 0.0655 0.7063 0.0003 . . . . . . . . . . . . . . . . . . . . .
C a r b o n dioxide CO2 "i ... 8.63 . . . . . . . . . . . . . . . . . . . . .
Hydrogen Sulfide H2S ...... 9.06 27.2295 0.i447 . . . . . . . . . . . . 260.1 4.3"0 45.50 ...
Nitrogen N2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Oxygen 02 ...... 4.78 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ammonia NH3 ... 12.26 14.5632 0.1990 0.1373 . . . . . . . . . 650.9 16.00 25.00 ..,
Carbon monoxide CO "'i . . . . . . . . . . . . . . . - . . . . . . . . . 608.9 12.50 74.00 ...
Hydrogen H2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 400.1 4.00 75.00 ...
Air N2+O2 . . . . . . . . . . . . . . . . . . . . . . . . .
Methanol CH3OH "1 45~52 10.60 0.3189 0.5'900 0.3269 2.98 lb185 463.85 ... 2~3 36.0 130
Ethanol C2H5OH 2 46.48 10.80 0.1603 1.0995 0.4579 3.97 12.85 422.85 ... 4.3 19.0 125
Isobutanol C4H9OH 4 44.0 10.94 0.0335 2.8700 0.7527 4.77 27.85 . . . . . . 1.7 10.9 ...
Nc: carbon number; psat : true vapor pressure at 100~ (311 K); v: kinematic viscosity; CH: carbon-to-hydrogen weight ratio; TM: freezing point; AI: autoignition temperature; AP: aniline point; RON: research
octane number. Data sources: API-TDB [2]; for autoignition temperature: DIPPR [20].
TABLE 2.3--Additional data on refractive index and viscosity of heavy hydrocarbons.
Components Formula M SG d v V VGC
Paraffins
n-Tetracosane C24H50 338.6 0.8028 0.7980 2.79 35.5 1.4504 1.051 0.745
11-n-Butyldocosane C26H54 366.7 0.8081 0.8041 2.78 35.5 1.4461 1.044 0.749
5-n-Bytyldocosane C26H54 366.7 0.8099 0.8059 3.02 36.3 1.4466 1.044 0.750
9-Ethyl-9-n-Heptyloctadecane C27H56 380.7 0.8117 0.8076 3.48 37.8 1.4513 1.048 0.746
7-n-Hexyldocosane C28H58 394.7 0.8121 0.8080 3.32 38.2 1.4517 1.048 0.749
11-n-Decylhenecicosane C31H64 436.8 0.8158 0.8117 3.66 38.4 1.454 1.084 0.7481
n-Dotriacontane C32H66 450.8 0.8160 0.8119 5.01 42.7 1.455 1.049 0.730
11-n-Decyldocosane C32H66 450.8 0.8170 0.8129 3.91 39.2 1.4543 1.048 0.747
1 l-n-Decyltetracosane C34H70 478.9 0.8194 0.8153 4.45 40.9 1.4556 1.048 0.746
9-n-Octylhexacosane C38H78 478.9 0.8187 0.8156 4.70 14.7 1.456 1.049 0.743
13-n-Dodeecylhexacosane C38H78 507.0 0.8209 0.8168 4.90 42.4 1.4567 1.040 0.745
13-n-Undecylpentacosane C38H78 535.0 0.8230 0.8189 5.54 4.44 1.4577 1.048 0.744
Naphthenes
1,3 -Dicylohexylcyclohexane C18H32 248.4 0.9490 0.940 4.61 41.4 1.5065 a 1.036 0.917
1,2-Dicylohexylcyclohexane C18H32 248.4 0.949 0.9404 4.86 42.2 1.5062 1.036 0.917
2(ar)-n-Butyl-3 (ar)-n-hexyl-tetralin C20H32 272.5 0.9182 0.9086 2.92 35.9 1.5075 1.048 0.909
Perhydridibenzo-(a,i) fluorene C21H34 286.5 1.0081 0.9981 10.66 61.6 1.5277 1.0287 0.976
Di(alpha-decalyl)-methane C21H36 288.5 0.9807 0.9709 10.96 62.6 1.518 1.031 0.940
1,2-Di(alpha-decalyl)ethane C22H38 302.5 0.9748 0.9661 12.2 67.2 1.5176 1.034 0.930
1,1-Di(alpha-decalyl)ethane C22H38 302.5 0.9854 0.9765 16.5 83.9 1.5217 1.033 0.938
1,5-Dicyclopentyl-3(2-cyclopentylethyl)p e n t a n e C22H38 304.5 0.9040 0.895 3.96 39.5 1.4853 1.0378 0.866
1,7-Dicyclopentyl-4(2-phenylethy) heptane C25H40 340.6 0.9298 0.9205 3.92 39.2 1.507 1.046 0.899
1-Phenyl-3 (2-cyclohexylethyl)-6-cyclopentylhexane C25H40 340.6 0.9318 0.9230 4.80 42 1.5086 1.0465 0.893
6-n-Ocylperhydrobenz(de) anthracene C25H44 344.6 0.9530 0.9432 7.91 52.1 1.5081 1.036 0.909
Cholestane C27H48 372.8 0.9578 0.9482 21.15 103.4 1.52 a 1.050 0.897
1,1-Di(alphadecalyl)hendecane C31H56 428.8 0.9451 0.9356 20.25 99.5 1.5062 1.038 0.885
Aromatics
1,1-Diphenylethylene C14H12 180.2 1.0330 1.0235 3.96 b 39.1 b 1.6078 1.097 1.033
4,5-Dimethyl-9,10-dihydropenanthrene C16H14 206.3 1.1090 1.0998 4.93 42.2 1.6433 a 1.094 1.127
4,5-Dimethyl-9,10-dihydropenanthrene C16H16 208.3 1.0738 1.0684 4.43 40.8 1.6286 1.095 1.089
1,2-Diphenylbenze C18H18 230.3 1.0923 1.0814 4.53 41.2 1.6447 1.105 1.107
9-n-Butylanthracene C16H16 239.3 1.0600 1.0530 4.61 41.4 1.6006 1.074 1.064
1,2,3,4,5,6,8,13,14,15,15-Dodecahydrochrysene C18H24 240.4 1.0572 1.0492 8.16 52.9 1.5752 1.051 1.049
1,5-Di-phenyl-3(2-phenylethyl)2-pentene C25H28 326.5 1.0210 1.0159 3.42 37.6 1.5816 1.074 1.030
1,5-Di-phenyl-3(2-phenylethyl)pentane C25H40 328.5 1.0175 1.0076 3.81 38.8 1.5725 1.070 1.015
1,7-Dicyclopentyl-4-(2-phenylethyl)heptane C25H28 340.6 0.9251 0.9205 3.92 39.2 1.5070 t.047 0.919
1,2,3,4,5,6,8,9,10,17,18-Dodecahydro-9(n-octyl)naphthecene C26H40 352.6 0.9845 0.9796 13.09 70.5 1.5467 1.060 0.950
1,10-Di-(alphanaphyl)decane C30H34 394.6 1.0400 1.0296 12.27 67.4 1.6094 1.095 1.016
M: molecular weight;
SG: specific gravity at 15.56~ (60/60~
d: density at 20~ g/cm3;
v: kinematic viscosity at 210~ (99~ cSt;
V: Saybolt viscosity at 210~ (99~ SUS;
n: refractive index at 20~
Ra: refractivity intercept;
VGC: viscosity gravity constant. Data source: API-RP 42 [ 18].
a Calculated values of n.
bKinematic or Saybolt viscosities at 100~ (38~
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 45
are calculated by Eqs. (2.14) and (2.15) and will be used to without any theoretical explanation. Boiling point and spe-
develop predictive methods for the composition of heavy frac- cific gravity were used in most correlations based on expe-
tions discussed in Chapter 3 (Section 3.5). rience and their availability. However, the characterization
methods proposed by Riazi and Daubert [28, 29, 35] are based
on the theory of intermolecular forces and EOS parameters
2.3 CHARACTERIZATION [36]. Although EOS are discussed in Chapter 5, their applica-
OF HYDROCARBONS tion in the development of analytical correlations to charac-
terize hydrocarbons are discussed here. In the following parts
The work on characterization of pure hydrocarbons began in in this section several characterization schemes developed by
1933 when Watson and Nelson for the first time developed Riazi et al. [28, 29, 31, 35, 37] are presented along with other
two empirical charts relating molecular weight to either boil- methods.
ing point and Kw or boiling point and API gravity [24]. In
these charts boiling point and specific gravity (or API grav-
ity) are used as the two independent input parameters. Since 2.3.1 Development of a Generalized Correlation
then the work on characterization and methods of estimation for Hydrocarbon Properties
of basic properties of pure hydrocarbons and petroleum frac- Properties of a fluid depend on the intermolecular forces that
tions has continued to the present time. Methods developed exist between molecules of that fluid [38, 39]. As summarized
in 1930s till 1960s were mainly graphical, while with the use by Prausnitz et al. [39] these forces are grouped into four cat-
of computer, methods developed in 1970s till present time are egories. (1) Electrostatic forces between charged molecules
in the forms of analytical correlations. The best example of (ions) and between permanent dipoles or higher multipoles.
chart-type correlations, which has been in use by the industry, These forces result from the chemical structure of molecules
is the Winn nomogram that relates molecular weight, critical and are important in polar compounds (i.e., water, methanol,
pressure, aniline point, and CH weight ratio to boiling point ethanol, etc.). (2) Induction forces on molecules that are po-
and specific gravity [25]. A version of Winn nomograph as larizable when subjected to an electric field from polar com-
used by the API [2] is presented in Fig. 2.12. Some of the an- pounds. These forces are also called dipole forces and are de-
alytical correlations that are used in the industry are Cavett termined by dipole moment of molecules 0z), which is propor-
[26], Kesler-Lee [12], Lee-Kesler [27], Riazi-Daubert [28, 29], tional to polarizability factor, ~, and the field strength. These
Twu [30], and Riazi-Sahhaf [31]. Most of these correlations forces are proportional to/z 2 x ~. (3) The third type of forces
use boiling point and specific gravity as the input parame- are attraction (dispersion forces) and repulsion between non-
ters to estimate parameters such as molecular weight, critical polar molecules. These forces, also called London forces, are
constants, and acentric factor. Most recently Korsten [32] has static in nature and are proportional to u2. (4) The last are spe-
developed a characterization scheme that uses boiling point cial (chemical) forces leading to association or complex for-
and a parameter called double-bond equivalent (DBE) as the mation such as chemical bonds. According to London these
input parameters. DBE can be estimated from H/C atomic forces are additive and except for very polar compounds, the
ratio. In another paper Korsten [33] lists and evaluates vari- strongest forces are of the London type. For light and medium
ous correlations for estimation of critical properties of pure hydrocarbon compounds, London forces are the dominant
hydrocarbons. Tsonopolus et al. [34] give the list of correla- force between the molecules.
tions developed for characterization of coal liquids in terms The intermolecular force, F, is related to the potential en-
of boiling point and specific gravity. There are some meth- ergy, F, according to the following relation:
ods of estimation of properties of pure compounds that are
based on various group contribution techniques. The most dF
(2.19) F --
accurate methods of group contributon for various proper- dr
ties with necessary recommendations are given in the fifth
edition of Properties of Gases and Liquids [4]. Even some of where r is the distance between molecules. The negative of
these group contribution methods require properties such as the potential energy, -F(r), is the work required to sepa-
molecular weight or boiling point. Examples of such proce- rate two molecules from the intermolecular distance r to in-
dures are the Lydersen and Ambrose methods [4]. The prob- finite separation. Equation of state parameters can be esti-
lem with group contribution methods is that the structure of mated from the knowledge of the potential energy relation
the compound must be known. For this reason they are not [39]. Most hydrocarbon compounds, especially the light and
appropriate for undefined petroleum fractions. However, they medium molecular weight hydrocarbons, are considered as
can be used to predict properties of pure compounds when nonpolar substances. There are two forces of attraction (dis-
experimental data are not available (i.e., critical properties persion forces) and repulsion between nonpolar molecules.
of heavy pure hydrocarbons). In fact, on this basis the prop- The common convention is that the force of attraction is neg-
erties of hydrocarbons heavier than C18 have been predicted ative and that of repulsion is positive. As an example, when
and reported by the API [2]. molecules of methane are 1 nm apart, the force of attraction
As discussed, during the past 70 years m a n y methods in between them is 2 x 10 -s dyne [39]. The following relation
the forms of charts and equations were proposed to esti- was first proposed by Mie for the potential energy of nonpolar
mate the basic properties of hydrocarbons from the knowl- molecules [39]:
edge of the boiling point and specific gravity or the molecular Ao /3o
weight. Nearly all of these correlations are empirical in nature (2.20) F-
r n r m
46 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
where
No
(2.27) k OV2} =0
Prep --
- - -r n-
Application of Eqs. (2.26) and (2.27) to any two-parameter
Bo
Patt = --- EOS would result in relations for calculation of parameters
r m
A and B in terms of Tc and Pc, as shown in Chapter 5. It should
be noted that EOS parameters are generally designated by
Prep = repulsive potential lower case a and b, but here they are shown by A and B. No-
Fatt= attractive potential
tation a, b, c . . . . are used for correlation parameters in various
Ao = parameter characterizing the repulsive force (>0)
equations in this chapter. Applying Eqs. (2.26) and (2.27) to
/3o = parameter characterizing the attractive force (>0)
Eq. (2.25) results into the following three relations for To, Pc,
r --- distance between molecules
and Vc:
n, m = positive numbers, n > m
The main characteristics of a two-parameter potential energy (2.28) Tc = f4(A, B)
function is the m i n i m u m value of potential energy, P m i n , desig-
(2.29) Pc = fs(A, B)
nated by e = - Pmin and the distance between molecules where
the potential energy is zero (F = 0) which is designated by a. (2.30) Vc = f6(A, B)
L o n d o n studied the theory of dispersion (attraction) forces
Functions f4, f5, and f6 are universal functions and are the
and has shown that m = 6 and it is frequently convenient for
same for all fluids that obey the potential energy relation ex-
mathematical calculations to let n = 12. It can then be shown
pressed by Eq. (2.20) or Eq. (2.21). In fact, if parameters A
that Eq. (2.20) reduces to the following relation k n o w n as
and B in a two-parameter EOS in terms of Tc and Pc are rear-
Lennard-Jones potential [39]:
ranged one can obtain relations for Tc and Pc in terms of these
(2.21) P = 48 [ ( ~ ) 1 2 - ( ~ ) 6 1 two parameters. For example, for van der Waals and Redlich-
Kwong EOS the two parameters A and B are given in terms
of Tc and Pc [21] as shown in Chapter 5. By rearrangement of
In the above relation, e is a parameter representing molec-
the vdW EOS parameters we get
ular energy and a is a parameter representing molecular
size. Further discussion on intermolecular forces is given in
Section 5.3. To= AB-1 Pc= A B -2 Vc = 3B
According to the principle of statistical thermodynamics
there exists a universal EOS that is valid for all fluids that and for the Redlich-Kwong EOS we have
follow a two-parameter potential energy relation such as
Eq. (2.21) [40]. [ (0"0867)5R] 2/3 A2/3B-2/3
rC=k~ J
(2.22) Z = fl(g, a, T, P)
[ (0.0867) 5R 1 '/3 A 2/3B -s/3
Pc = L ~ J vc = 3.847B
(2.23) Z- PVr,~
RT
Similar relations can be obtained for the parameters of other
where
two-parameter EOS. A generalization can be made for the
Z = dimensionless compressibility factor
relations between To, Pc, and V~ in terms of EOS parameters
Vr, e = molar volume at absolute temperature, T, and pres-
A and B in the following form:
sure, P
/'1 = universal function same for all fluids that follow (2.31) [Tc, Pc, Vc] = aAbB c
Eq. (2.21).
By combining Eqs. (2.20)-(2.23) we obtain where parameters a, b, and c are the constants which differ
for relations for To, Pc, and Vc. However, these constants are
(2.24) Vr, v = f2(Ao, ]3o, T, P) the same for each critical property for all fluids that follow
where Ao and/3o are the two parameters in the potential en- the same two-parameter potential energy relation. In a two-
ergy relation, which differ from one fluid to another. Equation parameter EOS such as vdW or RK, Vc is related to only one
(2.24) is called a two-parameter EOS. Earlier EOS such as van parameter B so that Vc/B is a constant for all compounds.
der Waals (vdW) and Redlich-Kwong (RK) developed for sim- However, formulation of Vc t h r o u g h Eq. (2.30) shows that
ple fluids all have two parameters A and B [4] as discussed in Vc must be a function of two parameters A and B. This is
Chapter 5. Therefore, Eq. (2.24) can also be written in terms one of the reasons that two-parameters EOS are not accurate
of these two parameters: near the critical region. Further discussion on EOS is given
in Chapter 5.
(2.25) Vr, p = f3(A, B, T, P) To find the nature of these two characterizing parameters
The three functions fl, f2, and f3 in the above equations vary one should realize that A and B in Eq. (2.31 ) represent the two
in the form and style. The conditions at the critical point for parameters in the potential energy relation, such as e and c~ in
any PVT relation are [41 ] Eq. (2.22). These parameters represent energy and size char-
acteristics of molecules. The two parameters that are readily
(2.26) (0~) =0 measurable for h y d r o c a r b o n systems are the boiling point,
re.Pc Tb, and specific gravity, SG; in fact, Tb represents the energy
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF PURE H Y D R O C A R B O N S 47
parameter and SG represents the size parameter. Therefore, parameters and transport properties are discussed in Chap-
in Eq. (2.31) one can replace parameters A and B by Tb and ters 5, 6, and 8. It is shown by various investigators that the
SG. However, it should be noted that Tb is not the same as ratio of Tb/Tc is a characteristic of each substance, which is
parameter A and SG is not the same as parameter B, but it related to either Tr or Tb [36, 43]. This ratio will be used to
is their combination that can be replaced. There are many correlate properties of pure hydrocarbons in Section 2.3.3.
other parameters that may represent A and B in Eq. (2.31). Equation (2.31) can be written once for Tr in terms of V and I
For example, if Eq. (2.25) is applied at a reference state of To and once for parameter Tb/Tr Upon elimination of parameter
and P0, it can be written as V between these two relations, a correlation can be obtained
to estimate T~ from Tb and I. Similarly through elimination
of Tc between the two relations, a correlation can be derived
(2.32) Vr0,P0 = f3(A, B, To, Po)
to estimate V in terms of Tb and I [42].
where Vro,eo is the molar volume of the fluid at the reference It should be noted that although both density and refrac-
state. The most convenient reference conditions are tempera- tive index are functions of temperature, both theory and ex-
ture of 20~ and pressure of 1 atm. By rearranging Eq. (2.32) periment have shown that the molar refraction (Rm = VI) is
one can easily see that one of the parameters A or B can be nearly independent of temperature, especially over a narrow
molar volume at 20~ and pressure of 1 atm [28, 36, 42]. range of temperature [38]. Since V at the reference tempera-
To find another characterization parameter we may con- ture of 20~ and pressure of 1 atm is one of the characteriza-
sider that for nonpolar compounds the only attractive force tion parameters, I at 20~ and 1 arm must be the other char-
is the London dispersion force and it is characterized by fac- acterization parameter. We chose the reference state of 20~
tor polarizability, a, defined as [38, 39] and pressure of 1 atm because of availability of data. Simi-
larly, any reference temperature, e.g. 25~ at which data are
(2.33) ct = ~ x x ~,r/2 q- 2 / available can be used for this purpose. Liquid density and re-
fractive index of hydrocarbons at 20~ and 1 atm are indicated
where by d20 and n20, respectively, where for simplicity the subscript
NA = Avogadro's number 20 is dropped in most cases. Further discussion on refractive
M = molecular weight index and its methods of estimation are given elsewhere [35].
p = absolute density From this analysis it is clear that parameter I can be used
n = refractive index as one of the parameters A or B in Eq. (2.31) to represent the
In fact, polarizability is proportional to molar refraction, Rm, size parameter, while Tb may be used to represent the energy
defined as parameter. Other characterization parameters are discussed
in Section 2.3.2. In terms of boiling point and specific gravity,
(2.34) Rrn----- (-~-) X {n2-l'~,tz
2 q- 2`/ Eq. (2.31) can be generalized as following:
(2.38) 0 = aT~SG c
M
(2.35) V = -- where Tb is the normal boiling point in absolute de-
p
grees (kelvin or rankine) and SG is the specific gravity at
n2-1 60 ~F(15.5 ~C). Parameter 0 is a characteristic property such as
(2.36) I = - - molecular weight, M, critical temperature, To, critical pres-
n2+2
sure, Pc, critical molar volume, Vc, liquid density at 20~
in which V is the molar volume and I is a characterization d20, liquid molar volume at 20~ and I atm, V20, or refrac-
parameter that was first used by Huang to correlate hydrocar- tive index parameter, I, at 20~ It should be noted that
bon properties in this way [ 10, 42]. By combining Eqs. (2.34)- 0 must be a temperature-independent property. As mentioned
(2.36) we get before, I at 20~ and 1 atm is considered as a character-
Rm actual molar volume of molecules istic parameter and not a temperature-dependent property.
(2.37) I- - Based on reported data in the 1977 edition of API-TDB, con-
V apparent molar volume of molecules
stants a, b, and c were determined for different properties
Rm, the molar refraction, represents the actual molar volume and have been reported by Riazi and Daubert [28]. The con-
of molecules, V represents the apparent molar volume and stants were obtained through linear regression of the loga-
their ratio, and parameter I represents the fraction of total rithmic form of Eq. (2.38). Equation (2.38) in its numerical
volume occupied by molecules. Rm has the unit of molar vol- form is presented in Sections 2.4-2.6 for basic characteri-
ume and I is a dimensionless parameter. Rm/M is the specific zation parameters. In other chapters, the form of Eq. (2.38)
refraction and has the same unit as specific volume. Parame- will be used to estimate the heat of vaporization and trans-
ter I is proportional to the volume occupied by the molecules port properties as well as interconversion of various distil-
and it is close to unity for gases (Ig ~ 0), while for liquids it lation curves. The form of Eq. (2.38) for T~ is the same as
is greater than zero but less than 1 (0 < /liq <Z 1). Parameter the form Nokay [44] and Spencer and Daubert [45] used to
I can represent molecular size, but the molar volume, V, is a correlate the critical temperature of some hydrocarbon com-
parameter that characterizes the energy associated with the pounds. Equation (2.38) or its modified versions (Eq. 2.42),
molecules. In fact as the molecular energy increases so does especially for the critical properties and molecular weight,
the molar volume. Therefore, both V and I can be used as two have been in use by industry for many years [2, 8, 34, 46-56].
independent parameters to characterize hydrocarbon proper- Further application of this equation will be discussed in
ties. Further use of molar refraction and its relation with EOS Section 2.9.
48 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
One should realize that Eq. (2.38) was developed based on characterizing molecular energy and the molecular size. This
a two-parameter potential energy relation applicable to non- means that Eq. (2.31) can be expressed in terms of two pa-
polar compounds. For this reason, this equation cannot be rameters 01 and 02:
used for systems containing polar compounds such as alco-
hol, water, or even some complex aromatic compounds that (2.39) 0 = aOblO~
are considered polar. In fact constants a, b, and c given in However, one should realize that while these two parame-
Eq. (2.38) were obtained based on properties of hydrocar- ters are independent, they should represent molecular energy
bons with carbon number ranging from Cs to C20. This is and molecular size. For example, the pairs such as (Tb, M)
almost equivalent to the molecular weight range of 70-300. or (SG, I) cannot be used as the pair of input parameters
In fact molecular weight of n-C20 is 282, but considering the (01, 02). Both SG and I represent size characteristics of
extrapolation power of Eq. (2.38) one can use this equation molecules and they are not a suitable characterization pair.
up to molecular weight of 300, which is roughly equivalent to In the development of a correlation to estimate the proper-
boiling point of 370~ (700~ Moreover, experimental data ties of hydrocarbons, all compounds from various hydrocar-
on the critical properties of hydrocarbons above C~8 were not bon groups are considered. Properties of hydrocarbons vary
available at the time of development of Eq. (2.38). For heav- from one hydrocarbon type to another and from one carbon
ier hydrocarbons additional parameters are required as will number to another. Hydrocarbons and their properties can
be shown in Section 2.3.3 and Chapter 3. The lower limit for be tabulated as a matrix of four columns with many rows.
the hydrocarbon range is C5, because lighter compounds of Columns represent hydrocarbon families (paraffins, olefins,
C1-C4 are mainly paraffinic and in the gaseous phase at nor- naphthenes, aromatics) while rows represent carbon num-
mal conditions. Equation (2.38) is mainly applied for unde- ber. Parameters such as Tb, M, or kinematic viscosity at 38~
fined petroleum fractions that have average boiling points (100~ v38(100), vary in the vertical direction with carbon
higher than the boiling point of C5 as will be seen in Chapter number, while parameters such as SG, I, and CH vary signif-
3. Methods of calculation of properties of natural gases are icantly with hydrocarbon type. This analysis is clearly shown
discussed in Chapter 4. in Table 2.4 for C8 in paraffin and aromatic groups and prop-
erties of C7 and C8 within the same group of paraffin family.
Example 2.3--Show that the molecular weight of hydrocar- As is clearly shown by relative changes in various properties,
bons, M, can be correlated with the boiling point, Tb, and parameters SG, I, and CH clearly characterize hydrocarbon
specific gravity, SG, in the form of Eq. (2.38). type, while Tb, M, and 1)38(100) are good parameters to charac-
terize the carbon number within the same family. Therefore,
Solution--It was already shown that molar volume at 20~ a correlating pair should be selected in a way that character-
and l atm, V20, can be correlated to parameters A and B of a izes both the hydrocarbon group and the compound carbon
potential energy function through Eq. (2.32) as follows: V20 = number. A list of properties that may be used as pairs of cor-
gl(A, B). In fact, parameter V2o is similar in nature to the relating parameters (01,02) in Eq. (2.39) are given below [29].
critical molar volume, Vr and can be correlated with Tb and
SG as Eq. (2.38): V20 = aTbbSGc. But V20 = M/d2o, where d20 Tc (K), Pc (bar), Vc (cm3/g), M, Tb (K), SG, I (20~ CH
is the liquid density at 20~ and 1 atm and is considered a (Tb, SG), (Tb, I), (Tb, CH), (M, SG), (M, I), (M, CH),
(01, 02)
size parameter. Since Tb is chosen as an energy parameter Pairs: (v38(100), SG), (v38(100),I), (v38(100), CH)
and SG is selected as a size parameter, then both d20 and SG
represent the same parameter and can be combined with the
energy parameter as M = aT~SG r which has the same form The accuracy of Eq. (2.39) was improved by modification of
of Eq. (2.38) for 0 = M. its a parameter in the following form:
TABLE 2.4--Comparison of properties of adjacent members of paraffin family and two families
of C8 hydrocarbons.
Hydrocarbongroup Tb,K M v38000),cSt SG 1 CH
Paraffin family
C7Hi6(n-heptane) 371.6 100.2 0.5214 0.6882 0.236 5.21
C8H18(n-octane) 398.8 114.2 0.6476 0.7070 0.241 5.30
% Difference in property +7.3 +14.2 +24.2 +2.7 +2.1 +1.7
Two Families (C8)
Paraffin (n-octane) 398.8 114.2 0.6476 0.7070 0.241 5.30
Aromatic (ethylbenzene) 409.4 106.2 0.6828 0.8744 0.292 9.53
% Difference in property +2.7 -7.0 5.4 +23.7 +21.2 +79.8
T A B L E 2.5--Coefficients o f Eq. (2.40) 0 = a exp (bO1 + c02 + dO102)O~Of for various properties.
0 01 02 a b c d e f AAD%
Tc Tb SG 9.5232 - 9 . 3 1 4 • 10 -4 -0.54444 6.48 x 10 -4 0.81067 0.53691 0.5
Tc Tb I 4.487601 • 105 - 1 . 3 1 7 1 • 10 -3 -16.9097 4.5236 • 10 -3 0.6154 4.3469 0.6
Tc Tb CH 8.6649 0 0 0 0.67221 0.10199 0.7
Tc M SG 308 - 1 . 3 4 7 8 • 10 -4 -0.61641 0 0.2998 1.0555 0.7
Tc M I 1.347444 • 106 2.001 • 10 -4 -13.049 0 0.2383 4.0642 0.8
Tc M CH 20.74 1.385 • 10 -3 -0.1379 -2.7 • 10 -4 0.3526 1.4191 1.0
Tc v38(100) I 2.4522 • 103 --0.0291 -- 1.2664 0 0.1884 0.7492 1.0
Pc Tb SG 3.195846 x 105 - 8 . 5 0 5 • 10 -3 -4.8014 5.749 x 10 -3 -0.4844 4.0846 2.7
Pc Tb I 8.4027 • 1023 --0.012067 --74.5612 0.0342 --1.0303 18.4302 2.6
Pc Tb CH 9.858968 - 3 . 8 4 4 3 x 10 -3 -0.3454 0 -0.1801 3.2223 3.5
Pc M SG 3116.632 - 1 . 8 0 7 8 • 10 -3 -0.3084 0 -0.8063 1.6015 2.7
Pc M I 2.025 x 1016 -0.01415 -48.5809 0.0451 -0.8097 12.9148 2.3
Pc M CH 56.26043 - 2 . 1 3 9 x 10 -3 -0.265 0 -0.6616 2.4004 3.1
Pc v38(lO0) I 393.306 0 0 0 -0.4974 2.052 4.2
Vc Tb SG 6.049 • 10 -2 - 2 . 6 4 2 2 x 10 3 -0.26404 1.971 • 10 -3 0.7506 -1.2028 1,8
Vc Tb I 6.712 • 10 -3 --2.720 • 10 -3 0.91548 7.920 x 10 -3 0.5775 --2.1548 1.8
Vc Tb CH 1.409 • 10 +1 - 1 . 6 5 9 4 x 10 -3 0.05345 2.6649 x 10 -4 0.1657 -1.4439 2.4
Vc M SG 7.529 x 10 -1 - 2 . 6 5 7 x 10 -3 0.5287 2.6012 • 10 -3 0.20378 -1,3036 1.9
Vc M I 6.3429 • 10 -5 - 2 . 0 2 0 8 • 10 -3 14.1853 4.5318 x 10 -3 0.2556 -4.60413 2.1
Vc M CH 1.597 x 10 +1 - 2 . 3 5 3 3 • 10 -3 0.1082 3.826 x 10 -4 0.0706 -1.3362 2.6
Vc v38(lO0) I 2.01 x 10 -1~ -0.16318 36.09011 0.4608 0.1417 - 10.65067 1.7
M Tb SG 1032.1 9.78 • 10 -4 -9.53384 2.0 • 10 -3 0 .9 7 4 7 6 6.51274 2.1
M Tb I 8.9205 x 10 -6 1.55833 x 10 -5 4.2376 0 2.0935 -1.9985 2.3
M Tb CH 1.81456 x 10 -3 0 0 0 1.9273 -0.2727 2.3
M v38(too) SG 1.51723 • 106 -0.195411 -9.638970 0.162470 0 .5 6 3 7 0 6.89383 3.4
M v38(lOO) I 4 x 10 -9 - 8 . 9 8 5 4 x 10 -2 38.106 0 0.6675 -10.6 3.5
Tb M SG 3.76587 3.77409 x 10 -3 2.984036 - 4 . 2 5 2 8 8 • 10 -3 4.0167 x 10 -1 -1.58262 1.0
Tb M I 75.775 0 0 0 0.4748 0.4283 1.1
Tb M CH 2 0. 253 4 7 - 1 . 5 7 4 1 5 x 10 -4 - 4 . 5 7 0 7 • 10 -2 9.22926 • 10 -6 0 .5 1 2 9 7 6 0.472372 1.1
Tb v38(i00) SG 237 9 . 86 - 1 . 3 0 5 1 • 10 -2 -1.68759 - 2 . 1 2 4 7 x 10 -2 0 .2 6 2 9 1 4 1.3489 1.7
Tb v38(lO0) I 0 . 0 506 2 9 - 6 . 5 2 3 6 x 10 -2 14.9371 6.029 x 10 -2 0.3228 -3.8798 1.6
SG Tb I 2.43810 • 107 -4.19400 • 10 -4 -23.553500 3.98736 • 10 -3 -0.341800 6.9195 0.5
SG Tb CH 2.86706 x 10 -3 -1.83321 • 10 -3 -0.081635 6.49168 x 10 -5 0.890041 0.73238 1.0
SG M I 1.1284 x 106 -1.58800 x 10 -3 -20.594000 7.34400 x 10 -3 -0.07710 6.30280 0.5
SG M CH 6.84403 • 10 -2 -1,48844 • 10 -3 -0.079250 4.92112 x 10 -5 0 .2 8 9 8 4 4 0.919255 1.3
SG v38(lOO) I 3.80830 • 107 -6.14060 x 10 -2 -26.393400 0.253300 -0.235300 8.04224 0.5
I Tb SG 0.02343 7.0294 • 10 -4 2.46832 -1.0268 x 10 -3 0.05721 -0.71990 0.5
I Tb CH 5.60121 • 10 -3 - 1 . 7 7 7 4 • 10 -4 - 6 . 0 7 3 7 x 10 -2 -7.9452 • 10 -5 0.44700 0 .9 8 9 6 0 0.5
I M SG 0.42238 3.1886 • 10 -4 -0.200996 -4.2451 x 10 -4 -0.00843 1.11782 0.5
I M CH 4.23900 x 10 2 - 5 . 6 9 4 6 • 10 -4 - 6 . 8 3 6 x 10 -2 0.0 0 .1 6 5 6 0 0,82910 0.9
I v38(loo SG 0. 26 37 60 1.7458 x 10 -2 0.231043 -1.8441 • 10 -2 -0.01128 0.77078 0.6
I v38(lOO CH 8.71600 • 10 -2 6.1396 • 10 -3 - 7 . 0 1 9 • 10 -2 -2.5935 • 10 -3 0.05166 0.84599 0.4
CH Tb SG 3.47028 1.4850 • 10 -2 16.94020 -0.01.2491 -2.72522 -6.79769 2.5
CH Tb I 8.39640 • 10 -13 7.7171 • 10 -3 71.65310 --0.02.088 --1.37730 --13.61390 1.7
CH M SG 2.35475 9.3485 x 10 -3 4.74695 - 8 . 0 1 7 2 • 10 -3 -0.68418 -0.76820 2.6
CH M I 2.9004 • 10 -23 7.8276 x 10 -3 60.34840 - 2 . 4 4 5 0 • 10 -2 -0.37884 -12.34051 1.7
CH v38(100 SG 2.52300 • 10-12 0.48281 29.98797 -0.55768 -0.14657 -20.31303 2.5
CH v38(lO0 I 2.1430 • 10 -12 0.28320 53.73160 -0.91085 -0.17158 -10.88065 1.6
Coefficients for Eq. (2.40) have been obtained from data on approximately 140 pure hydrocarbons in the molecular weight range of 70-300 and boiling point range of 300-620 K. Units: Vc in cma/g, Tb and
Tc in K, Pc in bar, and v38(100)in cSt. With permission from Ref. [29].
50 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
a similar approach as the constants of Eq. (2.38) were ob- Nc. Later, this type of correlation was used by other investiga-
tained. Similarly, Eq. (2.40) can be applied to hydrocarbon tors to correlate Tc and Pc for n-alkanes and alkanols [58-60].
systems in the molecular weight range of 70-280, which is Based on the above discussion, M or Tb may also be used in-
approximately equivalent to the boiling range of 30-350~ stead of Nc. Equation (2.41) suggests that for extremely high
("-80-650~ However, they may be used up to C22 o r molec- molecular weight hydrocarbons (M -+ or critical tempera-
ular weight of 300 (~boiling point 370~ with good accuracy. ture or pressure approaches a finite value (Tcoo, Pc~). While
In obtaining the constants, Eq. (2.40) was first converted into there is no proof of the validity of this claim, the above equa-
linear form by taking logarithm from both side of the equation tion shows a good capability for correlating properties of
and then using a linear regression program in a spreadsheet n-alkanes for the molecular weight range of interest in prac-
the constants were determined. The value of R 2 (index of cor- tical applications.
relation) is generally above 0.99 and in some cases near 0.999. Based on Eq. (2.41), the following generalized correlation
However, when viscosity or CH parameters are used the R 2 was used to characterize hydrocarbons within each homolo-
values are lower. For this reason use of kinematic viscosity or gous hydrocarbon group:
CH weight ratio should be used as a last option when other pa-
(2.42) ln(0~ - O) = a - b M c
rameters are not available. Properties of heavy hydrocarbons
are discussed in the next section. When Eq. (2.40) is applied The reason for the use of molecular weight was its avail-
to petroleum fractions, the choice of input parameters is de- ability for heavy fractions in which boiling point data may
termined by the availability of experimental data; however, not be available due to thermal cracking. For four groups
when a choice exists the following trends determine the char- of n-alkanes, n-alkylcycopentanes, n-alkylcyclohexanes, and
acterizing power of input parameters used in Eq. (2.39) or n-alkylbenzenes, constants in Eq. (2.42) were determined us-
(2.40): The first choice for 01 is Tb, followed by M, and then ing experimental data reported in the 1988 edition of API-TDB
v38(100), while for the parameter 02 the first choice is SG, fol- [2] and 1986 edition of TRC [21]. The constants for TM, Tb, SG,
lowed by parameters I and CH. Therefore the pair of (M, SG) dE0, I , Tbr (Tb/Tc), Pc, dc, w, and cr are given in Table 2.6 [31].
is preferable to (M, CH) when the choice exists. Carbon number range and absolute and average absolute de-
viations (AAD) for each property are also given in Table 2.6.
Errors are generally low and within the accuracy of the ex-
2.3.3 Prediction of Properties o f H e a v y
perimental data. Equation (2.42) can be easily reversed to
Pure Hydrocarbons
estimate M from Tb for different families if Tb is chosen as the
One of the major problems in characterization of heavy characterizing parameter. Then estimated M from Tb can be
petroleum fractions is the lack of sufficient methods to predict used to predict other properties within the same group (fam-
basic characteristics of heavy hydrocarbons. As mentioned in ily), as is shown later in this chapter. Similarly if Nc is chosen
the previous section, Eqs. (2.38) or (2.40) can be applied to as the characterization parameter, M for each family can be
hydrocarbons up to molecular weight of about 300. Crude oils estimated from Nc before using Eq. (2.42) to estimate various
and reservoir fluids contain fractions with molecular weights properties. Application and definition of surface tension are
higher than this limit. For example, products from vacuum discussed in Chapter 8 (Sec 8.6).
distillation have molecular weight above this range. For such Constants given in Table 2.6 have been obtained from the
fractions application of either Eq. (2.38) or (2.40) leads to properties of pure hydrocarbons in the carbon number ranges
some errors that will affect the overall property of the whole specified. For TM, Tb, SG, d, and I, properties of compounds
crude or fluid. While similar correlations may be developed up to C40 were available, but for the critical properties values
for higher molecular weight systems, experimental data are up to C20 were used to obtain the numerical constants. One
limited and most data (especially for critical properties for condition imposed in obtaining the constants of Eq. (2.42)
such compounds) are predicted values. As mentioned in the for the critical properties was the criteria of internal con-
previous section, the heavy hydrocarbons are more complex sistency at atmospheric pressure. For light compounds crit-
and two parameters may not be sufficient to correlate prop- ical temperature is greater than the boiling point (Tbr < 1)
erties of these compounds. and the critical pressure is greater than 1 atm (Pc > 1.01325
One way to characterize heavy fractions, as is discussed in bar). However, this trend changes for very heavy compounds
the next chapter, is to model the fraction as a mixture of pseu- where the critical pressure approaches 1 atm. Actual data
docompounds from various homologous hydrocarbon fami- for the critical properties of such compounds are not avail-
lies. In fact, within a single homologous hydrocarbon group, able. However, theory suggests that when Pc --~ 1.01325 bar,
such as n-alkanes, only one characterization parameter is suf- Tc ~ Tb or Tb~ ~ 1. And for infinitely large hydrocarbons
ficient to correlate the properties. This single characterization when Nc --~ ~ (M --~ oo), Pc --~ 0. Some methods developed
parameter should be one of those parameters that best char- for prediction of critical properties of hydrocarbons lead to
acterizes properties in the vertical direction such as carbon Tbr =- 1 as Nc -~ oo[43]. This can be true only if both Tc and
number (Nc), Tb, or M. As shown in Table 2.4, parameters Tb approach infinity as Nc ~ oo. The value of carbon num-
SG, I20, and CH weight ratio are not suitable for this purpose. ber for the compound whose Pc = 1 atm is designated by N~.
Kreglewiski and Zwolinski [57] used the following relation to Equation (2.42) predicts values of Tb~ = 1 at N~ for different
correlate critical temperature of n-alkanes: homologous hydrocarbon groups. Values of N~ for different
hydrocarbon groups are given in Table 2.7. In practical ap-
(2.41) ln(0~ - O) = a - bN2c/3
plications, usually values of critical properties of hydrocar-
where 0~ represents value of a property such as Tc at Nc -~ ~ , bons and fractions up to C45 or C50 are needed. However,
and 0 is the value of Tc for the n-alkane with carbon number of accurate prediction of critical properties at N~* ensures that
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 51
1.2
~n-alkanes
1.1
. . . . n-all~,lcyclope~taaes
. . . . . ,-alkylbenzenes ~ / ~;~. . . _ .
1.0
b.
o 0.9
~ 0.8
o
o
o
0.7
./
0.6
5 L i i ~ i j J~l i i i I i i ill K i J , , J ,I
Molecular Weight
do~ = 0.8541 and no~ = 1.478 (Ioo = 0.283), while the values Pan et al. [63, 64] also r e c o m m e n d use of the following re-
obtained t h r o u g h Eq. (2.42) (see Table 2.6) are Tboo= 1070, lation for the acentric factor of aromatics for hydrocarbons
doo = 0.859, and Ioo = 0.2833. One can see h o w close the val- with M < 800:
ues are although they have been derived by two different
(2.44) In to = -36.1544 + 30.94M 0"026261
methods. However, these values are of little practical applica-
tion as long as a proposed correlation satisfies the condition and when M > 800, to = 2.0. Equation (2.42) is r e c o m m e n d e d
of Tbr = 1 at Pc = 1.0133 bar. Equation (2.42) will be used later for calculation of other t h e r m o d y n a m i c properties based on
in Chapter 4 to develop physical properties of single carbon the evaluation made on t h e r m o d y n a m i c properties of waxes
n u m b e r (SCN) cuts up to C50 for the estimation of properties and asphaltenes [63, 64].
of heavy crude oils and reservoir oils. Graphical presentation For homologous h y d r o c a r b o n groups, various correlations
of Eq. (2.42) for Tbr and Tc versus molecular weight of dif- m a y be found suitable for the critical properties. For example,
ferent hydrocarbon families is shown in Figs. 2.1 and 2.2 for another relation that was found to be applicable to critical
molecular weights up to 3000 (Nc ~ 214). pressure of n-alkyl families is in the following form:
One direct application of critical properties of homologous
hydrocarbons is to calculate phase equilibrium calculations (2.45) Pc = (a + bM)-"
for wax precipitation and cloud point of reservoir fluids and where Pc is in bar and M is the molecular weight of pure hy-
crude oils as shown by Pan et al. [63, 64]. These investigators drocarbon from a homologous group. Constant n is greater
evaluated properties calculated through Eq. (2.42) and mod- than unity and as a result as M ~ oo we have Pc ~ 0, which
ified this equation for the critical pressure of PNA hydrocar- satisfies the general criteria for a Pc correlation. Based on data
bons with molecular weight above 300 t h r o u g h the following on Pc of n-alkanes from C2 to C22, as given in Table 2.1, it was
relation: found that n = 1.25, a = 0.032688, and b = 0.000385, which
(2.43) Pc = a - b e x p ( - c M ) gives R 2 = 0.9995 with average deviation of 0.75% for 21 com-
pounds. To show the degree of extrapolation of this equation,
where a, b, and c are given for the three h y d r o c a r b o n groups if data from C2 to C10 (only nine compounds) are used to
in Table 2.8 [64]. However, Eq. (2.43) does not hold the inter-
nal consistency at Pc of I atm, which was imposed in deriving
the constants of Eq. (2.42). But this m a y not affect results for TABLE 2.8---Coefficients of Eq. (2.43).
practical calculations as critical pressures of even the heaviest Coefficient Paraffins Naphthenes Aromatics
a 0.679091 2.58854 4.85196
c o m p o u n d s do not reach to atmospheric pressure. A compar- b -22.1796 -27.6292 -42.9311
ison between Eq. (2.42) and (2.43) for the critical pressure c 0.00284174 0.00449506 0.00561927
of paraffins, naphthenes, and aromatics is shown in Fig. 2.3. Taken from PanetaL[63,64].
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 53
700
* API-TDB
--n-ahkanes
. . . . . . . n-alkylcyclopentanes
. . . . n-alkylbenzenes
500
?
o)
#
-~ 300
100
lO 100 1000
Molecular Weight
50 o
",x
*" t . DIPPR Data
.~ R-S: Eq. 2.42
40 ., ~ - - n-alk~es
~ 20
10
Molecular Weight
TABLE 2.9--Constants in Eqs. (2.46a and 2.46b) 0 = al exp(bl01 + ClSG + dl01SG) 0~SGf for various properties of heavy hydrocarbons.
o o1 al bi cl dl el fl AAD%
Tc Tb 35.9413 -6.9 • 10-4 -1.4442 4.91 • 10-4 0.7293 1.2771 0.3
Pc Tb 6.9575 -0.0135 -0.3129 9.174 • 10 3 0.6791 -0.6807 5.7
Vc Tb 6.1677 x 101~ -7.583 • 10-3 -28.5524 0.01172 1.20493 17.2074 2.5
I Tb 3.2709 • 10-3 8.4377 • 10-4 4.59487 -1.0617 • 10-3 0.03201 -2.34887 0.l
d20 Tb 0.997 2.9 • 10-4 5.0425 -3.1 • 10-4 -0.00929 1.01772 0.07
0 01 32 b2 c2 d2 e2 ~ AAD%
Tb M 9.3369 1.65 • 10-4 1.4103 -7.5152 x 10-4 0.5369 -0.7276 0.3
Tc M 218.9592 -3.4 • 10-4 -0.40852 -2.5 • 10-5 0.331 0.8136 0.2
Pc M 8.2365 • 104 -9.04 • 10-3 -3.3304 0.01006 -0.9366 3.1353 6.2
Vc M 9.703 • 106 -9.512 • 10-3 -15.8092 0.01111 1.08283 10.5118 1.6
I M 1.2419 x 10-2 7.27 • 10-4 3.3323 -8.87 x 10-4 6.438 • 10-3 -1.61166 0.2
d20 M 1.04908 2.9 • 10 4 -7.339 • 10-2 -3.4 • 10 - 4 3.484 • 10-3 1.05015 0.09
Data generatedfrom Eq. (2.42) have been used to obtain these constants. Units:Vcin cm3/mol;To,and Tbin K;
Pc in bar; d20 in g/cm3 at 20~ Equations are recommendedfor the carbon range of C20-C50;however, they may be used for the C5-C20with lesser degree of
accuracy.
o b t a i n a a n d b in the above e q u a t i o n we get a = 0.032795 a n d extended in terms of three parameters, Tb, d20, a n d M, to es-
b = 0.000381. These coefficients give R 2 ----0.9998 b u t w h e n it timate the critical properties of both h y d r o c a r b o n s a n d n o n -
is used to estimate Pc from C2 to C22 AAD of 0.9% is obtained. h y d r o c a r b o n s [37].
These coefficients estimate Pc of t/-C36 a s 6.45 b a r versus value
given in DIPPR as 6.8. This is a good extrapolation power. I n Tc, Pc, Vc = exp[a + bM + cTb + dd20 + eTbd2o] MfT~+hMdi2o
Eq. (2.45) one m a y replace M by Tb or Nc a n d o b t a i n n e w (2.47)
coefficients for cases that these parameters are known.
Based o n the critical properties of more t h a n 170 hydrocar-
Properties of pure c o m p o u n d s predicted t h r o u g h Eqs.
bons from C1 to C18 a n d more t h a n 80 n o n h y d r o c a r b o n s , such
(2.42) a n d (2.43) have b e e n used to develop the following gen-
as acids, sulfur c o m p o u n d s , nitriles, oxide gases, alcohols,
eralized correlations in terms of (Tb, SG) or (M, SG) for the
halogenated c o m p o u n d s , ethers, amines, a n d water, the n i n e
basic properties of heavy h y d r o c a r b o n s from all h y d r o c a r b o n
parameters i n Eq. (2.47) were d e t e r m i n e d a n d are given in
groups in the C6-C50 range [65].
Table 2.10. I n using Eq. (2.47), the c o n s t a n t d should not be
Tc, Pc, Vc, I, d 2 0 - - a l [exp(blTb + c l S G + d l T b S G ) ] T~1 SG fi m i s t a k e n with p a r a m e t e r d20 used for liquid density at 20~
(2.46a) As in the other equations i n this chapter, values of Tb a n d Tc
are in kelvin, Pc is in bar, a n d Vc is in cma/g. P a r a m e t e r d20 is
Tb, Tc, Pc, Vc, I, d20 =a2[exp(bEM+CESG the liquid density at 20~ a n d 1.0133 b a r in g/cm 3. For light
(2.46b) +d2MSG)] M e2 SG f2 gases such as m e t h a n e (C1) or ethane (C2) in which they are in
the gaseous state at the reference conditions, a fictitious value
where Vc in these relations is in cm3/mol. Constants a l - f l of dE0 was o b t a i n e d t h r o u g h the extrapolation of density val-
a n d a2-f2 in these relations are given i n Table 2.9. These cor- ues at lower t e m p e r a t u r e given by Reid et al. [4]. The values
relations are r e c o m m e n d e d for h y d r o c a r b o n s a n d p e t r o l e u m of d20 for some gases f o u n d i n this m a n n e r are as follows:
fractions in the c a r b o n n u m b e r range of C20-C50. Although a m m o n i a , NH3 (0.61); n i t r o u s oxide, N20 (0.79); methane,
these equations m a y be used to predict physical properties C1 (0.18); ethane, C2 (0.343); propane, C3 (0.5); n-butane, nC4
of h y d r o c a r b o n s in the range of C6-C20, if the system does (0.579); isobutane, iC4 (0.557); nitrogen, N2 (0.135); oxygen,
n o t c o n t a i n heavy h y d r o c a r b o n s Eqs. (2.38) a n d (2.40) are 02 (0.22); hydrogen sulfide, H2S (0.829); a n d hydrogen chlo-
recommended. ride, HC1 (0.837). I n some references different values for liq-
uid densities of some of these c o m p o u n d s have been reported.
For example, a value of 0.809 g/cm 3 is reported as the den-
2.3.4 Extension of Proposed Correlations
sity of N2 at 15.5~ a n d 1 a t m by several authors in reservoir
to Nonhydrocarbon Systems
engineering [48, 51]. This value is very close to the density of
E q u a t i o n s (2.38) a n d (2.40) c a n n o t be applied to systems con- N2 at 78 K [4]. The critical t e m p e r a t u r e of N2 is 126.1 K a n d
t a i n i n g hydrocarbons, such as m e t h a n e a n d ethane, or hydro-
gen sulfide. These equations are useful for h y d r o c a r b o n s with
c a r b o n n u m b e r s above Cs a n d are not applicable to n a t u r a l TABLE 2.10--Constants for Eq. (2.47).
o~ Vc
gases or refinery gases. E s t i m a t i o n of the properties of n o n h y - Constants Tc. K Pc, MPa cm3'/g
d r o c a r b o n systems is b e y o n d the objective of this book. But a 1.60193 10.74145 -8.84800
in reservoir fluids, c o m p o u n d s such as light h y d r o c a r b o n s b 0.00558 0.07434 -0.03632
or H2S a n d CO2 m a y be present. To develop a generalized c -0.00112 -0.00047 -0.00547
correlation in the form of Eq. (2.40) that includes n o n h y d r o - d -0.52398 -2.10482 0.16629
e 0.00104 0.00508 -0.00028
carbons, usually a third p a r a m e t e r is needed to consider the f -0.06403 -1.18869 0.04660
effects of polarity. I n fact Vetere [66] has defined a polarity g 0.93857 -0.66773 2.00241
factor in terms of the molecular weight a n d boiling p o i n t to h -0.00085 -0.01154 0.00587
predict properties of polar c o m p o u n d s . E q u a t i o n (2.40) was i 0.28290 1.53161 -0.96608
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 55
therefore at temperature 288 or 293 K it cannot be a liquid specific gravity. Equation (2.38) for molecular weight is [28]
and values reported for density at these temperatures are fic-
titious. In any case the values given here for density of N2, (2.50) M = 1.6607 x 10-4T21962SG -1"0164
CO2, C1, Ca, and H2S should not be taken as real values and
they are only recommended for use in Eq. (2.47). It should This equation fails to properly predict properties for hydro-
be noted that dE0 is the same as the specific gravity at 20~ carbons above C2s. This equation was extensively evaluated
in the SI system (d4a~ This equation was developed based on for various coal liquid samples along with other correlations
the fact that nonhydrocarbons are mainly polar compounds by Tsonopoulos et al. [34]. They recommended this equation
and a two-parameter potential energy relation cannot rep- for the estimation of the molecular weight of coal liquid frac-
resent the intermolecular forces between molecules, there- tions. Constants in Eq. (2.40) for molecular weight, as given
fore a third parameter is needed to characterize the system. in Table 2.5, were modified to include heavy hydrocarbons up
This method would be particularly useful to estimate the bulk to molecular weight of 700. The equation in terms of Tb and
properties of petroleum fluids containing light hydrocarbons SG becomes
as well as nonhydrocarbon gases. Evaluation of this method
is presented in Section 2.9. M = 42.965[exp(2.097 x 10-4Tb -- 7.78712SG
(2.51) + 2.08476 x 10-3TbSG)]Tbl26~176 4"98308
2.4 PREDICTION OF MOLECULAR WEIGHT, This equation can be applied to hydrocarbons with molecular
BOILING POINT, AND SPECIFIC GRAVITY weight ranging from 70 to 700, which is nearly equivalent to
boiling point range of 300-850 K (90-1050~ and the API
Molecular weight, M, boiling point, Tb, and specific gravity,
gravity range of 14.4-93. These equations can be easily con-
SG, are perhaps the most important characterization param-
verted in terms of Watson K factor (Kw) and API degrees
eters for petroleum fractions and many physical properties
using their definitions through Eqs. (2.13) and (2.4). A graph-
may be calculated from these parameters. Various methods
ical presentation of Eq. (2.51) is shown in Fig. 2.4. (Equation
commonly used to calculate these properties are presented
(2.51) has been recommended by the API as it will be dis-
here. As mentioned before, the main application of these cor-
cussed later.) Equation (2.51) is more accurate for light frac-
relations is for petroleum fractions when experimental data
tions (M < 300) with an %AAD of about 3.5, but for heavier
are not available. For pure hydrocarbons either experimental
fractions the %AAD is about 4.7. This equation is included in
data are available or group contribution methods are used to
the API-TDB [2] and is recognized as the standard method
estimate these parameters [4]. However, methods suggested
of estimating molecular weight of petroleum fractions in the
in Chapter 3 to estimate properties of petroleum fractions are
industry.
based on the method developed from the properties of pure
For heavy petroleum fractions boiling point may not be
hydrocarbons in this chapter.
available. For this reason Riazi and Daubert [67] developed
a three-parameter correlation in terms of kinematic viscosity
2.4.1 Prediction of Molecular Weight based on the molecular weight of heavy fractions in the range
of 200-800:
For pure hydrocarbons from homologous groups, Eq. (2.42)
can be reversed to obtain the molecular weight from other [. (-1.2435+1.1228SG) .(3.4758-3.038SG)'1 SG-0.6665
M = 223.56 l_v38(loo) u99(21o) j
properties. For example, if Tb is available, M can be estimated
from the following equation: (2.52)
800
o n-alkane
700
600
500
400
300
200
100 y _.
~ _ _
200~
150~
30oC 100~
0 i ! t , I } i
_10000
"~ooo API G r a v i t y
-8000
_~7ooo
-6000 20
=5000 4 BOO
~'4000
~ 3000
--- 2 0 0 0
700
m
-'1000
900
800
7OO
600
500
600
400
o
3OO
oT - -
-200
._~ 5 0 0
~0 (D
100
90
o
o
.(~
>
%
80
70 o 0
60 o
50 E
c 4OO
40 5
30 10
-20
300
!0
g
8
7
6
5
200
4
3
_=
100
E ,
FIG. 2.5---Estimation of molecular weight from Eq. (2.52). Taken from Ref. [67] with per-
mission.
fraction distillation data is not usually available. When d20 is This relation is based on the assumption that the boiling point
not available it may be estimated from SG using the method of extremely large molecules (M --~ c~) approaches a finite
given in Section 2.6.1. value of 1071.28 K. Soreide [52] compared four methods for
the prediction of the boiling point of petroleum fractions:
2.4.1.6 Other Methods (1) Eq. (2.56), (2) Eq. (2.58), (3) Eq. (2.50), and (4) Twu
Twu [30] proposed a set of correlations for the calculation of method given by Eqs. (2.89)-(2.92). For his data bank on
M, Tc, Pc, and Vc of hydrocarbons. Because these correlations boiling point of petroleum fractions in the molecular weight
are interrelated, they are all given in Section 2.5.1. The com- range of 70-450, he found that Eq. (2.50) and the Twn cor-
puterized Winn method is given by Eq. (2.93) in Section 2.5.1 relations overestimate the boiling point while Eqs. (2.56)
and in the form of chart in Fig. 2.12. and (2.58) are almost identical with AAD of about 1%. Since
Eq. (2.56) was originally based on hydrocarbons with a molec-
Example 2.5--For n-Butylbenzene estimate the molecular ular weight range of 70-300, its application to heavier com-
weight from Eqs. (2.50), (2.51), (2.54), and (2.55) using the pounds should be taken with care. In addition, the database
input data from Table 2.1 for evaluations by Soreide was the same as the data used to
derive constants in his correlation, Eq. (2.58). For heavier hy-
drocarbons (M > 300) Eq. (2.57) may be used.
Solution--From Table 2.1, n-butylbenzene has Tb = 183.3~
For pure hydrocarbons from different homologous families
SG-- 0.8660, d = 0.8610, and M = 134.2. Applying various
equations we obtain the following: from Eq. (2.50), M = 133.2 Eq. (2.42) should be used with constants given in Table 2.6 for
with AD-- 0.8%, Eq. (2.51) gives M = 139.2 with AD = 3.7%, Tb to estimate boiling point from molecular weight. A graph-
Eq. (2.54) gives M = 143.4 with AD-- 6.9%, and Eq. (2.55) ical comparison of Eqs. (2.42), (2.56), (2.57), and (2.58) for
gives M = 128.7 with AD = 4.1%. For this pure and light hy- n-alkanes from C5 to C36 with data from API-TDB [2] is shown
drocarbon, Eq. (2.50) gives the lowest error because it was in Fig. 2.6.
mainly developed from the molecular weight of pure hydro-
carbons while the other equations cover wider range of molec- 2.4.3 Prediction of Specific Gravity/API Gravity
ular weight because data from petroleum fractions were also
used in their development, t Specific gravity of hydrocarbons and petroleum fractions is
normally available because it is easily measurable. Specific
gravity and the API gravity are related to each other through
2.4.2 Prediction of Normal Boiling Point Eq. (2.4). Therefore, when one of these parameters is known
2.4.2.1 Riazi-Daubert Correlations the other one can be calculated from the definition of the API
gravity. Several correlations are presented in this section for
These correlations are developed in Section 2.3. The best in-
the estimation of specific gravity using boiling point, molec-
put pair of parameters to predict boiling point are (M, SG) or
ular weight, or kinematic viscosity as the input parameters.
(M, 1). For light hydrocarbons and petroleum fractions with
molecular weight in the range of 70-300, Eq. (2.40) may be 2.4.3.1 Riazi-Daubert Methods
used for boiling point:
These correlations for the estimation of specific gravity re-
Tb = 3.76587[exp(3.7741 x 10-3M + 2.98404SG quire Tb and I or viscosity and CH weight ratio as the input
(2.56) -4.25288 • 10-3MSG)]M~176 -1'58262 parameters (Eq. 2.40). For light hydrocarbons, Eq. (2.40) and
Table 2.5 can be used to estimate SG from different input
For hydrocarbons or petroleum fractions with molecular parameters such as Tb and I.
weight in the range of 300-700, Eq. (2.46b) is recommended:
Tb = 9.3369[exp(1.6514 x 10-4M + 1.4103SG SG = 2.4381 x 107 exp(-4.194 x 10-4Tb -- 23.55351
(2.57) - 7.5152 x 10-4MSG)]M~ -0"7276 (2.59) + 3.9874 x lO-3Tbl)Tb~
Equation (2.57) is also applicable to hydrocarbons having where Tb is in kelvin. For heavy hydrocarbons with molecular
molecular weight range of 70-300, with less accuracy. Esti- weight in the range 300-700, the following equation in terms
mation of the boiling point from the molecular weight and of M and I can be used [65]:
refractive index parameter ( I ) is given by Eq. (2.40) with con-
stants in Table 2.5. The boiling point may also be calculated SG = 3.3131 x 104 exp(-8.77 x 1 0 - a M - 15.0496I
through Kw and API gravity by using definitions of these pa- (2.60) + 3.247 x lO-3MI)M-~176 4"9557
rameters given in Eqs. (2.13) and (2.4).
Usually for heavy fractions, Tb is not available and for this rea-
2.4.2.2 Soreide Correlation son, M and I are used as the input parameters. This equation
Based on extension of Eq. (2.56) and data on the boiling point also may be used for hydrocarbons below molecular weight
of some C7+ fractions, Soreide [51, 52] developed the follow- of 300, if necessary. The accuracy of this equation is about 0.4
ing correlation for the normal boiling point of fractions in the %AAD for 130 hydrocarbons in the carbon n u m b e r range of
range of 90-560~ C7-C50 (M ~ 70-700).
For heavier fractions (molecular weight from 200 to 800)
Tb = 1071.28 -- 9.417 x 104 exp(--4.922 x 10-aM and especially when the boiling point is not available the fol-
(2.58) - 4.7685SG + 3.462 x 10-3MSG) M-~176 TM lowing relation in terms of kinematic viscosities developed by
2. CHARACTERIZATION AND P R O P E R TIE S OF PURE H YD R O CA R B O N S 59
600
o
o API Data 9 r
. . . . . . . Eq. 2.42 ~ ,,~,~
500 .... Eq. 2.56
400
P
300
200
I00
0 1 i I ii i I i I I L J I I I I I I I i I I I i
0 5 10 15 20 25 30 35 4O
Carbon Number
50 0.78
40
0,88
30
o
o
20
0.98
l0
0 1.08
10 100 1000
Kinematic Viscosity at 37.8 ~ cSt
FIG. 2.7--API gravity and viscosity of heavy hydrocarbon fractions by Eq. (2.61).
60 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
Riazi and Daubert may be used [67]: Riazi-Daubert or Riazi methods. These equations are recom-
(2.61) SG ---~ 0.1157 r -0.16161
mended only for hydrocarbons in the molecular weight range
0.7717 [v38000) ] • [1299(210
)] of 70-300 and have been widely used in industry [2, 47, 49, 51,
in which v38~100~and ~d99(210) are kinematic viscosities in cSt at 54, 70]. However, these correlations were replaced with more
i00 and 210~ (37.8 and 98.9~ respectively. Equation (2.61) accurate correlations presented by Eq. (2.40) and Table 2.5 in
is shown in Fig. 2.7 and has also been adopted by the API and terms of Tb and SG as given below:
is included in 1987 version of API-TDB [2]. This equation
Tc = 9.5233[exp(-9.314 x 10-4Tb - - 0 . 5 4 4 4 4 2 S G
gives an AAD of about 1.5% for 158 fractions in the molecular
weight range of 200-500 (~SG range of 0.8-1.1). (2.65) +6.4791 • 1 0 - 4 T b S G ) ] T ~ 1 7 6 0"53691
For coal liquids and heavy residues that are highly aro-
matic, Tsonopoulos et al. [58] suggest the following relation Pc = 3.1958 x 105[exp(-8.505 x 10-3Tb - - 4 . 8 0 1 4 S G
in terms of normal boiling point (Tb) for the estimation of (2.66) + 5.749 x 10-3TbSG)]Tb~ 4"0846
specific gravity.
These correlations were also adopted by the API and have
SG = 0.553461 + 1.15156To - 0.708142To 2 + 0.196237T3 been used in m a n y industrial computer softwares under the
(2.62) API method. The same limitations and units as those for Eqs.
(2.63) and (2.64) apply to these equations. For heavy hydro-
where To = (1.8Tb- 459.67) in which Tb is in kelvin. This
carbons (>C20) the following equations are obtained from
equation is not recommended for pure hydrocarbons or
Eq. (2.46a) and constants in Table 2.9:
petroleum fractions and has an average relative deviation of
about 2.5% for coal liquid fractions [58]. For pure homolo- Tc = 35.9413[exp(-6.9 x 10-4Tb -- 1.4442SG
gous hydrocarbon groups, Eq. (2.42) with constants given in (2.67) +4.91 X 10-4TbSG)]Tb~ 1"2771
Table 2.6 for SG can be used. Another approach to estimate
specific gravity is to use the Rackett equation and a known
Pc = 6.9575[exp(-1.35 • 10-2Zb -- 0.3129SG
density data point at any temperature as discussed in Chap-
ter 5 (Section 5.8). A very simple and practical method of (2.68) + 9.174 x 1 0 - 3 T b S G ) ] T ~ -0"6807
estimating SG from density at 20~ d, is given by Eq. (2.110),
If necessary these equations can also be used for hydrocar-
which will be discussed in Section 2.6.1. Once SG is esti-
bons in the range of C5-C20 with good accuracy. Equation
mated the API gravity can be calculated from its definition,
(2.67) predicts values of Tc from C5 to Cs0 with %AAD of 0.4%,
i.e., Eq. (2.4).
but Eq. (2.68) predicts Pc with AAD of 5.8%. The reason for
this high average error is low values of Pc (i.e, a few bars) at
higher carbon numbers which even a small absolute deviation
2.5 P R E D I C T I O N OF CRITICAL shows a large value in terms of relative deviation.
P R O P E R T I E S A N D A C E N T R I C FACTOR
2.5.1.2 API Methods
Critical properties, especially the critical temperature and
The API-TDB [2] adopted methods developed by Riazi and
pressure, and the acentric factor are important input param-
Daubert for the estimation of pseudocritical properties of
eters for EOS and generalized correlations to estimate phys-
petroleum fractions. In the 1982 edition of API-TDB, Eqs.
ical and thermodynamic properties of fluids. As shown in
(2.63) and (2.64) were recommended for critical temperature
Chapter 1 even small errors in prediction of these proper-
and pressure of petroleum fractions, respectively, but in its
ties greatly affect calculated physical properties. Some of the
editions from 1987 to 1997, Eqs. (2.65) and (2.66) are included
methods widely used in the petroleum industry are given in
after evaluations by the API-TDB Committee. For pure hy-
this section. These procedures, as mentioned in the previous
drocarbons, the methods recommended by API are based on
sections, are mainly developed based on critical properties of
group contribution methods such as Ambrose, which requires
pure hydrocarbons in which validated experimental data are
the structure of the compound to be known. These methods
available only up to C18. The following correlations are given
are of minor practical use in this book since properties of
in terms of boiling point and specific gravity. For other in-
pure compounds of interest are given in Section 2.3 and for
put parameters, appropriate correlations given in Section 2.3
petroleum fractions the bulk properties are used rather than
should be used.
the chemical structure of individual compounds.
w h e r e Tb a n d Tc are in kelvin a n d Pc is in bar. I n these equa- i n c l u d e d in s o m e references [49Z1. However, the Twu correla-
tions a t t e m p t s were m a d e to keep internal consistency a m o n g tions a l t h o u g h b a s e d on the s a m e f o r m a t as the KLS o r LC
Tc a n d Pc that at Pc equal to 1 atm, Tc is c o i n c i d e d with nor- require i n p u t p a r a m e t e r s of Tb a n d SG a n d are a p p l i c a b l e to
m a l boiling point, Tb. The correlations were r e c o m m e n d e d h y d r o c a r b o n s b e y o n d C20. F o r heavy h y d r o c a r b o n s s i m i l a r
by the a u t h o r s for the m o l e c u l a r range of 70-700 (~C5-C50). to the a p p r o a c h of L e e - K e s l e r [12], Twu [30] used the crit-
However, the values of Tc a n d Pc for c o m p o u n d s with c a r b o n ical p r o p e r t i e s b a c k calculated f r o m v a p o r p r e s s u r e d a t a to
n u m b e r s greater t h a n Cls used to develop the above correla- e x p a n d his d a t a b a n k on the critical c o n s t a n t s of p u r e hydro-
tions were not b a s e d on e x p e r i m e n t a l evidence. c a r b o n c o m p o u n d s . F o r this r e a s o n the Twu correlations have
f o u n d a w i d e r range of application. The Twu correlations for
2.5.1.4 Cavett Method the critical properties, specific gravity, and m o l e c u l a r weight
Cavett [26] developed e m p i r i c a l correlations for Tc a n d Pc in of n-alkanes are as follows:
t e r m s of boiling p o i n t a n d API gravity, w h i c h are still available
T~~ = Tb(0.533272 + 0.34383 x 10 -3 • Tb
in s o m e process s i m u l a t o r s as an o p t i o n a n d in s o m e cases
give g o o d estimates of Tc a n d Pc for light to m i d d l e distillate +2.52617 x 10 -7 • T~ - 1.658481 • 10 -1~ x T3
p e t r o l e u m fractions. (2.73) +4.60773 x 1024 x Tb-13)-1
(2.85) ASGv = exp[4(SG ~ - SG2)] - 1 where Tb and Tc are in kelvin and Pc is in bar. Comparing
values estimated from these correlations with the values from
Critical pressure the original figures gives AAD of 2, 1, and 1.5% for M, To, and
Pc, respectively, as reported in Ref. [36]. In the literature these
(2.86) Pc = P~(TclT~) x (Vc/Vc)[(1 + 2fv)l(1 -2fp)] 2 equations are usually referred as Winn or Sim-Daubert and
are included in some process simulators. The original Winn
fe = ASGe[(2.53262 - 34.4321/T~/2 - 2.30193Tb/1000) nomograph for molecular weight and some other properties
is given in Section 2.8.
+( -- 11.4277 + 187.934/Tb1/2 + 4.11963Tb/lOOO)ASGv]
(2.87) 2.5.1.7 Tsonopoulos Correlations
Based on the critical properties of aromatic compounds,
(2.88) ASGv = exp[0.5(SG ~ - SG)] - 1 Tsonopoulos et al. [34] proposed the following correlations
for estimation of Tc and Pc for coal liquids and aromatic-rich
Molecular weight
fractions.
(2.89) ln(M) = (ln M~ + 2 f~)/(1 - 2fM)] 2
lOgl0 Tc = 1.20016 + 0.61954(log10 Tb)
f?a = ASGm[x + (-0.0175691 + 0.143979/T~/2)ASGM] (2.96) + 0.48262(1og10 SG) + 0.67365(log10 SG) 2
(2.90) log10 Pc = 7.37498 - 2.15833(log10 Tb)
(2.91) X --- ]0.012342 - 0.244541/T1/2[ (2.97) + 3.35417(log10 SG) + 5.64019(log10 SG) 2
where Tb and Tc are in kelvin and Pc is in bar. These correla-
(2.92) ASGM = exp[5(SG ~ - SG)] - 1 tions are mainly recommended for coal liquid fractions and
In the above relations Tb and Tc are in kelvin, Vc is in cm3/mol, they give average errors of 0.7 and 3.5% for the estimation of
and Pc is in bar. One can see that these correlations should critical temperature and pressure of aromatic hydrocarbons.
be solved simultaneously because they are highly interrelated
to each other and for this reason relations for estimation of 2.5.2 Prediction o f Critical Volume
M and Vcbased on this method are also presented in this part.
Critical volume, Vo is the third critical property that is not
Example 2.6---Estimate the molecular weight of n-eicosane directly used in EOS calculations, but is indirectly used to
(C20H42) from its normal boiling point using Eq. (2.49) and estimate interaction parameters (kii) needed for calculation
the Twu correlations. of mixture pseudocritical properties or EOS parameters as
will be discussed in Chapter 5. In some corresponding state
Solution--n-Eicosane is a normal paraffin whose molecular correlations developed to estimate transport properties of flu-
weight and boiling point are given in Table 2. I as M = 282.55 ids at elevated pressure, reduced density (Vc/V) is used as the
and Tb = 616.93 K. Substituting Tb in Eq. (2.49) gives M = correlating parameter and values of Vc are required as shown
282.59 (%AD = 0.01%). Using the Twu method, first an initial in Chapter 8. Critical volume is also used to calculate critical
guess is calculated through Eq. (2.79) as M ~ = 238 and from compressibility factor, Zc, as shown by Eq. (2.8).
iteration the final value of M ~ calculated from Eq. (2.78) is 2.5.2.1 Riazi-Daubert Methods
281.2 (%AD = 0.48%). Twu method for estimation of proper-
ties of hydrocarbons from other groups is shown later in the A simplified equation to calculate Vc of hydrocarbons in the
next example. r range of C5-C20 is given by Eq. (2.38) as follows.
(2.98) Vc = 1.7842 x 10-4T2"a829SG-1"683
2.5.1.6 Winn-Mobil Method
in which Vc is in cma/mol and Tb is in kelvin. When evalu-
Winn [25] developed a convenient nomograph to estimate var- ated against more than 100 pure hydrocarbons in the carbon
ious physical properties including molecular weight and the range of C5--C20 an average error of 2.9% was observed. This
pseudocritical pressure for petroleum fractions. Mobil [73] equation may be used up to Css with reasonable accuracy. For
proposed a similar nomograph for the estimation of pseudo- heavier hydrocarbons, Vc is given by Eq. (2.46a) and in terms
critical temperature. The input data in both nomographs are of Tb and SG is given as
boiling point (or Kw) and the specific gravity (or API gravity).
As part of the API project to computerize the graphical meth- Vr = 6.2 x 101~ x 10-3Tb -- 28.5524SG
ods for estimation of physical properties, these nomographs (2.99) + 1.172 x aO-2TbSG)]Tl2~ 17"2074
were reduced to equation forms for computer applications by
where Vc is in cm3/mol. Although this equation is recom-
Riazi [36] and were later reported by Sire and Daubert [74].
mended for hydrocarbons heavier than C20 it may be used, if
These empirically developed correlations have forms similar
necessary, for the range of C5-C50 in which the AAD is about
to Eq. (2.38) and for M, To and Pc are as follows.
2.5%. To calculate Vc from other input parameters, Eqs. (2.40)
(2.93) M = 2.70579 x 10-5T~4966SG-1'174 and (2.46b) with Tables 2.5 and 2.9 may be used.
2. CHARACTERIZATION AND PROPERTIES OF PUAE HYDROCARBONS 63
0.30
9 DIPPR Data
....... Eq. 2.103
..... Eq. 2.104
~ Eq. 2.42
84
--,.-.-:,
@
r.)
9 @ ~176
9 "..
020. ~ i t s I i , , . i I , , , i i , , , ,
5 l0 15 20 25
Carbon Number
m e t h o d for M, To, a n d Pc are given by Eqs. (2.93)-(2.95). I n with use of Eq. (2.10). Attempts to correlate co with p a r a m e -
part h, Vc is calculated from the Hall-Yarborough t h r o u g h ters such as Tb a n d SG all have failed. However, for homolo-
Eq. (2.100) a n d Zc is calculated t h r o u g h Eq. (2.8). S u m m a r y gous h y d r o c a r b o n groups the acentric factor can be related to
of results is given i n Table 2.11. No j u d g e m e n t can be m a d e molecular weight as given by Eqs. (2.42) or (2.44). For other
o n accuracy of these different methods t h r o u g h this single- c o m p o u n d s the acentric factor should be calculated t h r o u g h
p o i n t evaluation. However, methods of R i a z i - S a h h a f (Part d) its definition, i.e., Eq. (2.10), with the use of a correlation to
a n d Twu (Part g) give the m o s t accurate results for this par- estimate vapor pressure. Use of a n accurate correlation for
ticular case. The reason is that these m e t h o d s have specific vapor pressure would result i n a more accurate correlation
relations for n-alkanes family a n d n-hexatriacontane is hydro- for the acentric factor. Methods of the calculation of the vapor
c a r b o n from this family. I n addition, the values for the critical pressure are discussed in Chapter 7. There are three simple
properties from DIPPR [20] are estimated values rather t h a n correlations for the e s t i m a t i o n of vapor pressure that can be
true experimental values. # used in Eq. (2.10) to derive corresponding correlations for the
acentric factor. These three methods are presented here.
2.5.4 Prediction of Acentric Factor 2.5.4.1 Lee-Kesler M e t h o d
Acentric factor, w, is a defined p a r a m e t e r that is n o t directly They proposed the following relations for the e s t i m a t i o n of
measurable. Accurate values of the acentric factor can be ob- acentric factor based o n their proposed correlation for vapor
tained t h r o u g h accurate values of T~, Pc, a n d vapor pressure pressure [27].
TABLE 2.11--Prediction of critical properties of n-hexatriacontane from different methods a (Example 2.7).
M Tc, K APe,bar Vc, cma/mol Zc
Part Method(s) Est.** %D Est. %D Est. %D Est. %D Est. %D
Data from DIPPR [20] 507.0 --. 874.0 ... 6.8 .-- 2090.0 --. 0.196 -.-
a R-D: Eq. (2.38) 445.6 -12.1 885.8 1.3 7.3 7.4 1894.4 -9.3 0.188 -4.2
b API Methods 512.7 1.1 879.3 0.6 7.37 8.4 1 8 4 9 . 7 -11.5 0.205 4.6
c R - D (ext.): Eq. (2.46a) ...... 870.3 -0.4 5.54 -18.5 1964.7 -6.0 0.150 -23.3
d R-S: Eqs. 2.42 &2.43 506.9 0 871.8 -0.3 5.93 -12.8 1952.5 -6.6 0.16 -18.4
e L - K Methods 508.1 0.2 935.1 7.0 5.15 -24.3 2425.9 16.0 0 . 1 6 1 -18.0
f Cavett & Eq. (2.104) ...... 915.5 4.7 7.84 15.3 . . . . . . . . . . . .
g Twu 513.8 1.3 882.1 0.9 6.02 -11.4 2010.0 -3.8 0.165 -15.8
h Winn and H - Y 552.0 8.9 889.5 1.77 7.6 11.8 2362.9 13.1 0.243 24.0
aThe references for the methods are (a) R-D: Riazi-Daubert [28]; (b) API: Methods in the API-TDB[2]; (c) Extended Riazi-Dubert [65]; (d) Riazi-Sahhaf [31];
(e) Kesler-Lee[12] and Lee-Kesler[27]; (f) Cavett[26];Twu [31]; (h) Winn [25] and Hall-Yarborough[75]. Est.: Estimatedvalue.%D:% relativedeviationdefined
in Eq. (2.134).
2. CHARACTERIZATION AND PROPERTIES OF PURE HYDROCARBONS 65
f. Lee-Kesler method with To Pc obtained from Part a in density generally decreases with temperature. Variation of
Example 2.6 density with temperature is discussed in Chapter 6. However,
g. Lee-Kesler method with Tc, Pc obtained from Part b in in this section methods of estimation of density at 20~ d20,
Example 2.6 are presented to be used for the characterization methods dis-
h. Lee-Kesler method with To Pc obtained from Part c in cussed in Chapter 3. The most convenient way to estimate d20
Example 2.6 is through specific gravity. As a rule of thumb d20 = 0.995 SG.
i. Lee-Kesler method with To Pc obtained from Part d in However, a better approximation is provided through calcu-
Example 2.6 lation of change of density with temperature (Ad/AT), which
j. Edmister method with Tc, Pc obtained from Part h in Ex- is negative and for hydrocarbon systems is given as [7]
ample 2.6
(2.110) Ad/AT = - 1 0 -3 x (2.34 - 1.9dr)
k. Lee-Kesler method with Tc, Pc obtained from Part e in
Example 2.6 where dr is density at temperature T in g/cm 3. This equation
1. Lee-Kesler method with T~, Pc obtained from Part g in may be used to obtain density at any temperature once a value
Example 2.6 of density at one temperature is known. This equation is quite
m. Tabulate %D for estimated value of acentric factor in each accurate within a narrow temperature range limit. One can
method. use the above equation to obtain a value of density, d20, at
20~ (g/cm 3) from the specific gravity at 15.5~ as
Lee-Kesler method refers to Eq. (2.105) and Kesler-Lee to
Eq. (2.107). (2.111) d20 = SG - 4.5 x 10-3(2.34 - 1.9SG)
Equation (2.111) may also be used to obtain SG from density
SolutionkAll three methods of Lee-Kesler, Edmister, and at 20 or 25~
Korsten require Tb, To, and Pc as input parameters. The
method of Kesler-Lee requires Kw in addition to Tbr. From SG -- 0.9915d20 + 0.01044
(2.112)
definition of Watson K, we get Kw --- 13.64. Substituting these SG = 0.9823d25 + 0.02184
values from various methods one calculates the acentric fac-
Similarly density at any other temperature may be calculated
tor. A summary of the results is given in Table 2.12. The least
through Eq. (2.110). Finally, Eq. (2.38) may also be used to
accurate method is the Kesler-Lee correlations while the most
accurate method is Korsten combined with Eqs. (2.67) and estimate d20 from TB and SG in the following form:
(2.68) for the critical constants, t (2.113) d20 = 0.983719Tb~176176 1~176
This equation was developed for hydrocarbons from C5 to C20;
however, it can be safely used up to C40 with AAD of less than
2.6 P R E D I C T I O N OF DENSITY,
0.1%. A comparison is made between the above three meth-
R E F R A C T I V E I N D E X , CH W E I G H T
ods of estimating d for some n-paraffins with actual data taken
RATIO, A N D F R E E Z I N G P O I N T
from the API-TDB [2]. Results of evaluations are given in
Table 2.13. This summary evaluation shows that Eqs. (2.111)
Estimation of density at different conditions of temperature,
and (2.113) are almost equivalent, while as expected the rule
pressure, and composition (p) is discussed in detail in Chap-
of thumb is less accurate. Equation (2.111) is recommended
ter 5. However, liquid density at 20~ and 1 atm designated by
for practical calculations.
d in the unit of g/cm 3 is a useful characterization parameter
which will be used in Chapter 3 for the compositional analy-
sis of petroleum fractions especially in conjunction with the 2.6.2 Prediction of Refractive Index
definition of refractivity intercept by Eq. (2.14). The sodium
The refractive index of liquid hydrocarbons at 20~ is corre-
D line refractive index of liquid petroleum fractions at 20~
lated through parameter I defined by Eq. (2.14). If parameter
and 1 atm, n, is another useful characterization parameter.
Refractive index is needed in calculation of refractivity inter- I is known, by rearranging Eq. (2.14), the refractive index, n,
cept and is used in Eq. (2.40) for the estimation of various can be calculated as follows:
properties through parameter I defined by Eq. (2.36). More- (2.114) n = (11+~2//) 1/2
over refractive index is useful in the calculation of density and
transport properties as discussed in Chapters 5 and 8. Carbon-
to-hydrogen weight ratio is needed in Chapter 3 for the esti- For pure and four different homologous hydrocarbon com-
mation of the composition of petroleum fractions. Freezing pounds, parameter I is predicted from Eq. (2.42) using molec-
point, TF, is useful for analyzing solidification of heavy com- ular weight, M, with constants in Table 2.6. If boiling point is
ponents in petroleum oils and to determine the cloud point available, M is first calculated by Eq. (2.48) and then I is cal-
temperature of crude oils and reservoir fluids as discussed in culated. Prediction of I through Eq. (2.42) for various hydro-
Chapter 9 (Section 9.3.3). carbon groups is shown in Fig. 2.9. Actual values of refractive
index from API-TDB [2] are also shown in this figure.
For all types of hydrocarbons and narrow-boiling range
2.6.1 Prediction of Density at 20~ petroleum fractions the simplest method to estimate param-
eter I is given by Riazi and Daubert [28] in the form of
Numerical values of d20 for a given compound is very close
Eq. (2.38) for the molecular weight range of 70-300 as follows:
to the value of SG, which represents density at 15.5~ in the
unit of g/cm 3 as can be seen from Tables 2.1 and 2.3. Liquid (2.115) I = 0.3773Tb-~176 0~9182
2. CHARACTERIZATION AND PROPERTIES OF PURE HYDROCARBONS 67
1.6
n-alkylbenzenes
1.5
n-alkylcyclopentanes
4~
1.4
of -0.0004 in Eq. (2.118)) as a function of n20 rather than determined the atomic HC ratio can be calculated from their
a constant. Another approach to estimate refractive index definitions as described in Section 2.1.18:
at temperatures other than 20~ is to assume that specific
refraction is constant for a given hydrocarbon: 11.9147
(2.122) HC (atomic ratio) =
CH(weight ratio)
lr I2o
(2.119) Specific refraction - dr - d2o - constant
E x a m p l e 2.9--Estimate the values of CH (weight) and HC
where I2o is the refractive index parameter at 20~ and Ir is its (atomic) ratios for n-tetradecylbenzene (C20H34) from Eqs.
value at temperature T. Similarly dr is density at temperature (2.120) and (2.121) and compare with the actual value. Also
T. In fact the value of specific refraction is the same at all tem- draw a graph of CH values from C6 to C50 for the three homol-
peratures [38]. If/20, d2o, and dT are known, then IT can be esti- ogous hydrocarbon groups from paraffins, naphthenes, and
mated from the above equation. Value of nr can be calculated aromatics based on Eq. (2.121) and actual values.
from IT and Eq. (2.114). Equation (2.119) has the same accu-
racy as Eq. (2.118), but at the temperatures far from the refer- Solution--The actual values of CH weight and HC
ence temperature of 20 ~C accuracy of both methods decrease. atomic ratios are calculated from the chemical formula
Because of simplicity, Eq. (2.118) is recommended for calcu- and Eq. (2.122) as CH --- (20 x 12.011)/(34 x 1.008) -- 7.01,
lation of refractive index at different temperatures. It is obvi- HC(atomic) = 34/20 = 1.7. From Table 2.1, for n-tetradecyl-
ous that the reference temperature in both Eqs. (2. I 18) and b e n z e n e (C20H34), Tb ~ 627 K and S G = 0.8587. Substitut-
(2.119) can be changed to any desired temperature in which ing these values into Eq. (2.120) gives CH-- 7.000, and from
refractive index is available. Refractive index is also related to Eq. (2.122) atomic HC ratio--1.702. The error from
another property called dielectric constant, e, which for non- Eq. (2.134) is %D = 0.12%. Equation (2.121) gives CH--
polar compounds at any temperature is e -- n 2. For example, 6.998, which is nearly the same as the value obtained from
at temperature of 20~ a paraffinic oil has dielectric constant Eq. (2.120) with the same error. Similarly CH values are cal-
of 2.195 and refractive index of 1.481 (n 2 -- 2.193). Dielectric culated by Eq. (2.121) for hydrocarbons ranging from C6 to
constants of petroleum products may be used to indicate the C50 in three homologous hydrocarbon groups and are shown
presence of various constituents such as asphaltenes, resins, with actual values in Fig. 2.10. #
etc. [11]. However, for more complex and polar molecules
such as muhiring aromatics, this simple relation between e
2.6.4 Prediction of Freezing/Melting Point
and n 2 is not valid and they are related through dipole mo-
ment. Further discussion on the methods of estimation of For pure compounds, the normal freezing point is the same
refractive index is given by Riazi and Roomi [37]. as the melting point, TM. Melting point is mainly a param-
eter that is needed for predicting solid-liquid phase behav-
ior, especially for the waxy oils as shown in Chapter 9. All
2.6.3 Prediction of CH Weight Ratio
attempts to develop a generalized correlation for TM in the
Carbon-to-hydrogen weight ratio as defined in Section 2.1.18 form of Eq. (2.38) have failed. However, Eq. (2.42) developed
is indicative of the quality and type of hydrocarbons present by Riazi and Sahhaf for various homologous hydrocarbon
in a fuel. As will be shown in Chapter 3 from the knowledge groups can be used to estimate melting or freezing point of
of CH value, composition of petroleum fractions may be es- pure hydrocarbons from C7 to C40 with good accuracy (error
timated. CH value is also related to carbon residues as it is of 1-1.5%) for practical calculations [31]. Using this equation
discussed in the next chapter. For hydrocarbons with molec- with appropriate constants in Table 2.6 gives the following
ular weight in the range of 70-300, the relations to estimate equations for predicting the freezing point of n-alkanes (P),
CH values are given through Eq. (2.40) and Table 2.5. In terms n-alkycyclopentanes (N), and n-alkybenzenes (A) from molec-
of TD and SG the relation is also given by Eq. (2.120) which ular weight.
is also recommended for use in prediction of composition of
petroleum fractions [78]. (2.123) TMp = 397 -- exp(6.5096 -- 0.14187M ~
10.0
O8
8.0
O8
7.0
0
5.0
5 15 25 35 45 55
Carbon Number
FIG. 2.10--Estimation of CH weight Ratio from Eq. (2.121) for various families.
50
0
. Y..P-
-50
o Data for n-alkanes
Predicted: R-S Method
N
Predicted: P-F Method
~ -100 A Data for n-alkyl~'clopenlanes
. . . . . . . Predicted: R-S Method
,, K
~
iI .... Predicted: P-F Method
It I Data for n-alkylbenzenes
-150 - - - - - Predicted: R-S Method
m Predicted: P-F Method
-200
5 10 15 20 25
Carbon Number
correlation for n-alkanes is estimation of reference kinematic viscosities and are also in-
cluded in the API-TDB [2]:
(2.126) Trap = 374.5 + 0.02617M - 20172/M
log v38000) =
where TMe is in kelvin. For naphthenes, aromatics, and 4.39371 - 1.94733Kw + 0.12769K~v
isoparaffins the melting point temperature may be esti-
+3,2629 x 10-4API 2 - 1.18246 x 10-ZKwAPI
mated from the following relation given by Pan-Firrozabadi-
0.171617K2w + 10.9943(API) + 9.50663 x 10-2(API) 2 - 0.860218Kw(API)
Fotland [63]. +
(API) + 50.3642 -4.78231 Kw
(2.127) Tra(iP,N,A)= 333.45 -- 419 exp(-0.00855M) (2.128)
or their equivalent parameters Kw and API gravity. Relations Various forms of Eq. (2.130) are given in other sources
developed by Abbott et al. [80] are commonly used for the [2, 11, 17]. Errors arising from use of Eq. (2.130) are better or
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 71
12.5--
165
50
IZO-
1.0~ 40
I/I
1.2
0 x/
u2 ~" 11..5"
30 O
t-
1.6- O
0
g N
1.8
u~ 0 0_
<(
t-
a E
9 ~ O
t-
.~o O
O) 20
0 ~ .! 1.0"
u~
0
4 3o
0
!0
10.5 -
s--t L-2o0
I soo
1 7oo
70~
lO.O-J--
FIG. 2.12--Prediction of kinematic viscosity from Kw and the API gravity. With
permission from Ref. [2],
at least in the same range of errors for the prediction of viscos- from Fig. 2.12 or Eq. (2.130) should be taken with caution.
ity from Eqs. (2.128) and (2.129). Similarly constants AI and An application of this method to estimate kinematic viscosity
B1 in Eq. (2.130) can be determined when values of viscosity of petroleum fractions is demonstrated in Chapter 3. Further
at two temperatures other than 100 and 210~ are known. discussion on the estimation of viscosity is given in Chapter 8.
When vr is being calculated from Eq. (2.130) at temperature
T, a trial and error procedure is required to determine param- Consistency Text--One way to check reliability of a predicted
eter cr. The first estimate is calculated by assuming vr > 1.5 physical property is to perform a consistency test through
cSt and thus Cr = 0. If calculated value is less than 1.5 cSt, different procedures. For example, laboratory reports may
then Cr is calculated from Eq. (2.131). Extrapolated values consist of viscosity data at a temperature other than 38 or
Temperature, Degrees Celsius
u)
O
(D
t-
(9 3
C~
u) <
O (..)
>
O ~z
03 (l)
E
.c_
v 0
~0
Temperature, Degrees Celsius
99~ In many cases kinematic viscosity at 40~ (122~ or than two input parameters are involved. By drawing a straight
60~ (140~ is reported. One may estimate the kinematic line between values of input parameters, a reading can be
viscosities at 38 and 99~ through Eqs. (2.128) and (2.129) made where the straight line intersects with the line (or curve)
and then use the ASTM chart (or Eq. 2.128) to obtain the of the desired property. The best example and widely used
value of viscosity at 40 or 60~ If the interpolated value is far nomogram is the one developed by Winn in 1957 [25]. This
from laboratory data then the estimation method cannot be nomogram, which is also included in the API-TDB [2], re-
trusted and other methods should be considered. lates molecular weight (M), CH weight ratio, aniline point,
Another consistency test can be made through estimation and Watson K to boiling point and specific gravity (or API
of the molecular weight by Eq. (2.52) using estimated vis- gravity) on a single chart and is shown in Fig. 2.14.
cosities by Eqs. (2.128) and (2.129). If value of M calculated Application of this figure is mainly for petroleum fractions
through Eq. (2.52) is near the value of M estimated from Tb and the mean average boiling point defined in Chapter 3
and SG by Eqs. (2.51) or (2.50), then all estimated values can is used as the boiling point, Tb. If any two parameters are
be trusted. Such consistency tests can be extended to all other available, all other characterization parameters can be deter-
physical properties. The following example demonstrates the mined. However, on the figure, the best two input parameters
test method. are Tb and SG that are on the opposite side of the figure.
Obviously use of only M and Tb as input parameters is not
Example 2.11--The viscosity of a pure multiring hydrocar- suitable since they are near each other on the figure and an
bon from an aromatic group (naphthecene type compound) accurate reading for other parameters would not be possible.
with formula C 2 6 H 4 0 has been i~eported in the API RP-42 Similarly CH and SG are not suitable as the only two input
[ 18]. Data available are M = 352.6, SG = 0.9845, and 1299(210) = parameters. Previously the computerized form of the Winn
13.09 cSt. Estimate the kinematic viscosity of this hydrocar- nomogram for molecular weight was given by Eq. (2.95).
bon at 38 and 99~ (100 and 210~ by Eqs. (2.128) and Use of the nomogram is not common at the present time
(2.129). How can you assess the validity of your estimated especially with availability of personal computers (PCs) and
kinematic viscosity at 38~ simulators, but still some process engineers prefer to use a
nomogram to have a quick estimate of a property or to check
Solution--To estimate the viscosity through the Abbott cor- their calculations from analytical correlations and computer
relations, Kw and API gravity are needed. However, Tb is not programs.
available and should be estimated from M and SG. Since If the boiling point is not available, methods discussed in
M > 300, we use Eq. (2.57) in terms of M and SG to esti- Section 2.4.2 may be used to estimate the boiling point be-
mate Tb as follows: Tb = 720.7 K. Using Eqs. (2.13) and (2.4), fore using the figure. Equation (2.50) for molecular weight
Kw and API are calculated as Kw = 11.08 and API = 12.23. may be combined with Eq. (2.13) to obtain a relation for the
Using Eqs. (2.128) and (2.129) the viscosities are calculated estimation of Kw from M and SG [51].
as 1)38(100) = 299.76 cSt. 1399(210) ~ I 1.35 cSt. At 99~ the esti- (2.133) Kw = 4 . 5 5 7 9 M ~ -0"84573
mated value can be directly evaluated against the experimen-
tal data: %D = [(11.08 - 13.09)/13.09)] x 100 = -15.4%. To This equation gives an approximate value for Kw and should
evaluate accuracy of estimated viscosity at 38~ a consistency be used with care for hydrocarbons heavier than C30. A more
test is required. Since the actual value of molecular weight, accurate correlation for estimation of Kw can be obtained if
M, is given, one can estimate M through Eq. (2.52) using the boiling point is calculated from Eq. (2.56) or (2.57) and
estimated values of 1)38(100), 1)99(210), and SG as the input pa- used in Eq. (2.13) to calculate Kw.
rameter. The estimated M is 333.8 which in comparison with
actual value of 352.6 gives %AD of 11.36%. This error is ac- Example 2.12--Basic properties of n-tridecylcyclohexane
ceptable considering that Eq. (2.52) has been developed based (CIgH38) a r e given in Table 2.1. Use M and SG as available
on data of petroleum fractions and the fact that input param- input parameters to calculate
eters are estimated rather than actual values. Therefore, we
a. Kw from Eq. (2.133).
can conclude that the consistency test has been successful
b. Kw from most accurate method.
and the value of 299.8 cSt as viscosity of this hydrocarbon at
c. Kw from Winn Nomogram.
38~ is acceptable. The error on estimated viscosity at 99~
d. CH weight ratio from M and SG.
is 15.4%, which is within the range of errors reported for the
e. CH weight ratio from Winn method.
method. It should be realized that the equations for prediction
f. %D for each method in comparison with the actual values.
of kinematic viscosity and estimation of molecular weight by
Eq. (2.52) were originally recommended for petroleum frac-
Solution--From Table 2.1, M = 266.5, Tb = 614.7 K,
tions rather than pure compounds. #
and SG = 0.8277. HC atomic ratio -- 38/19 = 2.0. Using
Eq. (2.122), CH weight ratio = 11.9147/2.0 = 5.957. From
definition of Kw, i.e., Eq. (2.13), the actual value of Kw is
2.8 T H E W I N N N O M O G R A M
calculated as Kw = (1.8 x 614.7)1/3/0.8277 = 12.496.
Development of estimation techniques through graphical a. From Eq. (2.133), Kw = 4.5579 x [(266.5) ~ x
methods was quite common in the 1930s through the 1950s [(0.8277) -~ = 12.485. %D = -0.09%.
when computational tools were not available. Nomogram or b. The compound is from the n-alkylcyclohexane family and
homograph usually refers to a graphical correlation between the most accurate way of predicting its boiling point is
different input parameters and desired property when more through Eq. (2.42) with constants given in Table 2.6 which
74 CHARACTERIZATIONAND PROPERTIES OF PETROLEUMFRACTIONS
90-i_ _--1200
L
ss --- -~s
:l 10O
8o-
-~ooo
75-2
70-_ -0.70
65" ~00 r
F-
L800
500
60-
- -0.75 O
55-- ~too
u.
"
240~ ~ ~00 70O
50-
.~ 22o..2 r .
c~
,3- '
5-300 -600 ~.
Lsoo~~
.-
,~
E ID
< 35- -0.85
.._ c-
.~ 150
3o-2
-- ~
2s- -o.9o ~a- ~-~ :_
:I: -
- -300
- o9-2 '
--0.95 -9o
15-
-80 -20O
10 tOO
5-
~t05
- -100
0-
gives Tb = 1100 -- exp[7.00275 -- 0.01977 x (266.5) 2/3] = w h i c h result in CH value of 6.2 with % D = + 4% ( C H a c t u a i =
615.08 K. Using Tb = 615.08 a n d SG = 0.8277 in Eq. (2.13) 5.96).
gives Kw = 12.498. The %D is +0.016%.
c. W h e n using Winn n o m o g r a m (Fig. 2.14) it is easier to con- CH = 2.35475 x [exp(9.3485 • 10 -3 • 266.5 + 4.74695
vert SG to API gravity, w h i c h t h r o u g h Eq. (2.4) is 39.46. x 0.8277 - 8.01719 x 10 -3 x 266.5 x 0.8277)]
A straight line b e t w e e n points 266.5 on the M line an d x [(266,5) -0'68418] x [(0.8277) -0'7682] = 6.2
39.5 on the API gravity line intersects the Watson K line
at Kw = 12.27 an d the intersection with the CH line is at e, CH weight ratio f r o m the Winn m e t h o d was o b t a i n e d in
CH = 6.1. The %D for Kw is - 1 . 8 % . Part c as C H = 6 . 1 , w h i c h gives %D = +2.3%.
d. CH weight ratio can be e s ti m a t e d f r o m Eq. (2.40) using f. F o r all parts % D is calculated f r o m Eq. (2.134). Part b
M and SG as i n p ut p a r a m e t e r s with constants in Table 2.5, gives the m o s t a c c u r a t e Kw value b ecau se Tb w a s calculated
2. CHARACTERIZATION AND PROPERTIES OF PURE HYDROCARBONS 75
accurately. However, Part b in this case is also accurate. Es- of compounds selected, the source of data, number of data
timation of the CH value is less accurate than prediction of points, and the basis for the evaluation all affect evaluation
boiling point and gives errors higher than Kw. outcome. The number of numerical constants in a correla-
tion and number of input parameters also affect the accu-
racy. Usually older methods are based on a fewer and less
2.9 ANALYSIS AND COMPARISON OF accurate data than newer methods. It would be always useful
VARIOUS CHARACTERIZATION METHODS to test different methods on a set of data that have not been
used in obtaining the correlation coefficients. The roost ap-
Generally there are a large number of pure hydrocarbons propriate procedure would be to compare various methods
and their properties can be used for evaluation purposes. with an independent data set not used in the development of
However, hydrocarbons from certain groups (i.e., paraffins, any methods considered in the evaluation process. Another
naphthenes, and aromatics) are more abundant in petroleum fair comparison of two different correlation would be to use
fractions and can be used as a database for evaluation pur- the same database and reobtain the numerical constants in
poses. Molecular weight, critical properties, and acentric fac- each correlation from a single database. This was done when
tor are important properties and their predictive methods Eqs. (2.52) and (2.53) were compared, as discussed in Sec-
are presented in this chapter. Errors in any of these proper- tion 2.4.1. These are the bases that have been used to compare
ties greatly influence the accuracy of the estimated physical some of the correlations presented in this chapter.
property. Methods of estimation of these properties from bulk Basically there are two parameters for the evaluation of
properties such as boiling point and specific gravity that are a correlation. One parameter is the percent average absolute
presented in this chapter have been in use in the petroleum deviation (%AAD). Average errors reported in this chapter and
industry for many years. In some process simulators a user throughout the book are based on percent relative deviation
should select a characterization method out of more than a (%D). These errors are defined as following:
dozen methods included in the simulator [56]. In each ap-
plication, the choice of characterization method by the user estimated value - actual value)
strongly influences the simulation results. Although there has (2.134) %D = \ ac~u~ualva~ue x 100
not been a general and comprehensive evaluation of various
characterization methods, a conclusion can be made from
individual's experiences reported in the literature. In this sec-
tion first we discuss criteria for evaluation of various methods
and then different predictive methods for molecular weight, where N is the total number of data points and summation is
critical constants and acentric factor are compared and eval- made on all the points. I%D[ is called percent absolute devi-
uated. ation and it is shown by %AD. The maximum value of I%D[
in a data set is referred as %MAD. The second parameter is
called R squared (R2) that is considered as an index of the cor-
2.9.1 Criteria for Evaluation
o f a Characterization M e t h o d relation when parameters of a correlation are obtained from
a data set. A value of 1 means perfect fit while values above
Methods of characterization and correlations presented in 0.99 generally give good correlation. For a set of data with
this chapter are mainly based on properties of pure hydrocar- X column (independent variable) and Y column (dependent
bons. However, some of these correlations such as Eq. (2.52) variable) the parameter is defined as
for estimation of the molecular weight of heavy fractions or
the correlations presented for prediction of the kinematic vis- (2.136) R2= [N(y~xY)-(~X)(Y~Y)]2
cosity are based mainly on the properties of fractions rather [ N y ~ X 2 - ( ~ X ) 2] x [ N ~ Y 2 - (y~y)2]
than pure compounds. The main application of these correla-
tions is for basic properties of undefined petroleum fractions where X and Y are values of the independent and correspond-
in which bulk properties of a fraction are used to estimate ing dependent variables and N is the number of data points.
a desired parameter. Therefore, the true evaluation of these The ~ is the summation over all N values of X, X 2, Y, y2, and
characterization methods should be made through properties XY as indicated in the above equation. The R2 value can be
of petroleum fractions as will be discussed in upcoming chap- interpreted as the proportion of the variance in y attributable
ters. However, evaluation of these methods with properties of to the variance in x and it varies from 0 to maximum value
pure hydrocarbons can be used as a preliminary criteria to of 1.
judge the accuracy of various correlations. A method, which For most of the correlations presented in this chapter such
is more accurate than other methods for pure hydrocarbons, as Eqs. (2.40), (2.42), or (2.46a) the %AAD for various prop-
is not necessarily the best method for petroleum mixtures. A erties is usually given in the corresponding tables where the
database for pure hydrocarbons consists of many compounds constants are shown. Most of these properties have been cor-
from different families. However, evaluations made by some related with an R2 value of minimum 0.99. Some of these
researchers are based primarily on properties of limited pure properties such as kinematic viscosity or CH weight ratio
hydrocarbons (e.g., n-alkanes). The conclusions through such showed lower values for R2. Evaluation of some of the other
evaluations cannot be generalized to all hydrocarbons and correlations is made through various examples presented in
petroleum fractions. Perhaps it is not a fair comparison if this chapter.
a data set used to develop a method is also used to evalu- Nowadays with access to sophisticated mathematical tools,
ate the other methods that have used other databases. Type it is possible to obtain a very accurate correlation from any
76 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
data set. For example, when the method of neural network is TABLE 2.14--Evaluation of methods for estimation of molecular
used to obtain correlations for estimation of critical proper- weight of petroleum fractions,a
Abs Dev%**
ties, a very accurate correlation can be obtained for a large
number of compounds [82]. However, such correlations con- Method Equation(s) AAD% MAD%
tain as m a n y as 30 numerical values, which limit their power API (Riazi-Daubert) (2.51 ) 3.9 18.7
of extrapolatability. It is our experience that when a corre- Twu (2.89)-(2.92) 5.0 16.1
Kesler-Lee (2.54) 8.2 28.2
lation is based on some theoretical foundation, it has fewer Winn (2.93) 5.4 25.9
constants with a wider range of application and better ex- ~ N u r n b e r o f d a t a p o i n t s : 625; R a n g e s o f d a t a : M ~ 7 0 - 7 0 0 , Tb ~ 3 0 0 - 8 5 0 , S G
trapolatability. This is particularly evident for the case of 0.63-0.97
b D e f i n e d b y E q s . (2.134) a n d (135). R e f e r e n c e [29].
Eq. (2.38) developed based on the theory of intermolecular
forces and EOS parameters. Equation (2.38) has only three
parameters that are obtained from data on properties of pure fractions. This equation has been included in most process
hydrocarbons from C5 to C20. This equation for various prop- simulators [54-56]. Whitson [51, 53] has used this equation
erties can be safely used up to C30. Tsonopoulos et aL [34] and its conversion to Kw (Eq. 2.133) for fractions up to C25
and Lin et al. [83] have extensively evaluated Eq. (2.50) for in his characterization methods of reservoir fluids. A more
estimation of the molecular weight of different samples of general form of this equation is given by Eq. (2.51) for the
coal liquids, which are mainly aromatics, and compared with molecular weight range of 70-700. This equation gives an
other sophisticated multiparameter correlations specifically average error of 3.4% for fractions with M < 300 and 4.7%
developed for the molecular weight of coal liquids. Their con- for fractions with M > 300 for 625 fractions from Penn State
clusion was that Eq. (2.50) gave the lowest error even though database on petroleum fractions. An advantage of Eq. (2.51)
only pure component data were used to develop this equation. over Eq. (2.50) is that it is applicable to both light and heavy
Further evaluation of characterization methods for molecu- fractions. A comparative evaluation of various correlations
lar weight and critical properties are given in the following for estimation of molecular weight is given in Table 2.14 [29].
parts. Process simulators [55] usually have referred to Eq. (2.50)
as Riazi-Daubert method and Eq. (2.51) as the API method.
The Winn method, Eq. (2.93), has been also referred as S i m -
2.9.2 Evaluation o f Methods o f Estimation
Daubert method in some sources [55, 84].
o f Molecular Weight
For pure hydrocarbons the molecular weight of three ho-
As mentioned above most of the evaluations made on mologous hydrocarbon groups predicted from Eq. (2.51) is
Eq. (2.50) for the molecular weight of petroleum fractions drawn versus carbon number in Fig. 2.15. For a given car-
below 300 suggest that it predicts quite well for various bon number the difference between molecular weights of
3000
o A P I Data
..... n-Alkanes
2500 Naphthenes
- - n-Alk3,1benzenes
2000 /
Y~
e~ /
1500
o
1000
500
I i f i I i f
0
200 400 600 800 1000 1200
Boiling Point, K
FIG. 2.15--Evaluation of Eq. (2.51) for molecular weight of pure compounds.
2. CHARACTERIZATION AND P R O P E R TIE S OF PURE H YD R O CA R B O N S 77
400
o API D a t a ..M~" ,
-- R-D:Eq. 2.50 ~ A~."~
350
. . . . API: Eq. 2.51 ~ r
. . . . . . R.s: 248 /. G..?::
-- - --Twu: Eq. 2.89-2.92 ~ 7
300 Lee-Kcsler:E q ~
250
200 "'"
O
E
150
100
50 , , .., ~ , , , , r , , , , f , , , t , , , ,
5 10 15 20 25 30
Carbon Number
FIG. 2.16--Evaluation of various methods for prediction o1 molecular weight
of n-alkylcycohexanes. Riazi-Daubert: Eq. (2,50); API: Eq. (2.51); Riazi-Sahhaf:
Eq. (2.48); Lee-Kesler: Eq. (2.54); Twu: Eqs, (2,89)-(2,92).
hydrocarbons from different groups is small. Actual values of lated through generalized correlations which require critical
molecular weight of n-alkylbenzenes up to C20 as reported by properties as shown in Chapters 6 and 7. The phase behav-
API-TDB [2] are also shown on the figure. Equation (2.51) is ior prediction of reservoir fluids also requires critical prop-
not the best method for the prediction of molecular weight of erties of petroleum cuts that make up the fluid as discussed
pure compounds as it was primarily developed for petroleum in Chapter 9. These two indirect methods are the basis for
fractions. Various methods for the estimation of molecular the evaluation of correlations for estimation of critical prop-
weight for n-alkylcylohexanes with the API data (up to C26) erties. These evaluations very much depend on the type of
are shown in Fig. 2.16 for the range of C6-Cs0. At higher car- fractions evaluated. For example, Eqs. (2.63)-(2.66) for esti-
bon numbers the deviation between the methods increases. mation of Tc and Pc have been developed based on the critical
The Twu method accurately estimates molecular weight of data from C5 to Cls; therefore, their application to heavy frac-
low-molecular-weight pure hydrocarbons; however, at higher tions is not reliable although they can be safely extrapolated to
molecular weights it deviates from actual data. A compari- C25-C30 hydrocarbons. In the development of these equations,
son between evaluations presented in Fig. 2.16 and Table 2.14 the internal consistency between Tc and Pc was not imposed as
shows that a method that is accurate for prediction of proper- the correlations were developed for fractions with M < 300.
ties of pure hydrocarbons is not necessarily the best method These correlations were primarily developed for light frac-
for petroleum fractions. Evaluation of method of prediction tions and medium distillates that are produced from atmo-
of molecular weight from viscosity (Eqs. (2.52) and (2.53)) spheric distillation columns.
has been discussed in Section 2.4.1. For pure hydrocarbons from homologous families,
Eq. (2.42) with constants in Table (2.6) provide accurate val-
ues for Tc, Pc, and Vc. Prediction of Tc and Pc from this equa-
2.9.3 Evaluation o f M e t h o d s of E s t i m a t i o n
tion and comparison with the API-TDB data are shown in
of Critical Properties
Figs. 2.2 and 2.3, respectively. Evaluation of various methods
Evaluation of correlations for estimation of critical proper- for critical temperature, pressure, and volume of different hy-
ties of pure compounds can be made directly with the actual drocarbon families is demonstrated in Figs. 2.17-2.19 respec-
values for hydrocarbons up to C18. However, when they are tively. A summary of evaluations for Tc and Pc of hydrocarbons
applied to petroleum fractions, pseudocritical properties are from different groups of all types is presented in Table 2.15
calculated which are not directly measurable. These values [29]. Discontinuity of API data on Pc of n-alkylcyclopentanes,
should be evaluated through other properties that are mea- as seen in Fig. 2.18, is due to prediction of Pc for heav-
surable but require critical properties for their calculations. ier hydrocarbons (>C20) through a group contribution
For example, enthalpies of petroleum fractions are calcu- method.
78 CHARACTERIZATION AND P R O P E R TI E S OF PETROLEUM FRACTIONS
1050
o API Data
~ Winn
- - R - D _ , ~ . ~ , . ~. . ~. 9
- - - Twtl . . . ~ ~
900
~Z
E
750
600
450 l I I I I L I L L e i i i i i i i l i i i i i i i i i l
10 20 30 40 50 60
Carbon Number
50
o API Data
40 \~ - - - Winn
* ~ - - - - R-D
- - - - API
Ext. R-D
- - - - L - K
- - - P-F
30
- R-S
~ 20
(,9
10
r ~ i I i t L i ~ I I I I I I I I I ~ I i I I I I I i
5 10 15 20 25 30 35 40 45 50
Carbon Number
3000
o API Data /
S
/ .
.
2500 TWO /'i /
.... Ext.R-D /. "~
1500
= 1000
500
i i I I r i i I i i i i i i i l i i
0 10 20 30 40 50
Carbon Number
FIG. 2.19--Comparison of various methods for estimation of critical volume of
n-alkylbenzenes. API Data: API-TDB [2]; Twu: Eqs. (2.83)-(2.85); APh Eq. (2.101);
R-S: Riazi-Sahhaf, Eq. (2.42) and Table 2.6; R-D: Riazi-Daubert, Eq. (2.98); H-Y:
HalI-Yarborough, Eq. (2.100).
TABLE 2.15--Evaluation of various methods for prediction of critical temperature and pressure
of pure hydrocarbons from C5 to C20.
Abs Dev%**
T~ Pc
Method Equation(s) AD% MAD% AD% MAD%
API (2.65)-(2.66) 0.5 2.2 2.7 13.2
Twu (2.73)-(2.88) 0.6 2.4 3.9 16.5
Kesler-Lee (2.69)-(2.70) 0.7 3.2 4 12.4
Cavett (2.71 )-(2.73) 3.0 5.9 5.5 31.2
Winn (Sim-Daubert) (2.94)-(2.95) 1.0 3.8 4.5 22.8
Riazi-Daubert (2.63)-(2.64) 1.1 8.6 3.1 9.3
Lin & Chao Reference [72] 1.0 3.8 4.5 22.8
aData on Tc and Pc of 138 hydrocarbons from different families reported in API-TDBwere used for the evalua-
tion process [29].
bDefinedby Eqs. (2.134) and (2.135).
80 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
for Tc and Pc as observed by some researchers [51, 85]. How- error associated with critical properties of heavy fractions is
ever, Eq. (2.42) and subsequently derived Eqs. (2.67) and to back-calculate the critical properties of the heaviest end of
(2.68) have the internal consistency and can be used from the reservoir fluid from an EOS based on a measured physical
C20
C5 to C50 although they are developed for hydrocarbons from
to C50
The 1980 Riazi-Daubert correlations for Tc and Pc were
property such as density or saturation pressure [51, 52, 70].
Firoozabadi et al. [63, 64] have studied extensively the wax
and asphaltene precipitation in reservoir fluids. They ana-
generally used and recommended by m a n y researchers for lyzed various methods of calculating critical properties of
light fractions (M < 300, carbon number < C22). Yu et al. [84] heavy petroleum fractions and used Eq. (2.42) for the critical
used 12 different correlations to characterize the C7+ plus properties and acentric factor of paraffins, naphthenes, and
fraction of several samples of heavy reservoir fluids and bi- aromatics, but they used Eq. (2.43) for the critical pressure of
tumens. Based on the results presented on gas-phase compo- various hydrocarbon groups with M > 300. Their evaluation
sition, GOR, and saturation pressure, Eqs. (2.63) and (2.64) was based on the calculation of the cloud point of different
showed better or equivalent predictions to other methods. oils. It is believed that fractions with molecular weight greater
Whitson [53] made a good analysis of correlations for the crit- than 800 (NC57) mainly contain aromatic hydrocarbons [63]
ical properties and their effects on characterization of reser- and therefore Eq. (2.42) with constants given in Table 2.6 for
voir fluids and suggested the use of Eqs. (2.63) and (2.64) for aromatics is an appropriate correlation to estimate the prop-
petroleum cuts up to C25. But later [51] based on his observa- erties of such fractions.
tion for phase behavior prediction of heavy reservoir fluids, More recently Jianzhong et al. [87] reviewed and evalu-
he recommended the use of Kesler-Lee or Twu for estimation ated various methods of estimation of critical properties of
of Tc and Pc of such fluids, while for estimation of critical petroleum and coal liquid fractions. Their work followed
volume he uses Eq. (2.98). Soreide [52] in an extensive eval- the work of Voulgaris et al. [88], who recommended use of
uation of various correlations for the estimation of critical Eq. (2.38) for estimation of critical properties for the purpose
properties recommends use of the API-TDB [2] method for of prediction of physical properties of petroleum fractions
estimation Tc and Pc (Eqs. (2.65) and (2.66)) but he recom- and coal liquids. They correctly concluded that complexity
mends Twu method for the critical volume. His recommenda- of correlations does not necessarily increase their accuracy.
tions are based on phase behavior calculations for 68 samples They evaluated Lee-Kesler, Riazi-Daubert, and Twu methods
of North Sea reservoir fluids. In a recently published Hand- with more than 318 compounds (> C5) including those found
book o f Reservoir Engineering [48], and calculations made on in coal liquids with boiling point up to 418~ (785 ~F) and spe-
phase behavior of reservoir fluids [86], Eqs. (2.65), (2.66) have cific gravity up to 1.175 [87]. They suggested that Eq. (2.38)
been selected for the estimation of critical properties of unde- is the most suitable and accurate relation especially when
fined petroleum fractions. Another possibility to reduce the the coefficients are modified. Based on their database, they
3.00
/
API Data: n-Alkanes /
-- -- Predicted: n-Alkanes
9 API D auc n-Alkylcyclopentanes:
2.50
~j..//
.... Predicted: n-Alkylcyclopentanes
9 API Data: n-Alkylbenz enes
Preclicted~n-~dkylbeazenes
2.03 J
1.50 : f5
~9
<
1.03
(150
I t I I I r I I
(103 r
0 10 ~!0 30 40 5o 03
Carbon Number
obtained the coefficients for Tc, Pc, and Vo in Eq. (2.38) with 3.000
use of d20 (liquid density at 20~ and 1 atm in g/cm 3) instead o DIPPR Data
of SG (Tc, Pc, Vc = aTbbd~0). They reported the coefficients as Riazi-Sahhaf
[87] T j K (a = 18.2394, b = 0.595251, c = 0.347420), Pc/bar 2.500 P ~ et al. /
(a = 2.95152, b = -2.2082, c = 2.22086), and Vc/cma/mol (a = 9- Korstea /
8.22382 x 10-5, b = 2.51217, c = -1.62214). Equation (2.38) 2.O00 /
m ~ Lee-Kesler /
with these coefficients have not been extensively tested
against data on properties of petroleum fractions as yet but
for more than 300 pure hydrocarbons gives average errors of
0.7, 3.8, and 2.9% for To Pc, and Vc, respectively [87].
/.:.--:...---
< 1.000
2.9.4 E v a l u a t i o n o f M e t h o d s of Estimation
of Acentric Factor and Other Properties 0.500
For the calculation of the acentric factor of pure hydrocar-
bons Eq. (2.42) is quite accurate and will be used in Chapter 3 0.000
for the pseudocomponent method. Firoozabadi suggests that
0 10 20 30 40 50
for aromatics with M > 800, co = 2. Generally there are three
methods for the estimation of the acentric factor of undefined Carbon Number
petroleum fractions. Perhaps the most accurate method is to
estimate the acentric factor through its definition, Eq. (2.10), FIG. 2.22--Prediction of acentric factor of n-alkylbenzenes
and vapor pressure estimated from a reliable method [86]. from various methods. DIPPR Data: DIPPR [20]; Riazi-Sahhaf:
This method will be further discussed in Chapter 7 along with Eq. (2.42) and Table 2.6; Pan et al.: Ref. [63, 64], Eq. (2.44);
methods of calculation of vapor pressure. For pure hydrocar- Korsten: Eq. (2.109); Lee-Kesler: Eq. (2.105); Kesler-Lee:
bons the Lee-Kesler method is more accurate than the Edmis- Eq. (2.107).
ter method [36]. The Korsten method for estimating acentric
factor is new and has not yet been evaluated extensively. For
three different homologous hydrocarbon families from C6 to are presented in Figs. 2.21 and 2.22, respectively. The Riazi-
C50, values of acentric factor calculated from Eq. (2.42) are Sahhaf method refers to Eq. (2.42) and coefficients given
compared with values reported in the API-TDB [2] and they in Table 2.6 for different hydrocarbon families. In Fig. 2.22
are shown in Fig. 2.20. Prediction of acentric factors from dif- the Pan et al. [63, 64] method refers to Eq. (2.44), which
ferent methods for n-allcylcyclopentanes and n-alkylbenzenes has been recommended for n-alkylbenzenes (aromatics). The
Lee-Kesler method, Eq. (2.105), has been generally used for
the estimation of accentric factor of undefined petroleum
fractions [27]. The Kesler-Lee method refers to Eq. (2.107),
2.0 which was recommended by Kesler-Lee [12] for estimation
o API Data of the acentric factor of hydrocarbons with Tbr > 0.8, which
Riazi- 5ahhaf / is nearly equivalent to hydrocarbons with molecular weights
.......Lee-Kesler / / greater than 300. However, our experience shows that this
1.5 Kesler-Lee ~,,".. ~. -~ equation is accurate for pure compounds when true critical
temperatures are used and high errors can occur when the
.... ~'= .~7:-:'~.""
- -
predicted critical temperature is used in the equation. For
O
heavy hydrocarbons and petroleum fractions (M > 300) with
O
1.0 @~176~ estimated critical properties, either the method of pseudo-
component discussed in Chapter 3 or the Lee-Kesler may be
< the most appropriate method. The accuracy of a method to
estimate acentric factor also depends on the values of Tc and
0.5 Pc used to calculate co as was shown in Example 2.7. Usually
the Cavett correlations for Tc and Pc are used together with
the Edmister method. Evaluation of these methods for the
prediction of properties of undefined petroleum fractions is
- 0.0
discussed in Chapter 3.
0 10 20 30 40 50 The accuracy of correlations presented for estimation of
other properties such as density, refractive index, boiling
Carbon Number
point, and CH has been discussed in the previous section
FIG. 2.21--Prediction of acentric factor of n-alkylcyclo- where these methods are presented. Prediction of the refrac-
pentanes from various methods. API Data: API-TDB [2]; R-S: tive index for pure hydrocarbons is shown in Fig. 2.9. Predic-
Riazi-Sahhaf, Eq. (2.42) and Table 2.6; L-K: Lee-Kesler, tion of viscosity at 38~ (100~ I ) 3 8 , through Eq. (2.128) for
Eq. (2.105); K-L: Kesler-Lee, Eq. (2.107); Edmister: Eq. (2.108); pure hydrocarbons from three hydrocarbon groups is shown
Korsten: Eq. (2.109). in Fig. 2.23. Further assessment of accuracy of these methods
82 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
30
i
/
9 API Data: n-alkylcbenzenes / /
20 /
-- -- Predieted:n-allq,lcbenzenes 1/.,
15 /
/../
/ ,' /
o
10
/.'o
//o
. . . . ~ . i . . . . i . . . . J . . . .
0 5 10 15 20 25 30 35
Carbon Number
FIG. 2.23--Prefliction of kinematic viscosity of pure hydrocarbons at 37.8~
from Eq. (2.128).
is discussed in the following chapters where properties of parameters may be used as the last option for prediction of
petroleum fractions are calculated. properties of hydrocarbons. Various methods of estimation
of these parameters as well as critical properties and acen-
tric factor used in corresponding state correlations and a de-
2.10 CONCLUSIONS AND tailed review of their application for different purposes and
RECOMMENDATIONS recommendations made in the literature are presented. Basic
properties of more than 100 selected compounds are given in
In this chapter methods of characterization of pure hydro- Tables 2.1 and 2.2 and will be used frequently throughout the
carbons have been presented. These methods will be used book.
in Chapters 3 and 4 for the characterization of petroleum The most important information presented in this chap-
fractions and crude oils, respectively. This chapter is an im- ter is the methods of estimation of molecular weight, critical
portant chapter in this book as the method selected for the constants, and acentric factor for pure hydrocarbons. These
characterization of hydrocarbons affects the accuracy of es- methods are also recommended to estimate properties of nar-
timation of every physical property throughout the book. In row boiling range petroleum fractions as discussed in Chapter
this chapter the basic characterization parameters have been 3. A summary of evaluations made by various researchers was
introduced and based on the theory of intermolecular forces, reviewed in Section 2.9. Based on these evaluations it is clear
a generalized correlation for the characterization of hydro- that theoretically based correlations such as Eq. (2.38) or its
carbon systems has been developed. It is shown that funda- modified version Eq. (2.40), while simpler than other empiri-
mentally developed correlations are simpler with a wider field cally developed correlations, have a wide range of application
of application and accuracy. For light fractions (M < 300), with reasonable accuracy. Based on these evaluations a list
generally two-parameter correlations are sufficient for prac- of recommended methods for different properties of various
tical calculations, while for heavier hydrocarbons or nonhy- types of hydrocarbons and narrow boiling range fractions is
drocarbons the use of a third parameter is needed. The two given in Table 2.16. Estimation of wide boiling range frac-
characterization parameters should represent the energy and tions is discussed in the next chapter. The choice for meth-
size characteristics of molecules. Characterization parame- ods of calculation of properties not presented in Table 2.16
ters such as Tb, M, and v3s000) may be used as energy, param- is generally narrow and comments have been made where
eters while SG, I, and CH could be used as size parameters. the methods are introduced in each section. The information
Boiling point and specific gravity are the most easily measur- presented in this chapter should help practical engineers to
able and appropriate characterization parameters followed develop new correlations or to select an appropriate charac-
by molecular weight and refractive index. Viscosity and CH terization scheme when using a process simulator.
2. CHARACTERIZATION AND PROPERTIES OF PURE HYDROCARBONS 83
TABLE 2.16--Recommended methods for the prediction of the basic properties of pure hydrocarbons
and narrow boiling range petroleum fractionsa.
Property Range of M Method Equation
M 70-700 API [2] 2.51
70-300 Riazi-Dabubert [28] 2.50
200-800 API [2] 2.52(b)
70-700 Twu [30] 2.89-2.92(c)
Tc 70-300 API [2] 2.65
70-700 Lee-Kesler [ 12] 2.69
70-700 Extended API [65] 2.67
70-800 Riazi-Sahhaf [31 ] 2.42 d
<70 e Riazi et al. [37] 2.47 e
Pc 70-300 API [2] 2.66
70-700 Lee-Kesler [12] or Twu [30] 2.70
70-700 Extended API [65] 2.68
70-300 Riazi-Sahhaf [31 ] 2.42 d
300-800 Pan-Firoozabadi-Fotland [63] 2.43 d
<70 e Riazi et al. [37] 2.47 e
Vc 70-350 Riazi-Daubert [28] 2.98
300-700 Extended R-D [65] 2.99
70-700 Riazi-Sahhaf [31] 2.42 d
<70 e Riazi et al. [37] 2.47 e
Zc 70-700 By definition of Zc 2.8
w 70-300 Lee-Kesler [27] 2.105
300-700 Korsten [77] 2.109
70-700 Riazi-Sahhaf [31 ] 2.42 f
300-700 Pan-Firoozabadi-Fofland [63] 2.44g
Tb 70-300 Riazi-Daubert [29] 2.56
300-700 Extended R-D [65] 2.57
70-700 Riazi-Sahhaf [31 ] 2.42 d
SG All range Denis et al. [8] 2.112
70-300 Riazi-Daubert [29] 2.59
70-700 Extended R-D [65] 2.60
200-800 API [2] 2.61 d
I 70-300 Riazi-Daubert [29] 2.116
300-700 Extended R-D [65] 2.117
70-700 Riazi-Sahhaf [31 ] 2.42 d
d All range Denis et al. [8] 2.111
70-350 Riazi-Daubert [28] 2.113
TM 70-700 Pan-Firoozabadi-Fotland [63] 2.126 h
Riazi-Sahhaf [31] 2.124 and 2.125 i
Methods recommended for pure homologous hydrocarbons (designated by c-i) are also recommended for the pseu-
docomponent method discussed in Chapter 3 for petroleum fractions. The 300 boundaryis approximate and methods
recommended for the range of 70-300 may be used safely up to molecular weight of 350 and similarly methods rec-
ommended for the range 300-700 may be used for molecular weight as low as 250.
a For narrow boiling range fractions a midpoint distillation temperature can be used as Tb.
bOnly when Tb is not available.
CRecommended for pure hydrocarbons from all types.
'/Recommended for pure homologous hydrocarbon groups.
eFor compounds and fractions with molecular weight less than 70 and those containing nonhydrocarbon compounds
(H2S,CO2, N2, etc.) Eq. (2.47) is recommended.
fEquation (2.42) is applicable to acentric factor of n-alkylbenzenes up to molecular weight of 300.
gEquation (2.44) is applicable to acentric factor of aromatics for 300 < M < 800 and for M > 800, w - 2 should be
used.
hFor pure hydrocarbons from n-alkanes family.
i For pure hydrocarbons from n-alkylcylopentanes (naphthenes) and n-alkylbenzenes (aromatics) families.
2.4. A tank contains pure h y d r o c a r b o n liquid from the [2] Daubert, T. E. and Danner, R. E, Eds., API Technical Data
n-paraffin group. Determine the lightest h y d r o c a r b o n Book--Petroleum Refining, 6th ed., American Petroleum
from the n-alkane family that can exist in an open ves- Institute (API), Washington, DC, 1997.
sel at the environment of 38~ (100~ and 1 atm with- [3] Nikitin, E. D., Pavlov, E A., and Popov, A. E, '%'apor-Liquid
Critical Temperatures and Pressures of Normal Alkanes from
out danger of flammability in the vapor phase near the
19 to 36 Carbon Atoms, Naphthalene and m-Terphenyl
vessel. Determined by the Pulse-Heating Technique," Fluid Phase
2.5. Develop three relations for estimation of CH weight Equilibria, Vol. 141, 1997, pp. 155-164.
ratio of n-paraffins, n-alkylcyclopentanes, and n-alkyl- [4] Reid, R. C., Pransnitz, J. M., and Poling, B. E., Properties of
benzene in terms of their respective molecular weight. Gases and Liquids, 4th ed., Mc-Graw Hill, New York, 1987.
For each group calculate CHo~ (and HCo~). Show graph- Poling, B. E., Pransnitz, J. M., O'Connell, J. E, Properties of
ical presentation of the predicted values versus actual Gases and Liquids, 5th ed., Mc-Graw Hill, New York, 2001.
values of CH for the three families on a single graph. [5] Pitzer, K. S., "The Volumetric and Thermodynamic Properties
2.6. Predict the refractive index of n-paraffins, n-alkyl- of Fluids, I: Theoretical Basis and Virial Coefficients," Journal
cyclopentanes, and n-alkylbenzene versus carbon num- of American Chemical Society, Vol. 77, 1955, pp. 3427-3433.
[6] Pitzer, K. S., Lippmann, D. Z., Curl, Jr., R. E, Huggins, C. M.,
ber from C6 to Cs0 using Eq. (2.46a) and c o m p a r e graph-
and Petersen, D. E., "The Volumetric and Thermodynamic
ically with values from Eq. (2.42). In using Eq. (2.46a) Properties of Fluids, II: Compressibility Factor, Vapor Pressure,
it is necessary to obtain M from Nc in each family, and and Entropy of Vaporization," Journal of American Chemical
then from Eq. (2.42) Tb and SG m a y be estimated for Society, Vol. 77, 1955, pp. 3433-3440.
each carbon n u m b e r in each family. [7] Denis, J., Briant, J., and Hipeaux, J. C., Lubricant Properties
2.7. A pure hydrocarbon has molecular weight of 338.6 and Analysis and Testing, Translated to English by G. Dobson,
specific gravity of 0.8028. Using appropriate methods Editions Technip, Paris, 1997.
calculate 18] Wauquier, J.-P., Petroleum Refining, Vol. 1: Crude Oil, Petroleum
a. boiling point, Tb. Products, Process Flowsheets, Translated from French by David
b. refractivity intercept, Ri. H. Smith, Editions Technip, Paris, 1995.
[9] Watson, K. M., Nelson, E. E, and Murphy, G. B.,
c. kinematic viscosity at 38 and 99~
"Characterization of Petroleum Fractions," Industrial and
d. VGC f r o m three different methods. Engineering Chemistry, Vol. 27, 1935, pp. 1460-1464.
2.8. For n-butylcyclohexane, critical properties and molecu- [10] Daubert, T. E., "Property Predictions," Hydrocarbon Processing,
lar weight are give in Table 2.1. Use Tb and SG as the Vol. 59, No. 3, 1980, pp. 107-112.
input partameters and calculate [11] Speight, J. G., The Chemistry and Technology of Petroleum, 3rd
a. M, To, Pc, de, and Zc from the API-TDB-87 methods. ed., Marcel Dekker, New York, 1998.
b. M, To, Pc, dc, and Zc the Lee-Kesler correlations. [12] Kesler, M. G. and Lee, B. I., "Improve Prediction of Enthalpy of
c. M, To, Pc, de, and Zc from the Riazi-Daubert correla- Fractions," Hydrocarbon Processing, Vol. 55, 1976,
tions (Eq. 2.38). pp. 153-158.
d. M, Tc, Pc, de, and Zc from the Twu correlations. [13] Kurtz, Jr., S. S. and Ward, A. L., "The Refractivity Intercept and
the Specific Refraction Equation of Newton, I: Development of
e. Compare values from each m e t h o d with actual values
the Refractivity Intercept and Composition with Specific
and tabulate the %D. Refraction Equations," Journal of Franklin Institute, Vol. 222,
2.9. Use calculated values of Tc and Pc in Problem 2.8 to cal- 1936, pp. 563-592.
culate acentric factor from the Lee-Kesler and Korsten [14] Hill, J. B. and Coats, H. B., "The Viscosity Gravity Cosntant of
correlations for each part, then obtain the errors (%D) Petroleum Lubricating Oils," Industrial and Engineering
for each method. Chemistry, Vol. 20, 1928, pp. 641-4544.
2.10. Estimate the acentric factor of isooctane from Lee- [15] Riazi, M. R. and Daubert, T. E., "Prediction of the Composition
Kesler, Edmister, and Korsten correlations using input of Petroleum Fractions," Industrial and Engineering Chemistry,
data from Table 2.1. Calculate the %D for each method. Process Design and Development, Vol. 19, No. 2, 1980,
2.11. Estimate the kinematic viscosity of n-heptane at 38 pp. 289-294.
and 99~ and c o m p a r e with the experimental values [16] Kurtz, S. S., King, R. W., Stout, W. J., Partikian, D. G., and
Skrabek, E. A., "Relationsship Between Carbon-Type
reported by the API-TDB [2]. Also estimate viscosity Composition, Viscosity Gravity Constant, and Refractivity
of n-heptane at 50~ from Eq. (2.130) and the ASTM Intercept of Viscous Fractions of Petroleum," Analytical
viscosity-temperature chart. Chemistry, Vol. 28, 1956, pp. 1928-1936.
2.12. For n-alkylcylopentanes from C5 to C10, estimate d20 [17] Altgelt, K. H. and Boduszynski, M. M., Composition and
from SG using the rule of thumbs and a more accurate Analysis of Heavy Petroleum Fractions, Marcel Dekker, New
method. Compare the results with actual values from York, 1994.
Table 2.1. For these c o m p o u n d s also estimate refractive [I 8] API Research Project 42: Properties of Hydrocarbons of High
index at 25~ using M as the only input data available. Molecular Weight, American Petroleum Institute, New York,
Use both methods for the effect of temperature on re- 1966.
fractive index as discussed in Section 2.6.2 and c o m p a r e
[19] Daubert, T. E., Danner, R. E, Sibul, H. M., and Stebbins, C. C.,
Physical and Thermodynamic Properties of Pure Compounds:
your results with the values reported by the API-TDB [2].
Data Compilation, DIPPR-AIChE, Taylor & Francis, Bristol, PA,
1994 (extant) (www.aiche.org/dippr), Updated reference:
Rowley, R. L., Wilding, W. V., Oscarson, J. L., Zundel, N. A.,
REFERENCES Marshall, T. L., Daubert, T. E., and Danner, R. E, DIPPR Data
Compilation of Pure Compound Properties, Design Institute for
[1] Coplen, T. B., "Atomic Weights of the Elements 1999," Pure and Physical Properties (DIPPR), Taylor & Francis, New York, 2002
Applied Chemistry, Vol. 73, No. 4, 2001, pp. 667-683. (http://dippr.byu.edu).
2. C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P U R E H Y D R O C A R B O N S 85
[20] AIChE DIPPR| Design Institute for Physical [41] Smith, J. M., Van Ness, H. C., and Abbott, M. M., Introduction
Property Data (DIPPR), EPCON International, Houston, 1996. to Chemical Engineering Thermodynamics, 5th ed.,
[21] Hall, K. R., Ed., TRC Thermodynamic Tables--Hydrocarbons, McGraw-Hill, New York, 1996.
Thermodynamic Research Center, The Texas A&M University [42] Huang, E K., Characterization and Thermodynamic Correlations
System, College Station, TX, 1993. Updated reference: TRC for Undefined Hydrocarbon Mixtures, Ph.D. Dissertation,
Thermodynamic Tables--Hydrocarbons, Thermodynamic Pennsylvania State University, University Park, PA, 1977.
Research Center, National Institute of Standards and [43] Ambrose, D., "Correlation and Estimation of Vapor-Liquid
Technology, Boulder, CO, 2001. (http://www.trc.nist.gov/). Critical Properties, I: Critical Temperatures of Organic
[22] API Research Project 44: Selected Values of Properties of Compounds," National Physical Laboratory, Teddington,
Hydrocarbons and Related Compounds, Tables of Physcial and Middlesex, UK, TW11 0LW, NPL Report, Vol. 92, 1978, NPL
Thermodynamic Properties of Hydrocarbons, A&M Press, Report 98, 1979.
College Station, TX, 1978 (extant). [44] Nokay, R., "Estimate Petrochemical Properties," Chemical
[23] Dortmund Data Bank (DDB), Laboratory for Thermophysical Engineering, Vol. 6, No. 4, 1959, pp. 147-148.
Properties, LTP GmbH, Institute at the University of Oldenburg [45] Spencer, C. E and Daubert, T. E., "A Critical Evaluation of
(www.ddbst.de). Methods for the Prediction of Critical Propertis of
[24] Watson, K. M. and Nelson, E. E, "Improved Methods for Hydrocarbons," American Institute of Chemical Engineers
Approximating Critical and Thermal Properties of Petroleum Journal, Vol. 19, No. 3, 1973, pp. 482486.
Fractions," Industrial and Engineering Chemistry, Vol. 25, 1933, [46] Edmister, W. C. and Lee, 13. I., Applied Hydrocarbon
pp. 880-887. Thermodynamics, 2nd ed., Gulf Publishing, Houston, TX,
[25] Winn, E W., "Physical Properties by Nomogram," Petroleum 1985.
Refiners, Vol. 36, No. 21, 1957, pp. 157-159. [47] Stange, E. and Johannesen, S. O., "HYPO*S, A Program for
[26] Cavett, R. H., "Physical Data for Distillation Calculations, Heavy End Characterization," An Internal Program from the
Vapor-Liquid Equilibria," Proceeding of 27th API Meeting, API Norsk Hydro Oil and Gas Group (Norway), Paper presented at
Division of Refining, Vol. 42, No. 3, 1962, pp. 351-366. the 65th Annual Convection, Gas Processors Association,
[27] Lee, B. I. and Kesler, M. G., "A Generalized Thermodynamic San Antonio, TX, March 10, 1986.
Correlation Based on Three- Parameter Corresponding States," [48] Ahmed, T., Reservoir Engineering Handbook, Gulf Publishing,
American Institute of Chemical Engineers Journal, Vot. 21, 1975, Houston, TX, 2000.
pp. 510-527. [49] Ahmed, T., Hydrocarbon Phase Behavior, Gulf Publishing,
[28] Riazi, M. R. and Daubert, T. E., "Simplify Property Predictions," Houston, TX, 1989.
Hydrocarbon Processing, Vol. 59, No. 3, 1980, pp. 115-116. [50] "API Tech Database Software," EPCON International, Houston,
[29] Riazi, M. R. and Daubert, T. E., "Characterization Parameters TX, 2000 (www.epcon.com).
for Petroleum Fractions," Industrial and Engineering Chemistry [51] Whitson, C. H. and Brule, M. R., Phase Behavior, Monograph
Research, Vol. 26, 1987, pp. 755-759. (Corrections, p. 1268.) Vol. 20, Society of Petroleum Engineers, Richardson, TX,
[30] Twu, C. H., "An Internally Consistent Correlation for Predicting 2000.
the Critical Properties and Molecular Weights of Petroleum and [52] Soreide, I., Improved Phase Behavior Predictions of Petroleum
Coal-Tar Liquids," Fluid Phase Equilbria, Vol. 16, 1984, Reservoir Fluids from a Cubic Equation of State, Doctor of
pp. 137-150. Engineering Dissertation, Norwegian Institute of Technology,
[31] Riazi, M. R. and A1-Sahhaf, T., "Physical Properties of n-Alkanes Trondheim, Norway, 1989.
and n-Alkyl Hydrocarbons: Application to Petroleum Mixtures," [53] Whitson, C. H., "Effect of C7+ Properties on Equation-of-State
Industrial and Engineering Chemistry Research, Vol. 34, 1995, Predictions," Society of Petroleum Engineers Journal, December
pp. 4145-4148. 1984, pp. 685-696.
[32] Korsten, H., "Characterization of Hydrocarbon Systems by [54] Aspen Plus, "Introductory Manual Software Version," Aspen
DBE Concept," American Institute of Chemical Engineers Technology, Inc., Cambridge, MA, December 1986.
Journal, Vol. 43, No. 6, 1997, pp. 1559-1568. [55] PRO/II, "Keyword Manual," Simulation Sciences Inc.,
[33] Korsten, H., "Critical Properties of Hydrocarbon Systems," Fullerton, CA, October 1992.
Chemical Engineering and Technology, Vol. 21, No. 3, 1998, [56] HYSYS, "Reference Volume 1, Version 1.1," HYSYS Reference
pp. 229-244. Manual for Computer Software, HYSYS Conceptual Design,
[34] Tsonopoulos, C., Heidman, J. L., and Hwang, S.-C., Hyprotech Ltd., Calgary, Alberta, Canada, 1996.
Thermodynamic and Transport Properties of Coal Liquids, An [57] Kreglewski, A. and Zwolinski, B. J., "A New Relation for
Exxon Monograph, Wiley, New York, 1986. Physical Properties of n-Alkanes and n-Alkyl Compounds,"
[35] Riazi, M. R. and Roomi, Y., "Use of the Refractive Index in the Journal of Physical Chemistry, Vol. 65, 1961, pp. 1050-1052.
Estimation of Thermophysical Properties of Hydrocarbons and [58] Tsonopoulos, C., "Critical Constants of n-Alkanes from
Their Mixtures," Industrial and Engineering Chemistry Methane to Polyethylene," American Institute of Chemical
Research, Vol. 40, No. 8, 2001, pp. 1975-1984. Engineers Journal, Vo. 33, No. 12, 1987, pp. 2080-2083.
[36] Riazi, M. R., Prediction of Thermophysical Properties of [59] Gray, Jr., R. D., Heidman, J. L., Springer, R. D., and
Petroleum Fractions, Doctoral Dissertation, Department of Tsonopolous, C., "Characterization and Property Prediction for
Chemical Engineering, Pennsylvania State University, Heavy Petroleum and Sysnthetic Liquids," Fluid Phase
University Park, PA, 1979. Equilibria, Vol. 53, 1989, pp. 355-376.
[37] Riazi, M. R., Al-Sahhaf, T. A., and A1-Shammari, M. A., [60] Teja, A. S., Lee, R. J., Rosenthal, D. J., and Anselme, M.,
"A Generalized Method for Estimation of Critical Constants," "Correlation of the Critical Properties of n-Alkanes and
Fluid Phase Equilibria, Vol. 147, 1998, pp. 1-6. Alkanols," Fluid Phase Equilibria, Vol. 56, 1990,
[38] Hirschfelder, O. J., Curtiss, C. E, and Bird, R. B., Molecular pp. 153-169.
Theory of Gases and Liquids, Wiley, New York, 1964. [61] Goossens, A. G., "Prediction of Molecular Weight of Petroleum
[39] Prausnitz, J. M., Lichtenthaler, R. N., and de Azevedo, E. G., Fractions," Industrial and Engineering Chemistry Research,
Molecular Thermodynamics of Fluid-Phase Equilibria, Vol. 35, 1996, pp. 985 988.
Prentice-Hall, New Jersey, 1986. [62] Goossens, A. G., "Prediction of the Hydrogen Content of
[40] Hill, T. L., An Introduction to Statistical Thermodynmaics, Petroleum Fractions," Industrial and Engineering Chemistry
Addison Wesley, Boston, MA, 1960. Research, Vol. 36, 1997, pp. 2500-2504.
86 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
[63] Pan, H., Firoozabadi, A., and Fotland, E, "Pressure and [77] Korsten, H., "Internally Consistent Prediction of Vapor
Composition Effect on Wax Precipitation: Experimental Data Pressure and Related Properties," Industrial and Engineering
and Model Results," Society of Petroleum Engineers Chemistry Research, 2000, Vol. 39, pp. 813-820.
Production and Facilities, Vol. 12, No. 4, 1997, [78] Riazi, M. R. and Daubert, 1". E., "Prediction of Molecular Type
pp. 250-258. Analysis of Petroleum Fractions and Coal Liquids," Industrial
[64] Firoozabadi, A., Thermodynamics of Hydrocarbon Reservoirs, and Engineering Chemistry, Process Design and Development,
McGraw Hill, New York, 1999. Vol. 25, No. 4, 1986, pp. 1009-1015.
[65] Riazi, M. R. and Adwani, H. A., "Some Guidelines for Choosing [79] Won, K. W., Thermodynamics for Solid Solution-Liquid-Vapor
a Characterization Method for Petroleum Fractions in Process Equilibria: Wax Formation from Heavy Hydrocarbon
Simulators," Chemical Engineering Research and Design, Mixtures," Fluid Phase Equilibria, Vol. 30, 1986, pp. 265-279.
IchemE, Vol. 83, 2005. [80] Abbott, M. M., Kaufrnann, 17.G. and Domash, L., "A Correlation
[66] Vetere, A., "Methods to predict the critical constants of organic for Predicting Liquid Viscosities of Petroleum Fractions,"
compounds," Fluid Phase Equilibria, Vol. 109, 1995, Canadian Journal of Chemical Engineering, Vol. 49, No. 3, 197l,
pp. 17-27. pp. 379-384.
[67] Riazi, M. R., Daubert, 1". E., "Molecular Weight of Heavy [81] Twu, C. H., "Internally Consistent Correlation for Predicting
Fractions from Viscosity," Oil and Gas Journal, Vol. 58, No. 52, Liquid Viscosities of Petroleum Fractions," Industrial and
1987, pp. 110-113. Engineering Chemistry, Process Design and Development,
[68] ASTM, Annual Book of Standards, Section Five: Petroleum Vol. 24, 1985, 1287-1293.
Products, Lubricants, and Fossil Fuels (in Five Volumes), [82] Riazi, M. R. and Elkamel, A. "Using Neural Network Models to
ASTM International, West Conshohocken, PA, 2002. Estimate Critical Constants", Simulators International,
[69] Hirschler, A. E., "Molecular Mass of Viscous Hydrocarbon Oils: International Society for Computer Simulation, Vol. 29, No. 3,
Correlation of Density with Viscosities," Journal of the Institute 1997, pp. 172-177.
of Petroleum, Vol. 32, No. 1946, 267, pp. 133-161. [83] Lin, H. M., Kim, H., Gut, T., and Chat, K. C., "Equilibrium
[70] Chorn, L. G. and Mansoori, G. A., C7+Characterization, Taylor Vaporization of a Coal Liquid from a Kentucky No. 9 Coal,"
& Francis, New York, 1989. Industrial and Engineering Chemistry, Process Design and
[71] Kesler, M. G., Lee, B. I., and Sandler, S. I., "A Third Parameter Development, Vol. 24, No. 4, 1985, pp. 1049-1055.
for Use in Generalized Thermodynamic Correlations," [84] Yu, J. M., Wu, R. S., and Batycky, J. E, "A C7+ Characterization
Industrial and Engineering Chemistry Fundamentals, Vol. 18, Method for Bitumen Mixture Phase Behavior Prediction,"
1979, pp. 49-54. C7+Characterization, L. G. Chorn and G. A. Mansoori Eds.,
[72] Lin, H. M. and Chao, K. C., "Correlation of Critical Properties Taylor & Francis, New York, 1989, pp. 169-196.
and Acentric Factor of Hydrocarbons and Derivatives," [85] Danesh, A., PVT and Phase Behavior of Petroleum Reservoir
American Institute of Chemical Engineers Journal, Vol. 30, Fluids, Elsevier, Amsterdam, 1998.
No. 6, 1984, pp. 981-983. [86] Manafi, H., Mansoori, G. A. and Ghotbi, S., "Phase Behavior
[73] Mobil Oil Company, "Engineering Data Book," New York, Prediction of Petroleum Fluids with Minimum
1961. Characterization Data," Journal of Petroleum Science and
[74] Sire, W. J. and Daubert, T. E., "Predcition of Vapor-Liquid Engineering, Vol. 22, 1999, pp. 67-93.
Equilibria of Undefined Mixtures," Industrial and Engineering [87] Jianzhong, Zh., Biao, Zh., Suoqi, Zh., Renan, W., and
Chemistry, Process Design and Development, Vol. 19, 1980, Guanghua, Y., "Simplified Prediction of Critical Properties of
pp. 386-393. Nonpolar Compounds, Petroleum, and Coal Liquid Fractions,"
[75] Hall, K. R. and Yarborough, L., "New Simple Correlation for Industrial and Engineering Chemistry Research, Vol 37, No. 5,
Predicting Critical Volume," Chemical Engineering, Vol. 78, 1998, pp. 2059-2060.
No. 25, 1971, p. 76-77. [88] Voulgaris, M., Stamatakis, S., Magoulas, K., and Tassios, D.,
[76] Edmister, W. C., "Applied Hydrocarbon Thermodynamics, Part "Prediction of Physical Properties for nonpolar Compounds,
4: Compressibility Factors and Equations of State," Petroleum Petroleum, and Coal Liquid Fractions," Fluid Phase Equilibria,
Refiner, Vol. 37, 1958, pp. 173-179. Vol. 64, 1991, pp. 73-106.
MNL50-EB/Jan. 2005
Characterization of Petroleum
Fractions
AP Aniline point, ~ (unless specified otherwise) p Density at a given temperature and pressure, g/cm 3
API API gravity defined in Eq. (2.4) a Surface tension, dyn/cm [=mN/m]
A,B, . . . . F Correlation coefficients in various equations o~ Acentric factor defined by Eq. (2.10), dimensionless
a,b..i Correlation coefficients in various equations
CABP Cubic average boiling point, K Superscript
CH Carbon-to-hydrogen weight ratio
~ Properties of n-alkanes from Twu correlations
d Liquid density at 20~ and 1 atm, g/cm 3
Kw Watson (UOP) K factor defined by Eq. (2.13)
I Refractive index parameter defined in Eq. (2.36)
Subscripts
M Molecular weight, g/mol [kg/kmol] A Aromatic
MABP Molal average boiling point, K N Naphthenic
MeABP Mean average boiling point, K P Paraffinic
n Sodium D line refractive index of liquid at 20~ T Value of a property at temperature T
and 1 atm, dimensionless ~ A reference state for T and P
Nc Carbon number (number of carbon atoms in a c~ Value of a property at M --~ c~
hydrocarbon molecule) 20 Value of a property at 20~
Critical pressure, bar 39(100) Value of kinematic viscosity at 37.8~ (100~
Refractivity intercept in Eq. (2.14) 99(210) Value of kinematic viscosity at 98.9~ (210~
SG Specific gravity of liquid substance at 15.5~
(60~ defined by Eq. (2.2), dimensionless Acronyms
SGg Specific gravity of gas substance at 15.5~ (60~
defined by Eq. (2.6), dimensionless API-TDB American Petroleum Institute--Technical Data
SL ASTM D 86 slope between 10 and 90% points, Book
~ (K)/vol% ASTM D ASTM International (test methods by D com-
S% Weight percent of sulfur in a fraction mittee)
Tb Boiling point, K %AD Absolute deviation percentage defined by
Eq. (2.134)
Tc Critical temperature, K
TF Flash point, K %AAD Average absolute deviation percentage defined by
TM Melting (freezing point) point, K Eq. (2.135)
EFV Equilibrium flash vaporization
T~o Temperature on distillation curve at 10% volume
vaporized, K EOS Equation of state
Ts0 Temperature on distillation curve at 50% volume FBP Final boiling point (end point)
vaporized, K GC Gas chromatography
V Molar volume, cma/gmol GPC Gel permeation chromatography
V Saybolt universal viscosity, SUS HPLC High performance liquid chromatography
v~ Critical volume (molar), cm3/gmol KISR Kuwait Institute for Scientific Research
VGC Viscosity gravity constant defined by Eqs. (2.15)- IBP Initial boiling point
(2.18) IR Infrared
VABP Volume average boiling point, K MA Monoaromatic
WABP Weight average boiling point, K MS Mass spectroscometry
PA Poly (di- tri-, and higher) aromatic
PIONA Paraffin, isoparaffin, olefin, naphthene,
Greek Letters and aromatic
RVP Reid vapor pressure
F G a m m a function RS R squared (R2), defined by Eq. (2.136)
# Absolute (dynamic) viscosity, cp [mPa.s]. Also used for SD Simulated distillation
dipole moment TBP True boiling point
v Kinematic viscosity defined by Eq. (2.12), cSt [mm2/s] UV Ultraviolet
87
for mixtures containing very light gases or very heavy com- complete component separation is usually achieved. Mea-
pounds that cannot be vaporized. The test is conducted at surement of TBP data is more difficult than ASTM D 86 data
atmospheric pressure with 100 mL of sample and the result in terms of both time and cost. TBP and ASTM D 86 curves
is shown as a distillation curve with temperatures at 0, 5, 10, for a kerosene sample are shown in Fig. 3.3 based on data
20, 30, 40, 50, 60, 70, 80, 90, 95, and 100% volume vapor- provided by Lenoir and Hipkin [ 1]. As shown in this figure
ized. The final boiling point (at 100%) is the least accurate the IBP from TBP curve is less than the IBP from ASTM D
value and it is usually less than the true final boiling point. In 86 curve, while the FBP of TBP curve is higher than that of
many cases only a few temperatures are reported. An exposed ASTM curve. Therefore, the boiling range based on ASTM D
thermometer is used and temperatures are reported without 86 is less than the actual true boiling range. In TBP, the IBP
stem corrections. For heavy products, temperatures are re- is the vapor temperature that is observed at the instant that
ported at maximum of 90, 70, or even 50% volume vaporized. the first drop of condensate falls from the condenser.
This is due to the cracking of heavy hydrocarbons at high tem-
peratures in which vaporization temperatures do not repre- 3.1.1.3 Simulated Distillation by Gas Chromatography
sent boiling points of the original compounds in the mixture. Although ASTM D 86 test method is very simple and conve-
The cracking effect is significant at temperatures above 350 ~C nient, it is not a consistent and reproducible method. For this
(660~ however, ASTM D 86 temperatures reported above reason another method by gas chromatography (GC) is be-
250~ (480~ should be used with caution. Corrections ap- ing recommended to present distillation data. A distillation
plied to consider the effects of cracking are applicable from curve produced by GC is called a simulated distillation (SD)
250 to 500~ however, these procedures have not been widely and the method is described in ASTM D 2887 test method.
used and generally have not been confirmed. In the new revi- Simulated distillation method is simple, consistent, and
sions of API-TDB-97 no correction for cracking in ASTM D 86
temperatures has been recommended [2]. An apparatus to
300 ...........................................................................................................................
measure distillation of petroleum fractions by ASTM D 86
method is shown in Fig. 3.2. . . . . . . . ASTM D86
250 TBP .-""
3.1.1.2 True Boiling Point ~'~ . . . . - " ~ 1 7 6 *" -'~
ASTM D 86 distillation data do not represent actual boiling
point of components in a petroleum fraction. Process engi-
neers are more interested in actual or true boiling point (TBP)
of cuts in a petroleum mixture. Atmospheric TBP data are 150
obtained through distillation of a petroleum mixture using a
distillation column with 15-100 theoretical plates at relatively
high reflux ratios (i.e., 1-5 or greater). The high degree of 100
fractionation in these distillations gives accurate component 0 20 40 60 80 100
distributions for mixtures. The lack of standardized appara-
Vol% Vaporized
tus and operational procedure is a disadvantage, but vari-
ations between TBP data reported by different laboratories FIG. 3.3--ASTM D 86 and TBP curves for a
for the same sample are small because a close approach to kerosene sample.
90 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
250 ...................................................................................................................................................................... gas such as helium or nitrogen. The stationary phase is either
solid or liquid. A component that is more strongly attracted to
200 ....... ASTM D86 the mobile phase than to the stationary phase is swept along
with the mobile phase more rapidly than a component that is
SD(ASTMD ~
more strongly attracted to the stationary phase. The mobile
150 phase can be a liquid phase as well; in this case the chro-
matography method is called liquid chromatography (LC).
100 The basic elements of a GC are a cylinder of carrier gas,
flow controller and pressure regulator, sample injector, col-
umn, detector, recorder, and thermostats for cylinder, col-
50 umn, and detector. The sample after injection enters a heated
oven where it enters the GC column (stationary phase). The
__ i i i i eluted components by the carrier gas called effluents enter
0 20 40 60 80 100 a detector where the concentration of each component may
be determined. The presence of a component in the carrier
PercentVaporized
gas leaving the column is continuously monitored through
FIG. 3.4---Simulated and ASTM D 86 distilla- an electric signal, which eventually is fed to a recorder for a
tion curves for a petroleum fraction. (The per- visual readout.
cent is in vol% for ASTM D 86 and is in wt% for There are two types of columns, packed or capillary
ASTM D 2887.) columns, and two types of detectors, flame ionization detec-
tor or thermal conductivity detector. Packed columns have
reproducible and can represent the boiling range of a inner diameters of 5-8 mm and length of 1-5 m. Column
petroleum mixture without any ambiguity. This method is ap- and detector types depend on the nature of samples being
plicable to petroleum fractions with a FBP of less than 538~ analyzed by the GC. The capillary columns are equivalent to
(1000~ and a boiling range of greater than 55~ (100~ and hundreds of theoretical equilibrium stages and can be used in
having a vapor pressure sufficiently low to permit sampling at preference to packed columns. The inner diameter of capri-
ambient temperature. The ASTM D 2887 method is not appli- lary columns is about 0.25-0.53 m m and their length is about
cable to gasoline samples and the ASTM D 3710 test method 10-150 m. The stationary phase is coated on the inside wall
is recommended for such fractions. Distillation curves by SD of columns. The flame ionization detector (FID) is highly sen-
are presented in terms of boiling point versus wt% of mixture sitive to all organic compounds (10 -12 g) but is not sensitive
vaporized because as described below in gas chromatogra- to inorganic compounds and gases such as H20, CO2, N2,
phy composition is measured in terms of wt% or weight frac- and 02. The FID response is almost proportional to the mass
tion. Simulated distillation curves represent boiling points of concentration of the ionized compound. Hydrogen of high
compounds in a petroleum mixture at atmospheric pressure; purity is used as the fuel for the FID. The thermal conduc-
however, as will be shown later SD curves are very close to ac- tivity detector (TCD) is sensitive to almost all the compounds
tual boiling points shown by TBP curves. But these two types but its sensitivity is less than that of FID. TCD is often used
of distillation data are not identical and conversion methods for analysis of hydrocarbon gas mixtures containing nonhy-
should be used to convert SD to TBP curves. In comparison drocarbon gases. The retention time is the amount of time re-
with ASTM D 86, the IBP from a SD curve of a petroleum mix- quired for a given component spent inside the column from
ture is less than IBP from ASTM D 86 curve, while the FBP its entrance until its emergence from the column in the efflu-
from SD curve is higher than the FBP from ASTM D 86 of the ent. Each component has a certain retention time depending
same mixture (see Fig. 3.4). This is the same trend as that of on the structure of compound, type of column and station-
TBP curves in comparison with ASTM curves as was shown ary phase, flow rate of mobile phase, length, and tempera-
in Fig. 3.3. A typical SD curve for a gas oil sample is shown ture of column. More volatile compounds with lower boil-
in Fig. 3.4. Note that in this figure the percent vaporized for ing points have lower retention times. Detector response is
ASTM D 2887 (SD) is in wt% while for the ASTM D 86 curve measured in millivolts by electric devices. The written record
is in vol%. obtained from a chromatographic analysis is called a chro-
The gas chromatography technique is a separation method matograph. Usually the time is the abscissa (x axis) and mV
based on the volatility of the compounds in a mixture. The GC is the ordinate (y axis). A typical chromatograph obtained
is used for both generation of distillation curves as well as to to analyze a naphtha sample from a Kuwaiti crude is shown
determine the composition of hydrocarbon gas or liquid mix- in Fig. 3.5. Each peak corresponds to a specific compound.
tures, as will be discussed later in this chapter. For this reason Qualitative analysis with GC is done by comparing retention
in this part we discuss the basic function of chromatography times of sample components with retention times of reference
techniques and elements of GC. In an analysis of a mixture by compounds (standard sample) developed under identical ex-
a GC, the mixture is separated into its individual compounds perimental conditions. With proper flow rate and tempera-
according to the relative attraction of the components for ture, the retention time can he reproduced within I%. Ev-
a stationary and a mobile phase. Recent advances in chro- ely component has only one retention time; however, compo-
matography make it possible to identify and separate com- nents having the same boiling point or volatility but different
pounds with boiling points up to 750~ (1380~ A small fluid molecular structure cannot be identified through GC analysis.
sample (few microliters for liquid and 5 mL for gas samples) In Fig. 3.5, compounds with higher retention time (x coordi-
is injected by a needle injector into a heated zone in which nate) have higher boiling points and the actual boiling point
the sample is vaporized and carried by a high-purity carrier or the compound can be determined by comparing the peak
3. C H A R A C T E R I Z A T I O N OF P E T R O L E U M F R A C T I O N S 91
q~
J i
+
2~Q
i
i
J
!
+ti
r'o z~ ~ ~- ................../~ ~ ,45
TABLE 3.1--Calculation of composition of a naphtha sample with GC chromatograph shown in Fig. 3.4.
No. Name Time, min Area Area % Tb,~C
1 n-Hexane 11.16 1442160 3.92 68.7
2 Benzene 14.64 675785 1.84 80.1
3 Cyclohexane 15.55 3827351 10.40 80.7
4 n-Heptane 18.90 5936159 16.14 98.4
5 2,2,3-Trimethyl-pentane 20.38 8160051 22.18 109.8
6 Toluene 27.53 8955969 24.34 110.6
7 Ethylbenzene 42.79 1678625 4.56 136.2
8 p-Xylene 45.02 4714426 12.82 138.4
9 o-Xylene 49.21 1397253 3.8 144.4
Total 36787780 100
92 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
atmospheric boiling points are given in the ASTM Manual [4] 3.1.3 Molecular Weight
and will be discussed later in this chapter.
Molecular weight is another bulk property that is indicative of
molecular size and structure. This is an important property
3.1.2 Density, Specific Gravity, and API Gravity that usually laboratories do not measure and fail to report
when reporting various properties of petroleum fractions.
Specific gravity (SG) or relative density and the API gravity are
This is perhaps due to the low accuracy in the measurement
defined in Section 2.1.3 and for pure hydrocarbons are given
of the molecular weight of petroleum fractions, especially for
in Tables 2.1 and 2.2. Aromatic oils are denser than paraffinic
heavy fractions. However, it should be realized that experi-
oils. Once specific gravity is known, the API gravity can be de-
mental uncertainty in reported values of molecular weight
termined by Eq. (2.4), which corresponds to the ASTM D 287
is less than the errors associated with predictive methods for
method. The standard temperature to measure the specific
this very useful parameter. Since petroleum fractions are mix-
gravity is 15.56~ (60~ however, absolute density is usually
tures of hydrocarbon compounds, mixture molecular weight
reported at 20~ Specific gravity or density for a petroleum
is defined as an average value called number average molec-
mixture is a bulk property that can be directly measured for
ular weight or simply molecular weight of the mixture and it
the mixture. Specific gravity is a property that indicates the
is calculated as follows:
quality of a petroleum product and, as was shown in Chapter
2, is a useful property to estimate various physical proper- (3.1) M = ~x4Mi
ties of petroleum fluids. A standard test method for density i
and specific gravity of liquid petroleum products and distil-
lates in the range of 15-35~ through use of a digital density where xi and M/are the mole fraction and molecular weight of
meter is described in ASTM D 4052 method [4]. The appa- component i, respectively. Molecular weight of the mixture,
ratus must be calibrated at each temperature and this test M, represents the ratio of total mass of the mixture to the total
method is equivalent to ISO 12185 and IP 365 methods. An- moles in the mixture. Exact knowledge of molecular weight
other method using a hydrometer is described under ASTM D of a mixture requires exact composition of all compounds in
1298 test method. Hydrometer is a glass float with lead ballast the mixture. For petroleum fractions such exact knowledge is
that is floated in the liquid. The level at which hydrometer is not available due to the large number of components present
floating in the liquid is proportional to the specific gravity in the mixture. Therefore, experimental measurement of mix-
of the liquid. Through graduation of the hydrometer the spe- ture molecular weight is needed in lieu of exact composition
cific gravity can be read directly from the stalk of hydrometer. of all compounds in the mixture.
This method is simpler than the ASTM 4052 method but is There are three methods that are widely used to measure
less accurate. The French standard procedure for measuring the molecular weight of various petroleum fractions. These
density by hydrometer is described under NFT 60-101 test are cryoscopy, the vapor pressure method, and the size ex-
method. With some hydrometers densities with accuracy of clusion chromatography (SEe) method. For heavy petroleum
0.0005 g/mL can be measured. A digital density meter model fractions and asphaltenic compounds the SEC method is
DMA 48 from PARA (Austria) is shown in Fig. 3.8. commonly used to measure distribution of molecular weight
FIG. 3.8--PARA model DMA 48 digital density meter (courtesy of Chemical Engineering
Department at Kuwait University).
94 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
FIG. 3.9--A GPC from waters model 150-C plus (courtesy of Chemical Engineering
Department at Kuwait University).
in the fraction. The SEC method is mainly used to determine of about 0.001 ~ The relation to obtain molecular weight of
molecular weights of polymers in the range of 2000 to 2 • 106. a sample is [6]
This method is also called gel permeation chromatography 1000 x Kf x ml
(GPC) and is described in the ASTM D 5296 test method. In (3.2) M=
AT x m2
the GPC method, by comparing the elution time of a sample
with that of a reference solution the molecular weight of the where Kf is molal freezing point depression constant of the
sample can be determined. A GPC instrument is shown in solvent and is about 5.12~ AT is the freezing point de-
Fig. 3.9. The SEC experiment is usually performed for heavy pression and the reading from the cryoscope, rnl is the mass
residues and asphaltenes in crude oils and gives the wt% of of solute and m2 is the mass of solvent both in grams. It gener-
various constituents versus molecular weight as will be dis- ally consists of refrigerator, thermometer and the apparatus
cussed in Chapter 4. to hold the sample. A cryoscope is shown in Fig. 3.10.
The vapor pressure method is based on the measurement of
the difference between vapor pressure of sample and that of
3.1.4 Refractive Index
a known reference solvent with a vapor pressure greater than
that of the sample. A solution of about 1 g of sample in 25 mL Refractive index or refractivity is defined in Section 2.1.4 and
of the reference solvent is prepared. This solution, which has its values at 20~ for pure hydrocarbons are given in Table
vapor pressure less than that of the solvent, tends to condense 2.1. Refractive indexes of hydrocarbons vary from 1.35 to 1.6;
the vapors of solvent on the sample thus creating a temper- however, aromatics have refractive index values greater than
ature difference which is measured by two thermistors. The naphthenes, which in turn have refractive indexes greater
molarity of the solution is calculated using calibrated curves. than paraffins. Paraffinic oils have lower refractive index val-
This method is described by the ASTM D 2503 test method ues. It was shown in Chapter 2 that refractive index is a useful
and is applicable to oils with an initial boiling point greater parameter to characterize hydrocarbon systems and, as will
than 220~ [5]. A typical experimental error and uncertainty be seen later in this chapter, it is needed to estimate the com-
in measuring the molecular weight is about 5%. position of undefined petroleum fractions. Refractive index is
The third and most widely used method of determining the ratio of the speed of light in a vacuum to that of a medium.
the molecular weight of an unknown petroleum mixture is In a medium, the speed of light depends on the wavelength
by the cryoscopy method, which is based on freezing point and temperature. For this reason refractive index is usually
depression. The freezing point of a solution is a measure of measured and reported at 20~ with the D line sodium light.
the solution's concentration. As the concentration of the so- For mixtures, refractive index is a bulk property that can
lute increases, the freezing point of the solution will be lower. be easily and accurately measured by an instrument called
The relation between freezing point depression and concen- a refractometer Refractive index can be measured by digital
tration is linear. For organic hydrocarbons, benzene is usually refractometers with a precision of 4-0.0001 and temperature
used as the solvent. Special care should be taken when work- precision of 4-0. I~ The amount of sample required to mea-
ing with benzene [6]. Calibration curves can be prepared by sure refractive index is very small and ASTM D 1218 provides
measuring the freezing points of different solute concentra- a test method for clear hydrocarbons with values of refractive
tions with a known solute and a known solvent. A cryoscope indexes in the range of 1.33-1.5 and the temperature range
can measure the freezing point depression with an accuracy of 20-30~ In the ASTM D 1218 test method the Bausch and
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 95
FIG. 3.10~Model 5009 wide range cryoscope to measure molecular weight (courtesy of
Chemical Engineering Department at Kuwait University).
petroleum fractions, residues, and fossil fuels (i.e., coal liq- FIG. 3.12--An all-solvent fractionation procedure. Reprinted
uids), which have high contents of aromatics, resins, and as- from Ref. [7], p. 267, by courtesy of Marcel Dekker, Inc.
phaltenes. The elemental analysis gives information on hy-
drogen and sulfur contents as well as C/H ratio, which are removed from the product [7]. ASTM [4] provides several
indicative of the quality of petroleum products. methods based on solvent separation to determine amounts
of asphaltenes. In ASTM D 2007 test method n-pentane is used
3.1.5.2 Analytical Instruments as the solvent, while in ASTM D 4124 asphaltene is separated
Generally three methods may be used to analyze petroleum by n-heptane. Schematics of these test methods are shown in
fractions. These are Figs. 3.13 and 3.14, respectively, as given by Speight [7]. As-
phaltenes are soluble in liquids with a surface tension above
9 separation by solvents 25 dyne/cm such as pyridine, carbon disulfide, carbon tetra-
9 chromatography methods chloride, and benzene [7].
9 spectroscopic methods The principle of separation by chromatography technique
The method of separation by solvents is based on solubil- was described in Section 3.1.1.3. If the mobile phase is gas
ity of some compounds in a mixture in a particular solvent. the instrument is called a gas chromatograph (GC), while for
The remaining insoluble compounds may be in a solid or an-
other immiscible liquid phase. This method is particularly Feedstock/
useful for heavy petroleum fractions and residues contain- [ n-Pentane
ing asphahenes, resins, and saturate hydrocarbons. The de- /
gree of solubility of a compound in a solvent depends on the |
chemical structure of both the solute and the solvent. If the
two structures are similar there is a greater degree of solubil-
ity. For example, high-molecular-weight asphaltenes are not Asphaltenes Deasphaltened0il
soluble in a low-molecular-weight paraffinic solvent such as
n-heptane. Therefore, if n-heptane is added to a heavy oil, as- Clay
phaltenes precipitate while the other constituents form a solu- /
l
Asphaltenes
1
DeasphaltenedOil
as heavy petroleum fractions and residues. Use of LC for sepa-
ration of saturated and aromatic hydrocarbons is described in
ASTM D 2549 test method. Various forms of chromatography
techniques have been applied to a wide range of petroleum
~ Alumina products for analysis, such as PONA, PIONA, PNA, and SARA.
One of the most useful types of liquid chromatography is
high performance liquid chromatography (HPLC), which can
be used to identify different types of hydrocarbon groups.
Polar Aromatics Oils One particular application of HPLC is to identify asphaltene
(Resins) ~ Alumina and resin type constituents in nonvolatile feedstocks such as
residua. Total time required to analyze a sample by HPLC is
just a few minutes. One of the main advantages of HPLC is
l
NaphtheneAromatics Saturates
that the boiling range of sample is immaterial. A HPLC ana-
lyzer is shown in Fig. 3.15.
The accuracy of chromatography techniques mainly
depends on the type of detector used [7]. In Section 3.1.1.3,
(Aromatics)
flame ionization (FID) and thermal conductivity (TCD) detec-
FIG. 3.14.--Separation of asphaltenes and resins from tors are described, which are widely used in GC. For LC the
petroleum fractions. Reprinted from Ref. [7], p. 281, by cour- most common detectors are refractive index detector (RID)
tesy of Marcel Dekker, Inc. and wavelength UV (ultraviolet) detector. UV spectroscopy is
particularly useful to identify the types of aromatics in asphal-
liquid mobile phase it is called a liquid chromatograph (LC). tenic fractions. Another spectroscopy method is conventional
As discussed earlier, components can be separated by their infrared (IR) spectroscopy, which yields information about
boiling points through GC analysis. In advanced petroleum the functional features of various petroleum constituents.
refineries automatic online GCs are used for continuous anal- For example, IR spectroscopy will aid in the identification of
ysis of various streams to control the quality of products. A N--H and O--H functions and the nature of polymethylene
FIG. 3,16--HP made GC-MS model 5890 Series II. (courtesy of Chemical Engineering
Department at Kuwait University),
chains (C--H) and the nature of any polynuclear aromatic compounds an integrated LC-MS unit may be suitable for
systems [7]. analysis of mixtures; however, use of LC-MS is more difficult
Another type of analysis of petroleum fractions to iden- than GC-MS because in LC-MS solvent must be removed
tify molecular groups is by spectrometric methods such as from the elute before it can be analyzed by MS. A GC-MS
mass spectrometry (MS). In general, there is a difference be- instrument from Hewlett Packard (HP) is shown in Fig. 3.16
tween spectroscopy and spectrometry methods although in Another type of separation is by SEC or GPC, which can
some references this difference is not acknowledged. Spec- be used to determine molecular weight distribution of heavy
troscopy refers to the techniques where the molecules are ex- petroleum fractions as discussed in Section 3.1.3. Fractions
cited by various sources, such as UV and IR, to return to their are separated according to their size and molecular weight
normal state. Spectrometry refers to the techniques where and the method is particularly useful to determine the amount
the molecules are actually ionized and fragmented. Evolution of asphaltenes. Asphaltenes are polar multiring aromatic
of spectroscopic methods comes after chromatography tech- compounds with molecular weight above 1000 (see Fig. 1.2).
niques, nonetheless, and in recent decades they have received It is assumed that in this molecular weight range only aro-
considerable attention. While volatile and light petroleum matics are present in a petroleum fraction [8].
products can be analyzed by gas chromatography, heavier
and nonvolatile compounds can be analyzed and identified by 3.1.5.3 PNA Analysis
spectrometric methods. One of the most important types of As determination of the exact composition of a petroleum
spectrometry techniques in analysis of petroleum fractions is fraction is nearly an impossible task, for narrow boiling range
mass spectrometry (MS). In this method, masses of molecular petroleum fractions and products a useful type of compo-
and atomic components that are ionized in the gaseous state sitional analysis is to determine the amounts of paraffins
by collision with electrons are measured. The advantage of (P), naphthenes (N), and aromatics (A). As mentioned be-
MS over other spectrometric methods is its high reproducibil- fore, most petroleum products are olefin free and PNA anal-
ity of quantitative analysis and information on molecular type ysis provides a good knowledge of molecular type of mixture
in complex mixtures. Mass spectrometry can provide the most constituents. However, some analyzers give the amount of
detailed quantitative and qualitative information about the isoparaffins and olefins as well. These analyzers are called
atomic and molecular composition of organic and inorganic PIONA analyzer, and a Chrompack Model 940 PIONA ana-
compounds. However, use of MS is limited to organic com- lyzer is shown in Fig. 3.17. An output of this type of analyzer
pounds that are stable up to 300~ (570~ At higher tem- is similar to the GC output; however, it directly gives wt% of
peratures thermal decomposition may occur and the anal- n-paraffins, isoparaffins, olefins, naphthenes, and aromatics.
ysis will be biased [7]. Through MS analysis, hydrocarbons The composition is expressed in wt%, which can be converted
of similar boiling points can be identified. In the MS analy- to mole, weight, and volume fractions as will be shown later
sis, molecular weight, chemical formula of hydrocarbons, and in this chapter.
their amounts can be determined. The most powerful instru-
ment to analyze petroleum distillates is the combination of 3.1.5.4 Elemental Analysis
a GC and an MS called GC-MS instrument, which separates The main elements present in a petroleum fraction are car-
compounds both through boiling point and molecular weight. bon (C), hydrogen (H), nitrogen (N), oxygen (O), and sulfur
For heavy petroleum fractions containing high-boiling-point (S). The most valuable information from elemental analysis
3. CHARACTERIZATION OF P E T R O L E U M FRACTIONS 99
FIG. 3,17mA chrompack model 940 PIONA analyzer (courtesy of Chemical Engineering
Department at Kuwait University),
that can be obtained is on the C/H ratio and sulfur content tion. Metallic constituents are associated with heavy com-
of a petroleum mixture from which one can determine the pounds and mainly appear in residues. There is no general
quality of oil. As boiling points of fractions increase or their method to determine the composition of all metals at once
API gravity decrease the amount of C/H ratio, sulfur content, but ASTM [4] provides test methods for determination of
nitrogen content, and the metallic constituents increase, sig- various metallic constituents (i.e., ASTM D 1026, D 1262,
nifying a reduction in the quality of an oil. Sulfur content of D 1318, D 1368, D 1548). Another method is to burn the oil
very heavy fractions can reach 6-8% and the nitrogen content sample in which metallic compounds appear in inorganic
can reach 2.0-2.5 wt%. There are specific methods to measure ashes. The ash should be digested by an acid and the so-
these elements individually. However, instruments do exist lution is examined for metal species by atomic absorption
that measure these elements all together; these are called ele- spectroscopy [7].
mental analyzers. One of these apparatuses is CHN analyzers
in which there is a simultaneous combustion in pure oxygen
3.1.6 Viscosity
at 1000~ Carbon is reduced to CO2, H is reduced to H20,
and N is converted to nitrogen oxides. Nitrogen oxides are Absolute and kinematic viscosities are defined in Section
then reduced over copper at 650~ to nitrogen by eliminating 2.1.8 and experimental data for the kinematic viscosity of
oxygen. A mixture of CO2, H/O, and N2 is separated by gas some pure hydrocarbons are given in Table 2.2. Viscosity of
chromatography with TCD. In a similar fashion, sulfur is ox- petroleum fractions increase with a decrease in the API grav-
idized to SO2 and is detected by TCD after detection of CO2, ity and for residues and heavy oils with the API gravity of
Na, and H20. Oxygen is determined by passing it over carbon less than 10 (specific gravity of above 1), viscosity varies from
at high temperature and converted to CO, which is measured several thousands to several million poises. Viscosity is a bulk
by a GC [5]. ASTM test methods for elemental analysis of property that can be measured for all types of petroleum frac-
petroleum products and crude oils include hydrogen content tions in liquid form. Kinematic viscosity is a useful character-
(ASTM D 1018, D 3178, D 3343), nitrogen content (ASTM ization parameter for heavy fractions in which boiling point
D 3179, D 3228, D 3431), and sulfur content (ASTM D 129/IP data are not available due to thermal decomposition during
61, D 1266/IP 107, D 1552, D 4045). An elemental analyzer distillation. Not only is viscosity an important physical prop-
Model CHNS-932 (Leco Corp., St. Joseph, MI, USA) is shown erty, but it is a parameter that can be used to estimate other
in Fig. 3.18. In this analyzer, the CO2, H20, and SO2 are physical properties as well as the composition and quality of
detected by infrared detector (IRD) while N2 is determined undefined petroleum fractions as shown later in this chap-
by the TCD method. ter. Since viscosity varies with temperature, values of viscos-
Another group of heteroatoms found in petroleum mix- ity must be reported with specified temperature. Generally,
tures are metallic constituents. The amount of these met- kinematic viscosity of petroleum fractions are measured at
als are in the range of few hundreds to thousand ppm and standard temperatures of 37.8~ (100~ and 98.9~ (210~
their amounts increase with increase in boiling points or de- However, for very heavy fractions viscosity is reported at tem-
crease in the API gravity of oil. Even a small amount of these peratures above 38~ i.e., 50~ (122~ or 60~ (140~ When
metals, particularly nickel, vanadium, iron, and copper, in viscosity at two temperatures are reported from the method
the feedstocks for catalytic cracking have negative effects on of Section 2.7 one can obtain the viscosity at other tempera-
the activity of catalysts and result in increased coke forma- tures. Measurement of viscosity is easy but the method and
100 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
FIG. 3.18--Leco made CHNS-932 model elemental analyzer (courtesy of Chemical Engi-
neering Department at Kuwait University).
the instrument depend on the type of sample. For Newto- for simplicity in experimental measurement and the need for
nian and high-shear fluids such as engine oils, viscosity can another type for its application requires conversion methods
be measured by a capillary U-type viscometer. An example of between various distillation curves. The tedious procedures
such viscometers is the Cannon-Fenske viscometer. The test necessary to obtain experimental EFV data have given im-
method is described in ASTM D 445, which is equivalent to petus to the development of correlations for predicting EFV
ISO 3104 method, and kinematic viscosity is measured at tem- data from the analytical ASTM and TBP distillations. Sim-
peratures from 15 to 100~ (~60 to 210~ In this method, ulated distillation by gas chromatography appears to be the
repeatability and reproducibility are 0.35 and 0.7%, respec- most simple, reproducible, and consistent method to describe
tively [5]. Another type of viscometer is a rotary viscometer, the boiling range of a hydrocarbon fraction unambiguously.
which is used for a wide range of shear rates, especially for TBP is the most useful distillation curve, while available data
low shear rate and viscous fluids such as lubricants and heavy might be ASTM D 86, ASTM D 2887, or ASTM D 1160 distilla-
petroleum fractions. In these viscometers, fluid is placed be- tion curves. ASTM [4] has accepted this technique as a tenta-
tween two surfaces, one is fixed and the other one is rotating. tive method for the "Determination of Boiling Range Distribu-
In these viscometers absolute viscosity can be measured and tion of Petroleum Fractions by Gas Chromatography" (ASTM
an example of such viscometers is the Brookfield viscome- D 2887). In most cases distillation data are reported in terms
ter. Details of measurement and prediction of viscosity of of ASTM D 86 or SD. In this section methods of calculation of
petroleum fractions are given in Chapter 8. As the viscosity average boiling points, interconversion of various distillation
of petroleum fractions, especially the heavy oils, is one of the curves, and prediction of complete distillation curves from a
most difficult properties to estimate, its experimental value is limited data are presented.
highly useful and desirable.
3.2.1 Average Boiling Points
Boiling points of petroleum fractions are presented by distil-
3.2 PREDICTION AND CONVERSION lation curves such as ASTM or TBR However, in prediction of
OF DISTILLATION DATA physical properties and characterization of hydrocarbon mix-
tures a single characteristic boiling point is required. Gener-
Various distillation curves are introduced in Section 3.1.1. ally an average boiling point for a fraction is defined to de-
For simplicity ASTM is used to refer to ASTM D 86 distilla- termine the single characterizing boiling point. There are five
tion curve, similarly TBP, SD, and EFV refer to true boiling average boiling points defined by the following equations [9].
point, simulated distillation (ASTM D 2887), and equilibrium Three of these average boiling points are VABP (volume aver-
flash vaporization, respectively. Petroleum fractions have a age boiling point), MABP (molal average boiling point) and
range of boiling points. To use the correlations introduced WABP (weight average boiling point), defined for a mixture
in Chapter 2, a single value for boiling point is required. For of n components as
this reason there is a need for the definition of an average
tt
boiling point or a characteristic boiling point based on a dis-
(3.3) ABP = ~-~ x4Tb~
tillation curve. Availability of one type of distillation curve i=1
3. C H A R A C T E R I Z A T I O N OF PETROLEUM FRACTIONS 101
w h e r e ABP is the VABP, MABP, o r WABP a n d xi is the corre- VABP a n d SL are defined in Eqs. (3.6) a n d (3.7). Once AT
s p o n d i n g volume, mole, o r weight fraction of c o m p o n e n t i. is calculated for each case, c o r r e s p o n d i n g ABP is calculated
Tbi is the n o r m a l boiling p o i n t of c o m p o n e n t i in kelvin. Two from Eq. (3.8). E q u a t i o n s (3.8)-(3.12) calculate values of var-
o t h e r average boiling points are CABP (cubic average boiling ious ABP very close to those o b t a i n e d from e m p i r i c a l plot in
point) a n d MeABP ( m e a n average boiling point) defined as the API-TDB [2]. The following example shows a p p l i c a t i o n of
these equations in calculation of various ABE
(3.4) CABP = x~ (1.8Tbi -- 459.67) 1/3 + 255.37
E x a m p l e 3 . 1 - - A low boiling n a p h t h a has the ASTM D 86 tem-
p e r a t u r e s of 77.8, 107.8, 126.7, 155, a n d 184.4~ at 10, 30, 50,
MABP + CABP
(3.5) MeABP = 70, a n d 90 vol% distilled [11]. Calculate VABE WABE MABP,
2 CABP, a n d MeABP for this fraction.
w h e r e T~ in Eq. (3.4) is in kelvin. The conversion factors in
Eq. (3.4) c o m e from the fact t h a t the original definition of S o l u t i o n - - F r o m Eqs. (3.6) a n d (3.7) VABP a n d SL are cal-
CABP is in degrees Fahrenheit. F o r p e t r o l e u m fractions in culated as follows: VABP = (77.8 + 107.8 + 126.7 + 155 +
w h i c h volume, weight, o r m o l e fractions of c o m p o n e n t s are 184.4)/5 = 130.3~ = 403.5 K, a n d SL = ( 1 8 4 . 4 - 7 7 . 8 ) /
n o t known, the average boiling p o i n t s are calculated t h r o u g h 80 = 1.333~ (K)/%. F r o m Eqs. (3.9)-(3.12) various c o r r e c t i o n
ASTM D 86 distillation curve as t e m p e r a t u r e s are calculated: ATw = -3.3~ ATu = 13.8~
T10 + T30 + T50 + T70 + T90 A T c = 3.2~ a n d ATMe = 8.6~ F r o m Eq. (3.8) various av-
(3.6) VABP = erage boiling p o i n t s are calculated: WABP = 133.7, MABP =
5
116.5, CABP = 127.1, a n d MeABP = 121.7~
w h e r e 1"10, T30, Ts0, I"70, a n d T90 are ASTM t e m p e r a t u r e s at
10, 30, 50, 70, a n d 90 vol% distilled. ASTM distillation curves
Application a n d e s t i m a t i o n of various boiling p o i n t s are
c a n be c h a r a c t e r i z e d by the m a g n i t u d e s of t e m p e r a t u r e s a n d
discussed b y Van Winkle [12]. Since the m a t e r i a l s boil over a
overall slope of the curve. A p a r a m e t e r t h a t a p p r o x i m a t e l y
range of t e m p e r a t u r e , any one average boiling p o i n t fails to be
c h a r a c t e r i z e s slope o f a distillation curve is the slope o f a
useful for correlation of all properties. The m o s t useful type
l i n e a r line between 10 and 90% points. This slope shown by
of ABP is MeABP, which is r e c o m m e n d e d for correlation of
SL is defined as
m o s t physical p r o p e r t i e s as well as calculation of Watson K
T90 - Tl0 as will be discussed later in this chapter. However, for calcula-
(3.7) SL -
80 tion of specific heat, VABP is r e c o m m e n d e d [ 12]. I n E x a m p l e
w h e r e T10 a n d T90 are the ASTM D 86 t e m p e r a t u r e s at 10 a n d 3.1, MeABP is 121.7~ w h i c h varies from 126.7 for the ASTM
90% of v o l u m e vaporized. The 10-90 slope, SL, in s o m e refer- D 86 t e m p e r a t u r e at 50 vol% distilled (T50). However, b a s e d on
ences is referred to as the Engler slope a n d is indicative of a o u r experience, for n a r r o w boiling r a n g e fractions with SL <
variety of c o m p o u n d s in a p e t r o l e u m fraction. W h e n the boil- 0.8~ the MeABP is very close to 50% ASTM t e m p e r a t u r e .
ing points of c o m p o u n d s are n e a r each o t h e r the value of SL As a n example, for a gas oil s a m p l e [11] with ASTM t e m p e r a -
a n d the boiling range of the fraction are low. F o r p e t r o l e u m tures of 261.7, 270, 279.4, 289.4, a n d 307.2~ at I0, 30, 50, 70,
fractions, WABP, MABP, CABP, a n d MeABP are c o r r e l a t e d a n d 90 vol%, the MeABP is calculated as 279, w h i c h is very
t h r o u g h an e m p i r i c a l plot to VAPB a n d SL in C h a p t e r 2 of the close to 50% ASTM t e m p e r a t u r e of 279.4~ F o r this fraction
API-TDB [2]. Analytical correlations b a s e d on the API plot the value of SL is 0.57~ w h i c h indicates the boiling range
were developed b y Zhou [10] for use in a digital computer. is quite narrow. Since n o n e of the average boiling p o i n t s de-
F o r heavy fractions a n d v a c u u m distillates in w h i c h distilla- fined here r e p r e s e n t the true boiling p o i n t of a fraction, the
tion d a t a by ASTM D i 160 are available, they should first be 50% ASTM t e m p e r a t u r e m a y be u s e d as a c h a r a c t e r i s t i c boil-
converted to ASTM D 86 a n d then average boiling p o i n t s are ing p o i n t instead of average boiling point. I n this case it is
calculated. Analytical correlations for e s t i m a t i o n of average a s s u m e d that the difference b e t w e e n these t e m p e r a t u r e s is
boiling p o i n t s are given by the following equations in t e r m s within the range of e x p e r i m e n t a l u n c e r t a i n t y for the r e p o r t e d
of VABP a n d SL [2, 10]. distillation d a t a as well as the correlation u s e d to e s t i m a t e a
physical property.
(3.8) ABP = VABP - AT
l n ( - A T w ) = -3.64991 - 0.02706(VABP - 273.15) 0.6667
3.2.2 Interconversion o f Various Distillation Data
(3.9) + 5.163875SL ~
W o r k to develop e m p i r i c a l m e t h o d s for converting ASTM dis-
ln(ATM) = - 1.15158 - 0.01181 (VABP - 273.15) 0"6667
tillations to TBP a n d EFV distillations b e g a n in the late 1920s
(3.10) + 3.70612SL 0'333 a n d c o n t i n u e d t h r o u g h the 1950s a n d 1960s by a large n u m b e r
ln(ATc) = - 0 . 8 2 3 6 8 - 0.08997(VABP - 273.15) 0.45 of r e s e a r c h e r s [13-18]. All of the correlations were b a s e d on
d i s c o r d a n t e x p e r i m e n t a l d a t a from the literature. E x p e r i m e n -
(3.1 I) + 2.456791SL ~ tal ASTM, TBP, a n d EFV d a t a on w h i c h the e m p i r i c a l correla-
ln(ATue) = - 1.53181 - 0.0128(VABP - 273.15) 0.6667 tions are b a s e d suffer a lack of r e p r o d u c i b i l i t y b e c a u s e there
(3.12) + 3.646064SL ~ were no s t a n d a r d i z e d p r o c e d u r e s o r a p p a r a t u s available. All
of these correlations were evaluated a n d c o m p a r e d to each
w h e r e ABP is an average boiling p o i n t such as WABP, MABP, o t h e r by H o u s e et al. [ 19] to select m o s t a p p r o p r i a t e m e t h o d s
CABP, o r MeABP a n d AT is the c o r r e s p o n d i n g c o r r e c t i o n for inclusion in the API-TDB. As a result of their evaluations,
t e m p e r a t u r e for each ABP. All t e m p e r a t u r e s are in kelvin. the following m e t h o d s were a d o p t e d in the API Data Book
102 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
as the best method: E d m i s t e r - P o l l o c k [14] for ASTM to TBP, TABLE 3.2---Correlation constants for Eq. (3.14).
E d m i s t e r - O k a m o t o [15-17] for ASTM to EFV, a n d Maxwell ASTM D 86
Vol% a b range,a~C
for conversion of TBP to EFV [19]. Most of these correlations
0 0.9177 1.0019 20-320
were in g r a p h i c a l forms a n d inconvenient for c o m p u t e r appli- 10 0.5564 1.0900 35-305
cations. Later, Arnold c o m p u t e r i z e d these g r a p h i c a l m e t h o d s 30 0.7617 1.0425 50-315
t h r o u g h a set of nth o r d e r p o l y n o m i a l s [20]. Correlation to 50 0.9013 1.0176 55-320
convert ASTM D 2887 (SD) to ASTM D 86 were first developed 70 0.8821 1.0226 65-330
by F o r d using m u l t i p l i e r regression analysis [21]. In the m i d 90 0.9552 1.0110 75-345
95 0.8177 1.0355 75-400
1980s Riazi a n d D a u b e r t [22] developed analytical m e t h o d s Source: Ref. [22].
for the conversion of distillation curves b a s e d on the general- aTemperatures are approximated to nearest 5.
ized correlation for h y d r o c a r b o n p r o p e r t i e s given b y Eq. (2.2).
These m e t h o d s were a d o p t e d by the API in the fifth edition of lation curve is available then ASTM curve can be e s t i m a t e d as
API-TDB-88 [2] to replace the previous methods. Continued
interests f r o m the p e t r o l e u m i n d u s t r y for these conversion (3.15) ASTM D 86 = (TBP) 1/b
m e t h o d s led to d e v e l o p m e n t of further methods. The latest
m e t h o d s for the conversion of distillation curves were devel- w h e r e constants a a n d b are given in Table 3.2 as for Eq. (3.14).
o p e d by D a u b e r t in m i d 1990s [23] t h r o u g h m o d i f y i n g R i a z i -
D a u b e r t correlations. In this section the API m e t h o d s ( R i a z i - 3.2.2.1.2 A S T M D 86 and E F V Conversions--Application
D a u b e r t a n d Daubert) for conversion of distillation d a t a are of Eq. (2.13) to this type of conversion gives
presented, w h i c h are also r e c o m m e n d e d a n d used in o t h e r
(3.16) EFV = a(ASTM D 86)b(SG) c
references a n d industrial software [24, 25].
w h e r e constants a, b, a n d c were o b t a i n e d from m o r e t h a n 300
3.2.2.1 Riazi-Daubert M e t h o d d a t a points a n d are given in Table 3.3. E q u a t i o n (3.16) was
Riazi a n d D a u b e r t m e t h o d s for the interconversion of vari- evaluated with m o r e t h a n 300 d a t a points f r o m 43 different
ous distillation d a t a are b a s e d on the generalized c o r r e l a t i o n s a m p l e s a n d gave AAD of 6~ while the m e t h o d of E d m i s t e r -
for p r o p e r t y e s t i m a t i o n of h y d r o c a r b o n s in the form of Eq. O k a m o t o [15] gave an AAD of 10~ [22, 26].
(2.38). Available distillation t e m p e r a t u r e a n d specific gravity In using these equations if specific gravity of a fraction is
of the fraction are used as the i n p u t p a r a m e t e r s to e s t i m a t e n o t available, it m a y be e s t i m a t e d from available distillation
the d e s i r e d distillation d a t a in the following form [22]: curves at 10 a n d 50% p o i n t s as given by the following equa-
tion:
(3.13) T/ (desired) = a [T/ (available)] b SG c
(3.17) SG = aTbmoT(o
w h e r e T/ (available) is the available distillation t e m p e r a t u r e
w h e r e constants a, b, a n d c for the three types of distillation
at a specific vol% distilled a n d T/(desired) is the desired distil-
data, namely, ASTM D 86, TBP, a n d EFV, are given in Table
lation d a t a for the s a m e vol% distilled, b o t h are in kelvin. SG
3.4. T e m p e r a t u r e s at 10 a n d 50% are b o t h in kelvin.
is the specific gravity of fraction at 15.5~ a n d a, b, a n d c are
correlation p a r a m e t e r s specific for each conversion type a n d
3.2.2.1.3 SD to A S T M D 86 Conversions--The equation
each vol% p o i n t on the distillation curve. F o r example, if this
derived from Eq. (3.13) for the conversion of s i m u l a t e d dis-
e q u a t i o n is u s e d to convert ASTM to EFV at 10%, T,.(available)
tillation (ASTM D 2887) to ASTM D 86 distillation curve has
is ASTM t e m p e r a t u r e at 10% a n d T/(desired) is the EFV tem-
the following form:
p e r a t u r e at 10% a n d constants a, b, a n d c are specific for this
conversion type at 10% of volume vaporized. (3.18) ASTM D 86 = a(SD)b(F) C
w h e r e c o n s t a n t F is a p a r a m e t e r specifically used for this type
3.2.2.1.1 A S T M D 86 and TBP Conversion--If distillation of conversion a n d is given by the following equation:
d a t a available are in the form of ASTM D 86 a n d d e s i r e d dis-
tillation is TBP, Eq. (3.13) can be used, b u t for this p a r t i c u l a r (3.19) F = 0.01411(SD 10~ 50~
type of conversion value of c o n s t a n t c for all p o i n t s is zero in w h i c h SD 10% a n d SD 50% are the SD t e m p e r a t u r e s in
a n d the e q u a t i o n reduces to kelvin at 10 a n d 50 wt% distilled, respectively. P a r a m e t e r F
(3.14) TBP = a(ASTM D 86) b calculated f r o m Eq. (3.19) m u s t be substituted in Eq. (3.18) to
estimate ASTM D 86 t e m p e r a t u r e at c o r r e s p o n d i n g p e r c e n t
w h e r e b o t h TBP a n d ASTM t e m p e r a t u r e s are for the s a m e p o i n t expressed in v o l u m e basis. E q u a t i o n (3.18) c a n n o t be
vol% distilled a n d are in kelvin. Constants a a n d b at various
p o i n t s along the distillation curve with the r a n g e of applica- TABLE 3.3---Correlation constants for Eq. (3.16).
tion are given in Table 3.2. ASTM D 86
F o r a total of 559 d a t a p o i n t s for 80 different samples, Eq. Vol% a b c range,a~C
(3.14) gives an average absolute deviation (AAD) of a b o u t 0 2.9747 0.8466 0.4209 10-265
10 1.4459 0.9511 0.1287 60-320
5~ while the E d m i s t e r - P o l l o c k m e t h o d [14] gives an AAD 30 0.8506 1.0315 0.0817 90-340
of a b o u t 7~ Generally p r e d i c t i o n s at 0% give higher errors 50 3.2680 0.8274 0.6214 110-355
a n d are less reliable. Details of evaluations are given in o u r 70 8.2873 0.6871 0.9340 130-400
previous p u b l i c a t i o n s [22, 26]. E q u a t i o n (3.14) can be easily 90 10.6266 0.6529 1.1025 160-520
100 7.9952 0.6949 1.0737 190-430
reversed to p r e d i c t ASTM f r o m TBP data, b u t this is a rare Source: Ref. [22].
a p p l i c a t i o n as usually ASTM d a t a are available. If TBP distil- aTemperatures are approximated to nearest 5.
3. C H A R A C T E R I Z A T I O N O F P E T R O L E U M F R A C T I O N S 103
used in a reverse form to predict SD from ASTM D 86, b u t this calculated ASTM a n d SG, the EFV t e m p e r a t u r e s can be es-
type of conversion is usually n o t desired as most predictive timated from Eq. (3.16) with constants given in Table 3.2.
m e t h o d s use ASTM D 86 data while laboratories report SD EFV = 2.9747 (32 + 273) 0.8466 (0.7862)~176 : 340.9 K ---
data. Constants a, b, a n d c in Eq. (3.18) were obtained from 340.9 - 273 = 67.9~ The calculated value is very close to the
81 different samples a n d 567 data points a n d are given i n actual value of 68.3~ (see Table 3.5). Similarly EFV values at
Table 3.5 with the range of SD data at each percentage along other points are calculated a n d results are s h o w n i n Fig. 3.19.
the distillation curve. Predicted EFV curve from TBP are very close to the actual
E q u a t i o n (3.18) a n d the m e t h o d of Ford published by EFV curve. The AAD between predicted EFV a n d experimen-
ASTM, included in the earlier editions of API-TDB [21], were tal data is 2.6 K. It should be n o t e d that if experimental ASTM
evaluated by some 570 data points a n d gave AAD of 5 a n d data a n d specific gravity were used, the predicted values of
5.5~ respectively [22, 26]. Larger errors were observed at EFV would be even closer to the experimental values. #
the initial a n d final boiling points (0 a n d 100%) b u t excluding
these points the AAD reduces to a b o u t 3~ for conversions 3.2.2.2 Daubert's M e t h o d
within the range of 10-90% distilled. Daubert a n d his group developed a different set of equations
The procedures given in this section should be used with the to convert ASTM to TBP, SD to ASTM, a n d SD to TBP [23].
range of data specified in Tables 3.1-3.4. Use of these equa- These m e t h o d s have b e e n included in the sixth edition of API-
tions outside the specified ranges could cause large errors. TDB [2] a n d are given in this section. I n these methods, first
Graphical forms of these equations for conversion of various conversion should be m a d e at 50% p o i n t a n d then the differ-
distillation curves are given in Reference [22] as well as in the ence between two cut points are correlated i n a form similar
fourth edition of the API-TDB-88 [2]. One of the advantages of to Eq. (3.14). I n this m e t h o d SD data can be converted di-
these equations is that they can be used i n reversed form. This rectly to TBP without calculating ASTM as was needed i n the
m e a n s one m a y estimate EFV from TBP data t h r o u g h conver- R i a z i - D a u b e r t method.
sion of TBP to ASTM by Eq. (3.15) a n d t h e n u s i n g Eq. (3.16)
to estimate EFV from calculated ASTM curve. The example 3.2.2.2.1 A S T M a n d T B P C o n v e r s i o n - - T h e following
below shows this conversion process. e q u a t i o n is used to convert a n ASTM D 86 distillation at 50%
p o i n t t e m p e r a t u r e to a TBP distillation 50% p o i n t tempera-
E x a m p l e 3 . 2 - - F o r a b l e n d of n a p h t h a - k e r o s e n e sample, ture.
ASTM, TBP, a n d EFV distillation curves are given in the API-
TDB [2]. These data are represented in Table 3.6. Use the TBP(50%) = 255.4 + 0.8851[ASTM D 86(50%) - 255.4] 1~
R i a z i - D a u b e r t m e t h o d s to predict EFV curve from TBP curve. (3.20)
where ASTM (50%) a n d TBP (50%) are t e m p e r a t u r e s at 50%
$ o l u t i o n - - T B P data are used as available i n p u t data. Equa-
volume distilled in kelvin. E q u a t i o n (3.20) c a n also be used
tion (3.15) should be used to estimate ASTM D 86 from TBP.
i n a reverse form to estimate ASTM from TBP. The following
For the initial p o i n t at 0%, the calculations are as follows.
e q u a t i o n is used to d e t e r m i n e the difference between two cut
ASTM D 86 = (1/0.9177) 1/1~176(10 + 273) 1/1~176= 305 K =
points:
305 - 273 = 32~ The actual data for the initial ASTM tem-
perature is 35~ which is close to the calculated value. (3.21) Y / = AX/B
Now to estimate EFV from Eq. (3.16), specific gravity, is re-
where
quired which is not given by the problem. SG can be esti-
difference in TBP t e m p e r a t u r e b e t w e e n two cut
m a t e d from Eq. (3.17) a n d constants given in Table 3.3 for Y/=
points, K (or ~
the TBP. F r o m Table 3.6, T10(TBP) = 71.1 a n d Ts0(TBP) =
observed difference in ASTM D 86 t e m p e r a t u r e be-
204.4~ Using these values i n Eq. (3.17) gives SG = 0.10431 Xi --
tween two cut points, K (or ~
(71.1 + 273) 0"1255 (204.4 + 273)020862= 0.7862. Now from
A , B = constants varying for each cut p o i n t a n d are given
in Table 3.7
TABLE 3.5---Correlation constants for Eq. (3.18).
SD
Vol% a b c rangefl~C TABLE 3.6---Data on various distillation curves
for a naphtha-kerosene blend [2].
0 5.1764 0.7445 0.2879 -20-200
10 3.7452 0.7944 0.2671 25-230 Vol% ASTMD 86, TBP, EFV,
distilled ~C ~C ~C
30 4.2749 0.7719 0.3450 35-255
50 1 . 8 4 4 5 0.5425 0.7132 55-285 0 35.0 10.0 68.3
70 1 . 0 7 5 1 0.9867 0.0486 65-305 10 79.4 71.1 107.2
90 1 . 0 8 4 9 0.9834 0.0354 80-345 30 145.6 143.3 151.1
100 1.799l 0.9007 0.0625 95--405 50 201.7 204.4 182.2
Source:Ref.[22]. 70 235.6 250.6 207.2
~Ternperaturesare approximatedto nearest 5. 90 270.6 291.7 228.3
104 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 3.8--Prediction of TBP from ASTM for a kerosene sample of Example 3.3.
Vol% ASTM D 86 TBP Eq. (3.14) Eqs. (3.20)-(3.22)
distilled exp,~ exp,~ TBP calc,~ AD,~ TBP calc,~ AD,~
0 165.6 146.1 134.1 12.0 133.1 13.0
10 176.7 160.6 160.6 0.0 158.1 2.5
30 193.3 188.3 188.2 0.1 189.2 0.9
50 206.7 209.4 208.9 0.5 210.6 1.2
70 222.8 230.6 230.2 0.4 232.9 2.3
90 242.8 255.0 254.7 0.3 258.1 3.1
Overall AAD,~ 2.2 3.8
R i a z i - D a u b e r t a n d Daubert's m e t h o d s a n d calculate AAD for d a t a are similar to the equations developed b y D a u b e r t [2, 23]
each method. to convert ASTM o r SD to TBP as given in this section. The
equations are s u m m a r i z e d as following:
Solution--The R i a z i - D a u b e r t m e t h o d s do n o t provide a di-
rect conversion f r o m SD to TBP, but one c a n use Eqs. (3.18) ASTM D 86(50vo1%) = 255.4 + 0.79424
a n d (3.19) to convert SD to ASTM D 86 a n d t h e n Eq. (3.14) (3.26) [SD(50 wt%) - 255.4] 1.0395
s h o u l d be used to convert ASTM to TBP data. F r o m Eq. (3.19)
a n d use of SD at 10% a n d 50% points, the value of p a r a m - w h e r e SD (50 wt%) a n d ASTM D 86 (50 vo1%) a r e t e m p e r a -
eter F is calculated as 0.8287. Value of SD t e m p e r a t u r e at tures at 50% distilled in kelvin. The difference b e t w e e n adja-
50 wt% is 168.9~ from Eq. (3.18) with a p p r o p r i a t e c o n s t a n t s cent cut points is calculated from the following e q u a t i o n as
in Table 3.5 one c a n o b t a i n ASTM D 86 (50%) = 166.3~ given b y the API-TDB [2].
S u b s t i t u t i n g this value for ASTM into Eq. (3.14) gives TBP
(50 vol%) = 167.7~ while the e x p e r i m e n t a l value as given (3.27) Ui = E T f
in Table 3.10 is 166.7~ The AD is t h e n calculated as
where
167.7-166.7 = 1~C. Daubert's m e t h o d for conversion of SD Ui = difference in ASTM D 86 t e m p e r a t u r e s between
to TBP is direct a n d is p r e s e n t e d t h r o u g h Eqs. (3.23)-(3.25).
two cut points, K (or ~
According to Eq. (3.23), TBP (50%) = SD (50%) = 168.9~
T / = observed difference in SD t e m p e r a t u r e s between
w h i c h gives a n AD of 2.2~ for this point. A s u m m a r y of com-
two cut points, K (or ~
plete calculation results is given in Table 3.10. The overall AAD
E, F = constants varying for each cut p o i n t a n d are given
for Eqs. (3.14) a n d (3.18) is 4.8, while for Eqs. (3.23)-(3.25)
in Table 3.11
is 2.2~
To d e t e r m i n e the ASTM D 86 t e m p e r a t u r e at any p e r c e n t
distilled, calculations should begin with 50% ASTM D 86 tem-
Results p r e s e n t e d in E x a m p l e 3.4 show t h a t Eqs. (3.23)- p e r a t u r e a n d a d d i t i o n o r s u b t r a c t i o n of the p r o p e r t e m p e r a -
(3.25) are m o r e a c c u r a t e t h a n Eqs. (3.14) a n d (3.18) for the ture difference Ui.
conversion of SD to TBE One of the reasons for such a result
is t h a t the s a m p l e p r e s e n t e d in Table 3.10 to evaluate these ASTM D 86(0%) = ASTM D 86(50%)
m e t h o d s is t a k e n from the s a m e d a t a b a n k used to develop
- U4 - Us - U6
correlations of Eqs. (3.23)-(3.25). In a d d i t i o n these equations
provide a direct conversion of SD to TBE However, one should ASTM D 86(10%) -- ASTM D 86(50%)
realize t h a t Eqs. (3.23)-(3.25) are b a s e d on only 19 datasets - u4 - u5
a n d this limits the a p p l i c a t i o n of these equations. While Eqs.
ASTM D 86(30%) -- ASTM D 86(50%) - U4
(3.14), (3.18), a n d (3.19) are b a s e d on m u c h larger d a t a b a n k s (3.28)
with w i d e r range of application. As available d a t a on b o t h ASTM D 86(70%) = ASTM D 86(50%) + U3
SD a n d TBP are very limited, a concrete r e c o m m e n d a t i o n on ASTM D 86(90%) = ASTM D 86(50%)
s u p e r i o r i t y of these two a p p r o a c h e s c a n n o t be m a d e at this
.-~ U3 -~ U 2
time.
ASTM D 86(100%) = ASTM D 86(50%)
3,2.2.2.3 SD to A S T M D 86 Conversion--Equations to + U3 + U2 + Ul
convert SD (ASTM D 2887) distillation d a t a to ASTM D 86
This m e t h o d is a p p l i c a b l e to fractions with ASTM D 86 50%
TABLE 3.9--Correlation constants for Eq. (3.24). points in the r a n g e o f 65-315~ (150-600~ The average ab-
Cutpoint M a x i m u m allowable solute deviation for this m e t h o d as r e p o r t e d b y the API-TDB
i range, % C D W/,~ [2] is a b o u t 6~ for s o m e 125 s a m p l e s a n d a p p r o x i m a t e l y 850
1 100-95 0.03849 1.9733 15 d a t a points. P r e d i c t e d ASTM t e m p e r a t u r e s at 0 a n d 100% are
2 95-90 0.90427 0.8723 20 the least a c c u r a t e values followed by values at 10 a n d 90%
3 90-70 0.37475 1.2938 40
4 70-50 0.25088 1.3975 40 points as is s h o w n in the following example.
5 50-30 0.08055 1.6988 40
6 30-10 0.02175 2.0253 40
7 10-0 0.20312 1.4296 20 Example 3 . 5 - - E x p e r i m e n t a l ASTM D 2887 (SD) a n d ASTM
Source: Taken with permission from Refs. [2, 23]. D 86 distillation d a t a for a p e t r o l e u m fraction are given in
aTemperatures are approximated to nearest 5. Table 3.11 as t a k e n f r o m the API-TDB [2]. Predict the ASTM
106 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
3.2.2.3.2 Conversion of a Distillation Curve from Sub- or in the above relations all t e m p e r a t u r e s are either in ~ o r in
Super- Atmospheric Pressures to a Distillation Curve at kelvin.
Atmospheric Pressure--The m e t h o d of conversion of boil-
ing p o i n t s t h r o u g h Eqs. (3.29)-(3.32) c a n b e used to every 3.2.2.4 Summary Chart for Interconverion of Various
p o i n t on a distillation curve u n d e r either sub- o r s u p e r a t m o - Distillation Curves
spheric p r e s s u r e conditions. I n these equations Tb o r T (760 A s u m m a r y of all conversion m e t h o d s is s h o w n in Fig. 3.20.
m m H g ) r e p r e s e n t a p o i n t along the distillation curve at at- It s h o u l d be n o t e d that any distillation curve at low p r e s s u r e
m o s p h e r i c pressure. It can be a p p l i e d to any of TBP, EFV, (i.e., ASTM D 1160 o r EFV at 1, 10, 50, m m H g o r TBP at
o r ASTM D 1160 distillation curves. However, it s h o u l d be 1 m m Hg) should be first converted to TBP distillation curve
n o t e d t h a t these equations convert distillation curves from at 10 m m Hg before they are converted to TBP at a t m o s p h e r i c
one p r e s s u r e to a n o t h e r w i t h i n the s a m e type. F o r example, it pressure.
is n o t possible to use these equations to directly convert ASTM
D 1160 at 10 m m H g to TBP at 760 m m H g . S u c h conversions Example 3 . 6 - - F o r a p e t r o l e u m fraction the ASTM D 1160
require two steps that are discussed in the following section. distillation d a t a at 10 m m Hg are given in Table 3.13. P r e d i c t
The only distillation curve type t h a t m i g h t be r e p o r t e d un- the TBP curve at a t m o s p h e r i c pressure.
d e r s u p e r a t m o s p h e r i c p r e s s u r e (P > 1.01325 bar) c o n d i t i o n
is the EFV distillation curve. TBP curve m a y be at I, 10, 100, Solution--ASTM D 1160 d a t a have b e e n converted to TBP
o r 760 m m Hg pressure. E x p e r i m e n t a l d a t a on ASTM D 1160 at 10 m m H g by Eq. (3.34). Then Eq. (3.29) with P - - 10
are usually r e p o r t e d at i, 10, o r 50 m m Hg. ASTM D 86 distil- m m H g a n d Q = 0.001956 is u s e d to convert TBP from 10
l a t i o n is always r e p o r t e d at a t m o s p h e r i c pressure. It should to 760 m m H g . A s u m m a r y of results is given in Table 3.13.
be n o t e d t h a t w h e n ASTM D 1160 distillation curve is con-
The second a n d less a c c u r a t e m e t h o d to convert TBP from 10
verted to o r r e p o r t e d at a t m o s p h e r i c p r e s s u r e (760 m m Hg)
to 760 m m Hg is t h r o u g h Eq. (3.32), w h i c h in its reverse form
it is not equivalent to or the s a m e as ASTM D 86 distillation
b e c o m e s T (760 m m H g ) = 1.17T (10 m m H g ) + 67.51. Esti-
data. They are different types of distillation curves a n d there m a t e d TBP at 760 m m Hg t h r o u g h this relation is p r e s e n t e d
is no direct conversion b e t w e e n these two curves.
in the last c o l u m n of Table 3.13. #
Superatmospheric
I
ASTMD2887
Simulated EFV
Distillation (P >760 mmHg)
(SD)
TBP ! --'-I
TBP
1, 50, 100mm 6 760 mmHg
Iv
TBP EFV
10 mmHg 10 mrnHg
ASTMD-1160
10 mmHg
4 V
ASTMD-1160 ASTMD-1160
6 Reported
at
1, 30, 50 mmHg ~" 760 mmHg
does not give a finite value for T at x = 1 (end point at 100% following linear form:
distilled). According to this model the final boiling point is
(3.36) Y = Cz + C2X
infinite (oo), which is true for heavy residues. Theoretically,
even for light products with a limited boiling range there is a where Y = In [(T - To) / To] and X - - In In [1 / (1 - x ) ] . Con-
very small a m o u n t of heavy c o m p o u n d since all c o m p o u n d s stants C1 and Ca are determined from linear regression of Y
in a mixture cannot be completely separated by distillation. versus X with an initial guess for To. Constants A and B are
For this reason predicted values from Eq. (3.35) are reliable determined from C1 and C2 as B = 1/C2 and A = B exp (C1B).
up to x = 0.99, but not at the end point. Parameters A, B, and Parameter To can be determined by several estimates to maxi-
To in Eq. (3.35) can be directly determined by using Solver mize the R squared (RS) value for Eq. (3.36) and minimize the
(in Tools) in Excel spreadsheets. Another way to determine AAD for prediction of T form Eq. (3.35). If the initial boiling
the constants in Eq. (3.35) is t h r o u g h its conversion into the point in a distillation curve is available it can be used as the
110 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 3.13---Conversion of ASTM D 1160 to TBP at 760 mm Hg Development of these relations is discussed in C h a p t e r 4.
for the petroleum fraction of Example 3.6. I n Eq. (3.35), i f x is v o l u m e fraction, t h e n Tar c a l c u l a t e d f r o m
XZol% ASTM
distilled D 1160, ~ TBP%,~ TBl~76O,
~ TBP~76o,~ Eq. (3.37) w o u l d be volume average boiling p o i n t (VABP)
a n d if x is the weight fraction t h e n Tar w o u l d b e equivalent
10 150 142.5 280.1 280.7
30 205 200.9 349.9 349.0 to the weight average boiling p o i n t (WABP). S i m i l a r l y m o l e
50 250 250 407.2 406.5 average boiling p o i n t c a n be e s t i m a t e d from this e q u a t i o n
70 290 290 453.1 453.3 if x is in m o l e fraction. However, the m a i n a p p l i c a t i o n of
90 350 350 520.4 523.5 Eq. (3.35) is to p r e d i c t c o m p l e t e distillation curve from a lim-
aEq. (3.34).
bEq. (3.29). ited d a t a available. It can also be used to p r e d i c t boiling p o i n t
eEq. (3.32). of residues in a crude oil as will be s h o w n in C h a p t e r 4. Equa-
tion (3.35) is also perfectly a p p l i c a b l e to density o r specific
initial guess, b u t value of To s h o u l d always be less t h a n value gravity d i s t r i b u t i o n along a distillation curve for a p e t r o l e u m
of T for the first d a t a point. F o r fractions with final boiling fraction a n d c r u d e oils. F o r the case of density, p a r a m e t e r T
p o i n t very high a n d uncertain, such as a t m o s p h e r i c o r vac- is r e p l a c e d b y d or SG and density of the m i x t u r e m a y be cal-
u u m residues a n d h e p t a n e - p l u s fraction of crude oils, value o f c u l a t e d from Eq. (3.37). W h e n Eqs. (3.35)-(3.38) are u s e d for
B c a n be set as 1.5 a n d Eq. (3.35) reduces to a t w o - p a r a m e t e r p r e d i c t i o n of d e n s i t y of p e t r o l e u m fractions, the value of RS
equation. However, for various p e t r o l e u m fractions with fi- is less t h a n t h a t of distillation data. While the value of B for
nite boiling r a n g e p a r a m e t e r B s h o u l d be d e t e r m i n e d from the case of density is greater t h a n that of boiling p o i n t a n d is
the regression analysis a n d value of B for light fractions is usually 3 for very heavy fractions (C7+) a n d h i g h e r for lighter
h i g h e r t h a n t h a t of heavier fractions a n d is n o r m a l l y g r e a t e r mixtures. It s h o u l d be n o t e d t h a t w h e n Eqs. (3.35)-(3.38) are
t h a n 1.5. E q u a t i o n (3.35) can be a p p l i e d to a n y type of distil- a p p l i e d to specific gravity o r density, x s h o u l d b e cumulative
lation data, ASTM D 86, ASTM D 2887 (SD), TBP, EFV, a n d v o l u m e fraction. F u r t h e r p r o p e r t i e s a n d a p p l i c a t i o n of this
ASTM D 1160 as well as TBP at r e d u c e d pressures o r EFV d i s t r i b u t i o n function as well as m e t h o d s of calculation of av-
at elevated pressures. I n the case of SD curve, x is c u m u l a - erage p r o p e r t i e s for the m i x t u r e are given in C h a p t e r 4. Here-
tive weight fraction distilled. The average boiling p o i n t of the in we d e m o n s t r a t e use for this m e t h o d for p r e d i c t i o n of dis-
fraction c a n be d e t e r m i n e d from the following relation: tillation curves of p e t r o l e u m fractions t h r o u g h the following
example.
Tav -= To(1 + Tar)
(3.37' (A):~ ( 1) Example 3 . 7 - - A S T M D 86 distillation d a t a from initial to
T~*v= F 1+
final boiling p o i n t for a gas oil s a m p l e [ i] are given in the first
two c o l u m n s of Table 3.14. Predict the distillation curve for
in w h i c h F is the g a m m a function a n d m a y be d e t e r m i n e d
the following four cases:
f r o m the following relation w h e n value of p a r a m e t e r B is
g r e a t e r t h a n 0.5.
a. Use d a t a points at 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, a n d
F (l +l) =o.992814-O.504242B-l +O.696215B -z 95 vo1% distilled.
b. Use all d a t a p o i n t s from 5 to 70 vol% distilled.
- 0 . 2 7 2 9 3 6 B -3 + 0.088362B -4 c. Use three d a t a p o i n t s at 10, 30, a n d 50%.
(3.38) d. Use three d a t a p o i n t s at 30, 50, a n d 70%.
TABLE 3.14--Prediction of ASTM D 86 distillation curve for gas oil sample of Example 3.7.
Data Set A Data Set B Data Set C Data Set D
Vol% distilled Temp.exp, K Pred, K AD, K Pred, K AD, K Pred, K AD, K Pred, K AD, K
0 520.4 526.0 5.6 525.0 4.6 530.0 9.6 512.0 8.4
5 531.5 531.6 0.1 531.5 0.0 532.7 1.2 526.4 5.1
10 534.8 534.6 0.2 534.7 0.1 534.8 0.0 531.2 3.6
20 539.8 539.5 0.4 539.6 0.2 538.9 0.9 537.8 2.0
30 543.2 543.8 0.7 544.0 0.8 543.1 0.0 543.1 0.0
40 548.2 548.1 0.1 548.1 0.0 547.6 0.5 547.9 0.3
50 552.6 552.5 0.1 552.4 0.2 552.6 0.0 552.6 0.0
60 557.0 557.3 0.3 557.0 0.1 558.5 1.4 557.4 0.4
70 562.6 562.9 0.3 562.2 0.4 565.6 3.0 562.6 0.0
80 570.4 569.9 0.5 568.7 1.7 575.2 4.8 568.8 1.6
90 580.4 580.4 0.0 578.2 2.2 590.6 10.3 577.5 2.9
95 589.8 589.6 0.2 586.6 3.3 605.2 15.4 584.8 5.1
100 600.4 608.3 7.9 603.1 2.7 637.1 36.7 598.4 2.0
AAD (total), K 1.3 1.3 6.5 2.4
No. of data used 11 8 3 3
To 526 525 530 512
A 0.01634 0.0125 0.03771 0.00627
B 1.67171 1.80881 1.21825 2.50825
RS 0.9994 0.999 1 1
AAD (data used), K 0.25 0.23 0 0
VABP, K 554.7 555.5 555 557 554
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 111
For each case give parameters To, A, and B in Eq. (3.35) as and 70% points are used and the predicted values are more
well as value of RS and AAD based on all data points and accurate than values obtained in Case C. However, for this last
based on data used for the regression. Also calculate VABP case the highest error for the IBP is obtained because the first
from Eq. (3.37) and compare with actual VABP calculated data point used to obtain the constants is at 30%, which is far
from Eq. (3.6). from 0% point. Summary of results for predicted distillation
curves versus experimental data are also shown in Fig. 3.2 I.
Solution--Summary of calculation results for all four cases As can be seen from the results presented in both Table 3.14
are given in Table 3.14. For Case A all experimental data given and Fig. 3.21, a good prediction of the entire distillation curve
on the distillation curve (second column in Table 3.14) from is possible through use of only three data points at 30, 50, and
5 to 95% points are used for the regression analysis by Eq. 70%. r
(3.36). Volume percentages given in the first column should be
converted to cumulative volume fraction, x, (percent values
divided by 100) and data are converted to X and Y defined in 3.3 P R E D I C T I O N OF P R O P E R T I E S
Eq. (3.36). The first data point used in the regression process OF P E T R O L E U M F R A C T I O N S
is at x = 0.05 with T = 531.5 K; therefore, the initial guess
(To) should be less than 531. With a few changes in To values, As discussed in Chapter 1, petroleum fractions are mixtures of
the maximum RS value of 0.9994 is obtained with minimum many hydrocarbon compounds from different families. The
AAD of 0.25 K (for the 11 data points used in the regression most accurate method to determine a property of a mixture
process). The AAD for the entire data set, including the IBP is through experimental measurement of that property. How-
and FBP, is 1.3 K. As mentioned earlier the experimentally re- ever, as this is not possible for every petroleum mixture, meth-
ported IBP and especially the value of FBP are not accurate. ods of estimation of various properties are needed by process
Therefore, larger errors for prediction of IBP and FBP are ex- or operation engineers. The most accurate method of esti-
pected from Eq. (3.35). Since values of FBP at x = 1 are not mating a property of a mixture is through knowledge of the
finite, the value of T at x = 0.99 may be used as an approx- exact composition of all components existing in the mixture.
imate predicted value of FBP from the model. These values Then properties of pure components such as those given in
are given in Table 3.14 as predicted values for each case at Tables 2.1 and 2.2 can be used together with the composi-
100 vol% vaporized. Estimated VABP from Eq. (3.37) for Case tion of the mixture and appropriate mixing rules to determine
A is 555.5 versus value of 554.7 from actual experimental data properties of the mixture. If experimental data on properties
and definition of VABP by Eq. (3.6). of pure compounds are not available, such properties should
For Case B, data from 5 to 70 vol% distilled are used for be estimated through the methods presented in Chapter 2.
the regression process and as a result the predicted values Application of this approach to defined mixtures with very
up to 70% are more accurate than values above 70% point. few constituents is practical; however, for petroleum mixtures
However, the overall error (total AAD) is the same as for Case with many constituents this approach is not feasible as the
A at 1.3 K. For Case C only three data points at 10, 30, and determination of the exact composition of all components
50% are used and as a result much larger errors especially in the mixture is not possible. For this reason appropriate
for points above 50% are observed. In Case D, data at 30, 50, models should be used to represent petroleum mixtures by
some limited number of compounds that can best represent
the mixture. These limited compounds are different from the
650 real compounds in the mixture and each is called a "pseudo-
component" or a "pseudocompound". Determination of these
pseudocompounds and use of an appropriate model to de-
o Exp. data "l scribe a mixture by a certain number of pseudocompounds
-- Pred.(data set A) '1/ is an engineering art in prediction of properties of petroleum
. . . . . . . Pred. (data set B) ; "t mixtures and are discussed in this section.
600
..... Pred. (data set C) 1'~'
m ~ Perd. (data set D) ~o4'(.~q? 3.3.1 M a t r i x o f P s e u d o c o m p o n e n t s Table
As discussed in Chapter 2, properties of hydrocarbons vary
~9
by both carbon number and molecular type. Hydrocarbon
properties for compounds of the same carbon number vary
550
from paraffins to naphthenes and aromatics. Very few frac-
tions may contain olefins as well. Even within paraffins fam-
ily properties of n-paraffins differ from those of isoparaffins.
Boiling points of hydrocarbons vary strongly with carbon
number as was shown in Table 2.1; therefore, identification of
500 . . . . . . . . hydrocarbons by carbon number is useful in property predic-
0 20 40 60 80 100 tions. As discussed in Section 3,1.5.2, a combination of GS-
MS in series best separate hydrocarbons by carbon number
Vol% Distilled
and molecular type. If a mixture is separated by a distillation
FIG. 3.211Prediction of distillation curves for the gas oil column or simulated distillation, each hydrocarbon cut with a
sample of Example 3.7. single carbon number contains hydrocarbons from different
112 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
groups, which can be identified by a PIONA analyzer. As an are rich in one hydrocarbon type such as coal liquids that
example in Table 1.3 (Chapter 1), carbon number ranges for may have up to 90% aromatics, it would be appropriate to di-
different petroleum products are specified. For a diesel fuel vide the aromatics into further subgroups of monoaromatics
sample, carbon number varies from Cll to C16 with a boiling (MA) and polyaromatics (PA). Therefore, creation of a matrix
range of 400-550~ If each single carbon number hydrocar- of pseudocomponents, such as Table 3.15, largely depends on
bon cut is further separated into five pseudocomponents from the nature and characteristics of the petroleum mixture as
different groups, the whole mixture may be represented by a well as availability of experimental data.
group of 30 pseudocomponents as shown in Table 3.15. Al-
though each pseudocomponent is not a pure hydrocarbon but
3.3.2 Narrow Versus Wide Boiling Range Fractions
their properties are very close to pure compounds from the
same family with the same carbon number. If the amounts of In general, regardless of molecular type, petroleum fractions
all these 30 components are known then properties of the mix- may be divided into two major categories: narrow and wide
ture may be estimated quite accurately. This requires exten- boiling range fractions. A narrow boiling range fraction was
sive analysis of the mixture and a large computation time for defined in Section 3.2.1 as a fraction whose ASTM 10-90%
estimation of various properties. The number of pseudocom- distillation curve slope (SL) is less than 0.8~ although
ponents may even increase further if the fraction has wider this definition is arbitrary and may vary from one source to
boiling point range such as heptane plus fractions as will be another. Fractions with higher 10-90% slopes may be consid-
discussed in Chapter 4. However, many petroleum fractions ered as wide boiling range. However, for simplicity the meth-
are olefin free and groups of n-paraffins and isoparaffins may ods presented in this section for narrow fractions may also
be combined into a single group of paraffins. Therefore, the be applied to wider fractions. For narrow fractions, only one
number of different families reduces to three (paraffins, naph- carbon number is considered and the whole fraction may be
thenes, and aromatics). In this case the number of compo- characterized by a single value of boiling point or molecu-
nents in Table 3.15 reduces to 6 x 3 or 18. If a fraction is nar- lar weight. For such fractions, if molecular type is known
row in boiling range then the number of rows in Table 3.15 (PNA composition), then the number of pseudocomponents
decreases indicating lower carbon number range. In Chapter in Table 3.15 reduces to three and if the composition is not
4, boiling points of various single carbon number groups are known the whole mixture may be considered as a single pseu-
given and through a TBP curve it would be possible to deter- docomponent. For this single pseudocomponent, properties
mine the range of carbon number in a petroleum fraction. In of a pure component whose characteristics, such as boiling
Table 3.15, if every two carbon number groups and all paraf- point and specific gravity, are the same as that of the fraction
fins are combined together, then the whole mixture may be can be considered as the mixture properties. For mixtures the
represented by 3 x 3 or 9 components for an olefin-free frac- best characterizing boiling point is the mean average boiling
tion. Similarly if all carbon numbers are grouped into a single point (MeABP); however, as mentioned in Section 3.2.1, for
carbon number group, the mixture can be represented by only narrow fractions the boiling point at 50 vol% distilled may
three pseudocomponents from paraffins (P), naphthenes (N), be considered as the characteristic boiling point instead of
and aromatics (A) groups all having the same carbon number. MeABP. The specific gravity of a fraction is considered as the
This approach is called pseudocomponent technique. second characteristic parameter for a fraction represented by
Finally the ultimate simplicity is to ignore the difference in a single pseudocomponent. Therefore, the whole mixture may
properties of various hydrocarbon types and to present the be characterized by its boiling point (Tb) and specific gravity
whole mixture by just a single pseudocomponent, which is (SG). In lieu of these properties other characterization pa-
the mixture itself. The simplicity in this case is that there is rameters discussed in Chapter 2 may be used.
no need for the composition of the mixture. Obviously the Treatment of wide boiling range fractions is more compli-
accuracy of estimated properties decreases as the number of cated than narrow fractions as a single value for the boiling
pseudocomponents decreases. However, for narrow boiling point, or molecular weight, or carbon number cannot rep-
range fractions such as a light naphtha approximating the resent the whole mixture. For these fractions the number of
mixture with a single pseudocomponent is more realistic and constituents in the vertical columns of Table 3.15 cannot be
more accurate than a wide boiling range fraction such as an reduced to one, although it is still possible to combine var-
atmospheric residuum or the C7+ fraction in a crude oil sam- ious molecular types for each carbon number. This means
ple. As discussed in Chapter 2, the differences between prop- that the minimum number of constituents in Table 3.14 for
erties of various hydrocarbon families increase with boiling a wide fraction is six rather than one that was considered
point (or carbon number). Therefore, assumption of a single for narrow fractions. The best example of a wide boiling
pseudocomponent for a heavy fraction (M > 300) is less ac- range fraction is C7+ fraction in a crude oil or a reservoir
curate than for the case of light fractions. For fractions that fluid. Characterization of such fractions through the use of a
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 113
distribution model that reduces the mixture into a n u m b e r of The values of To and SGo determined from regression of
p s e u d o c o m p o n e n t s with k n o w n characterization parameters data through Eq. (3.35) do not m a t c h well with the exper-
will be discussed in detail in Chapter 4. However, a simpler imental initial values. This is due to the maximizing value
a p p r o a c h based on the use of TBP curve is outlined in Ref. of RS with data used in the regression analysis. Actually
[32]. In this approach the mixture property is calculated from one can imagine that the actual initial values are lower than
the following relation: experimentally measured values due to the difficulty in such
measurements. However, these initial values do not affect
1
subsequent calculations. Predicted values at all other points
(3.39) 0 = [ O(x) dx from 5 up to 95% are consistent with the experimental values.
1/
0 F r o m calculated values of SGo, A, and B for the SG curve,
in which 0 is the physical property of mixture and O(x) is the one can determine the mixture SG for the whole fraction
value of property at point x on the distillation curve. This t h r o u g h use of Eqs. (3.37) and (3.38). For SG, B = 7.1957 and
approach may be applied to any physical property. The in- from Eq. (3.38), F (I + 1 / B) = 0.9355. From Eq. (3.37) we
get SG~v= ( ~ ) 1 / 7 1 9 5 7 F ( 1 + 7.1~57) ---- 0.5269 x 0.9355 =
tegration should be carried out by a numerical method. The
fraction is first divided into a n u m b e r of p s e u d o c o m p o n e n t s 0.493. Therefore, for the mixture: SGav -- 0.5(1 + 0 . 4 9 3 ) =
along the entire range of distillation curve with k n o w n boiling 0.746. Comparing with experimental value of 0.74, the
points and specific gravity. Then for each c o m p o n e n t physi- percent relative deviation (%D) with experimental value is
cal properties are calculated from methods of Chapter 2 and 0.8%. In Chapter 4 another m e t h o d based on a distribution
finally the mixture properties are calculated through a sim- function is introduced that gives slightly better prediction
ple mixing rule. The procedure is outlined in the following for the density of wide boiling range fractions and crude
example. oils.
To calculate a mixture property such as molecular weight,
the mixture is divided to some narrow pseudocomponents,
Example 3 . 8 - - F o r a low boiling naphtha, TBP curve is pro-
Np. If the mixture is not very wide such as in this exam-
vided along with the density at 20~ as tabulated below [32].
ple, even Nv = 5 is sufficient, but for wider fractions the mix-
Estimate specific gravity and molecular weight of this fraction
ture m a y be divided to even larger n u m b e r of pseudocompo-
using the wide boiling range approach. Compare the calcu-
nents (10, 20, etc.). If Nv = 5, then values of T and SG at x =
lated results with the experimental values reported by Lenior
0, 0.2, 0.4, 0.6, 0.8, and 0.99 are evaluated t h r o u g h Eq. (3.35)
and Hipkin and others [1, 11, 32] as SG = 0.74 and M = 120.
and parameters determined above. Value of x = 0.99 is used
instead o f x = 1 for the end point as Eq. (3.35) is not defined
vo1% 0 (IBP) 5 10 20 30 50 70 90 95
at x = 1. At every point, molecular weight, M, is determined
TBP, K 283.2 324.8 348.7 369.3 380.9 410.4 436.5 467.6 478.7
from methods of Chapter 2. In this example, Eq. (2.50) is quite
d20, g/cm 3 ... 0.654 0.689 0.719 0.739 0.765 0.775 0.775 0.785
accurate and m a y be used to calculate M since all compo-
nents in the mixture have M < 300 (~Nc < 20) and are within
Solution--For this fraction the 10-90% slope based on TPB the range of application of this method. Equation (2.50) is
curve is about 1.49~ This value is slightly above the slope M = 1.6604 x 10-4Tb2"1962 SG -1-~ Calculations are s u m m a -
based on the ASTM D 86 curve but still indicates h o w wide rized in the following table.
the fraction is. For this sample based on the ASTM distilla-
tion data [1], the 10-90% slope is 1.35~ which is above
x Tb, K SG M
the value of 0.8 specified for narrow fractions. To use the
m e t h o d by Riazi-Daubert [32] for this relatively wide frac- 0 240.0 0.500 56.7
0.2 367.1 0.718 99.8
tion, first distribution functions for both boiling point and 0.4 396.4 0.744 114.0
specific gravity should be determined. We use Eqs. (3.35)- 0.6 421.0 0.764 126.7
(3.38) to determine the distribution functions for both prop- 0.8 448.4 0.785 141.6
erties. The molecular weight, M, is estimated for all points on 0.99 511.2 0.828 178.7
the curve t h r o u g h appropriate relations in Chapter 2 devel-
oped for pure hydrocarbons. The value of M for the mixture The trapezoidal rule for integration is quite accurate to
then m a y be estimated from a simple integration over the en- estimate the molecular weight of the mixture. May -- (1/5) x
tire range o f x as given by Eq. (3.39): May = fd M(x)dx, where [(56.7+ 178.7)/2+(99.8 + 114 + 126.7+ 141.6+ 178.7)] =
M(x) is the value of M at point x determined from Tb(X) and 119.96 - 120. This is exactly the same as the experimental
SG(x). May is the average molecular weight of the mixture. value of molecular weight for this fraction [ 1, 11].
For this fraction values of densities given along the distilla- If the whole mixture is considered as a single pseudocom-
tion curve are at 20~ and should be converted to specific ponent, Eq. (2.50) should be applied directly to the mixture
gravity at 15.5~ (60~ through use of Eq. (2.112) in Chap- using the MeABP and SG of the mixture. For this fraction
ter 2: SG = 0.9915d20 + 0.01044. Parameters of Eq. (3.35) for the Watson K is given as 12.1 [1]. F r o m Eq. (3.13) using
both temperature and specific gravity have been determined experimental value of SG, average boiling point is calcu-
and are given as following. lated as Tb ----(12.1 x 0.74)3/1.8 ----398.8 K. F r o m Eq. (2.50),
the mixture molecular weight is 116.1, which is equivalent to
Parameters in Eq. (3.35) To, K SGo A B RS %D = -3.25%. For this sample, the difference between 1 and
TBP curve 240 1.41285 3.9927 0.996 5 p s e u d o c o m p o n e n t s is not significant, but for wider fraction
SG curve 0.5 0.07161 7.1957 0.911 the improvement of the proposed method is m u c h larger.
114 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
Example 3.9--It is assumed for the same fraction of Ex- For narrow boiling range fractions, ASTM D 86 temperature
ample 3.8, the only information available is ASTM D 86 data: at 50 vol% vaporized may be used as the characterizing boil-
temperatures of 350.9, 380.9, 399.8, 428.2, and 457.6 K at 10, ing point for the whole mixture. However, for a wide boiling
30, 50, 70 and 90 vol% distilled, respectively. State how can range fraction if it is treated as a single pesudocomponent
you apply the proposed approach for wide boiling range frac- the MeABP should be calculated and used as the character-
tions to calculate the molecular weight of the fraction. izing parameter for Tb in the correlations of Chapter 2. If
for a fraction TBP distillation data are available the average
Solution--Since distillation data are in terms of ASTM boiling point calculated through Eq. (3.37) with parameters
D 86, the first step is to convert ASTM to TBP through determined from TBP curve would be more appropriate than
Eq. (3.14). The second step is to determine the TBP distri- MeABP determined from ASTM D 86 curve for use as the
bution function through Eqs. (3.35) and (3.36). The third characterizing boiling point. For cases where only two points
step is to generate values of T at x -- 0, 0.2, 0.4, 0.6, 0.8, and on the distillation curve are known the interpolated value at
0.99 from Eq. (3.35) and parameters determined from TBP 50% point may be used as the characterizing boiling point of
distillation curve. Since the specific gravity for this frac- the fraction.
tion is not known it may be estimated from Eq. (3.17) and For heavy fractions (M > 300) in which atmospheric dis-
constants in Table 3.4 for the ASTM D 86 data as follows: tillation data (ASTM D 86, SD, or TBP) are not available, if
SG = 0.08342 • (350.9)~176176 = 0.756. The Wat- ASTM D 1160 distillation curve is available, it should be con-
son K is calculated as Kw = (1.8 x 399.8)1/3/0.756 = 11.85. verted to ASTM D 86 or TBP through methods outlined in
Now we assume that Kw is constant for the entire range of Section 3.2. In lieu of any distillation data, molecular weight
distillation curve and on this basis distribution of SG can be or viscosity may be used together with specific gravity to esti-
calculated through distribution of true boiling point. At every mate basic parameters from correlations proposed in Chap-
point that T is determined from Eq. (3.35) the specific grav- ter 2. If specific gravity is not available, refractive index or
ity can be calculated as SG : (1.8 x T)I/a/Kw, where T is the carbon-to-hydrogen weight ratio (CH) may be used as the
temperature on the TBP curve. Once TBP temperatures and second characterization parameter.
SG are determined at x = 0, 0.2, 0.4, 0.6, 0.8, and 0.99 points,
molecular weight may be estimated from Eq. (2.50). Numer-
3.3.4 Method of Pseudocomponent
ical integration of Eq. (3.39) can be carried out similar to the
(Defined Mixtures)
calculations made in Example 3.8 to estimate the molecular
weight. In this approach the result may be less accurate than A defined mixture is a mixture whose composition is known.
the result in Example 3.8, as ASTM distillation curve is used For a petroleum fraction if at least the PNA composition
as the only available data. # is known it is called a defined fraction. Huang [11,33] used
the pseudocompounds approach to estimate enthalpies of
Although the method outlined in this section improves narrow and defined petroleum fractions. This technique has
the accuracy of prediction of properties of wide boiling been also used to calculate other physical properties by
range fraction, generally for simplicity in calculations most other researchers [34, 35]. According to this method all com-
petroleum products are characterized by a single value of boil- pounds within each family are grouped together as a sin-
ing point, molecular weight, or carbon number regardless of gle pseudocomponent. An olefin-free fraction is modeled
their boiling range. The proposed method is mainly applied into three pseudocomponents from three homologous groups
to crude oils and C7+ fraction of reservoir fluids with an ap- of n-alkanes (representing paraffins), n-alkylcyclopentanes
propriate splitting technique as is shown in the next chapter. or n-alkylcyclohexanes (representing naphthenes), and n-
However, as shown in the above example, for very wide boiling alkylbenzenes (representing aromatics) having the same boil-
range petroleum products the method presented in this sec- ing point as that of ASTM D 86 temperature at 50% point.
tion may significantly improve the accuracy of the estimated Physical properties of a mixture can be calculated from prop-
physical properties. erties of the model components by the following mixing rule:
(3.40) 0 : XpOp@XNON+ XAOA
3.3.3 Use of Bulk Parameters
where 0 is a physical property for the mixture and 0p, ON, and
(Undefined Mixtures)
0A are the values of 0 for the model pseudocomponents from
An undefined petroleum fraction is a fraction whose com- the three groups. In this equation the composition presented
position (i.e., PNA) is not known. For such fractions infor- by xA, XN, and XAshould be in mole fraction, but because the
mation on distillation data (boiling point), specific gravity, molecular weights of different hydrocarbon groups having
or other bulk properties such as viscosity, refractive index, the same boiling point are close to each other, the compo-
CH ratio, or molecular weight are needed. If the fraction is sition in weight or even volume fractions may also be used
considered narrow boiling range then it is assumed as a sin- with minor difference in the results. If the fraction contains
gle component and correlations suggested in Chapter 2 for olefinic compounds a fourth term for contribution of this
pure hydrocarbons may be applied directly to such fractions. group should be added to Eq. (3.40). Accuracy of Eq. (3.40)
All limitations for the methods suggested in Chapter 2 should can be increased if composition of paraffinic group is known
be considered when they are used for petroleum fractions. in terms of n-paraffins and isoparaffins. Then another pseudo-
As mentioned in Chapter 2, the correlations in terms of Tb component contributing the isoparaffinic hydrocarbons may
and SG are the most accurate methods for the estimation of be added to the equation. Similarly, the aromatic part may be
various properties (molecular weight, critical constants, etc.). split into monoaromatics and polyaromatics provided their
3. CHARACTERIZATION OF P E T R O L E U M FRACTIONS 115
amount in the fraction is known. However, based on our expe- deviation of 9%. From the pseudocomponent approach, Mp,
rience the PNA three-pseudocomponent model is sufficiently MN, and MA are calculated from Eqs. (3.41)-(3.43) as 79.8,
accurate for olefin-free petroleum fractions. For coal liquids 76.9, and 68.9, respectively. The mixture molecular weight is
with a high percentage of aromatic content, splitting aromat- calculated through Eq. (3.40) as M = 0.82 x 79.8 + 0.155 x
ics into two subgroups may greatly increase the accuracy of 76.9 + 0.025 • 68.9 = 79, with relative deviation of 1.3%. If
model predictions. In using this method the minimum data values of Alp, MN, and MA are substituted in Eq. (2.42) for
needed are at least one characterizing parameter (Tb or M) the specific gravity, we get SGp = 0.651, SGN = 0.749, and
and the PNA composition. SGA -----0.895. From Eq. (3.40) the mixture specific gravity is
Properties of pseudocomponents may be obtained from in- SG = 0.673, with AD of 2.3%. It should be noted that when
terpolation of values in Tables 2.1 and 2.2 to match boiling Eq. (3.40) is applied to molecular weight, it would be more
point to that of the mixture. As shown in Section 2.3.3, prop- appropriate to use composition in terms of mole fraction
erties of homologous groups can be well correlated to only rather than volume fraction. The composition can be con-
one characterization parameter such as boiling point, molec- verted to weight fraction through specific gravity of the three
ular weight, or carbon number, depending on the availability components and then to mole fraction through molecular
of the parameter for the mixture. Since various properties of weight of the components by equations given in Section
pure homologous hydrocarbon groups are given in terms of 1.7.15. The mole fractions are Xmp = 0.785, XmN = 0.177, and
molecular weight by Eq. (2.42) with constants in Table 2.6, Xmn=0.038 and Eq. (3.39) yields M = 78.8 for the mixture
if molecular weight of a fraction is known it can be used di- with deviation of 1%. The difference between the use of
rectly as the characterizing parameter. But if the boiling point volume fraction and mole fraction in Eq. (3.40) is minor and
is used as the characterizing parameter, molecular weights of within the range of experimental uncertainty. Therefore, use
the three model components may be estimated through rear- of any form of composition in terms of volume, weight, or
rangement of Eq. (2.42) in terms of boiling point as following: mole fraction in the pseudocomponent method is reasonable
without significant effect in the results. For this reason, in
(3.41) Mp = [ 1_____2___[6.98291 _ ln(1070 - Tb)]/3/2/ most cases the PNA composition of petroleum fractions are
| 0.02013 simply expressed as fraction or percentage and they may
considered as weight, mole, or volume.
(3.42) MN = { ~ [ 6 . 9 5 6 4 9 -- ln(1028 - Tb)]} 3/2
In the above example the method of pseudocomponent pre-
(3.43) MA = / 1 - - - - ~ [ 6 . 9 1 0 6 2 -- ln(1015 -- Tb)]} 3/2 dicts molecular weight of the fraction with much better ac-
| 0.02247 curacy than the use of Eq. (2.50) with bulk properties (%AD
where Mp, MN, and MA are molecular weights of paraffinic, of 1.3% versus 9%). This is the case for fractions that are
naphthenic, and aromatic groups, respectively. Tb is the char- highly rich in one of the hydrocarbon types. For this frac-
acteristic boiling point of the fraction. Predicted values of Mp, tion paraffinic content is nearly 80%, but for petroleum frac-
MN, and MAversus Tb were presented in Fig. 2.15 in Chapter 2. tions with normal distribution of paraffins, naphthenes, and
aromatics both methods give nearly similar results and the
As shown in this figure the difference between these molecu-
advantage of use of three pseudocomponents from different
lar weights increase as boiling point increases. Therefore, the
groups over the use of single pseudocomponent with mixture
pseudocomponent approach is more effective for heavy frac-
bulk properties is minimal. For example, for a petroleum frac-
tions. If ASTM D 86 distillation curve is known the tempera-
ture at 50% point should be used for Tb, but if TBP distillation tion the available experimental data are [36] M = 170, Tb =
data are available an average TBP would be more suitable to 487 K, SG = 0.802, xp --- 0.42, XN = 0.41, andxn = 0.17. Equa-
tion (2.50) gives M = 166, while Eq. (3.40) gives M = 163 and
be used for Tb. Once Mr, MN, and MA are determined, they
SG = 0.792. Equation (3.40) is particularly useful when only
should be used in Eq. (2.42) to determine properties from cor-
responding group to calculate other properties. The method one bulk property (i.e., Tb) with the composition of a frac-
tion is available. For highly aromatic (coal liquids) or highly
is demonstrated in the following example.
paraffinic mixtures the method of pseudocomponent is rec-
ommended over the use of bulk properties.
Example 3.10---A petroleum fraction has ASTM D 86 50%
temperature of 327.6 K, specific gravity of 0.658, molecular
weight of 78, and PNA composition of 82, 15.5, and 2.5 in
3.3.5 Estimation of Molecular Weight, Critical
vol% [36]. Estimate molecular weight of this fraction using
Properties, and Acentric Factor
bulk properties of Tb and SG and compare with the value esti-
mated from the pseudocomponent method. Also estimate the Most physical properties of petroleum fluids are calculated
mixture specific gravity of the mixture through the pseudo- through corresponding state correlations that require pseu-
component technique and compare the result with the exper- docritical properties (Tpc, Pr~, and Vpc) and acentric factor
imental value. as a third parameter. In addition molecular weight (M) is
needed to convert calculated mole-based property to mass-
Solution--For this fraction the characterizing parameters based property. As mentioned in Section 1.3, the accuracy
are Tb ----327.6 K and SG = 0.672. To estimate M from these of these properties significantly affects the accuracy of es-
bulk properties, Eq. (2.50) can be applied since the boiling timated properties. Generally for petroleum fractions these
point of the fraction is within the range of 40-360~ (~C5- basic characterization parameters are calculated through ei-
C22). The results of calculation is M = 85.0, with relative ther the use of bulk properties and correlations of Chapter 2
116 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
or by the pseudocomponent approach as discussed in Sec- The Twu method gives the highest error (AAD of 14.3%) fol-
tions 3.3.2-3.3.4. lowed by the Goossens with average deviation of 11.4%. The
For petroleum fractions, pseudocritical properties are not Twu and Goossens methods both underestimate the molecu-
directly measurable and therefore it is not possible to make lar weight of these heavy fractions. The Lee-Kesler method is
a direct evaluation of different methods with experimental more accurate for lighter fractions, while the API method is
data. However, these methods can be evaluated indirectly more accurate for heavier fractions. The pseudocomponent
through prediction of other measurable properties (i.e., en- method gives generally a consistent error for all fractions and
thalpy) through corresponding state correlations. These cor- the lowest AAD%. Errors generated by the API, Lee-Kesler,
relations are discussed in Chapters 6-8. Based on more than and the pseudocomponent methods are within the experi-
700 data points for enthalpies of eight petroleum fractions mental uncertainty in the measurement of molecular weight
over a wide range of temperature and pressure [1], different of petroleum fractions. r
methods of estimation of pseudocritical temperature, pres-
sure (Tpo Pp~), and acentric factor (w) have been evaluated In summary, for light fractions (M < 300) methods recom-
and compared [37]. These petroleum fractions ranging from mended by the API for Tr and PC (Eqs. 2.65 and 2.66) [2] or the
naphtha to gas oil all have molecular weights of less than 250. simple method of Riazi-Daubert (Eqs. 2.63 and 2.64) [38] are
Details of these enthalpy calculations are given in Chapter 7. suitable, while for heavier fractions the Lee-Kesler method
Summary of evaluation of different methods is given in Ta- (Eqs. 2.69 and 2.70) [40] may be used. The pseudocompo-
ble 3.16. As shown in Table 3.16, the methods of pseudocom- nent method may also be used for both Tc and Pc when the
ponent, Lee-Kesler, and Riazi-Daubert have nearly similar composition is available. For all fractions methods of calcu-
accuracy for estimating the critical properties of these light lation of acentric factor from the pseudocomponent or the
petroleum fractions. However, for heavier fractions as it is method of Lee-Kesler [39] presented by Eq. (2.105) may be
shown in Example 3.11, the methods of pseudocomponent used. Molecular weight can be estimated from the API method
provide more accurate results. [2] by Eq. (2.51) from the bulk properties; however, if the PNA
composition is available the method of pseudocomponent is
preferable especially for heavier fractions.
Example 3.1 l k E x p e r i m e n t a l data on molecular weight and
composition of five heavy petroleum fractions are given in
Table 3.17, In addition, normal boiling point, specific grav- 3.3.6 E s t i m a t i o n o f D e n s i t y , S p e c i f i c G r a v i t y ,
ity, density, and refractive index at 20~ are also given [36]. Refractive Index, and Kinematic Viscosity
Calculate the molecular weight of these fractions from the fol-
lowing five methods: (1) API method [2, 42] using Eq. (2.51), Density (d), specific gravity (SG), and refractive index (n) are
(2) Twu method [42] using Eqs. (2.89)-(2.92), (3) Goossens all bulk properties directly measurable for a petroleum mix-
method [43] using Eq. (2.55), (4) Lee-Kesler method [40] us- ture with relatively high accuracy. Kinematic viscosity at 37.8
ing Eq. (2.54), and (5) the pseudocomponent method using or 98.9~ (1)38(100), 1,'99(210))are usually reported for heavy frac-
Eqs. (3.40)-(3.43). Calculate the %AAD for each method. tions for which distillation data are not available. But, for
light fractions if kinematic viscosity is not available it should
be estimated through measurable properties. Methods of es-
Solution--Methods 1, 2, and 4 require bulk properties of Tb timation of viscosity are discussed in Chapter 8; however, in
and SG, while the method of pseudocomponent requires Tb this chapter kinematic viscosity at a reference of temperature
and the PNA composition as it is shown in Example 3.10. of 37.8 or 98.9~ (100~ or 210~ is needed for estimation
Method 2 requires Tb and density at 20~ (d20). Results of of viscosity gravity constant (VGC), a parameter required for
calculations are given in Table 3.18. prediction of composition of petroleum fractions. Generally,
TABLE3.17--Molecular weight and composition of five heavy petroleum fractions of Example 3.11 [36].
No. M Tb,~ SG d20, g/ml n20 P% N% A%
1 233 298.7 0.9119 0.9082 1.5016 34.1 45.9 20.0
2 267 344.7 0.9605 0.9568 1.5366 30.9 37.0 32.1
3 325 380.7 0.8883 0.8845 1.4919 58.4 28.9 12.7
4 403 425.7 0.9046 0.9001 1.5002 59.0 28.0 13.0
5 523 502.8 0.8760 0.8750 1.4865 78.4 13.3 8.3
3. C H A R A C T E R I Z A T I O N O F P E T R O L E U M F R A C T I O N S 117
TABLE 3.18---Comparison of various methods of predicting molecular weight of petroleum fractions of Table 3.17 (Example 3.12).
(1) API (2) Twu, (3) Goossens, (4) Lee-Kesler, (5) Pseudocomp.,
Eq. (2.51) Eqs. (2.89)-(2.92) Eq. (2.55) Eq. (2.54) Eqs. (3.40)-(3.43)
No. M, exp, M, calc AD% M, catc AD% M, calc AD% M, calc AD% M, calc AD%
1 233 223.1 4.2 201.3 13.6 204.6 12.2 231.7 0.5 229.1 1.7
2 267 255.9 4.2 224.0 16.1 235.0 12.0 266.7 0.1 273.2 2.3
3 325 320.6 1.4 253.6 16.8 271.3 11.0 304.3 0.2 321.9 1,0
4 403 377.6 6.3 332.2 17.6 345.8 14.2 374.7 7.0 382.4 5.1
5 523 515.0 1.5 485.1 7.2 483.8 7.5 491,7 6.0 516.4 1.3
Total, AAD% 3.5 14.3 11.4 2.8 2.3
density, which is required in various predictive methods mea- to Eq. (2.111). If the specific gravity is not available, then it is
sured at 20~ is shown by d or d20 in g/mL. These properties necessary to estimate the SG at first step and then to estimate
can be directly estimated through bulk properties of mixtures the density at 20 ~C. The liquid density decreases with increase
using the correlations provided in Chapter 2 with good accu- in temperature according to the following relationship [24].
racy so that there is no need to use the pseudocomponent
d = dlss - k(T - 288.7)
approach for their estimation.
Specific gravity (SG) of petroleum fractions may be es- where d15.5 is density at 15.5~ which may be replaced by
timated from methods presented in Sections 2.4.3 and 2.6. 0.999SG. T is absolute temperature in kelvin and k is a con-
If API gravity is known, the specific gravity should be di- stant for a specific compound. This equation is accurate
rectly calculated from definition of API gravity using Eq. (2.5). within a narrow range of temperature and it may be applied
If density at one temperature is available, then Eq. (2.110) to any other reference temperature instead of 15.5~ Value
should be used to estimate the increase in density with de- of k varies with hydrocarbon type; however, for gasolines it is
crease in temperature and therefore density at 15.5~ (60~ close to 0.00085 [24].
may be calculated from the available density. A simpler re- Refractive index at 20~ n20, is an important character-
lation between SG and da0 based on the rule of thumb is ization parameter for petroleum fractions. It is needed for
SG = 1.005d20. If density at 20~ (d20) is available, the fol- prediction of the composition as well as estimation of other
lowing relation developed in Section 2.6.1 can be used: properties of petroleum fractions. If it is not available, it may
be determined from correlations presented in Section 2.6 by
(3.44) SG -- 0.01044 + 0.9915d20
calculation of parameter I. Once I is estimated, n20 can be cal-
where d20 is in g/mL and SG is the specific gravity at 15.5~ culated from Eq. (2.114). For petroleum fractions with molec-
This equation is quite accurate for estimating density of ular weights of less than 300, Eq. (2.115) can be used to esti-
petroleum fractions. If no density data are available, then mate I from Tb and SG [38]. A more accurate relation, which
SG may be estimated from normal boiling point and refrac- is also included in the API-TDB [2], is given by Eq. (2.116).
tive index or from molecular weight and refractive index for For heavier fraction in which Tb may not be available, Eq.
heavy fractions in which boiling point may not be available. (2.117) in terms of M and SG may be used [44]. Most re-
Equation (2.59) in Section 2.4.3 gives SG from Tb and I for cently Riazi and Roomi [45] made an extensive analysis of
fractions with molecular weights of less than 300, while for predictive methods and application of refractive index in pre-
heavier fractions Eq. (2.60) can be used to estimate specific diction of other physical properties of hydrocarbon systems.
gravity from M and I. Parameter I is defined in terms of re- An evaluation of these methods for some petroleum fractions
fractive index at 20~ n20, by Eq. (2.36). If viscosity data are is demonstrated in the following example.
available Eq. (2.61) should be used to estimate specific grav-
ity, and finally, if only one type of distillation curves such as E x a m p l e 3.12--Experimental data on M, Tb, SG, d2o, and n20
ASTM D 86, TBP, or EFV data are available Eq. (3.17) may be for five heavy petroleum fractions are given in Table 3.17, Es-
used to obtain the specific gravity. timate SG, dzo, and n2o from available methods and calculate
Density (d) of liquid petroleum fractions at any temper- %AAD for each method with necessary discussion of results.
ature and atmospheric pressure may be estimated from the
methods discussed in Section 2.6. Details of estimation of S o l u t i o n - - T h e first two fractions in Table 3.17 may be consid-
density of petroleum fractions are discussed in Chapter 6; ered light (M < 300) and the last two fractions are considered
however, for the characterization methods discussed in this as heavy (M > 300). The third fraction can be in either cat-
chapter, at least density at 20~ is needed. If specific gravity egory. Data available on d20, Tb, n20, and M may be used to
is available then the rule of thumb with d = 0.995SG is the estimate specific gravity. As discussed in this section, SG can
simplest way of estimating density at 20~ For temperatures be calculated from d20 or from Tb and / or from M and I.
other than 20~ Eq. (2.110) can be used. Equation (2.113) In this example specific gravity may be calculated from four
may also be used to estimate d20 from Tb and SG for light methods: (1) rule of thumb using d20 as the input parameter;
petroleum fractions (M < 300) provided that estimated den- (2) from d20by Eq. (3.44); (3) from Tb and n20 by Eq. (2.59); (4)
sity is less than the value of SG used in the equation. This from M and n20 by Eq. (2.60). Summary of results is given in
is an accurate way of estimating density at 20~ especially Table 3.19. Methods 1 and 2, which use density as the input
for light fractions. However, the simplest and most accurate parameter, give the best results. Method 4 is basically devel-
method of estimating d20 from SG for all types of petroleum oped for heavy fractions with M > 300 and therefore for the
fractions is the reverse form of Eq. (3.44), which is equivalent last three fractions density is predicted with better accuracy.
1 18 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 3.19--Comparison of various methods of predicting specific gravity of petroleum fractions of Table 3.17 (Example 3.12).
(1) Rule of thumb, (2) Use of d20, (3) Use of Tb & n20, (4) Use of M & n20,
SG = 1.005 d20 Eq. (3.44) Eq. (2.59) for M < 300 Eq. (2.60) for M >300
No. M, exp SG, exp SG, calc AD% SG, calc AD% SG, calc AD% SG, calc AD%
1 233 0.9119 0.9127 0.09 0.9109 0.11 0.8838 3.09 0.8821 3.27
2 267 0.9605 0.9616 0.11 0.9591 0.15 0.9178 4.44 0.9164 4.59
3 325 0.8883 0.8889 0.07 0.8874 0.10 0.8865 0.20 0.8727 1.76
4 403 0.9046 0.9046 0.00 0.9029 0.19 0.9067 0.23 0.8867 1.98
5 523 0.8760 0.8794 0.39 0.8780 0.23 0.9062 3.45 0.8701 0.67
Total, AAD% 0.13 0.15 2.28 2.45
TABLE 3.20--Comparison of various methods of predicting refractive index of petroleum fractions of Table 3.17
(Example 3.12).
(1) Use of Tb & SG, (2) Use of Tb & SG, (4) Use o f M & SG,
Eq. (2.115) for M < 300 Eq. (2.116) for M < 300 Eq. (2.117) for M > 300
No. M, exp n20 exp n2o exp. AD% n20 calc AD% n2o calc AD%
1 233 1.5016 1.5122 0.70 1.5101 0.57 1.5179 1.08
2 267 1.5366 1.5411 0.29 1.5385 0.13 1.5595 1.49
3 325 1.4919 1.4960 0.28 1.4895 0.16 1.4970 0.34
4 403 1.5002 1.5050 0.32 1.4952 0.34 1.5063 0.41
5 523 1.4865 1.4864 0.01 1.4746 0.80 1.4846 0.13
Total, AAD% 0.32 0.40 0.69
3. C H A R A C T E R I Z A T I O N OF P E T R O L E U M F R A C T I O N S 119
since 1)38(100) > 1.5 a n d 1)99(210) < 1.5 c S t w e have C38(100) -~- 0 PNA composition in terms of weight, volume, or mole does
and c99(210)= 0.0392. From Eq. (2.132), A1 = 10.4611, B1 = not seriously affect the predicted mixture properties. Use of
-4.3573, D1 = -0.4002, DE = --0.7397, and from Eq. (2.130) bulk properties such as Tb and SG to calculate mixture prop-
at T -- 140~ (60~ we calculate the kinematic viscosity. It erties as described for petroleum fractions cannot be used for
should be noted that in calculation of v60(140)from Eq. (2.130) a synthetic and ternary mixture of C5-C10-C25. Another exam-
trial and error is required for calculation of parameter c. At ple of a mixture that bulk properties directly cannot be used to
first it is assumed that c = 0 and after calculation of 1)60(140)if calculate its properties is a crude oil or a reservoir fluid. For
it is less than 1.5 cSt, parameter c should be calculated from such mixtures exact knowledge of composition is required
Eq. (2.131) and substituted in Eq. (2.130). Results of calcula- and based on an appropriate mixing rule a certain physical
tions are as follows: 1)38(100) = 1.8 and 1)60(140)= 1.27 cSt. Com- property for the mixture may be estimated. The most simple
paring with the experimental values, the percent relative de- and practical mixing rule that is applicable to most physical
viations for kinematic viscosities at 100 and 140~ are 8.4 and properties is as follows:
3.3%, respectively. The result is very good, but usually higher N
errors are observed for estimation of kinematic viscosity of (3.45) 0m = E XiOi
petroleum fractions from this method. i=1
where Xwi is the weight fraction and the equation can be ap-
3.4.1 L i q u i d M i x t u r e s
plied to d, SG, or parameter I. In calculation of these proper-
In liquid systems the distance between molecules is much ties for a mixture, using Eq. (3.45) with volume fraction and
smaller than in the case of gases and for this reason the inter- Eq. (3.46) with weight fraction gives similar results. Applica-
action between molecules is stronger in liquids. Therefore, the tion of these equations in calculation of mixture properties
knowledge of types of molecules in the liquid mixtures is more will be demonstrated in the next chapter to calculate proper-
desirable than in gas mixtures, especially when the mixture ties of crude oils and reservoir fluids.
constituents differ significantly in size and type. For example, For liquid mixtures the mixing rule should be applied to the
consider two liquid mixtures, one a mixture of a paraffinic hy- final desired property rather than to the input parameters. For
drocarbon such as n-eicosane (n-C20) with an aromatic com- example, a property such as viscosity is calculated through
pound such as benzene (C6) and the second one a mixture of a generalized correlation that requires critical properties as
benzene and toluene, which are both aromatic with close the input parameters. Equation (3.45) may be applied first
molecular weight and size. The role of composition in the to calculate mixture pseudocritical properties and then mix-
n-C20-benzene mixture is much more important than the ture viscosity is calculated from the generalized correlation.
role of composition in the benzene-toluene mixture. Simi- An alternative approach is to calculate viscosity of individual
larly the role of type of composition (weight, mole, or volume components in the mixture for the generalized correlation
fraction) is more effective in mixtures of dissimilar con- and then the mixing rule is directly applied to viscosity. As
stituents than mixtures of similar compounds. It is for this it is shown in the following chapters the second approach
reason that for narrow-range petroleum fractions, use of the gives more accurate results for properties of liquid mixtures,
120 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
while for gaseous mixtures there is no significant difference where Mg is the gas molecular weight. Density of both a gas
between these two methods. mixture and air at STP can be calculated from Eq. (3.47).
Mgas Psc
3.4.2 Gas Mixtures (3.48) SGg - Pga~ _ 83.14Ts~ _ Mg~
Pair M~irPsc Mair
As discussed earlier the gases at atmospheric pressure condi-
83.14Tsc
tion have much larger free space between molecules than do
liquids. As a result the interaction between various like and where sc indicates the standard condition. Molecular weight
unlike molecules in a gaseous state is less than the molecular of air can be calculated from Eq. (3.48) with molecular
interactions in similar liquid mixtures. Therefore, the role of weight of its constituents obtained from Table 2.1 as MN2 =
composition on properties of gas mixtures is not as strong as 28.01, 3/lo2 = 32.00, and MAr = 39.94. With composition given
in the case of liquids. Of course the effect of composition on as YN2 = 0.78, Yo2 = 0.21, and YA~= 0.01, from Eq. (3.1) we
properties of gas mixtures increases as pressure increases and get Mair = 28.97 g/mol. Equation (2.6) can be derived from
free space between molecules decreases. The role of compo- substituting this value for Mair in Eq. (3.49). In practical cal-
sition on properties of dense gases cannot be ignored. Under culations molecular weight of air is rounded to 29. If for a gas
low-pressure conditions where most gases behave like ideal mixture, specific gravity is known its molecular weight can be
gases all gas mixtures regardless of their composition have the calculated as
same molar density at the same temperature and pressure. As
it will be discussed in Chapter 5, at the standard conditions (3.49) Mg = 29SGg
(SC) of 1.01325 bar and 298 K (14.7 psia and 60~ most gases
where SGg is the gas specific gravity. It should be noted that
behave like ideal gas and RT/P represents the molar volume
values of specific gravity given for certain gases in Table 2.1
of a pure gas or a gas mixture. However, the absolute density
are relative to density of water for a liquefied gas and are
varies from one gas to another as following:
different in definition with gas specific gravity defined from
Mmix P Eq. (2.6). #
(3.47) PmLx -- - -
83.14T
where Pmix is the absolute density of gas mixture in g/cm 3,
3.5 P R E D I C T I O N OF T H E C O M P O S I T I O N
Mmi~ is the molecular weight of the mixture in g/tool, P is
OF P E T R O L E U M F R A C T I O N S
pressure in bar, and T is the temperature in kelvin. Equation
(3.1) can be used to calculate molecular weight of a gas mix-
As discussed earlier the quality and properties of a petroleum
ture, MmL~.However, the mole fraction of component i in a
fraction or a petroleum product depend mainly on the
gas mixture is usually shown as Yi to distinguish from com-
mixture composition. As experimental measurement of the
position of liquid mixtures designated by x~. From definition
composition is time-consuming and costly the predictive
of mole and volume fractions in Section 1.7.15 and use of
methods play an important role in determining the quality
Eq. (3.47) it can be shown that for ideal gas mixtures the
of a petroleum product. In addition the pseudocomponent
mole and volume fractions are identical. Generally volume
method to predict properties of a petroleum fraction requires
and mole fractions are used interchangeably for all types of
the knowledge of PNA composition. Exact prediction of all
gas mixtures. Composition of gas mixtures is rarely expressed
components available in a petroleum mixture is nearly im-
in terms of weight fraction and this type of composition has
possible. In fact there are very few methods available in the
very limited application for gas systems. Whenever composi-
literature that are used to predict the composition. These
tion in a gas mixture is expressed only in percentage it should
methods are mainly capable of predicting the amounts (in
be considered as tool% or vol%. Gas mixtures that are mainly
percentages) of paraffins, naphthenes, and aromatic as the
composed of very few components, such as natural gases, it is
main hydrocarbon groups in all types of petroleum fractions.
possible to consider them as a single pseudocomponent and
These methods assume that the mixture is free of olefinic hy-
to predict properties form specific gravity as the sole param-
drocarbons, which is true for most fractions and petroleum
eter available. This method of predicting properties of nat-
products as olefins are unstable compounds. In addition to the
ural gases is presented in Chapter 4 where characterization
PNA composition, elemental composition provides some vital
of reservoir fluids is discussed. The following example shows
information on the quality of a petroleum fraction or crude
derivation of the relation between gas phase specific gravity
oil. Quality of a fuel is directly related to the hydrogen and
and molecular weight of gas mixture.
sulfur contents. A fuel with higher hydrogen or lower carbon
content is more valuable and has higher heating value. High
Example 3.14 Specific gravity of gases is defined as the ra- sulfur content fuels and crude oils require more processing
tio of density of gas to density of dry air both measured at cost and are less valuable and desirable. Methods of predict-
the standard temperature and pressure (STP). Composition ing amounts of C, H, and S% are presented in the following
of dry air in tool% is 78% nitrogen, 21% oxygen, and 1% ar- section.
gon. Derive Eq. (2.6) for the specific gravity of a gas mixture.
Solution--Equation (2.6) gives the gas specific gravity as 3.5.1 Prediction o f PNA Composition
Parameters that are capable of identifying hydrocarbon types
(2.6) SGg- Mg are called characterization parameters. The best example of
28.97
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 121
such a parameter is the Watson characterization factor, which aromatic content, amounts of sulfur, nitrogen, and other het-
along with other parameters is introduced and discussed in eroatoms also increase as shown in Fig. 3.22.
this section. However, the first known method to predict the
PNA composition is the n-d-M method proposed by Van Nes 3.5.1.1 Characterization Parameters for Molecular
and Van Westen [30] in the 1950s. The n-d-M method is Type Analysis
also included in the ASTM manual under ASTM D 3238 test A characterization parameter that is useful for molecular type
method. The main limitation of this method is that it can- prediction purposes should vary significantly from one hy-
not be applied to light fractions. Later in the 1980s Riazi and drocarbon type to another. In addition, its range of varia-
Daubert [36, 47] proposed a series of correlations based on tion within a single hydrocarbon family should be minimal.
careful analysis of various characterization parameters. The With such specifications an ideal parameter for character-
unique feature of these correlations is that they are appli- izing mo]ecular type should have a constant value within a
cable to both light and heavy fractions and identify various single family but different values in different families. Some
types of aromatics in the mixture. In addition various meth- of these characterization parameters (i.e., SG, I, VGC, CH,
ods are proposed based on different bulk properties of the and Kw), which are useful for molecular type analysis, have
mixture that might be available. The Riazi-Daubert methods been introduced and defined in Section 2. i. As shown in Table
have been adopted by the API Committee on characteriza- 2.4, specific gravity is a parameter that varies with chemical
tion of petroleum fractions and are included in the fourth structure particularly from one hydrocarbon family to an-
and subsequent editions of the API-TDB [2] since the early other. Since it also varies within a single family, it is not a
1980s. The other method that is reported in some literature perfect characterizing parameter for molecular type analysis
sources is the Bergamn's method developed in the 1970s [48]. but it is more suitable than boiling point that varies within a
This method is based on the Watson K and specific gravity single family but its variation from one family to another is
of the fraction as two main characterization parameters. One not significant. One of the earliest parameters to characterize
c o m m o n deficiency for all of these methods is that they do hydrocarbon molecular type was defined by Hill and Coats in
not identify n-paraffins and isoparaffins from each other. In 1928 [49], who derived an empirical relation between viscos-
fact compositional types of PIONA, PONA, and PINA can- ity and specific gravity in terms of viscosity gravity constant
not be determined from any of the methods available in the (VGC), which is defined by Eq. (2.15) in Section 2.1.17. Def-
literature. These methods provide m i n i m u m information on inition of VGC by Eqs. (2.15) or (2.16) limits its application
the composition that is predictive of the PNA content. This is to viscous oils or fractions with kinematic viscosity at 38~
mainly due to the complexity of petroleum mixtures and dif- (100~ above 38 SUS (~3.8 cSt.). For quick hand estima-
ficulty of predicting the composition from measurable bulk tion of VGC from viscosity and specific gravity, ASTM [4] has
properties. The method of Riazi-Daubert, however, is capable provided a homograph, shown in Fig. 3.23, that gives VGC
of predicting the monoaromatic (MA) and polyaromatic (PA) values close to those calculated from Eq. (2.15). Paraffinic
content of petroleum fractions. oils have low VGC, while napthenic oils have high VGC val-
In general low boiling point fractions have higher paraffinic ues. Watson K defined by Eq. (2.13) in terms of MeABP and
and lower aromatic contents while as boiling point of the frac- SG was originally introduced to identify hydrocarbon type
tion increases the amount of aromatic content also increases. [9, 50, 51], but as is shown later, this is not a very suitable
In the direction of increase in boiling point, in addition to parameter to indicate composition of petroleum fractions.
--41
4.6 is defined by Eq. (2.36). Paraffinic oils have low I values while
I
aromatics have high I values. Carbon-to-hydrogen weight
42 ratio defined in Section 2.1.18 is also a useful parameter
5- ---'13 that indicates degree of hydrocarbon saturation and its value
increases from paraffinic to naphthenic and aromatic oils.
--44-
---45 L02O
Methods of prediction of CH was discussed in Section 2.6.3.
6 - 1.02 Application of the hydrogen-to-carbon ratio in characteriza-
tion of different types of petroleum products is demonstrated
. LOoo .~
7 -7-5o by Fryback [54]. An extensive analysis and comparison of
6 1.00 various characterization parameters useful for prediction of
.980 the composition of petroleum fractions is presented by Ri-
azi and Daubert [36,47]. Comparison of parameters P~, VGC,
IO , 60 .$8
Kw, and I is presented in Table 3.21 and Fig. 3.24. From this
-70 ,960 1 analysis it is clear that parameters/~ and VHC best separate
-940 .96
hydrocarbon types, while parameters Kw and I show large
45, -ao variations for aromatic and naphthenic compounds making
them less suitable for prediction of composition of petroleum
fractions.
8 - o~ ~_ .goo 20 Another very useful parameter that not only separates
paraffins and aromatics but also identifies various hydrocar-
- o 8 89o.1 _~ .gz ,-" bon types is defined through molecular weight and refractive
30- rC3 .~ >
,
..
P-
index as [36]:
.880=- Eo: 25 i <
40- CO <
-200 _~ .~ =~-.90 ,-r
0
(3.50) m = M ( n - 1.475)
50 -840 ~ ~
CO
-_ -zso where n is the refractive index at 20~ Parameter m was de-
60 - aco . ,e2o -: " ~ .as fined based on the observation that refractive index varies lin-
70 ~_: 50
early with 1/M with slope of m for each hydrocarbon group
8 0 -~ ,800
9 0 -- 4tO0 [55]. Values of parameter m for different hydrocarbon groups
I00 -~ - -780 - .86 calculated from Eq. (3.50) are given in Table 3.22.
120 ~ 500 - As shown in Table 3.22, paraffins have low m values while
55
~. - 600 ,;"60 --
140 aromatics have high m values. In addition, paraffinic and
-- - 700
-- ~ _ ~ .84
160 ~ -740 naphthenic oils have negative m values while aromatic oils
180 .-800
zoo --9oo have positive m values. Parameter m nicely identifies vari-
L I000 .720 40
ous aromatic types and its value increases as the number of
- .82
-- 1200 .ZOO tings increases in an aromatic compound. A pure hydrocar-
bon whose m value is calculated as - 9 has to be paraffinic,
500 "-- t 4 0 0
it cannot be naphthenic or aromatic. This parameter is par-
i- 1600 ticularly useful in characterizing various aromatic types in
400 "-- 1800
aromatic-rich fractions such as coal liquids or heavy residues.
FIG. 3 . 2 3 ~ E s t i m a t i o n of V G C from kinematic viscosity and Besides the parameters introduced above there are a num-
specific gravity [4]. ber of other parameters that have been defined for the purpose
of characterizing hydrocarbon type. Among these parameters
Aromatic oils have low Kw values while paraffinic oils have viscosity index (VI) and correlation index (CI) are worth defin-
high Kw values. Kurtz and Ward in 1935 [52] defined refrac- ing. The viscosity index was introduced in 1929 by Dean and
tivity intercept, Ri, in terms of refractive index (n) and density Davis and uses the variation of viscosity with temperature
(d) at 20~ which is presented by Eq. (2.14). The definition as an indication of composition of viscous fractions. It is an
is based on this observation that a plot of refractive index empirical number indicating variation of viscosity of oil with
against density for any homologous hydrocarbon group is lin- temperature. A low VI value indicates large variation of vis-
ear. Ri is high for aromatics and low for naphthenic stocks. cosity with temperature that is a characteristic of aromatic
The most recent characterization parameter was introduced oils. Similarly, paraffinic hydrocarbons have high VI values.
in 1977 by Huang [53] in terms of refractive index and it The method is described under ASTM D 2270-64 [4] and in
N I[] I A I
Ri I I I I I I I I
1.02 1.04 1.06 1.08 1.00
[] I---'7-]1 A
VGC I I I I I I I I
0.76 0.84 0.92 1.00 1.08
N I
A I
K I I I I I I
10 11 12 13 14
N I
I A I
I I I I I I I I I I I
0.26 0.28 0.30 0.32 0.34 0.36
the API-TDB [2]. The VI is defined as [2] in tables i n the s t a n d a r d methods (ASTM D 2270, ISO 2909)
as well as in the API-TDB [2]. However, for viscous oils with
(3.51) VI - L - U 100 a k i n e m a t i c viscosity at 100~ of greater t h a n 70 mm2/s, the
D values of L a n d D are given by the following relationships [5]:
where
L = 0.8353Y 2 + 14.67Y - 216
L -- kinematic viscosity of reference oil at 40~ with 0
VI oil, cSt. (3.52) D = 0.6669Y 2 + 2.82Y - 119
U -- k i n e m a t i c viscosity of oil at 40~ whose VI is to be where Y is the k i n e m a t i c viscosity of oil whose VI is to be cal-
calculated, cSt. culated at 100~ i n cSt. I n English units in which reference
D = L - H, i n cSt. t e m p e r a t u r e s of 37.8 a n d 98.9~ (100 a n d 210~ are used,
H = kinematic viscosity of reference oil at 40 ~C with 100 values of n u m e r i c a l coefficients in Eq. (3.52) are slightly dif-
VI oil, cSt. ferent a n d are given i n the API-TDB [2].
The reference oils with 0 a n d 100 VI a n d the oil whose
VI is to be calculated have the same k i n e m a t i c viscosity at L = 1.01523Y2 + 1 2 . 1 5 4 Y - 155.61
(3.53)
100~ I n English units of system, VI is defined in terms of D ----0.8236Y 2 - 0.5015Y - 53.03
viscosity at 37.8 a n d 98.9~ which correspond to 100 a n d
210~ respectively [2]. However, in the SI units, viscosity at Viscosity index defined by Dean a n d Davis in the form of
reference t e m p e r a t u r e s of 40 a n d 100~ have b e e n used to Eq. (3.51) does not work very well for oils with VI values of
define the VI [5]. For oils with a kinematic viscosity of 100~ greater t h a n 100. For such oils ASTM D 2270 describes the
of less t h a n 70 mm2/s (cSt.), the values of L a n d D are given calculation m e t h o d a n d it is s u m m a r i z e d below for viscosity
at reference t e m p e r a t u r e s of 40 a n d 100~ [4, 5]:
10 N -- 1
TABLE 3.22--Values of parameter m for different types (3.54) VI- - - + 100
of hydrocarbons [36]. 0.00715
Hydrocarbontype m where N is given by the following relation:
Paraffins -8.79
Cyclopentanes - 5.41 (3.55) N = log H - log U
Cyclohexanes -4.43 logY
Benzenes 2.64
Naphthalenes 19.5 i n which U a n d Y are k i n e m a t i c viscosity of oil i n cSt whose
Condensed Tricyclics 43.6 VI is to be calculated at 40 a n d 100~ respectively. Values of
124 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 3.23---Comparison of VI, VGC, Kw, and CI for several groups of oils [29].
Type of oil VI VGC Kw CI
High VI distillate 100 0.800-0.805 12.2-12.5 < 15
Medium VI distillate 80 0.830-0.840 11.8-12.0 15-50
Low VI distillate 0 0.865-0.890 11.0-11.5 >50
Solvent Extracts ... 0.880-0.950 10.0-11.0 ...
Recycle Stock ... 0.900-0.950 10.0-11.0 ...
Cracked Residues ... 0.950-1.000 9.8-11.0 ...
H versus k i n e m a t i c viscosity of oil at 100 ~ (Y) w h e n Y is less to calculate the VI. However, since k i n e m a t i c viscosities are
t h a n 70 cSt are t a b u l a t e d in the ASTM D 2270 m e t h o d [4]. F o r given at 38 a n d 99~ Eqs. (3.57) a n d (3.58) s h o u l d be used.
oils with values of Y greater t h a n 70 cSt H m u s t be c a l c u l a t e d F r o m the i n f o r m a t i o n given U = i000 a n d Y = 100. Since Y
from the relation given b e l o w [5]. is greater t h a n 75 cSt, Eq. (3.58) m u s t be used to calculate
p a r a m e t e r H, w h i c h gives H -- 3080.6. F r o m Eqs. (3.55) a n d
(3.56) H = 0.1684Y 2 + 1 1 . 8 5 Y - 97
(3.57) we get N = 0.2443 a n d VI--- 200.7. Since calculated VI is
If kinematic viscosities are available in English units at 37.8 g r e a t e r t h a n 100, the initial a s s u m p t i o n is correct. Otherwise,
a n d 98.9~ (100 a n d 210~ t h e n Eqs. (3.54) a n d (3.56) should Eq. (3.51) m u s t be used. Since the value of VI is quite high
be r e p l a c e d with the following relations as given in the API- the oil is paraffinic as s h o w n in Table 3.24. If i)99(210) was less
TDB [2]: t h a n 70 cSt then tables p r o v i d e d by ASTM [4] o r API-TDB [2]
10 N - I should be used to calculate p a r a m e t e r s L a n d D.
(3.57) VI- - - + 100
0.0075
3.5.1.2 A P I (Riazi-Daubert) Methods
(3.58) H = 0.19176Y 2 + 1 2 . 6 5 5 9 Y - 102.58 To develop a m e t h o d for p r e d i c t i n g the c o m p o s i t i o n of olefin-
free p e t r o l e u m fractions three equations are r e q u i r e d to ob-
in w h i c h Eq. (3.58) m u s t be u s e d for fractions with k i n e m a t i c
tain fractions of paraffins (xp), n a p h t h e n e s (XN), a n d a r o m a t -
viscosity at 99~ (210~ greater t h a n 75 cSt [2].
ics (XA). The first a n d m o s t obvious e q u a t i o n is k n o w n f r o m
Finally the correlation index (CI) defined b y the U.S. B u r e a u
the m a t e r i a l balance:
of Mines is expressed b y the following e q u a t i o n [7]:
48640 (3.60) Xp + XN + XA = 1
(3.59) CI - + 473.7SG - 456.8
rb
Two o t h e r equations can be established by a p p l y i n g Eq. (3.40)
in w h i c h Tb is the v o l u m e average boiling p o i n t (VABP) in for two p a r a m e t e r s that can c h a r a c t e r i z e h y d r o c a r b o n types
kelvin. Values of CI b e t w e e n 0 a n d 15 indicate a p r e d o m i - a n d are easily m e a s u r a b l e . Analysis of various characteriza-
n a n t l y paraffinic oil. A value of CI greater t h a n 50 indicates a tion factors s h o w n in Table 3.21 a n d Fig. 3.24 indicates t h a t / ~
p r e d o m i n a n c e of a r o m a t i c c o m p o u n d s [7]. It has a t e n d e n c y a n d VGC are the m o s t suitable p a r a m e t e r s to identify hydro-
to increase with increasing boiling p o i n t in a given c r u d e oil. A c a r b o n type. F o r example, if for a p u r e h y d r o c a r b o n / ~ = 1.04,
c o m p a r i s o n b e t w e e n values ofVI, VGC, Kw, a n d CI for several it has to be a n a p h t h e n i c h y d r o c a r b o n , it c a n n o t be paraffinic
types of p e t r o l e u m fractions a n d p r o d u c t s is p r e s e n t e d in Ta- o r a r o m a t i c since only for the n a p h t h e n i c g r o u p /~ varies
ble 3.23. A c o m p l e t e c o m p a r i s o n of various c h a r a c t e r i z a t i o n f r o m 1.03 to 1.046. Refractivity intercept has been related
p a r a m e t e r s i n d i c a t i n g c o m p o s i t i o n of p e t r o l e u m fractions for to the p e r c e n t o f n a p h t h e n i c c a r b o n a t o m s (%CN) as /~ =
the three h y d r o c a r b o n groups is p r e s e n t e d in Table 3.24. All 1.05-0.0002 %CN [7]. Riazi a n d D a u b e r t [47] u s e d b o t h / ~
p a r a m e t e r s except Ri, Kw, a n d VI increase in the direction a n d VGC to develop a predictive m e t h o d for the c o m p o s i t i o n
from paraffinic to n a p h t h e n i c a n d a r o m a t i c oils. of viscous p e t r o l e u m fractions. Properties of p u r e hydrocar-
b o n s from the API RP-42 [56] have b e e n u s e d to c a l c u l a t e / ~
Example 3.15--Calculate viscosity index of an oil having a n d VGC for a n u m b e r of heavy h y d r o c a r b o n s with molecu-
k i n e m a t i c viscosities of 1000 a n d 100 cSt at 37.8 a n d 98.9~ lar weights greater t h a n 200 as shown in Table 2.3 a n d Table
(100 a n d 210~ respectively. 3.21. Based on the values o f / ~ a n d VGC for all h y d r o c a r b o n s ,
average values of these p a r a m e t e r s were d e t e r m i n e d for the
Solution--For this oil 1)99(210) > 70 cSt, thus we can use Eqs. three g r o u p s of paraffins, n a p h t h e n e s , a n d aromatics. These
(3.51)-(3.57) for calculation of VI. Since the VI is n o t k n o w n average values f o r / ~ are as follows: 1.0482 (P), 1.0138 (N),
we a s s u m e it is greater t h a n 100 a n d we use Eqs. (3.53)-(3.57) a n d 1.081 (A). S i m i l a r average values for VGC are 0.744 (P),
0.915 (N), a n d 1.04 (A). Applying Eq. (3.40) for Ri a n d VGC In these set of equations XA m u s t be calculated from Eq.
gives the following two relations. (3.72). F o r cases t h a t calculated XA is negative it s h o u l d be
set equal to zero a n d values of Xp a n d XN m u s t be n o r m a l i z e d
(3.61) /~ = 1.0482xp + 1.038XN + 1.081XA
in a w a y t h a t Xp + xr~ = 1. The s a m e p r o c e d u r e s h o u l d be
(3.62) VGC = 0.744Xp + 0.915XN + 1.04XA a p p l i e d to Xp o r Xn if one of t h e m calculated from the
A regression of 33 defined h y d r o c a r b o n mixtures changes above equations is negative. F o r 85 s a m p l e s Eqs. (3.70) a n d
the n u m e r i c a l constants in the above equations b y less t h a n (3.72) give average deviation of 0.04 a n d 0.06 for xv a n d
2% as follows [29, 47]: XN, respectively. F o r 72 heavy fractions, Eqs. (3.72)-(3.74)
p r e d i c t Xp, XN, a n d XA w i t h average deviations of 0.03, 0.04,
(3.63) Ri = 1.0486xv + 1.022XN + 1.1 lXA a n d 0.02, respectively [36]. These deviations are w i t h i n the
(3.64) VGC = 0.7426xv + 0.9XN + 1.112XA range of e x p e r i m e n t a l u n c e r t a i n t y for the PNA c o m p o s i t i o n .
E q u a t i o n s (3.70)-(3.74) are r e c o m m e n d e d to be u s e d if
S i m u l t a n e o u s solution of Eqs. (3.60), (3.63), a n d (3.64) e x p e r i m e n t a l d a t a on viscosity are available. F o r cases t h a t
gives the following equations for e s t i m a t i o n of the PNA com- n2o a n d d2o are not available, they can be accurately e s t i m a t e d
p o s i t i o n of fractions with M > 200. f r o m the m e t h o d s p r e s e n t e d in C h a p t e r 2.
(3.65) Xp = - 9 . 0 + 12.53/~ - 4.228VGC F o r fractions t h a t k i n e m a t i c viscosity is n o t available, Riazi
a n d D a u b e r t [36] developed a series of correlations in t e r m s of
(3.66) xN = 18.66 - 19.9/~ + 2.973VGC
o t h e r c h a r a c t e r i z a t i o n p a r a m e t e r s w h i c h are r e a d i l y available
(3.67) XA = --8.66 + 7.37/~ + 1.255VGC or predictable. These p a r a m e t e r s are SG, m, a n d CH a n d the
These equations c a n be a p p l i e d to fractions with m o l e c u l a r predictive equations for PNA c o m p o s i t i o n are as follows:
weights in the r a n g e of 200-600. As m e n t i o n e d earlier, Xp,
Xn, a n d XA calculated f r o m the above relations m a y r e p r e s e n t F o r fractions w i t h M < 200
volume, mole, or weight fractions. E q u a t i o n s (3.65)-(3.67) (3.75) Xp = 2.57 - 2.877SG + 0.02876CH
c a n n o t be a p p l i e d to light fractions having k i n e m a t i c viscosity
at 38~ of less t h a n 38 SUS (~3.6 cSt.). This is b e c a u s e VGC (3.76) XN = 0.52641 -- 0.7494Xp -- 0.021811m
c a n n o t be d e t e r m i n e d as defined by Eqs. (2.15) a n d (2.16). F o r or
such fractions Riazi a n d D a u b e r t [47] defined a p a r a m e t e r
s i m i l a r to VGC called viscosity gravity function, VGF, b y the (3.77) Xp = 3.7387 - 4.0829SG + 0.014772m
following relations: (3.78) XN = - 1 . 5 0 2 7 + 2.10152SG - 0.02388m
(3.68) VGF = - 1.816 + 3.484SG - 0.1156 In v38(100) F o r fractions with M > 200
(3.69) VGF = - 1 . 9 4 8 + 3.535SG - 0.1613 In i)99(210)
(3.79) xp = 1.9842 - 0.27722Ri - 0.15643CH
w h e r e 1)38(100) a n d 1)99(210) are k i n e m a t i c viscosity in mm2/s
(3.80) XN = 0.5977 -- 0.761745Ri + 0.068048CH
(cSt) at 38 a n d 99~ (100 a n d 210~ respectively. F o r a
p e t r o l e u m fraction, b o t h Eqs. (3.68) a n d (3.69) give nearly the or
(3.77) and (3.78) as: Xp -- 3 . 2 5 7 4 - 3.48148 SG + 0.011666m get XMA= --0.228 and XpA= -0.025. Since both numbers are
andxN = --1.9571 + 2.63853 SG - 0.03992m. Most of the data negative the actual estimated values are XMA= 0 and XpA= 0.
used in development of Eqs. (3.70)-(3.82) were in terms of From Eq. (3.85), ;cA = 0, which is consistent with the previous
volume fractions for xp, XN, and Xg. Therefore, generally esti- result from Eqs. (3.77), (3.78), and (3.72). t
mated values represent volume fractions; however, they can
be used as mole fractions without serious errors. 3.5.1.3 A P I M e t h o d
In all the above equations total aromatic content is cal- Since 1982 API has adopted the methods developed by Ri-
culated from Eq. (3.72). As discussed earlier for cases that azi and Daubert [36, 47] for prediction of the composition
aromatic content is high it should be split into two parts for of petroleum fractions. Equations (3.65)-(3.67) and similar
a more accurate representation of hydrocarbon types in a equations developed for light fractions in terms of Ri and
petroleum mixture. Aromatics are divided into monoaromat- VGF by Riazi and Daubert in 1980 [47] were included in
ics (XMA)and polyaromatics (XPA) and the following relations the fourth edition of the API-TDB-82. However, after devel-
have been derived for fractions with molecular weights of less opment of Eqs. (3.70)-(3.74) in terms of viscosity and Eqs.
than 250 [36]: (3.83)-(3.84) for prediction of the amount of different types
(3.83) XMA= --62.8245 + 59.90816R~ - 0.0248335m of aromatics in 1986 [36], they were included in the fifth and
subsequent editions of the API-TDB [2]. The API methods for
(3.84) XpA= 11.88175 -- 11.2213R~ + 0.023745m prediction of the composition of petroleum fractions require
(3.85) XA -~- XMA -~- XpA kinematic viscosity at 38 or 99~ and if not available, it should
be estimated from Eq. (2.128) or (2.129) in Chapter 2.
The equations may be applied to fractions with total aromatic
content in the range of 0.05-0.96 and molecular weight range 3.5.1.4 n-doM M e t h o d
of 80-250. Based on a data set for aromatic contents of 75
This method requires three physical properties of refractive
coal liquid sample, Eqs. (3.83)-(3.85) give AAD of about 0.055,
index (n20), density (d20), and molecular weight (M). For this
0.065, and 0.063 for XMA,XPA,and XA, respectively. Maximum
reason the method is called n-d-M method and is the oldest
AD is about 0.24 for Xpg. Equations (3.83) and (3.84) have not
method for prediction of the composition of petroleum frac-
been evaluated against petroleum fractions. For heavier frac-
tions. The method is described in the book by Van Nes and
tions no detailed composition on aromatics of fractions were
Van Westen in 1951 [30] and it is included in the ASTM man-
available; however, if such data become available expressions
ual [4] under ASTM D 3238 test method. The method does
similar to Eqs. (3.83) and (3.84) may be developed for heavier
not directly give the PNA composition, but it calculates the
fractions.
distribution of carbon in paraffins (%Cp), naphthenes (%CN),
and aromatics (%CA). However, since carbon is the dominant
Example 3.16--A gasoline sample produced from an Aus-
element in a petroleum mixture it is assumed that the %Cp,
tralian crude oil has the boiling range of C5-65~ specific
%CN, and %CA distribution is proportional to %P, %N, and %A
gravity of 0.646, and PNA composition of 91, 9, and 0 vol%
distribution. In this assumption the ratio of carbon to hydro-
(Ref. [46], p. 302). Calculate the PNA composition from a suit-
gen is considered constant in various hydrocarbon families.
able method and compare the results with the experimental
Errors caused due to this assumption are within the range of
values.
uncertainty in experimental data reported on the PNA com-
position. Another input data required for this method is sulfur
Solution--For this fraction the only information available
content of the fraction in wt% (%S) and should be known if
are boiling point and specific gravity. From Table 2.1 the boil-
it exceeds 0.206 wt%. The method should not be applied to
ing point of n-C5 is 36~ Therefore, for the fraction the
fractions with sulfur content of greater than 2%. This method
characteristic boiling point is Tb ----(36 + 65)/2 ----50.5~ The
is applicable to fractions with boiling points above gasoline.
other characteristic of the fraction is SG -- 0.646. This is
In addition this method should be applied to fractions with
a light fraction (low Tb) so we use Eq. (2.51) to calculate
ring percent, %CR (%CN + %CA) up to 75% provided that %CA
molecular weight as 84.3. Since viscosity is not known, the
(as determined from the n-d-M method) is not higher than
most suitable method to estimate composition is through
1.5 times %CN [7]. The n-d-M method also provides equations
Eqs. (3.77) and (3.78). They require parameter m, which
for calculation of total number of rings (RT), number of
in turn requires refractive index, n. From Eq. (2.115), I =
aromatic rings (Ra), and number of naphthenic rings (RN)
0.2216 and from Eq. (2.114), n = 1.3616. With use of M and
in an average molecule in the fractions. The method is
n and Eq. (3.50), m = -9.562. From Eq. (3.77) and (3.78),
expressed in two sets of equations, one for n2o, d20 (20~
xp = 0.96 and xN -- 0.083. From Eq. (3.72), ;cA = -4.3. Since
and another set for n70 and d7o (70~ as input data. In this
XA is negative thus ;CA----0 and Xp, XN should be normalized
section correlation in terms of n20 and d20 are presented. The
as xp = 0.96/(0.96 + 0.083) -- 0.92 and XN = l -- 0.92 ----0.08.
other set of correlations for measurement of n and d at 70~
Therefore, the predicted PNA composition is 92, 8, 0% versus
is given in the literature [7, 24].
the experimental values of 91, 9, and 0%.
The aromatic content for this fraction is zero and there %CA = av + 3660/M
is no need to estimate XMA and XpA from Eqs. (3.83) and %CN = %CR -- %CA
(3.84); however, to see the performance of these equations
(3.86) %Cp = 100 - %CR
for this sample we calculate ;CA from Eq. (3.85). From
Eq. (2.113), d = 0.6414 and from Eq. (2.14), R4 -- 1.0409. Us- Rn = 0.44 + bvM
ing these values o f / ~ and m in Eqs. (3.83) and (3.84), we RN = RT-- R,
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 127
82.952 - 65.341n 306 where xA is the fraction of aromatics in the mixture and
(3.92) %H = 30.346 + Tb is an average value of boiling points at 10, 50, and 90
d M
vol% vaporized in kelvin [Tb = (/'10 + Ts0 + %0)/3]. This cor-
where M is the molecular weight and n and d are refractive relation was developed based on 247 aviation fuels and 84
index and density at 20~ respectively. This method should pure hydrocarbons. This method is quite accurate if all the
be applied to fractions with molecular weight range of 84- input data are available from experimental measurement.
459, boiling point range of 60-480~ refractive index range
of 1.38-1.51, and hydrogen content of 12.2-15.6 wt%. In 3.5.2.1.4 Jenkins-Walsh M e t h o d - T h e y developed a sim-
cases that M is not available it should be estimated from the ple relation in terms of specific gravity and aniline point in
Goossens correlation given by Eq. (2.55). the following form [59]:
:1:: 20
...........................................................i.......................................................;................... ...........................................................~..............
.................................... ..................... ............ ........... ........ ~ .................. 7 ..............
6
~iiiiiiiil 4+ 1.27274%H) 1IIIIIIIIIIIII
~0
10
, J ; r i
0
5 10 15
(a) H y d r o g e n Weight Percent, % H
.< 70
6O
so
o 40
.2 30
o
20
0
11 12 13 14 15
There are a number of other methods reported in the liter- 8 ..... Middle East (5%) .-'"
....... West African (5%) .'"
ature. The Winn nomograph may be used to estimate the CH
World Average (5%) . "~
ratio and then %H can be estimated from Eq. (3.90). Fein- . . . . World Average (2%)
9"
..~/,-" . j.s
S"
Finally Fig. 3.25 is based on data from Ref. [57]. Analytical FIG. 3.26~Distribution of sulfur in straight-run prod-
correlation is also presented in Fig. 3.25(a), which represents ucts for several crude oils. Numbers in the parentheses
data with an average deviation of 0.5%. Equation (3.96) is pre- indicate sulfur content of crudes.
sented in Fig. 3.25(b). When CH ratio is available, %H can be
determined from Fig. 3.25(a) and then %A can be determined
from Fig. 3.25(b). As boiling point, specific gravity, or aromatic content of
a fraction increases the sulfur content also increases [7].
3.5.2.2 Prediction of Sulfur and Nitrogen Contents Parameters Ri, m, and SG have been successfully used to
Sulfur is the most important heteroatom that may be present predict the PNA composition especially aromatic content of
in a crude oil or petroleum products as dissolved free sulfur petroleum fractions as shown in Section 3.5.2.1. On this ba-
and hydrogen sulfide (H2S). It may also be present as organic sis the same parameters have been used for the estimation of
compounds such as thiophenes, mercaptanes, alkyl sulfates, sulfur content of petroleum fractions in the following form
sulfides (R--S--W), disulfides (R--S--S--R'), or sulfoxides for two ranges of molecular weight.
(R--SO--R'), where R and R' refer to any aliphatic or aromatic
group. Its presence is undesirable for the reasons of corrosion, For fractions with M < 200
catalysts poisoning, bad odor, poor burning, and air pollu-
(3.97) %S = 177.448 - 170.946Ri + 0.2258m+ 4.054SG
tion. In addition presence of sulfur in lubricating oils lowers
resistance to oxidation and increases solid deposition on en- and for fractions with M > 200
gine parts [62]. New standards and specifications imposed
by governments and environmental authorities in industrial (3.98) %S --- -58.02 + 38.463R~ - 0.023m+ 22.4SG
countries require very low sulfur content in all petroleum
products. For example, reformulated gasolines (RFG) require For light fractions in which Eq. (3.96) may give very small
sulfur content of less than 300 ppm (<0.03 wt%) [63]. Re- negative values, %S would be considered as zero. Squared cor-
cently a federal court has upheld an Environmental Protec- relation coefficients (R2) for these equations are above 0.99.
tion Agency (EPA) rule to cut pollution from tractor-trailers A summary of evaluation of these equations is presented in
and other large trucks and buses. The rule is expected to re- Table 3.25 as given in Ref. [62].
duce tailpipe emissions from tractors, buses, and other trucks In using these equations parameters n2o, d2o, M, and SG
up to 90%. The EPA also calls on refineries to reduce the sul- are required. For samples in which any of these parameters
fur content in diesel oils to 15 ppm by 2007 from the current are not known they can be estimated from the methods dis-
level of 500 ppm. The American Lung Association claims that cussed earlier in this chapter. In Chapter 4, it is shown how
low-sulfur fuel will reduce the amount of soot from larger this method can be used to estimate sulfur content of whole
trucks by 90%. This is expected to prevent 8300 premature crudes. The author is not familiar with any other analytical
deaths, 5500 cases of chronic bronchitis, and another 17600 method for estimation of sulfur content of petroleum frac-
cases of acute bronchitis in children as provided by the EPA tions reported in the literature so a comparison with other
[64]. Products with high-sulfur contents have low quality and methods is not presented. Generally amount of sulfur in var-
heating values. Generally, sulfur is associated with heavy and ious products is tabulated for various crudes based on the
aromatic compounds [7]. Therefore, high aromatic content sulfur content of each crude [61 ].
or high boiling point fractions (i.e., residues and coal liquids) Another heteroatom whose presence has adverse effect on
have naturally higher sulfur contents. Distribution of sulfur in the stability of the finished product and processing catalysts
straight-run products of several crude oils and the world aver- is nitrogen. High nitrogen content fractions require high hy-
age crude with 2 and 5% sulfur contents is shown in Fig. 3.26. drogen consumption in hydro processes. Nitrogen content of
Data used to generate this figure are taken from Ref. [61]. As crudes varies from 0.01 to 0.9 wt%. Most of the compounds
the boiling point of products increases the sulfur content in having nitrogen have boiling points above 400~ (~750~
the products also increases. However, the distribution of sul- and are generally within the aromatic group. Crudes with
fur in products may vary from one crude source to another. higher asphaltene contents have higher nitrogen content as
130 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
well. S i m i l a r to sulfui, n i t r o g e n content of various p e t r o l e u m Eqs. (2.15) a n d (3.50) we get Ri -- 1.0534, m - - 1.2195. Since
fractions is p r e s e n t e d in t e r m s of n i t r o g e n content of the M < 250, Eq. (3.97) is used to e s t i m a t e the sulfur content as
crude oil [61]. Ball et al. [65] have s h o w n that nitrogen con- %S = 1. i % versus the e x p e r i m e n t a l value of 0.8%. Therefore,
tent of crude oils for each geological p e r i o d is linearly related the e r r o r is calculated as follows: 1.1%-0.8% --- 0.3%.
to c a r b o n r e s i d u e o f the crude. However, the c o r r e l a t i o n does To calculate %C a n d % H from e x p e r i m e n t a l data, values of
not provide i n f o r m a t i o n on n i t r o g e n c o n t e n t of p e t r o l e u m CH = 6.69 with %S = 0.8 are used in Eqs. (3.90) a n d (3.91).
products. I n general n i t r o g e n c o n t e n t of fractions whose m i d This w o u l d result in %C -- 86.3 a n d %H -- 12.9. According to
boiling p o i n t is less t h a n 450~ have n i t r o g e n contents less the general m e t h o d p r e s e n t e d in this b o o k (author's p r o p o s e d
t h a n t h a t of crude a n d for heavier cuts the n i t r o g e n wt% in the method), CH is calculated f r o m Eq. (2.120) as CH = 6.75 a n d
fraction is g r e a t e r t h a n that of crude [61]. However, the value with e s t i m a t e d value of sulfur content as %S -- 1.1, %C a n d
of 450~ at w h i c h nitrogen c o n t e n t of the fraction is n e a r l y the %H are c a l c u l a t e d f r o m Eqs. (3.90) a n d (3.91) as %C = 86.1
s a m e as that of crude is a p p r o x i m a t e a n d it m a y vary slightly a n d %H = 12.8. In use of Goossens m e t h o d t h r o u g h Eq. (3.92),
with the type of the crude. D a t a r e p o r t e d in Ref. [61] for dis- e s t i m a t e d values of n, d, a n d M are r e q u i r e d w h e r e M s h o u l d
t r i b u t i o n o f n i t r o g e n c o n t e n t o f straight r u n distillates have be e s t i m a t e d f r o m Eq. (2.55) as M = 190. F o r this m e t h o d %C
been c o r r e l a t e d in the following form: m a y be calculated from Eq. (3.88) if %S is known. A s u m m a r y
o r results for calculation of % H with AD for various m e t h o d s
%N2 in fraction
= - 0 . 4 6 3 9 + 8.47267T - 28.9448T 2 is given in Table 3.26. The Goossens m e t h o d gives the high-
%N2 in crude
est e r r o r b e c a u s e all i n p u t d a t a r e q u i r e d are p r e d i c t e d values.
(3.99) + 27.8155T 3 The ASTM m e t h o d gives the s a m e value as e x p e r i m e n t a l value
w h e r e T = Tb/1000 in w h i c h Tb is the m i d boiling p o i n t of the b e c a u s e the e x p e r i m e n t a l values on all the i n p u t p a r a m e t e r s
cut in kelvin. This equation is valid for cuts with m i d boiling r e q u i r e d in Eq. (3.93) are available in this p a r t i c u l a r example.
points greater t h a n 220~ a n d is not a p p l i c a b l e to finished However, in m a n y cases a r o m a t i c c o n t e n t o r c o m p l e t e distilla-
p e t r o l e u m products. A m o u n t of n i t r o g e n in a t m o s p h e r i c tion curve as r e q u i r e d b y the ASTM m e t h o d are not available.
distillates is quite small on p e r c e n t basis. The wt% ratio in The general m e t h o d of a u t h o r p r e s e n t e d in this section b a s e d
Eq. (3.99) can be r e p l a c e d b y p p m weight ratio for small quan- on calculation of CH a n d %S gives g o o d results a l t h o u g h 50%
tities of nitrogen. E s t i m a t i o n of c o m p o s i t i o n of elements is ASTM distillation t e m p e r a t u r e a n d specific gravity have b e e n
d e m o n s t r a t e d in E x a m p l e s 3.17 a n d 3.18. used as the only available data. t
Example 3 . 1 7 ~ A p e t r o l e u m fraction with a boiling range of Example 3.18--A p e t r o l e u m cut has the boiling r a n g e of 370-
250-300~ is p r o d u c e d from a Venezuelan c r u d e oil (Ref. [46], 565~ a n d is p r o d u c e d from a c r u d e oil from Danish N o r t h
p. 360). E x p e r i m e n t a l l y m e a s u r e d p r o p e r t i e s are as follows: Sea fields (Ref. [46], p. 353). The nitrogen content of c r u d e
ASTM distillation 262.2,268.3, a n d 278.9~ at 10, 50, a n d 90 is 1235 p p m . Calculate n i t r o g e n content of the fraction a n d
vol% recovered, respectively; specific gravity 0.8597; c a r b o n - c o m p a r e with the e x p e r i m e n t a l value of 1625 p p m .
to-hydrogen weight ratio 6.69; aniline p o i n t 62~ a r o m a t i c
c o n t e n t 34.9%; a n d sulfur wt% 0.8. E s t i m a t e sulfur c o n t e n t Solution--Tb = (370 + 565)/2 = 467.5~ = 740.6 K. T = Tb/
of the fraction from the m e t h o d p r e s e n t e d in Section 3.5.2.2. 1000 = 0.7406. S u b s t i t u t i n g T in Eq. (3.99) gives %N2 in cut ----
Also calculate %C a n d % H from the following methods: exper- 1.23 x 1235 = 1525. The p e r c e n t relative deviation with the
i m e n t a l data, Riazi, Goossens, ASTM, Jenkins-Watsh, B u r e a u e x p e r i m e n t a l value is - 6 % . This is relatively a g o o d predic-
of Mines a n d Eq. (3.96). tion, b u t n o r m a l l y larger errors are o b t a i n e d especially for
lighter cuts. t
Solution--To e s t i m a t e the sulfur content, p a r a m e t e r s M, n2o,
a n d d20 are r e q u i r e d as the i n p u t data. The fraction is a nar- 3.6 P R E D I C T I O N OF O T H E R P R O P E R T I E S
r o w fraction a n d the boiling p o i n t at 50% distilled c a n be
c o n s i d e r e d as the characteristic average boiling point, Tb = In this section, predictive m e t h o d s for s o m e i m p o r t a n t p r o p -
268.3~ = 541.5 K. This is a light fraction w i t h M < 300; erties that are useful to d e t e r m i n e the quality of certain
therefore, M, d2o, a n d n20 a r e c a l c u l a t e d f r o m Eqs. (2.50) a n d p e t r o l e u m p r o d u c t s a r e presented. S o m e of these proper-
(2.112)-(2.114) as 195.4, 0.8557, a n d 1.481, respectively. F r o m ties such as flash p o i n t or p o u r p o i n t are useful for safety
consideration or storage and transportation of products. One part, methods of prediction of vapor pressure, fuel vapor liq-
of the most important properties of petroleum products re- uid (V/L) ratio, fuel volatility index, and flash points are pre-
lated to volatility after the boiling point is vapor pressure. sented.
For petroleum fractions, vapor pressure is measured by the
method of Reid. Methods of prediction of true vapor pressure 3.6.1. i Reid Vapor Pressure
of petroleum fractions are discussed in Chapter 7. However, Reid vapor pressure is the absolute pressure exerted by a mix-
Reid vapor pressure and other properties related to volatility ture at 37.8~ (311 K or 100~ at a vapor-to-liquid volume
are discussed in this section. The specific characteristics of ratio of 4 [4]. The RVP is one of the important properties of
petroleum products that are considered in this part are flash, gasolines and jet fuels and it is used as a criterion for blend-
pour, cloud, freezing, aniline, and smoke points as well as car- ing of products. RVP is also a useful parameter for estimation
bon residue and octane number. Not all these properties apply of losses from storage tanks during filling or draining. For
to every petroleum fraction or product. For example, octane example, according to Nelson method losses can be approxi-
number applies to gasoline and engine type fuels, while car- mately calculated as follows: losses in vol% = (14.5 RVP - 1)/6,
bon residue is a characteristic of heavy fractions, residues, where RVP is in bar [24, 66]. The apparatus and procedures
and crude oils. Freezing, cloud, and pour points are related for standard measurement of RVP are specified in ASTM
to the presence of heavy hydrocarbons and are characteristics D 323 or IP 402 test methods (see Fig. 3.27). In general, true
of heavy products. They are also important properties under vapor pressure is higher than RVP because of light gases dis-
very cold conditions. Predictive methods for some of these solved in liquid fuel. Prediction of true vapor pressure of pure
properties are rare and scatter. Some of these methods are hydrocarbons and mixtures is discussed in detail in Chapter 7
developed based on a limited data and should be used with (Section 7.3). The RVP and boiling range of gasoline governs
care and caution. ease of starting, engine warm-up, mileage economy, and ten-
dency toward vapor lock [63]. Vapor lock tendency is directly
3.6.1 Properties Related to Volatility related to RVP and at ambient temperature of 21~ (70~
the maximum allowable RVP is 75.8 kPa (11 psia), while this
Properties that are related to volatility of petroleum fraction limit at 32~ (90~ reduces to 55.2 kPa (8 psia) [63]. RVP
are boiling point range, density, Reid vapor pressure, and flash can also be used to estimate true vapor pressure of petroleum
point. Prediction of boiling point and density of petroleum fractions at various temperatures as shown in Section 7.3.
fractions have been discussed earlier in this chapter. In this True vapor pressure is important in the calculations related
to losses and rate of evaporation of liquid petroleum prod-
ucts. Because RVP does not represent true vapor pressure,
the current tendency is to substitute RVP with more modern
and meaningful techniques [24]. The more sophisticated in-
struments for measurement of TVP at various temperatures
are discussed in ASTM D 4953 test method. This method
can be used to measure RVP of gasolines with oxygenates
and measured values are closer to actual vapor pressures
E4, 24].
As will be discussed in Chapters 6 and 7, accurate calcu-
lation of true vapor pressure requires rigorous vapor liquid
equilibrium (VLE) calculations through equations of state.
The API-TDB [2] method for calculation of RVP requires a
tedious procedure with a series of flash calculations through
Soave cubic equation of state. Simple relations for estima-
tion of RVP have been proposed by Jenkins and White and
are given in Ref. [61]. These relations are in terms of tem-
peratures along ASTM D 86 distillation curve. An example of
these relations in terms of temperatures at 5, i0, 30, and 50
vol% distilled is given below:
where all temperatures are in~ and RVP is in bar. The diffi-
culty with this equation is that it requires distillation data up
to 50% point and frequently large errors with negative RVP
values for heavier fuels have been observed. Another method
FIG. 3.27--Apparatus to measure RVP of for prediction of RVP was proposed by Bird and Kimball [61 ].
petroleum products by ASTM D 323 test method In this method the gasoline is divided into a number (i.e.,
(courtesy of KISR). 28) of cuts characterized by their average boiling points. A
132 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
blending RVP of each cut is then calculated by the following and specific gravity of the fraction. Critical temperature and
equation: pressure may be estimated from Tb and SG using methods
presented in Chapter 2. This equation is based on data with
7.641 RVP in the range of 0.0007-1.207 bar (0.01-17.5 psia), normal
Bi=
exp (0.03402Tbi + 0.6048) boiling point range of 305-494 K, and specific gravity range
i=28 of 0.65-1.08. The average absolute deviation for 52 samples is
(3.101) Pa = ~ BiXvi 0.061 bar (0.88 psia). The above equation may be used for cal-
i=l culation of RVP to determine quality characteristics of a fuel.
f -- 1.0 + 0.003744 (VAPB - 93.3) The calculated RVP value should not be used for calculation
RVP = fPa of TVP when very accurate values are needed. (Appropriate
methods for direct estimation of TVP of petroleum fractions
where B~ = RVP is the blending number for cut i and T b i = are discussed in Section 7.3.3.) Vapor pressure of a petroleum
normal boiling point of cut i in ~ Xvi is the volume fraction mixture depends on the type of its constituents and with use
of cut i, VABP is the volume average boiling point in~ and of only two bulk properties to predict RVP is a difficult task.
RVP is the Reid vapor pressure in bars. The constants were ob- This equation is recommended for a quick and convenient
tained from the original constants given in the English units. estimation of RVP, but occasionally large errors may be ob-
The average error for this method for 51 samples was 0.12 tained in use of this equation. For more accurate estimation
bar orl.8 psi [67, 68]. of RVP the sophisticated method suggested in the API-TDB
Recently some data on RVP of gasoline samples have been [2] may be used. In this method RVP is calculated through a
reported by Hatzioznnidis et al. [69]. They measured vapor series of vapor-liquid-equilibrium calculations.
pressure according to ASTM D 5191 method and related to RVP is one of the main characteristics that is usually used
RVP. They also related their measured vapor pressure data to to blend a fuel with desired specifications. The desired RVP
TVP thus one can obtain RVP from TVP, but their relations of a gasoline is obtained by blending naphtha with n-butane
have not been evaluated against a wide range of petroleum (M = 58, RVP = 3.58 bar or 52 psia) or another pure hydro-
fractions. Other relations for calculation of TVP from RVP carbon with higher RVPs than the original fuel. For condi-
for petroleum fractions and crude oils are given in Section tions where RVP should be lowered (hot weather), heavier
7.3.3. TVP at 100~ (311 K) can be estimated from Eq. (3.33) hydrocarbons with lower RVP are used for blending purposes.
as RVP of several pure hydrocarbons are given as follows: i-C4:
4.896 bar (71 psia); n-C4:3.585 (52); i-Cs:1.338 (19.4); n-Cs:
(3.102) logl0(TVP)100=3.204x 1-4x l~_-~r b 1.0135 (14.7); i-C6:0.441 (6.4); n-C6:0.34 (5.0); benzene: 0.207
(3.0); and toluene: 0.03 (0.5), where all the numbers inside the
where Tb is the normal boiling point in K and TVP100 is the parentheses are in psia as given in Ref. [63]. However in the
true vapor pressure at 100~ (311 K). Once TVP is calculated same reference in various chapters different values of RVP
it may be used instead of RVP in the case of lack of sufficient for a same compound have been used. For example, values of
data. When this equation is used to estimate RVP of more than 4.14 bar (60 psi) for n-C4, 1.1 bar (16 psi) for n-Cs, and 0.48
50 petroleum products an average error of 0.13 bar (~ 1.9 psi) bar (7 psi) for i-C6 are also reported by Gary and Handwerk
and a maximum error of 5.9 psi were obtained [67, 68]. [63]. They also suggested two methods for calculation of RVP
RVP data on 52 different petroleum products (light and of a blend when several components with different RVPs are
heavy naphthas, gasolines, and kerosenes) from the Oil and blended. The first method is based on the simple Kay's mixing
Gas Journal data bank [46] have been used to develop a sim- rule using mole fraction (Xr~) of each component [63]:
ple relation for prediction of RVP in terms of boiling point
(3.104) RVP(blend) = E Xmi(RVP)/
and specific gravity in the following form [67]: i
RVP = Pc exp(Y) where (RVP)i is the RVP of component i in bar or psia. The
second approach is to use blending index for RVP as [63]:
r=- X (TbSG'~
',--g--~ / (1 - Tr)s (RVPBI)i = (RVP)] 2s
X = -276.7445 + 0.06444Tb + 10.0245SG - 0.129TbSG (3.105) RVPBI (blend) = EXvi(RVPBI)i
i
9968.8675 RVP (blend) = [RVPBI (blend)] ~
+ + 44.6778 In Tb + 63.6683 In SG
TbSG
where (RVPBI)/ is the blending index for (RVP)i and Xvi is
T~ = 311/Tc the volume fraction of component i. Both units of bar or
(3.103) psia may be used in the above equation. This relation was
originally developed by Chevron and is also recommended
where Tb is the mid boiling point and Tc is the pseudocritical in other industrial manuals under Chevron blending number
temperature of the fraction in kelvin. Pc is the pseudocritical [61]. Equations (3.104) and (3.105) may also be applied to
pressure and RVP is the Reid vapor pressure in bars. The basis TVP; however, methods of calculation of TVP of mixtures are
for development of this equation was to use Miller equation discussed in Section 7.3 through thermodynamic relations.
for TVP and its application at 311 K (100~ The Miller equa-
tion (Eq. 7.13) is presented in Section 7.3.1. The constants Example 3.19--Estimate RVP of a gasoline sample has
of vapor pressure correlation were related to boiling point molecular weight of 86 and API gravity of 86.
3. C H A R A C T E R I Z A T I O N OF P E T R O L E U M F R A C T I O N S 133
Solution--API = 86 and M = 86. From Eq. (2.4), SG = 0.65 ifications require that its value be limited to 900 in summer,
and from Eq. (2.56), Tb = 338 K. Since only Tb and SG are 1000 in fall/spring, and 1150 in the winter season. Automobile
known, Eq. (3.103) is used to calculate the RVP. From Eqs. manufacturers in France require their own specifications
(2.55) and (2.56) we get Tc -- 501.2 K a n d Pc = 28.82 bar. From that the value of FVI not be exceeded by 850 in summer [24].
Eq. (3.103), Tr = 0.6205, X = 1.3364, and Y = -3.7235. Thus
we calculate RVP = 0.696 bar or 10.1 psia. The experimental 3.6.1.3 Flash Point
value is I 1.1 psia [63]. # Flash point of petroleum fractions is the lowest temperature
at which vapors arising from the oil will ignite, i.e. flash,
3.6.1.2 V/L Ratio and Volatility Index when exposed to a flame under specified conditions. There-
fore, the flash point of a fuel indicates the maximum tem-
Once RVP is known it can be used to determine two other
perature that it can be stored without serious fire hazard.
volatility characteristics, namely vapor liquid ratio (V/L) and
fuel volatility index (FVI), which are specific characteristics Flash point is related to volatility of a fuel and presence of
light and volatile components, the higher vapor pressure cor-
of spark-ignition engine fuels such as gasolines. V/L ratio is a
responds to lower flash points. Generally for crude oils with
volatility criterion that is mainly used in the United States and
Japan, while FVI is used in France and Europe [24]. The V/L RVP greater than 0.2 bar the flash point is less than 20~ [24].
Flash point is an important characteristics of light petroleum
ratio at a given temperature represents the volume of vapor
fractions and products under high temperature environment
formed per unit volume of liquid initially at 0~ The proce-
and is directly related to the safe storage and handling of
dure of measuring V/L ratio is standardized as ASTM D 2533.
The volatility of a fuel is expressed as the temperature levels at such petroleum products. There are several methods of de-
termining flash points of petroleum fractions. The Closed
which V/L ratio is equal to certain values. Usually V/L values
Tag method (ASTM D 56) is used for petroleum stocks with
of 12, 20, and 36 are of interest. The corresponding tempera-
flash points below 80 ~C (175 ~ The Pensky-Martens method
tures may be calculated from the following relations [24]:
(ASTM D 93) is used for all petroleum products except waxes,
T(V/L)12 m_ 8 8 . 5 -- 0 . 1 9 E 7 0 - 42.5 RVP solvents, and asphalts. Equipment to measure flash point ac-
(3.106) T(v/L)2o= 90.6 - 0.25E70 - 39.2 RVP cording to ASTM D 93 test method is shown in Fig. 3.28.
The Cleveland Open Cup method (ASTM D 92) is used for
T~v/L)36 = 94.7 -- 0.36E70 - 32.3 RVP petroleum fractions with flash points above 80~ (175~ ex-
where T(v/L)x is the temperature in ~ at which V/L = x. cluding fuel oil. This method usually gives flash points 3-6~
Parameter E70 is the percentage of volume distilled at 70~ higher than the above two methods [61]. There are a number
E70 and RVP are expressed in percent distilled and bar, of correlations to estimate flash point of hydrocarbons and
respectively. Through Lagrange interpolation formula it is petroleum fractions.
possible to derive a general relation to determine temper- Buffer et al. [70] noticed that there is a linear relationship
ature for any V/L ratio. E70 can be calculated through a between flash point and normal boiling point of hydrocar-
distribution function for distillation curve such as Eq. (3.35) bons. They also found that at the flash point temperatures, the
in which by rearrangement of this equation we get product of molecular weight (M) and vapor pressure (pvap) for
pure hydrocarbons is almost constant and equal to 1.096 bar
(3.107) E70 = 1 0 0 - 100 exp -
E \
(3.108) T(V/L)2O= 52.5 + 0.2T10 + 0.17T50 - 33 RVP where pvap is the vapor pressure at 37.8~ (100~ in bar and
TF is the flash point in kelvin. For simplicity RVP may be used
where T10 and Ts0 are temperatures at 10 and 50 vol% distilled for pwp. Methods of calculation of vapor pressure are dis-
on the ASTM D 86 distillation curve. All temperatures are cussed in Chapter 7. Various oil companies have developed
in~ and RVP is in bar. For cases that T10 is not available it special relations for estimation of flash points of petroleum
may be estimated through reversed form of Eq. (3.17) with T50 fractions. Lenoir [72] extended Eq. (3.100) to defined mix-
and SG. Several petroleum refining companies in the United tures through use of equilibrium ratios.
States such as Exxon and Mobil use the critical vapor locking The most widely used relation for estimation of flash point
index (CVLI), which is also related to the volatility index [61 ]. is the API method [2], which was developed by Riazi and
(3.109) CVLI -- 4.27 + 0.24E70 + 0.069 RVP Daubert [73]. They used vapor pressure relation from Clasius-
Clapeyron (Chapter 6) together with the molecular weight
The fuel volatility index is expressed by the following relation relation form Eq. (2.50) in Eq. (3.111) to develop the following
[24]: relation between flash point and boiling point:
(3.110) FVI = 1000 RVP + 7E70 (3.113) 1/TF=a+b/Tb+clnTb+dlnSG
where RVP is in bar. FV/is a characteristic of a fuel for its per- where Tb is the normal boiling point of pure hydrocarbons.
formance during hot operation of the engine. In France, spec- It was observed that the coefficient d is very small and TF is
134 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
nearly independent of specific gravity. Based on data from (3.114) is adopted by the API as the standard method to esti-
pure hydrocarbons and some petroleum fractions, the con- mate flash point of petroleum fractions [2]. It was shown that
stants in Eq. (3.113) were determined as Eq. (3.114) can be simplified into the following linear form
1 2.84947 [73]:
(3.114) ~FF = --0.024209+ T1----~+3.4254 • 10 -3 lnTm
(3.115) TF = 15.48 + 0.70704/'1o
where for pure hydrocarbons T10 is normal boiling point,
while for petroleum fractions it is distillation temperature at where both Tx0 and TF are in kelvin. This equation is applica-
10 vol% vaporized (ASTM D 86 at 10%) and it is in kelvin. TF is ble to fractions with normal boring points (i.e., ASTM D 86
the flash point in kelvin determined from the ASTM D 93 test temperature at 50%) less than 260~ (500~ For such light
method (Pensky-Martens closed cup tester). This equation is fractions, Eq. (3.115) is slightly more accurate than Eq.
presented in Fig. 3.29 for a quick and convenient estimate (3.t14). For heavier fractions Eq. (3.114) should be used.
of flash point. For 18 pure hydrocarbons and 39 fractions, There are some relations in the literature that correlate flash
Eq. (3.114) predicts flash points with an average absolute de- points to either the initial boiling point (T10) or the distillation
viation (AD) of 6.8~ (12~ while Eq. (3.111) predicts the temperature at 50% point (T50). Such correlations are not ac-
flash points with AD of 18.3~ curate over a wide range of fractions, especially when they are
Equation (3.114) should be applied to fractions with nor- applied to fractions not used in obtaining their coefficients.
mal boring points from 65 to 590~ 150-1100~ Equation Generally reported initial boiling points for petroleum frac-
tions are not reliable and if mid boiling point temperature is
used as the characteristics boiling point it does not truly rep-
150 resent the boiling point of light components that are initially
being vaporized. For this reason the correlations in terms of
o ~176 ILILL distillation temperature at 10% point (7"10)are more accurate
than the other correlations for estimation of flash points of
petroleum fractions. Flash points of petroleum fractions may
also be estimated from the pseudocomponent method using
the PNA composition and values of flash points of pure hy-
drocarbons from Table 2.2. However, volumetric averaging
of component flash point through Eq. (3.40) generally over-
-50 predicts the flash point of the blend and the blending index
0 100 200 300 400 approach described below should be used to estimate flash
point of defined mixtures.
ASTM 10% Temperature,~
If the flash point of a petroleum fraction or a petroleum
FIG. 3.29--Prediction of flash product does not meet the required specification, it can be
point of petroleum fractions from adjusted by blending the fraction with other compounds hav-
Eq. (3.114). ing different flash points. For example in hot regions where
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 135
where TF is the flash point in kelvin and the best value of x Tp = 130.47[SG 297~ x [M (~176
is -0.06. However, they suggest that the exponent x be cus- F (0.310331-0.32834SG)1
tomized for each refinery to give the best results [61]. The (3.119) x lP38(100) J
following example shows application of these methods. where Tp is the pour point (ASTM D 97) in kelvin, M is the
molecular weight, and v38o00) is the kinematic viscosity at
Example 3.20---A kerosene product with boiling range of 37.8~ (100~ in eSt. This equation was developed with data
175-260~ from Mexican crude oil has the API gravity of 43.6 on pour points of more than 300 petroleum fractions with
(Ref. [46], p. 304). (a) Estimate its flash point and compare molecular weights ranging from 140 to 800 and API gravities
with the experimental value of 59~ (b) For safety reasons it from 13 to 50 with the AAD of 3.9~ [73]. This method is also
is required to have a m i n i m u m flash point of 65~ to be able accepted by the API and it is included in the API-TDB since
to store it in a hot summer. How much n-tetradecane should 1988 [2] as the standard method to estimate pour point of
be added to this kerosene for a safe storage? petroleum fractions. As suggested by Hu and Burns [75, 76],
Eqs. (3.117) and (3.118) used for blending index of flash point
can also be used for pour point blending index (TpB) with
Solutions(a) To estimate flash point we use either Eq.
x = 0.08 :
(3.114) or its simplified form Eq. (3.15), which require ASTM
10% temperature, T10. This temperature may be estimated (3.120) BIp = T1/0.08
~p
from Eq. (3.17) with use of specific gravity, SG = 0.8081, and
ASTM 50% temperature, Ts0. Since complete ASTM curve is where Tp is the pour point of fraction or blend in kelvin. The
not available it is assumed that the mid boiling point is the AAD of 2.8~ is reported for use of Eqs. (3.117) and (3.120)
same as Ts0; therefore, T50 = 217.5~ and from Eq. (3.17) with to estimate pour points of 47 blends [76].
coefficients in Table 3.4,/'10 = 449.9 K. Since Ts0 is less than
260~ Eq. (3.115) can be used for simplicity. The result is
3.6.3 C l o u d P o i n t
TF = 60.4~ which is in good agreement with the experimen-
tal value of 59~ considering the fact that an estimated value The cloud point is the lowest temperature at which wax crys-
of ASTM 10% temperature was used. tals begin to form by a gradual cooling under standard con-
(b) To increase the flash point from 59 to 65~ n-Ct4 with ditions. At this temperature the oil becomes cloudy and the
flash point of 1000C (Table 2.2) is used. If the volume frac- first particles of wax crystals are observed. The standard pro-
tion of ~/-C14 needed is shown by Xadd, then using Eq. (3.117) cedure to measure the cloud point is described under ASTM
we have BIFB = (1 --Xadd) X BIvK + Xadd X BIVadd where B I ~ , D 2500, IP 219, and ISO 3015 test methods. Cloud point
BIFK, and BIradd are the blending indexes for flash points is another cold characteristic of petroleum oils under low-
of final blend, kerosene sample, and the additive (n-C14), temperature conditions and increases as molecular weight of
respectively. The blending indexes can be estimated from oil increases. Cloud points are measured for oils that con-
Eq. (3.116) as 111.9, 165.3, and 15.3, respectively, which re- tain paraffins in the form of wax and therefore for light frac-
sult in xaaa = 0.356. This means that 35.6% in volume of rt-C14 tions, such as naphtha or gasoline, no cloud point data are
is required to increase the flash point to 65~ If the blending reported. Cloud points usually occur at 4-5~ (7 to 9~ above
indexes are calculated from Eq. (3.118), the amount of r/-C14 the pour point although the temperature differential could be
required is 30.1%. , in the range of 0-10~ (0-18~ as shown in Table 3.27. The
136 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 3.27--Cloud and pour points and their differences for some petroleum products.
Fraction API gravity T~,,~C TCL,~C Tp -TcL, ~
Indonesian Dist. 33.0 -43.3 -53.9 10.6
Australian GO 24.7 -26.0 -30.0 4.0
Australian HGO 22.0 -8.0 -9.0 1.0
Abu Dhabi LGO 37.6 -19.0 -27.0 8.0
Abu Dhabi HGO 30.3 7.0 2.0 5.0
Abu Dhabi Disst. 21.4 28.0 26.0 2.0
Abu Dhabi Diesel 37.4 - 12.0 - 12.0 0.0
Kuwaiti Kerosene 44.5 -45.0 -45.0 0.0
Iranian Kerosene 44.3 -46.7 -46.7 0.0
Iranian Kerosene 42.5 -40.6 -48.3 7.8
Iranian GO 33.0 - 11.7 - 14.4 2.8
North Sea GO 35.0 6.0 6.0 0.0
Nigerian GO 27.7 -32.0 -33.0 1.0
Middle East Kerosene 47.2 -63.3 -65.0 1.7
Middle East Kerosene 45.3 -54.4 -56.7 2.2
Middle East Kerosene 39.7 -31.1 -34.4 3.3
Middle East Disst. 38.9 -17.8 -20.6 2.8
Source: Ref. [46].
Tp: pour point; TcL:cloud point.
difference between cloud and p o u r point depends on the na- the oil [61]. Freezing point is one of the important charac-
ture of oil and there is no simplified correlation to predict this teristics of aviation fuels where it is determined by the pro-
difference. Cloud point is one of the important characteris- cedures described in ASTM D 2386 (U.S.), IP 16 (England),
tics of crude oils under low-temperature conditions. As tem- and NF M 07-048 (France) test methods. M a x i m u m freezing
perature decreases below the cloud point, formation of wax point of jet fuels is an international specification which is re-
crystals is accelerated. Therefore, low cloud point products quired to be at - 4 7 ~ (-53~ as specified in the "Aviation
are desirable u n d e r low-temperature conditions. Wax crys- Fuel Quantity Requirements for Jointly Operated Systems"
tals can plug the fuel system lines and filters, which could [24]. This m a x i m u m freezing point indicates the lowest tem-
lead to stalling aircraft and diesel engines u n d e r cold condi- perature that the fuel can be used without risk of separation
tions. Since cloud point is higher than p o u r point, it can be of solidified hydrocarbons (wax). Such separation can result
considered that the knowledge of cloud point is more impor- in the blockage in fuel tank, pipelines, nozzles, and filters [61 ].
tant than the p o u r point in establishing distillate fuel oil spec- Walsh-Mortimer suggest a t h e r m o d y n a m i c model based on
ifications for cold weather usage [61]. Table 3.27 shows the the solubility of n-paraffin hydrocarbons in a petroleum mix-
difference between cloud and p o u r points for some petroleum ture to determine the freezing point [71]. Accurate determi-
products. Cloud and pour points are also useful for predict- nation of freezing point requires accurate knowledge of the
ing the temperature at which the observed viscosity of an oil composition of a fuel which is normally not known. However,
deviates from the true (newtonian) viscosity in the low tem- the m e t h o d of determination of carbon n u m b e r distribution
perature range [7]. The a m o u n t of n-paraffins in petroleum along with solid-liquid equilibrium can be used to determine
oil has direct effect on the cloud point of a fraction [8]. Pres- freezing points of petroleum fractions and crude oils as will be
ence of gases dissolved in oil reduces the cloud point which is discussed in Chapter 9. A simpler but less accurate method to
desirable. The exact calculation of cloud point requires solid- determine freezing points of petroleum fractions is t h r o u g h
liquid equilibrium calculations, which is discussed in Chap- the p s e u d o c o m p o n e n t approach as shown in the following
ter 9. The blending index for cloud point is calculated from example.
the same relation as for p o u r point through Eq. (3.118) with
x = 0.05 :
Example 3.21--A kerosene sample produced from a crude
(3.121) BIcL = T~/~176 oil from North Sea Ekofisk field has the boiling range of 150-
204.4~ (302--400~ and API gravity of 48.7. Estimate the
where TCL is the cloud point of fraction or blend in kelvin.
freezing point of this kerosene and c o m p a r e with the experi-
Accuracy of this m e t h o d of calculating cloud point of blends is
mental value of - 6 5 ~ (-85~
the same as for the pour point (AAD of 2.8~ Once the cloud
point index for each c o m p o n e n t of blend, BIcLi, is determined
through Eq. (3.12 t), the cloud point index of the blend, Blczs, Solution--The mid boiling point is Tb = 177.2~ and the spe-
is calculated through Eq. (3.117). Then Eq. (3.121) is used in cific gravity is SG = 0.785. We use the method of pseudo-
its reverse form to calculate cloud point of the blend from its c o m p o n e n t using predicted composition. F r o m Eq. (2.50),
cloud point index [76]. M = 143 and since M > 143, we use Eqs. (3.77), (3.78), and
(3.72) to predict Xp, XN, and XA, respectively. F r o m Eqs. (2.114)
3.6.4 Freezing Point and (2.115), n = 1.439 and from Eq. (3.50), m = -5.1515. Us-
ing SG and m, we calculate the PNA composition as xp =
Freezing point is defined in Section 2.1.9 and freezing points 0.457,xN = 0.27, and XA = 0.273. F r o m Eqs. (3.41)-(3-43),
of pure hydrocarbons are given in Table 2.2. For a petroleum Mp = 144.3, MN = 132.9, and MA = 129.3. Using Eq. (2.42)
fraction, freezing point test involves cooling the sample until a for prediction of the freezing point for different families we
slurry of crystals form throughout the sample or it is the tem- get TFp = 242.3, TFN = 187.8, and TrA = 178.6 K. Using Eq.
perature at which all wax crystals disappear on rewarming (3.40) we get TF = 210.2 K or -63.1~ versus the measured
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 137
value of -65~ The result is quite satisfactory considering in which Me is the molecular weight of n-paraffin whose boil-
minimum data on Tb and SG are used as the only available ing point is the same as mid boiling point of the fraction which
parameters. can be determined from Eq. (3.41).
40.0 /
(3.132) CI = AP - 15.5
where AP is in ~ Calculated cetane index (CI) is also related
30.0 to n-paraffin content (%NP) of diesel fuels in the following
30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 110.0 120.0 from [87].
Experimental Aniline Point, ~
(3.133) %NP = 1.45CI - 57.5
FIG. 3.31--Error distribution for prediction of aniline point
The relation for calculation of cetane number blending in-
from Eq. (3.127). Taken with permission from Ref. [68].
dex is more complicated than those for pour and cloud point.
Blending indexes for cetane number are tabulated in various
by the fuel characterization factor called cetane number (CN). sources [61, 75]. Cetane number of diesel fuels can be im-
The behavior of a diesel fuel is measured by comparing proved by adding additives such as 2-ethyl-hexyl nitrate or
its performance with two pure hydrocarbons: n-cetane or other types of alkyl nitrates. Cetane number is usually im-
n-hexadecane (n-C16H34) which is given the number 100 and proved by 3-5 points once 300-1000 ppm by weight of such
a-methylnaphthalene which is given the cetane number of additives is added [24]. Equation (3.129) suggested for cal-
0. A diesel fuel has a cetane number of 60 if it behaves culating cetane number does not consider presence of addi-
like a binary mixture of 60 vol% cetane and 40 vol% ~- tives and for this reason calculated cetane index for some
methylnaphthalene. In practice heptamethylnonane (HMN) fuels differ with measured cetane index. Generally, CCI is less
a branched isomer of n-cetane with cetane number of 15 is than measured CN and for this reason in France automobile
used instead of a-methylnaphthalene [24, 61]. Therefore, in manufacturers have established minimum CN for both the
practice the cetane number is defined as: calculated CI (49) and the measured CN (50) for the quality
(3.128) CN = vol%(n-cetane) + 0.15(vo1% HMN) requirement of the fuels [24].
by ASTM D 908 and MON is measured by ASTM D 357 test 3.28--Octane numbers of some alcohols and
TABLE
methods. Generally there are three kinds of gasolines: regu- ethers (oxygenates).
Compound RON MON
lar, intermediate, and premium with PON of 87, 90, and 93,
Methanol 125-135 100-105
respectively. In France the minimum required RON for su- MTBE 113-I17 95-101
perplus gasoline is 98 [24]. Required RON of gasolines vary Ethanol 120-130 98-103
with parameters such as air temperature, altitude, humidity, ETBE 118-122 100-102
engine speed, and coolant temperature. Generally for every TBA 105-110 95-100
300 m altitude RON required decreases by 3 points and for TAME 110-114 96-100
M T B E : m e t h y l - t e r t i a r y - b u t y l ether; E T B E : e t h y l - t e r t i a r y - b u t y l ether;
every I I~ rise in temperature RON required increases by 1.5 TBA: t e r t i a r y - b u t y l alcohol; TAME: t e r t i a r y - a m y l - m e t h y l ether.
points [63]. Improving the octane number of fuel would result Source: Ref. [24],
in reducing power loss of the engine, improving fuel economy,
and a reduction in environmental pollutants and engine dam-
age. For these reasons, octane number is one of the important after addition of an additive. ONoxis the corresponding octane
properties related to the quality of gasolines. There are a num- number of oxygenate, which can be taken as the average val-
ber of additives that can improve octane number of gasoline ues for the ranges of RON and MON as given in Table 3.28. For
or jet fuels. These additives are tetra-ethyl lead (TEL), alco- example for MTBE, the range of RONox is 113-117; therefore,
hols, and ethers such as ethanol, methyl-tertiary-butyl ether for this oxygenate the value of RONox for use in Eq. (3.136)
(MTBE), ethyl-tertiary-butyl ether (ETBE), or tertiary-amyl is 115. Similarly the value MONox for this oxygenate is are
methyl ether (TAME). Use of lead in fuels is prohibited in 98. Equation (3.136) represents a simple linear relation for
nearly all industrialized countries due to its hazardous nature octane number blending without considering the interaction
in the environment, but is still being used in many third world between the components. This relation is valid for addition
and underdeveloped countries. For a fuel with octane number of additives in small quantities (low values of Xox,i.e., < 0.15).
(ON) of 100, increase in the ON depends on the concentra- However, when large quantities of two components are added
tion of TEL added. The following correlations are developed (i.e., two types of gasolines on 25:75 volume basis), linear mix-
based on the data provided by Speight [7]: ing rule as given by Eq. (3.136) is not valid and the interac-
tion between components should be taken into account [61 ].
Du Pont has introduced interaction parameters between two
TEL = -871.05 +2507.81 ~ - 2415.94 \ 1 0 0 / or three components for blending indexes of octane number
(oN 3 which are presented in graphical forms [89]. Several other
(3.134) +779.12 \ 1 0 0 ] blending approaches are provided in the literature [61]. The
simplest form of their tabulated blending indexes have been
ON = 100.35 + 11.06(TEL) - 3.406(TEL) 2 converted into the following analytical relations:
(3.135) + 0.577(TEL) 3 - 0.038(TEL) 4 BIRoN =
2
36.01+38.33X-99,8X +341.3X 3 -507.2X 4
+268.64X 5
11_<RON<76
where ON is the octane number and TEL is milliliter TEL - 2 9 9 . 5 + 1272X - 1552.9X 2 + 651X 3 76 _< RON _< 103
added to one U.S. gallon of fuel. These relations nearly repro-
2 2 0 6 . 3 - 4 3 1 3 . 6 4 X + 2178.57X 2 103 _< RON < 106
duce the exact data given by Speight and valid for ON above
/ x = RON/IO0
100. In these equations when clear octane number (without
TEL) is 100, TEL concentration is zero. By subtracting the cal- (3.137)
culated ON from 100, the increase in the octane number due
to the addition of TEL can be estimated, which may be used where BIRoNis the blending index for RON and should be used
to calculate the increase in ON of fuels with clear ON different together with Eq. (3.117) to calculate RON of a blend. Equa-
from 100. Equation (3.134) is useful to calculate amount of tion (3.137) reproduce the tabulated values of RON blending
TEL required for a certain ON while Eq. (3.135) gives ON of indexes with AAD of 0.06%.
fuel after a certain amount of TEL is added. For example, if Estimation of octane number of a fuel from its bulk proper-
0.3 mL of TEL is added to each U.S. gallon of a gasoline with ties is a challenging task, since ON very much depends on the
RON of 95, Eq. (3.135) gives ON of 104.4, which indicates an chemical structure of components of the mixture. Figure 3.32
increase of 4.4 in the ON. This increase is based on the refer- shows variation of RON with boiling point of pure hydrocar-
ence ON of 100 which can be used for ON different from 100. bons from different families as produced from data given in
Therefore, the ON of gasoline in this example will be 95 + 4.4 Table 2.2. If PIONA composition of a fuel is known, RON of a
or 99.4. Different relations for octane number of various fuels fuel may be estimated from the pseudocomponent techniques
(naphthas, gasolines, and reformates) in terms of TEL con- in the following form:
centration are given elsewhere [88]. RON = x~p(RON)Np + xI~(RON)Ip + xo(RON)o
Octane numbers of some oxygenates (alcohols and ethers)
(3.138) + xN(RON)N + XA(RON)A
are given in Table 3.28 [24]. Once these oxygenates are added
to a fuel with volume fraction of :Coxthe octane number of where x is the volume fraction of different hydrocarbon
product blend is [24] families i.e., n-paraffins (NP), isoparaffins (IP), olefins (O),
naphthenes (N), and aromatics (A). RONNp, RONIp, RONo,
(3.136) ON = Xox(ON)ox+ (1 - Xox)(ON)clear
RONN, and RONA are the values of RON of pseudocompo-
where ONclear is the clear octane number (RON or MON) of nents from n-paraffin, isoparaffins, olefins, naphthenes, and
a fuel and ON is the corresponding octane number of blend aromatics families whose boiling points are the same as the
140 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
8o
,2o ; 7 - : : .......... ...... ..................
" --omZ; 5l ! i I I I
,oo .
60 i ' - ' I , , o~ ,
.~ 40 n-paraft~s \ k~ \ l : ~ ~phthenes ~ so ,
~ 40 .
0 An-alkyIcyclopentanes : :x
On-alkylbenzenes
-20 . . . . . ' . . . . ' . . . .
-100 -50 0
' ....
50
i ....
100
,
150
,i;
200
$03~t J'"lol'oox:
"<~176r'""
BoilingPoint,~ j I . 1
II.3 11.4 1T.5 11.6 11.7 11.8 11,9 12.0 12.1 12.2 12.3
FIG. 3.32--Research octane number of different Characterization Factor
families of hydrocarbons. Taken with permission
from Ref. [68], FIG. 3.33--Research octane number of naphthas (~ = 1.8 x
~ +32). Taken with permission from Ref. [79].
TABLE 3.29--Coefficients for Eq. (3.139) for estimation of RON [68, 78].
Hydrocarbon family a b c d
n-Paraffins 92.809 -70.97 -53 20 10
isoparaffins
2-Methyl-pentanes 95.927 -157.53 561 -600 200
3-Methyl-pentanes 92.069 57.63 -65 0 0
2,2-Dimethyl-pentanes 109.38 -38.83 -26 0 0
2,3-Dirnethyl-pentanes 97.652 -20.8 58 -200 100
Naphthenes -77.536 471.59 -418 100 0
Aromatics 145.668 -54.336 16.276 0 0
Taken with permission from Ref. [68].
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 141
rr
4~
30
J\\" \ .-
equally b e t w e e n n-paraffins a n d isoparaffins as xNp = Xip =
0.928/2 = 0.464. I n this case RON = 72.36. The e r r o r o n
calculation of RON is 2.76.
c. To use Fig. 3.33 we need Tb = 42.8~ -- 109~ a n d Kw,
\ , w h i c h from Eq. (2.13) is c a l c u l a t e d as Kw -- 12.75. Since
20 I
ii \ Xx \~ !oo~F.
the Kw is outside the r a n g e of values in Fig. 3.33, a c c u r a t e
I r e a d i n g is n o t possible, b u t f r o m value of the boiling p o i n t
10 I i \"?50 ~ P. I it is obvious t h a t the RON from e x t r a p o l a t i o n of the curves
10 20 30 40 50 60 70 80 90 100 is above 70.
Wt% Paraffin Hydrocarbons d. To use Fig. 3.34 we need total paraffins w h i c h is % - -
49.33 + 41.45 = 90.78 a n d Tb = 109~ In this case Tb is out-
FIG. 3.34--Fiesearch octane number versus paraffin content side the range of values on the curves, b u t with extrapola-
(~ ---- 1 . 8 x ~ +32). Taken with permission from Fief. [7g]. tion a value of a b o u t 66 can be read. The e r r o r is a b o u t
-3.6.
e q u a t i o n sensitivity of gasoline can be d e t e r m i n e d . Gasolines e. To calculate MON we use Eq. (3.140) with R O N =
with lower sensitivity are desirable. 69.6, SG = 0.6501, a n d % 0 -- TML = TEL -----0. The esti-
m a t e d value is (MON)e~t. = 67.3, w h i c h is in g o o d agree-
Example 3.22~A n a p h t h a s a m p l e f r o m a n Australian c r u d e m e n t w i t h the r e p o r t e d value of 66.2 [46] w i t h e r r o r of + 1.1.
oil has the following characteristics: boiling p o i n t r a n g e f. If e s t i m a t e d RON value of 70.55 (from P a r t a) is u s e d in
15.5-70~ specific gravity 0.6501, n-paraffins 49.33%, Eq. (3.140), the p r e d i c t e d MON is 68 with deviation of + 1.8.
isoparaffins 41.45%, n a p h t h e n e s 9.14%, a r o m a t i c s 0.08%, g. E r r o r s are c a l c u l a t e d a n d given in each part. E q u a t i o n
clear RON 69.6, a n d MON 66.2 [Ref. [46], p. 359). (3.138) gives the lowest e r r o r w i t h deviation of less t h a n 1
w h e n e x p e r i m e n t a l PIONA c o m p o s i t i o n is used. F o r sam-
a. E s t i m a t e RON f r o m the p s e u d o c o m p o n e n t m e t h o d using ples in w h i c h the difference b e t w e e n a m o u n t s of n-paraffins
experimental composition. a n d isoparaffins is small, Eq. (3.138) gives even b e t t e r re-
b. E s t i m a t e RON f r o m the p s e u d o c o m p o n e n t m e t h o d using sults. I n the cases t h a t the c o m p o s i t i o n is n o t available the
p r e d i c t e d PNA c o m p o s i t i o n . p r o c e d u r e used in P a r t b using p r e d i c t e d c o m p o s i t i o n w i t h
c. E s t i m a t e RON from Fig. 3.33. m i n i m u m d a t a on boiling p o i n t a n d specific gravity gives
d. E s t i m a t e RON f r o m Fig. 3.34. an a c c e p t a b l e value for RON. *
e. E s t i m a t e MON f r o m actual r e p o r t e d RON.
f. E s t i m a t e MON f r o m p r e d i c t e d RON from Part a.
g. F o r each case calculate the e r r o r (deviation between esti-
3.6.8 Carbon Residue
m a t e d a n d r e p o r t e d values).
W h e n a p e t r o l e u m fraction is v a p o r i z e d in the a b s e n c e of air
Solution--For this fraction: Tb = (15.5 + 70)/2 = 42.8~ at a t m o s p h e r i c pressure, the nonvolatile c o m p o u n d s have a
SG = 0.6501,Xp = 0.4933, xw = 0.4145, XN = 0.0914, XA : c a r b o n a c e o u s residue k n o w n as carbon residue, w h i c h is des-
0.008. i g n a t e d b y CR. Therefore, heavier fractions w i t h m o r e aro-
m a t i c contents have h i g h e r c a r b o n residues while volatile a n d
a. RON can be e s t i m a t e d f r o m Eq. (3.138) t h r o u g h pseu-
light fractions such as n a p h t h a s a n d gasolines have n o car-
d o c o m p o n e n t m e t h o d using R O N values for p u r e hydro-
b o n residues. CR is p a r t i c u l a r l y a n i m p o r t a n t c h a r a c t e r i s t i c
c a r b o n s calculated f r o m Eq. (3.139) a n d Table 3.29
of c r u d e oils a n d p e t r o l e u m residues. H i g h e r CR values in-
with Tb = 315.9 K. Results of calculation are (RON)np =
dicate low-quality fuel a n d less h y d r o g e n content. There are
54.63, (RON)Ip = (90.94 + 104.83 + 88 + 87.05)/4 = 92.7,
two o l d e r different test m e t h o d s to m e a s u r e c a r b o n residues,
(RON)N -- 55.57, a n d (RON)A ----125.39. I n calculation of
R a m s b o t t o m (ASTM D 524) a n d the C o n r a d s o n (ASTM
(RON)w, an average value for RON of four families in
D 189). The r e l a t i o n s h i p between these m e t h o d s are also given
Table 3.29 is calculated. F r o m Eq. (3.139), clear RON can
b y the ASTM D 189 m e t h o d . Oils that have ash f o r m i n g c o m -
be calculated as:
p o u n d s have e r r o n e o u s l y high c a r b o n residues b y b o t h meth-
RON = 0.4933 • 54.63 + 0.4145 x 92.7 + 0.0914 x 55.57 ods. F o r such oils ash should be r e m o v e d before the m e a s u r e -
+ 0 . 0 0 0 8 • 125.39 = 70.55. ment. There is a m o r e recent test m e t h o d (ASTM D 4530) t h a t
requires s m a l l e r s a m p l e a m o u n t s a n d is often referred as mi-
In c o m p a r i s o n with the r e p o r t e d value of 69.6 the e r r o r is crocarbon residue (MCR) a n d as a result it is less precise in
70.55 - 69.6 = 0.95. practical t e c h n i q u e [7]. In m o s t cases c a r b o n residues are
142 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
reported in wt% by Conradson method, which is designated lamp without smoking. It is expressed in millimeters and a
by %CCR. high smoke point indicates a fuel with low smoke-producing
Carbon residue can be correlated to a number of other prop- tendency [61]. Measurement of smoke point is described un-
erties. It increases with an increase in carbon-to-hydrogen der ASTM D t322 (U.S.) or IP 57 (UK) and ISO 3014 test
ratio (CH), sulfur content, nitrogen content, asphaltenes con- methods. For a same fuel measured smoke point by IP test
tent, or viscosity of the oil. The most precise relation is method is higher than ASTM method by 0.5-1 m m for smoke
between CR and hydrogen content in which as hydrogen con- point values in the range of 20-30 m m [61].
tent increases the carbon residues decreases [7]. The hydro- Smoke point may be estimated from either the PNA com-
gen content is expressed in terms of H/C atomic ratio and position or from the aniline point. The SP of kerosenes from
the following relation may be used to estimate CCR from H/C IP test method may be estimated from the following relation
[81]. [90]:
(3.141) %CCR = 148.7 - 86.96 H/C SP = 1.65X - 0.0112X 2 - 8.7
if H/C _> 1.71 (%CCR < 0), set %CCR = 0.0 and if H/C < 0.5 (3.144) 100
X =
(%CCR > 100), set %CCR = 100. H/C ratio can be estimated 0.61Xp + 3.392xN + 13.518xA
from CH ratio methods given in Section 2.6.3. where SP is the smoke point by IP test method in m m and Xv,
The carbon residue is nearly a direct function of high boil- xN, and :cA are the fraction of paraffin, naphthene, and aro-
ing asphaltic materials and Nelson has reported a linear re- matic content of kerosenes. The second method is proposed
lation between carbon residue and asphalt yield [82]. One by Jenkins and Walsh as follows [83]:
of the main characteristic of residuum is its asphaltene con-
tent. Asphaltenes are insoluble in low molecular weight n- SP = -255.26 + 2.04AP - 240.8 ln(SG) + 7727(SG/AP)
alkanes including n-pentane. Knowledge of n-pentane insol- (3.145)
ubles in residual oils is quite important in determining yields
and products qualities for deasphalting, thermal visbreaking, where AP is the aniline point in ~ and SG is the specific grav-
and hydrodesulfurization processing. The relation between ity at 15.5~ Both Eqs. (3.144) and (3.145) estimate SP ac-
the normal pentane insolubles and carbon residue is as fol- cording to the IP test method. To estimate SP from the ASTM
lows [61]: D 1322 test method, 0.7 mm should be subtracted from the
calculated IP smoke point. Equations (3.144) and (3.145) are
(3.142) %NCs = 0.74195 (%CCR) + 0.01272 (%CCR) 2 based on data with specific gravity in the range of 0.76-0.82,
where %NC5 is the wt% of n-pentane insolubles and %CCR is and smoke points in the range of 17-39 mm. Based on some
the wt% of Conradson carbon residue. Once %NCs is known, preliminary evaluations, Eq. (3.133) is expected to perform
the asphaltene content (asphaltene wt%) of a residue can be better than Eq. (3.144), because smoke point is very much re-
determined from the following empirical relation: lated to the aromatic content of the fuel which is expressed in
terms of aniline point in the Jenkins-Walsh method. In addi-
(3.143) %Asphaltene = a(%NCs) tion the specific gravity, which is an indication of molecular
where a is 0.385 for atmospheric residue and 0.455 for vac- type, is also used in the equation. Equation (3.144) may be
uum residues [61,66]. These equations are approximate and used for cases that the aniline point is not available but ex-
do not provide accurate predictions. perimental PNA composition is available. Albahri et al [68]
also proposed the following relation for prediction of smoke
Example 3.23--A vacuum residue of an Australian crude oil point using API gravity and boiling point:
has carbon-to-hydrogen weight ratio of 7.83. Estimate its car- (3.146) SP = 0.839(API) + 0.0182634(Tb) -- 22.97
bon residue and asphaltene contents and compare the results
with the experimental values of 15.1 and 4.6%, respectively where SP is in m m (ASTM method) and Tb is the average
[46]. boiling point in kelvin. This equation when tested for 136
petroleum fractions gave an average error of about 2 m m [68].
Solution--With CH=7.83, from Eq. (2.122), HC atomic ratio
is calculated as HC --- 1.52. From Eq. (3.141), %CCR = 16.4% Example 3.24--A Nigerian kerosene has an API gravity of
and from Eq. (3.142), %NC5 = 15.6%. From Eq. (3.143) with 41.2, aniline point of 55.6~ and the PNA composition of
a = 0.455 (for vacuum residue) we calculate %Asphaltene = 36.4, 49.3, and 14.3%. Estimate the smoke point of this fuel
7.1. The results show that while Eq. (3.141) provides a from. Equations (3.144)-(3.146) and compare with the exper-
good prediction for %CCR, prediction of %Asphaltene from imental value of 20 mm (Ref. [46], p. 342).
Eq. (3.143) is approximate.
Solution--To estimate SP from Eq. (3.144), we have Xp =
0.364, XN = 0.493, and XA= 0.143 which give X = 26.13. Cal-
3.6.9 S m o k e P o i n t
culated SP is 26.8 m m according to the IP method or 26.1
Smoke point is a characteristic of aviation turbine fuels and mm according to the ASTM method. To use Eq. (3.145) we
kerosenes and indicates the tendency of a fuel to burn with have from API gravity, SG = 0.819, AP -- 55.6~ the calcu-
a smoky flame. Higher amount of aromatics in a fuel causes lated SP is SP = 20 mm. The ASTM smoke point is then 19.3
a smoky characteristic for the flame and energy loss due to m m which is in very good agreement with the experimental
thermal radiation. The smoke point (SP) is a maximum flame value of 20 with deviation of -0.7 mm. Predicted value from
height at which a fuel can be burned in a standard wick-fed Eq. (3.146) is 17.6 mm.
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 143
M e t h o d s p r e s e n t e d in this c h a p t e r c a n be u s e d to evaluate
the quality of p e t r o l e u m p r o d u c t s f r o m available p a r a m e t e r s .
The quality of a p e t r o l e u m p r o d u c t d e p e n d s on certain -5]
specifications o r p r o p e r t i e s of the fuel to satisfy r e q u i r e d
criteria set b y the m a r k e t d e m a n d . These characteristics are
specified for b e s t use of a fuel (i.e., highest engine perfor-
m a n c e ) or for cleaner e n v i r o n m e n t while the fuel is in use.
These specifications vary f r o m one p r o d u c t to a n o t h e r a n d -10
from one c o u n t r y to another. F o r example, for gasoline the
quality is d e t e r m i n e d b y a series of p r o p e r t i e s such as sulfur
a n d a r o m a t i c contents, octane number, v a p o r pressure,
h y d r o g e n content, a n d boiling range. Engine w a r m - u p t i m e
is affected by the p e r c e n t distilled at 70~ a n d the ASTM 90% -15
t e m p e r a t u r e . F o r the a m b i e n t t e m p e r a t u r e of 26.7~ (80~
a gasoline m u s t have ASTM 90% t e m p e r a t u r e of 188~ a n d
3% distilled at 70~ to give a c c e p t a b l e w a r m - u p t i m e [63].
S t a n d a r d o r g a n i z a t i o n s such as ASTM give such specifica-
-20
tions for various products. F o r r e f o r m u l a t e d gasoline sulfur f
HC CO NOx
c o n t e n t of less t h a n 300 p p m (0.03 wt%) is r e q u i r e d [63].
A m o u n t of p a r t i c u l a t e emissions is directly related to the FIG. 3.35--Influence of sulfur content in gasoline (from
a r o m a t i c a n d sulfur content of a fuel. Figure 3.35 shows the 500 to 50 ppm) in reduction of pollutant gases. Taken with
influence of sulfur r e d u c t i o n in gasoline f r o m 500 to 50 p p m permission from Ref. [24].
in the r e d u c t i o n of p o l l u t a n t emissions [24].
Vapor p r e s s u r e of gasoline of jet fuel d e t e r m i n e s their ig- p r o d u c e the r e q u i r e d final product. In E x a m p l e 3.20, it was
nition characteristics. While freezing p o i n t is i m p o r t a n t for s h o w n t h a t to have a p r o d u c t with certain flash point, one c a n
jet fuels it is n o t a m a j o r c h a r a c t e r i s t i c for gas oils. F o r lubri- d e t e r m i n e the v o l u m e of various c o m p o n e n t s in the blend.
cating oil p r o p e r t i e s such as viscosity a n d viscosity index are The s a m e a p p r o a c h c a n be e x t e n d e d to a n y o t h e r property. F o r
i m p o r t a n t in a d d i t i o n to sulfur a n d PNA c o m p o s i t i o n . Aniline example, to increase v a p o r p r e s s u r e of gasoline n-butane m a y
p o i n t is a useful characteristic to indicate p o w e r of solubil- be a d d e d d u r i n g w i n t e r season to i m p r o v e engine starting
ity of solvents as well as a r o m a t i c contents of certain fuels. characteristics of the fuel [63]. The a m o u n t of r e q u i r e d b u t a n e
F o r heavy p e t r o l e u m p r o d u c t s knowledge of p r o p e r t i e s such to r e a c h a certain v a p o r p r e s s u r e value can be d e t e r m i n e d
as c a r b o n residue, p o u r point, and cloud p o i n t are of inter- t h r o u g h calculation of v a p o r p r e s s u r e b l e n d i n g index for the
est. S o m e i m p o r t a n t specifications of jet fuels are given in c o m p o n e n t s a n d the p r o d u c t as discussed in Section 3.6.1.1.
Table 3.30.
One of the techniques u s e d in refining t e c h n o l o g y to pro-
duce a p e t r o l e u m p r o d u c t w i t h a certain characteristic is the 3.8 M I N I M U M LABORATORY DATA
b l e n d i n g m e t h o d . Once a certain value for a p r o p e r t y (i.e.,
viscosity, octane number, p o u r point, etc.) of a p e t r o l e u m As discussed earlier m e a s u r e m e n t of all p r o p e r t i e s of various
p r o d u c t is required, the mixture m a y be b l e n d e d with a cer- p e t r o l e u m fractions a n d p r o d u c t s in the l a b o r a t o r y is an
tain c o m p o n e n t , additive o r a n o t h e r p e t r o l e u m fraction to i m p o s s i b l e task due to the r e q u i r e d cost a n d time. However,
TABLE 3.31---Standard test methods for measurement of some properties of liquid petroleum products.
Property ASTMD IP ISO Property ASTMD IP ISO
Aniline Point 611 2/98 2977 Flash Point 93 34/97 2719
Carbon Residue (Ramsbottom) 524 14/94 4262 Freezing Point 2386 16/98 3013
Carbon Residue (Conradson) 189 13/94 6615 Hydrocarbon Types 1319 156/95 3837
Centane Number 4737 380/98 4264 Heating Value 240 12
Cloud Point 2500 219/94 3016 Kinematic Viscosity 445 71/97 3104
Color 1500 1 9 6 / 9 7 2049 Octane Number (Motor) 2700 236 5163
Density/Sp. Gr. 4052 365/97 2185 Refractive Index 1218
Distillation at Atm. Pressure 86 1 2 3 / 9 9 " 3405 Pour Point 97 15/95 3015
Distillation at Reduced Pressures 1160 6616" Sulfur Content 1266 1 0 7 / 8 6 2192
Distillation by Gas Chromatography 2887 406/99* Thermal Conductivity 2717-95
Distillation of Crude Oils 2892 8708* Vapor Pressure (Reid) 323 69/94 3007
Viscosity (Viscous Oils) 2983 370/85*
ASTM has test methods for certain properties for which other test methods do not suggest equivalentprocedures. Some of these methods
include heat of combustion:D 4809; smoke point: D 1322;surface tension: D 3825;vapor-liquidratio: D 2533; viscositytemperature chart: D
341; autoignition:D 2155 (ISO 3988). Further test methods for some specificproperties are givenin the text where the property is discussed.
ASTMmethods are taken from Ref. [4]. IP methods are taken from Ref. [85]. Methods specifiedby * are similar but not identical to other
standard methods. Most IP methods are also used as British Standard under BS2000methods [85]. The number after IP indicatesthe year of
]ast approval. ISO methods are taken from Refs. [24] and [85].
3. CHARACTERIZATION OF PETROLEUM FRACTIONS 145
the knowledge of PNA composition for prediction of prop- laboratory data for any physical property in terms of some
erties of heavy fractions is more useful than for light frac- other parameters. As the complexity and the n u m b e r of pa-
tions. For wide boiling range fractions knowledge of complete rameters increases the accuracy of the correlation also in-
distillation curve is quite useful to consider nature of different creases with respect to the data used in the development of
compounds and their effects on the properties of the mixture. the correlation. However, the main problem with empirical
As it is shown in Chapter 4, for wide and heavy fractions such correlations is their limited power of extrapolation and the
as C7+ fractions, distribution of carbon number in the frac- large number of numerical constants involved in the correla-
tion is the most useful information besides specific gravity. tion. For example, in Chapter 2, several correlations are pro-
Further analysis of minimum laboratory data for characteri- vided to estimate molecular weight of hydrocarbons in terms
zation of petroleum fractions is provided in our previous work of boiling point and specific gravity. Equation (2.50) derived
[84]. from Eq. (2.38) has only three numerical constants, which
Predictive methods of characterization must be used when are developed from molecular weight of pure hydrocarbons.
experimental data are not available. If possible, one should Tsonopoulos et al. [86] made an extensive analysis of various
make m a x i m u m use of available experimental data, A sum- methods of estimation of molecular weights of coal liquids.
mary of standard test methods for some specifications of Equation (2.50) was compared with several empirical corre-
liquid petroleum products is given in Table 3.31. For some lations specifically developed for coal liquids having as m a n y
properties equivalent test methods according to the interna- as 16 numerical constants. They concluded that Eq. (2.50)
tional standards organization (ISO) are also specified in this is the most accurate method for the estimation of molecular
table [24, 85]. weight of coal liquids. No data on coal liquid were used in
development of constants in Eq. (2.50). However, since it was
developed with some physical basis and properties of pure
3.9 ANALYSIS OF LABORATORY DATA A N D hydrocarbons were used to obtain the numerical constants
D E V E L O P M E N T OF P R E D I C T I V E M E T H O D S the equation has a wide range of applications from pure hy-
drocarbons to petroleum fractions and coal liquids, which are
In Chapter 2 and this chapter the predictive methods in terms mainly aromatics. This indicates the significance of develop-
of readily available parameters are presented for estimation of ment of correlations based on the physical understanding of
various properties related to basic characteristics and quality the nature of the system and its properties. The main advan-
of petroleum fractions. Generally these methods fall within tage of such correlations is their generality and simplicity.
two categories of empirical and semiempirical correlations. The main characteristics of an ideal predictive method for a
In an empirical correlation the structure of the correlation certain property are accuracy, simplicity, generality, and avail-
is determined through fitting the data and the type of input ability of input parameters. The best approach toward the
parameters in each correlation are determined through anal- development of such correlations would be to combine phys-
ysis of experimental data. While in a semi-empirical corre- ical and theoretical fundamentals with some modifications.
lation, the structure and functionality of the relation is de- An example of such type of correlations is Eq. (2.39), which is
termined from a theoretical analysis of parameters involved an extension of Eq. (2.38) derived from physical basis. A pure
and through analysis of existing theoretical relations. Once empirical correlation might be quite accurate to represent the
the main functionality and nature of a correlation between data used in its development but when it is applied to other
various physical properties is determined, the correlation co- systems the accuracy is quite low. In addition characterizing
efficients can be determined from experimental data. The best the systems according to their degree of complexity is helpful
example of such a predictive method is development of Eq. to develop more accurate correlations. For example, heavy
(2.38) in Chapter 2, which was developed based on the un- fractions contain heavy and nonpolar compounds, which dif-
derstanding of the intermolecular forces in hydrocarbon sys- fer with low-molecular-weight hydrocarbons present in light
tems. This generalized correlation has been successfully used petroleum fractions. Therefore, in order to increase the de-
to develop predictive methods for a variety of physical prop- gree of accuracy of a predictive method for a certain prop-
erties. An example of an empirical correlation is Eq. (2.54) erty it is quite appropriate to develop one correlation for
developed for estimation of molecular weight of petroleum light and one correlation for heavy fractions. For heavy frac-
fractions. Many other correlations presented in this chapter tions because of the nature of complex compounds in the
for estimation of properties such as aniline and smoke points mixture three input parameters are required. As variation in
or methods presented for calculation of octane numbers for a properties of pure hydrocarbons from one family to another
blend are also purely empirical in nature. In development of increases with increase in carbon numbers (i.e., see Figs. 2.15
an empirical relation, knowledge of the nature of properties and 2.23), the role of composition on estimation of such prop-
involved in the correlation is necessary. For example, aniline erties for heavy fractions is more than its effect on the prop-
point is a characteristic that depends on the molecular type erties of light fractions. Therefore, including molecular type
of hydrocarbons in the fraction. Therefore, it is appropriate in the development of predictive methods for such properties
to relate aniline point to the parameters that characterize hy- of heavy fractions is quite reasonable and useful and would
drocarbon types (i.e., Ri) rather than boiling point that char- enhance the accuracy of the method.
acterizes carbon number in a hydrocarbon series. Once the structure of a correlation is determined from
Mathematical functions can be expressed in the form of theoretical developments between various properties, ex-
polynomial series; therefore, it is practically possible to de- perimental data should be used to determine the numerical
velop correlations in the forms of polynomial of various de- constants in the correlation. If the data on properties of
grees. With powerful computational tools available at present pure hydrocarbons from different families are used to
it is possible to find an empirical correlation for any set of determine the constants, the resulting correlation would be
146 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
more general and applicable to various types of petroleum based on a minimum of three data points along the distillation
fractions within the same boiling point range. The accuracy curve. For prediction of each characteristic of a petroleum
of a correlation for a specific group of fractions could be fraction, several methods are provided that have been in use
increased if the coefficients in the correlation are obtained in the petroleum industry. Limitations, advantages, and dis-
from the same group of fractions. In obtaining the constants, advantages of each method are discussed.
many equations may be convertible into linear forms and Basically two approaches are proposed in characterization
spreadsheets such as Lotus or Excel programs may be used to of petroleum fractions. One technique is based on the use
obtain the constants by means of least squared method. Non- of bulk properties (i.e., Tb and SG) considering the whole
linear regression of correlations through these spreadsheets mixture as a single pseudocomponent. The second approach
is also possible. In analyzing the suitability of a correlation, called pseudocomponent technique considers the fraction as
the best criteria would be the R2 or correlation parameter a mixture of three pseudocomponents from the three fami-
defined by Eq. (2.136) in which values of above 0.99 indicate lies. This technique is particularly useful for heavy fractions.
an equation is capable of correlating data. Use of a larger data A third approach is also provided for wide boiling range frac-
bank and most recent published data in obtaining the nu- tions. However, since behavior of such fractions is similar to
merical constants would enhance accuracy and applicability that of crude oils the technique is mainly presented in the next
of the correlation. A fair way of evaluations and comparison chapter. Fractions are generally divided into light and heavy
of various correlations to estimate a certain property from fractions. For heavy fractions a minimum of three character-
the same input parameters would be through a data set not ization parameters best describe the mixture. Recommenda-
used in obtaining the coefficients in the correlations. In such tions on the use of various input parameters and advantages
evaluations AAD or %AAD can be used as the criterion to of different methods were discussed in Sections 3.8 and 3.9.
compare different methods. Average absolute deviation may For light fractions (M < 300, Nc < 22) and products of atmo-
be used when the range of variation in the property is very spheric distillation unit, Eq. (2.38) for M, I, and d is quite ac-
large and small values are estimated. For example, in predic- curate. For such light fractions, To, Pc, and cocan be estimated
tion of PNA composition, amount of aromatics varies from from Eqs. (2.65), (2.66), and (2.105), respectively. For pre-
1% in petroleum fractions to more than 90% in coal liquids. diction of the PNA composition for fractions with M < 200,
AAD of 2 (in terms of percentage) in estimating aromatic Eqs. (3.77) and (3.78) in terms of m and SG are suitable. For
content is quite reasonable. This error corresponds to 200% fractions with M > 200, Eqs. (3.71)-(3.74) in terms of P~ and
in terms of %AD for fractions with t% aromatic content. VGC are the most accurate relations; however, in cases that
Experience has shown that correlations that have fewer nu- viscosity data are not available, Eqs. (3.79) and (3.80) in terms
merical constants and are based on theoretical and physical of P~ an CH are recommended. Special recommendations on
grounds with constants obtained from a wide range of data use of various correlations for estimation of different prop-
set are more general and have higher power of extrapolation. erties of petroleum fractions from their bulk properties have
been given in Section 2.10 and Table 2.16 in Chapter 2.
In this chapter various characterization methods for differ- 3.1. List four different types of analytical tools used for com-
ent petroleum fractions and mixtures have been presented. positional analysis of petroleum fractions.
This is perhaps one of the most important chapters in the 3.2. What are the advantages/disadvantages and the differ-
book. As the method selected for characterization of a frac- ences between GC, MS, GC-MS, GPC and HPLC instru-
tion would affect prediction of various properties discussed in ments?
the remaining part of the book. As it is discussed in Chapter 4, 3.3. A jet naphtha has the following ASTM D 86 distillation
characterization and estimation of properties of crude oils de- curve [ 1]:
pend on the characterization of petroleum fractions discussed
in this chapter. Through methods presented in this chapter vol% distilled 10 30 50 70 90
one can estimate basic input data needed for estimation of ASTM D 86 temperature, ~ 151.1 156.1 160.6 165.0 171.7
thermodynamic and physical properties. These input param-
eters include critical properties, molecular weight, and acen- a. Calculate VABP, WABR MABP, CABP, and MeABP for
tric factor. In addition methods of estimation of properties this fraction. Comment on your calculated MeABR
related to the quality of a petroleum product such as distil- b. Estimate the specific gravity of this fraction and com-
lation curves, PNA composition, elemental composition, vis- pare with reported value of 0.8046.
cosity index, carbon residue, flash, pour, cloud, smoke, and c. Calculate the Kw for this fraction and compare with
freezing points as well as octane and cetane numbers are pre- reported value of 11.48.
sented. Such methods can be used to determine the quality 3.4. A kerosene sample has the following ASTM D 86, TBP,
of a fuel or a petroleum product based on the minimum labo- and SG distribution along distillation curve [1]. Convert
ratory data available for a fraction. Methods of conversion of ASTM D 86 distillation curve to TBP by Riazi-Daubert
various types of distillation curves help to determine neces- and Daubert's (API) methods. Draw actual TBP and pre-
sary information for process design on complete true boiling dicted TBP curves on a single graph in ~ Calculate the
point distillation curve when it is not available. In addition a average specific gravity of fraction form SG distribution
method is provided to determine complete distillation curve and compare with reported value of 0.8086.
3. C H A R A C T E R I Z A T I O N OF P E T R O L E U M F R A C T I O N S 147
3.5. F o r the k e r o s e n e s a m p l e of P r o b l e m 3.4, a s s u m e ASTM 3.13. F o r the fraction of P r o b l e m 3.12 calculate acentric fac-
t e m p e r a t u r e s at 30, 50, a n d 70% p o i n t s are the only tor, w from the following methods:
k n o w n i n f o r m a t i o n . Based on these d a t a points o n the a. L e e - K e s l e r m e t h o d w i t h Tc a n d Pc f r o m Parts (a), (b),
distillation curve p r e d i c t the entire distillation curve a n d (c), a n d (d) of P r o b l e m 3.12.
c o m p a r e with the a c t u a l values in b o t h t a b u l a t e d a n d b. E d m i s t e r m e t h o d w i t h Tc a n d Pc from Parts a n d (a)
g r a p h i c a l forms. a n d (e) of P r o b l e m 3.12.
3.6. F o r the kerosene s a m p l e of P r o b l e m 3.4 calculate the c. K o r s t e n m e t h o d with Tc a n d Pc f r o m P a r t (a) of Prob-
W a t s o n K factor a n d generate the SG d i s t r i b u t i o n b y lem 3.12.
c o n s i d e r i n g c o n s t a n t Kw along the curve. Calculate d. P s e u d o c o m p o n e n t m e t h o d
%AAD for the deviations b e t w e e n p r e d i c t e d SG a n d ac- 3.14. A p u r e h y d r o c a r b o n has a boiling p o i n t of 110.6~ a n d
tual SG for t h e 8 d a t a points. S h o w a g r a p h i c a l evalua- a specific gravity of 0.8718. W h a t is the type of this hy-
tion of p r e d i c t e d distribution. d r o c a r b o n (P, N, MA, o r PA)? H o w can you check y o u r
3.7. F o r the kerosene s a m p l e of P r o b l e m 3.4 generate SG answer? Can you guess the c o m p o u n d ? In y o u r analysis
d i s t r i b u t i o n using Eq. (3.35) a n d d r a w p r e d i c t e d SG dis- it is a s s u m e d that you do not have access to the table of
t r i b u t i o n w i t h actual values. Use Eq. (3.37) to calculate p r o p e r t i e s of p u r e h y d r o c a r b o n s .
specific gravity of the m i x t u r e a n d c o m p a r e with the ac- 3.15. E x p e r i m e n t a l d a t a on sulfur content of s o m e p e t r o l e u m
tual value of 0.8086. p r o d u c t s along with o t h e r b a s i c p a r a m e t e r s are given in
3.8. F o r the kerosene s a m p l e of P r o b l e m 3.4 calculate den- Table 3.32. It is a s s u m e d that the only d a t a available for
sity at 75~ a n d c o m p a r e with the r e p o r t e d value o f a p e t r o l e u m p r o d u c t is its m i d boiling point, Tb, ( c o l u m n
0.8019 g/cm 3 [1]. 1) a n d refractive index at 20~ n20 ( c o l u m n 2). Use ap-
3.9. F o r the kerosene s a m p l e of P r o b l e m 3.4, a s s u m e the only p r o p r i a t e m e t h o d s to c o m p l e t e c o l u m n s 4,6,8,9,10, a n d
d a t a available are ASTM D 86 t e m p e r a t u r e s at 30, 50, 12 in this table.
a n d 70% points. Based on these d a t a calculate the spe- 3.16. ASTM 50% t e m p e r a t u r e (Tb), specific gravity (SG), a n d
cific gravity at 10, 30, 50, 70, a n d 90 % a n d c o m p a r e w i t h the PNA c o m p o s i t i o n of 12 p e t r o l e u m fractions are given
the values given in P r o b l e m 3.4. in Table 3.33. Complete c o l u m n s of this table b y calculat-
3.10. A p e t r o l e u m fraction has the following ASTM D 1160 ing M, n, m, a n d the PNA c o m p o s i t i o n for each fraction
distillation curve at 1 m m H g : b y using a p p r o p r i a t e methods.
3.17. A gasoline p r o d u c t from N o r t h Sea c r u d e oil has boiling
vol% distilled 10 30 50 70 90 range of C5-85~ a n d specific gravity of 0.6771. Predict
ASTMD 1160 temperature, ~ 104 143 174 202 244 the PNA c o m p o s i t i o n a n d c o m p a r e with e x p e r i m e n t a l l y
d e t e r m i n e d c o m p o s i t i o n of 64, 25, a n d 11 in wt% [46].
Predict TBP, ASTM D 86, a n d EFV distillation curves all 3.18. A r e s i d u e f r o m a N o r t h Sea crude has the following ex-
at 760 m m Hg. p e r i m e n t a l l y d e t e r m i n e d characteristics: 1)99(210) = 14.77
3.11. A gas oil s a m p l e has the following TBP a n d density dis- cSt., SG ----0.9217. F o r this fraction e s t i m a t e the follow-
t r i b u t i o n [32]. E x p e r i m e n t a l values of M, n20, a n d SG ing p r o p e r t i e s a n d c o m p a r e with the e x p e r i m e n t a l val-
are 214, 1.4694, a n d 0.8475, respectively. ues w h e r e they are available [46]:
a. Use the m e t h o d o u t l i n e d for wide boiling range frac- a. K i n e m a t i c viscosity at 38~ (100~
tions a n d e s t i m a t e M a n d n20. b. M o l e c u l a r weight a n d average boiling p o i n t
b. Use Eqs. (2.50) a n d (2.115) to e s t i m a t e M a n d n
c. C o m p a r e %AD f r o m m e t h o d s a a n d b.
Vol% IBP 5 10 20 30 40 50 60 70 80 90 95
TBP, ~ 420 451 470 495 514 ... 544 ... 580 ... 604 621
d20 ... 0.828 0.835 0.843 0.847 ... 0.851 ... 0.855 ... 0.856 0.859
TABLE 3.32--Basic parameters and sulfur content of some undefined petroleum products (Problem 3.15). a
(i) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Tb, K n2o SG, SG, M, M, d20, d20, rr~ RI Sulfur %, Sulfur %,
Fraction exp. exp. exp. calc. exp. calc. exp. calc. calc. calc. exp. calc.
Kuwaiti kerosene 468 1 . 4 4 1 0.791 . . . . . . 0.01
Kuwaiti diesel oil 583 1.480 0.860 1.3
US jet naphtha 434 1.444 0.805 144 01861 1.0
US high boiling naphtha 435 1.426 0.762 142.4 0.759 0.0
US kerosene 480 1.444 0.808 162.3 0.804 0.0
US fuel off 559 1.478 0.862 227.5 0.858 1.3
aExperimental data for the first two Kuwaiti fractions are taken from Riazi and Roomi [62] and for the US fractions are taken from Lenoir
and Hipkin data set [ 1].
The specific gravity (at 60/60~ ) of fraction is 0.8597 [10] Zhou, E, "Correlation of the Average Boiling Points of
[46]. Predict the following p r o p e r t i e s for the p r o d u c t Petroleum Fractions with Pseudocritical Constants,"
a n d c o m p a r e with r e p o r t e d values. International Chemistry Engineering, Vol. 24, 1984, pp. 731-742.
a. K i n e m a t i c viscosities at 100, 140, a n d 210~ (experi- [11] Daubert, T. E., "Property Predictions," Hydrocarbon Processing,
March, 1980, pp. 107-112.
m e n t a l values are 3.26, 2.04, a n d 1.12 cSt.)
[12] Van Winkle, M., "Physical Properties of Petroleum Fraction,"
b. Molecular weight f r o m viscosity a n d c o m p a r e with
Petroleum Refiner, Vol. 34, No. 6, 1955, pp. 136-138.
m o l e c u l a r weight e s t i m a t e d f r o m the boiling point. [13] Nelson, W. L. and Hansburg, M., "Relation of ASTM and True
c. Boiling p o i n t a n d specific gravity from experimen- Boiling Point Distillations," Oil & Gas Journal, Vol. 38, No. 12,
tal values of k i n e m a t i c viscosities at 100 a n d 210~ 1939, pp. 45-47.
(Part a) a n d c o m p a r e with actual values. [14] Edmister, W. C. and Pllock, D. H., "Phase Relations for
d. Aniline p o i n t (experimental value is 143.5~ Petroleum Fractions," Chemical Engineering Progress, Vol. 44,
e. Cetane n u m b e r (experimental value is 43.2) 1948, pp. 905-926.
f. Freezing p o i n t (experimental value is 21 ~ [15] Edmister, W. C. and Okamoto, K. K., "Applied Hydrocarbon
g. F l a s h p o i n t (experimental value is 230~ Thermodynamics, Part 12: Equilibrium Flash Vaporization
Correlations for Petroleum Fractions," Petroleum Refiner,
h. C a r b o n - t o - h y d r o g e n weight ratio (experimental value
Vol. 38, No. 8, 1959, pp. 117-132.
is 6.69) [16] Edmister, W. C. and Okamoto, K. K., Applied Hydrocarbon
i. A r o m a t i c c o n t e n t f r o m e x p e r i m e n t a l viscosity at Thermodynamics, Gulf Publishing, Houston, TX, 1961,
100~ (experimental value is 34.9) pp. 116-133.
j. A r o m a t i c content from e x p e r i m e n t a l aniline p o i n t [17] Edmister, W. C. and Okamoto, K. K., "Applied Hydrocarbon
(experimental value is 34.9) Thermodynamics, Part 13: Equilibrium Flash Vaporization
k. Refractive index at 75~ (experimental value is Correlations for Heavy Oils Under Subatmospheric Pressures,"
1.4759) Petroleum Refiner, Vol. 38, No. 9, 1959, pp. 271-288.
3.25. A v a c u u m r e s i d u e has k i n e m a t i c viscosity of 4.5 mm2/s [18] Okamoto, K. K. and van Winkle, M., "Equilibrium Flash
at 100~ a n d specific gravity of 0.854. E s t i m a t e viscosity Vaporization of Petroleum Fractions," Industrial and
Engineering Chemistry, Vol. 45, 1953, pp. 429-439.
index (VI) of this fraction a n d c o m p a r e with r e p o r t e d
[19] House, H. G., Braun, W. G., Thompson, W. T. and Fenske,
value of 119 [24]. M. R , "Documentation of the Basis for Selection of the
3.26. A kerosene s a m p l e h a s boiling r a n g e of 180-225~ a n d Contents of Chapter 3, ASTM, TBP, and EFV Relationships for
specific gravity of 0.793. This p r o d u c t has a r o m a t i c Petroleum Fractions in Technical Data Book-Petroleum
c o n t e n t of 20.5%. Predict s m o k e p o i n t a n d freezing Refining," Documentation Report No. 3-66, Pennsylvania State
p o i n t of this p r o d u c t a n d c o m p a r e with the experi- University, 1966.
m e n t a l values of 19 m m a n d - 5 0 ~ [24]. H o w m u c h [20] Arnold, V. E., "Microcomputer Program Converts TBP, ASTM,
2 - m e t h y l n o n a n e (C10Hzz) with freezing p o i n t of - 7 4 ~ EFV Distillation Curves," Oil & Gas Journal, Vol. 83, No. 6,
s h o u l d be a d d e d to this kerosene to r e d u c e the freezing 1985, pp. 55-62.
p o i n t to -60~ [21] Ford, D. C., Miller, W. H., Thren, R. C. and Wertzler, R.,
Correlation of ASTM Method D 2887-73 Boiling Range
Distribution Data with ASTM Method D 86-67 Distillation Data,
STP 577, ASTM International, West Conshohocken, PA, 1973,
REFERENCES p. 29.
[22] Riazi, M. R. and Daubert, T. E., "Analytical Correlations
[1] Lenoir, J. M. and Hipkin, H. G., "Measured Enthalpies of Eight Interconvert Distillation Curve Types," Oil & Gas Journal,
Hydrocarbon Fractions," Journal of Chemical and Engineering Vol. 84, 1986, August 25, pp. 50-57.
Data, Vol. 18, No. 2, 1973, pp. 195-202. [23] Daubert, T. E.,"Petroleum Fraction Distillation
[2] Daubert, T. E. and Danner, R. E, Eds., API Technical Data Interconversion," Hydrocarbon Processing, Vol. 73, No. 9, 1994,
Book Petroleum Refining, 6th ed., American Petroleum pp. 75-78.
Institute (API), Washington, DC, 1997. [24] Wauquier, J.-E, "Petroleum Refining, Vol. 1: Crude Oil,
[3] Curvers, J. and van den Engel, E, "Gas Chromatographic Petroleum Products, Process Ftowsheets," Editions Technip,
Method for Simulated Distillation up to a Boiling Point of Paris, 1995.
750~ Using Temperature-Programmed Injection and High [25] HYSYS, "Reference Volume 1., Version 1.1," HYSYS Reference
Temperature Fused Silica Wide-Bore Columns," J. High Manual for Computer Software, HYSYS Conceptual Design,
Resolution Chromatography, Vol. 12, 1989, pp. 16-22. Hyprotech Ltd., Calgary, Alberta, Canada, 1996.
[4] ASTM, Annual Book of Standards, ASTM International, West [26] Daubert, T. E., Riazi, M. R. and Danner, R. E, "Documentation
Conshohocken, PA, 2002. of the Basis for Selection of the Contents of Chapters 2 and 3 in
[5] Denis, J., Briant, J. and Hipeaux, J. C., Lubricant Properties the API Technical Data Book-Petroleum Refining,"
Analysis and Testing, Translated to English by G. Dobson, Documentation Report No. 2,3-86, Pennsylvania State
Editions Technip, Paris, 1997. University, 1986.
[6] Cryette, W. R., Model 5009 Wide Range Cryoscope--Instruction [27] Maxwell, J. B. and Bonnel, L. S., Vapor Pressure Charts for
Manual Precision Systems Inc., Natick, MA, 1994. Petroleum Engineers, Esso Research and Engineering Company,
[7] Speight, J. G., The Chemistry and Technology of Petroleum, 3rd Linden, NJ, 1955.
ed., Marcel Dekker, New York, 1998. [28] Myers, H. S. and Fenske, M. R., "Measurement and Correlation
[8] Firoozabadi, A., Thermodynamics of Hydrocarbon Reservoirs, of Vapor Pressure Data for High-Boiling Hydrocarbons,"
Mc-Graw Hill, New York, 1999. Industrial and Engineering Chemistry, Vol. 47, 1955,
[9] Smith, R. L. and Watson, K. M., "Boiling Points and Critical pp. 1652-1658.
Properties of Hydrocarbon Mixtures," Industrial and [29] Riazi, M. R., Prediction of Thermophysical Properties of
Engineering Chemistry, Vol. 29, 1937, pp. 1408-1414. Petroleum Fractions, Doctoral Dissertation, Department of
150 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
Chemical Engineering, Pennsylvania State University, [50] Watson, K. M. and Nelson, E. F., "Improved Methods for
University Park, PA, 1979. Approximating Critical and Thermal Properties of Petroleum
[30] Van Nes, K. and Van Western, H. A., Aspects of the Constitution Fractions," Industrial and Engineering Chemistry, Vol. 25, 1933,
of Mineral Oils, Elsevier, New York, 1951. pp. 880-887.
[31] Riazi, M. R., "Distribution Model for Properties of [51] Watson, K. M., Nelson, E. E, and Murphy, G. B.,
Hydrocarbon-Plus Fractions," Industrial and Engineering "Characterization of Petroleum Fractions," Industrial and
Chemistry Research, Vol. 28, 1989, pp. 1731-1735. Engineering Chemistry, Vol. 27, 1935, pp. 1460-1464.
[32] Riazi, M. R. and Daubert, T. E., "Improved Characterization of [52] Kurtz, Jr., S. S. and Ward, A. L., "The Refractivity Intercept and
Wide Boiling Range Undefined Petroleum Fractions," Industrial the Specific Refraction Equation of Newton, I: Development of
and Engineering Chemistry Research, Vol. 26, 1987, pp. 629-632. the Refractivity Intercept and Composition with Specific
[33] Huang, E K. and Daubert, T. E., "Prediction of the Enthalpy of Refraction Equations," Journal of Franklin Institute, Vol. 222,
Petroleum Fractions: Pseudocompound Method," Industrial 1936, pp. 563-592.
and Engineering Chemistry, Process Design and Development, [53] Huang, P. K., Characterization and Thermodynamic Correlations
Vol. 13, No. 4, 1974, pp. 359-362. for Undefined Hydrocarbon Mixtures, Ph.D. Dissertation,
[34] Katinas, T. G., Prediction of the Viscosities of Petroleum Pennsylvania State University, University Park, PA, 1977.
Fractions, M. Sc. Thesis, Department of Chemical Engineering, [54] Fryback, M. G., "Synthetic Fuels: Promises and Problems,"
Pennsylvania State University, University Park, PA, Chemical Engineering Progress, Vol. 77, No. 5, 1981,
1977. pp. 39-43.
[35] Miqueu, C., Satherley, J., Mendiboure, B., Lachaise, J. and [55] Hersh, R. E., Fenske, M. R., Booser, E. R., and Koch, E. E,
Graciaa, A., "The Effect of P/N/A Distribution on the Parachors "Ring Analysis of Hydrocarbon Mixtures," Journal of Institute
of Petroleum Fractions," Fluid Phase Equilibria, Vol. 180, of Petroleum, Vol. 36, No. 322, 1950, pp. 624-668.
No.1/2, 2001, pp. 327-344. [56] API Research Project 42: Properties of Hydrocarbons of High
[36] Riazi, M. R. and Daubert, T. E., "Prediction of Molecular Type Molecular Weight, American Petroleum Institute, New York,
Analysis of Petroleum Fractions and Coal Liquids," Industrial 1966.
and Engineering Chemistry, Process Design and Development, [57] Goodger, E. and Vere, R., Aviation Fuels Technology, MacMillan,
Vol. 25, No. 4, 1986, pp. 1009-1015. London, 1985.
[37] Riazi, M. R. and A1-Sahhaf, T., "Physical Properties of n-Alkanes [58] Goossens, A. G., "Prediction of the Hydrogen Content of
and n-Alkyl Hydrocarbons: Application to Petroleum Mixtures," Petroleum Fractions," Industrial and Engineering Chemistry
Industrial and Engineering Chemistry Research, Vol. 34, 1995, Research, Vol. 36, 1997, pp. 2500-2504.
pp. 4145-4148. [59] Jenkins, G. I. and Walsh, R. E, "Quick Measure of Jet Fuel
[38] Riazi, M. R. and Daubert, T. E., "Simplify Property Predictions," Properties," Hydrocarbon Processing, Vol. 47, No. 5, 1968,
Hydrocarbon Processing, Vol. 59, No. 3, 1980, pp. 115-116. pp. 161-164.
[39] Lee, B. I. and Kesler, M. G., "A Generalized Thermodynamic [60] Fein, R. S., Wilson, H. I., and Sherman, J., "Net Heat of
Correlation Based on Three- Parameter Corresponding States," Combustion of Petroleum Fractions," Industrial and
American Institute of Chemical Engineers Journal, Vol. 21, 1975, Engineering Chemistry, Vol. 45, No. 3, 1953, pp. 610-614.
pp. 510-527. [61] Baird, C. T., Crude Oil Yields and Product Properties, Ch. De la
[40] Kesler, M. G. and Lee, B. I., "Improve Prediction of Enthalpy of Haute-Belotte 6, Cud Thomas Baird IV, 1222 Vezenaz, Geneva,
Fractions," Hydrocarbon Processing, Vol. 55, No. 3, March 1976, Switzerland, June, 1981.
pp. 153-158. [62] Riazi, M. R., Nasimi, N., and Roomi, Y., "Estimating Sulfur
[41] Winn, E W., "Physical Properties by Nomogram," Petroleum Content of Petroleum Products and Crude Oils," Industrial and
Refiners, Vol. 36, No. 21, 1957, pp. 157-159. Engineering Chemistry Research, Vol. 38, No. 11, 1999,
[42] Twu, C. H., "An Internally Consistent Correlation for Predicting pp. 4507-4512.
the Critical Properties and Molecular Weights of Petroleum [63] Gary, J. H. and Handwerk, G. E., Petroleum Refining, Technology
and Coal-Tar Liquids," Fluid Phase Equilbria, Vol. 16, 1984, and Economic, 3rd ed., Marcel Dekker, New York, 1994.
pp. 137-150. [64] Hebert, J., "Court Upholds Tougher Diesel Rules," Associated
[43] Goossens, A. G., "Prediction of Molecular Weight of Petroleum Press Writer, Yahoo News Website, May 3, 2002.
Fractions," Industrial and Engineering Chemistry Research, Vol. [65] Ball, S. J., Whisman, M. L., and Wenger, W. J., "Nitrogen
35, 1996, pp. 985-988. Content of Crude Petroleums," Industrial and Engineering
[44] Riazi, M. R. and Daubert, T. E., "Characterization Parameters Chemistry, Vol. 43, No. 11, 1951, pp. 2577-2581.
for Petroleum Fractions," Industrial and Engineering [66] Nelson, W. L., Petroleum Refinery Engineering, 4th ed.,
Chemistry Research, Vol. 26, 1987, pp. 755-759. (Corrections, McGraw-Hill, New York, 1969.
p. 1268.) [67] Riazi, M. R., Albahri, T. A., and Alqattan, A., "Prediction of Reid
[45] Riazi, M. R. and Roomi, Y., "Use of the Refractive Index in the Vapor Pressure of Petroleum Fuels," Petroleum Science and
Estimation of Thermophysical Properties of Hydrocarbons and Technology, Vol. 23, No. 1, 2005, pp. 1-12.
Their Mixtures," Industrial and Engineering Chemistry [68] Albahri, T., Riazi, M. R., and Alqattan, A., "Analysis of Quality
Research, Vol. 40, No. 8, 2001, pp. 1975-1984. of the Petroleum Fuels," Journal of Energy and Fuels, Vol. 17,
[46] Oil and Gas Journal Data Book, 2000 Edition, PennWell, Tulsa, No. 3, 2003, pp. 689-693.
OK, 2000. [69] Hatzioznnidis, I., Voutsas, E. C., Lois, E., and Tassios, D. P.,
[47] Riazi, M. R. and Daubert, T. E., "Prediction of the Composition "Measurement and Prediction of Reid Vapor Pressure of
of Petroleum Fractions," Industrial and Engineering Chemistry, Gasoline in the Presence of Additives," Journal of Chemical and
Process Design and Development, Vol. 19, No. 2, 1980, Engineering Data, Vol. 43, No. 3, 1998, pp. 386-392.
pp. 289-294. [70] Butler, R. M., Cooke, G. M., Lukk, G. G., and Jameson. B. G.,
[48] Ahmed, T., Hydrocarbon Phase Behavior, Gulf Publishing, "Prediction of Flash Points of Middle Distillates," Industrial
Houston, TX, 1989. and Engineering Chemistry, Vol. 48, No. 4, 1956, pp. 808- 812.
[49] Hill, J. B. and Coats, H. B., "The Viscosity Gravity Constant of [71] Walsh, R. E and Mortimer, J. V., "New Way to Test Product
Petroleum Lubricating Oils," Industrial and Engineering Quality," Hydrocarbon Processing, Vol. 50, No. 9, 1971,
Chemistry, Vol. 20, 1928, pp. 641-644. pp. 153-158.
3. C H A R A C T E R I Z A T I O N OF P E T R O L E U M F R A C T I O N S 151
[72] Lenoir, J. M., "Predict Flash Points Accurately," Hydrocarbon [82] Nelson, W. L., "Estimating Yields of Asphalt in Cmde
Processing, Vol. 54, No. 1, 1975, pp. 153-158. Oil-Again," Oil and Gas Journal, Vol. 66, No. 16, 1968,
[73] Riazi, M. R. and Daubert, T. E., "Predicting Flash and Pour p. 93.
Points," Hydrocarbon Processing, Vol. 66, No. 9, 1987, [83] Jenkins, G. I. and Walsh, R. E, "Quick Measure of Jet Fuel
pp. 81-83. Properties," Hydrocarbon Processing, Vol. 47, No. 5, 1968,
[74] Wickey, R. O. and Chittenden, D. I-I., "Flash Points of Blends pp. 161-164.
Correlated," Hydrocarbon Processing, Vol. 42, No. 6, 1963, [84] Riazi, M. R. and Roomi, Y. A., "Minimum Laboratory Data for
pp. 157-158. Properties of Petroleum Fractions and Crude Oils," Preprints of
[75] Hu, J. and Bums, A. M., "New Method Predicts Cloud, Pour Division of Petroleum Chemistry Symposia, American Chemical
and Flash Points," Hydrocarbon Processing, Vol. 49, No. 11, Society (ACS), Vol. 45, No. 4, 2000, pp. 570-573.
1970, pp. 213-216. [85] Standard Methods for Analysis and Testing of Petroleum and
[76] Hu, J. and Bums, A. M., "Index Predicts Cloud, Pour and Flash Related Products and British Standard 2000 Parts, IP 1-324
Points of Distillate Fuel Blends," Oil & Gas Journal, Vol. 68, Vol. 1, IP 325-448 Vol. 2, The Institute of Petroleum, London,
No. 45, 1970, pp. 66-69. 1999.
[77] Linden, H. R., "The Relationship of Physical Properties and [86] Tsonopoulos, C., Heidman, J. L., and Hwang, S.-C.,
Ultimate Analysis of Liquid Hydrocarbon," Oil & Gas Journal, Thermodynamic and Transport Properties of Coal Liquids, An
Vol. 48, No. 9, 1949, pp. 60-65. Exxon Monograph, Wiley, New York, 1986.
[78] Albahri, T. A., Riazi, M. R., and Qattan, A., "Octane Number [87] Knepper, J. I. and Hutton, R. E, "Blend for Lower Pour
and Aniline Point of Petroleum Fractions," Preprints, Division Point," Hydrocarbon Processing, Vol. 54, No. 9, 1975,
of Fuel Chemistry, 224 ACS National Meeting, Boston, August pp. 129-136.
18-22, 2002. [88] Auckland, M. H. T. and Charnock, D. J., "The Development
[79] Nelson, W. L., "Octane Numbers of Naphthas," Oil and Gas of Linear Blending indexes for Petroleum Products,"
Journal, Vol. 67, No. 25, 1969, p. 122, Journal of Institute of Petroleum, Vol. 55, No. 545, 1969,
[80] Jenkins, G. I., "Calculation of the Motor Octane Number from pp. 322-329.
the Research Octane Number," Journal of Institute of [89] Morris, W. E., "The Interaction Approach To Gasoline
Petroleum, Vol. 54, No. 529, 1968, pp. 14-18. Blending," Presented at the NPRA 73rd Annual Meeting, San
[81] Altgelt, K. H. and Boduszynski, M. M., Composition and Antonio, Texas, March 23-27, 1975.
Analysis of Heavy Petroleum Fractions, Marcel Dekker, New [90] Modern Petroleum Technology, 3rd ed., The Institute of
York, 1994. Petroleum, London, 1962.
MNL50-EB/Jan. 2005
Characterization of Reservoir
Fluids and Crude Oils
152
FIG. 4.2--A GC for measuring SCN distribution in crude oils and reservoir fluids (courtesy
of KISR [5]),
| ?w
~~]i
'!'~'
''"' i~"I "~ ' , r 0. r i~ a 0. ~ G 0 ~ 0 "
=1 ; 3 : 3 : 1 ~ - _ _ O , 0 . Q . ~ n O .
i t q i i ~ r I i
' I
Time, Min
FIG. 4.3--A sample of output from the GC of Fig. 4.2 for a Kuwait crude oil.
156 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 4.2--Extended compositional data for a light waxy crude oil and a Middle East crude.
Waxycrude oil Middle East crude oil
Component mol% M, g/tool wt% M Normalized
mol%
C2 0.0041 30 0.0076 30 0.0917
C3 0.0375 44 0.1208 44 0.9940
iC4 0.0752 58 0.0921 58 0.5749
NC4 0.1245 58 0.4341 58 2.7099
iC5 0.3270 72 0.4318 72 2.1714
NC5 0.2831 72 0.7384 72 3.7132
C6 0.3637 86 1.6943 82 7.4812
C7 3.2913 100 2.2346 95 8.5166
C8 8.2920 114 2.7519 107 9.3120
C9 10.6557 128 2.8330 121 8.4772
Clo 11.3986 142 2.8282 136 7.5294
Cll 10.1595 156 2.3846 149 5.7946
Ct2 8.7254 170 2.0684 163 4.5945
C13 8.5434 184 2.1589 176 4.4413
C14 6.7661 198 1.9437 191 3.6846
Cls 5.4968 212 1.9370 207 3.3881
C16 3.5481 226 1.5888 221 2.6030
C17 3.2366 240 1.5580 237 2.3802
C18 2.1652 254 1.5006 249 2.1820
C19 1.8098 268 1.5355 261 2.1301
C20 1.4525 282 1.5441 275 2.0330
C21 1.2406 296 1.1415 289 1.4301
C22 1.1081 310 1.4003 303 1.6733
C23 0.9890 324 0.9338 317 1.0666
C24 0.7886 338 1.0742 331 1.1750
C25 0.7625 352 1.0481 345 1.1000
C26 0.6506 366 0.9840 359 0.9924
C27 0.5625 380 0.8499 373 0.8250
C28 0.5203 394 0.9468 387 0.8858
C29 0.4891 408 0.8315 400 0.7527
C30 0.3918 422 0.8141 415 0.7103
C31 0.3173 436 0.7836 429 0.6613
C32 0.2598 450 0.7450 443 0.6089
C33 0.2251 464 0.7099 457 0.5624
C34 0.2029 478 0.6528 471 0.5018
C35 0.1570 492 0.6302 485 0.4705
C36 0.1461 506 0.5400 499 0.3918
C37 0.t230 520 0.5524 513 0.3899
C38 0.1093 534 0.5300 528 0.3634
C39 0.1007 548 0.4703 542 0.3142
C40+ 3.0994 700 C40:0.4942 C40:556 0.3217
C41+ 51.481
Source: waxy oil data from Ref. [7] and Middle East crude from Ref. [5]. For the Middle East crude the
GC output is shown in Fig. 4.3. Normalized mole% for the Middle East crude excludes C41+ fraction.
Based on calculated value of M41+ = 865, mole% of C41+ is 17.73%.
as its products from atmospheric or v a c u u m distillation p a r a m e t e r to quantify quality of a crude. API gravity of
columns. The Oil & Gas Journal Data Book has published a crudes varies from 10 to 50; however, most crudes have API
comprehensive set of data o n various crude oils from a r o u n d gravity between 20 a n d 45 [9]. A crude oil having API gravity
the world [8]. Characteristics of seven crude oils from a r o u n d greater t h a n 40 (SG < 0.825) is considered as light crude,
the world a n d their products are given i n Table 4.3. A crude while a crude with API gravity less t h a n 20 (SG > 0.934) is
assay dataset contains i n f o r m a t i o n on API gravity, sulfur a n d considered as a heavy oil. Crudes with API gravity between 20
metal contents, k i n e m a t i c viscosity, p o u r point, a n d Reid a n d 40 are called intermediate crudes. However, this division
vapor pressure (RVP). I n addition to boiling p o i n t range, API m a y vary from one source to a n o t h e r a n d usually there is
gravity, viscosity, sulfur content, PNA composition, a n d other no sharp division b e t w e e n light and heavy crude oils. Crude
characteristics of various products o b t a i n e d from each crude oils having API gravity of 10 or lower (SG > 1) are referred
are given. F r o m i n f o r m a t i o n given for various fractions, boil- as very heavy crudes a n d often have more t h a n 50 wt%
ing p o i n t curve can be obtained. residues. Some of Venezuelan crude oils are from this
Quality of crude oils are m a i n l y evaluated based o n higher category. Another p a r a m e t e r that characterizes quality of a
value for the API gravity (lower specific gravity), lower aro- crude oil is the total sulfur content. The total sulfur c o n t e n t
matic, sulfur, nitrogen a n d metal contents, lower p o u r point, is expressed i n wt% a n d it varies from less t h a n 0.1% to more
carbon-to-hydrogen (CH) weight ratio, viscosity, c a r b o n res- t h a n 5% [9]. Crude oils with total sulfur c o n t e n t of greater
idue, a n d salt a n d water contents. Higher API crudes usually t h a n 0.5% are t e r m e d as sour crudes while w h e n the sulfur
c o n t a i n higher a m o u n t of paraffins, lower CH weight ratio, c o n t e n t is less t h a n 0.5% they are referred as "sweet" crude
less sulfur a n d metals, a n d have lower c a r b o n residues a n d [9]. After sulfur content, lower n i t r o g e n a n d metal contents
viscosity. For this reason API gravity is used as the p r i m a r y signify quality of a crude oil.
TABLE 4.3--Crude assays data for selected oils [81.
Crude 1 Crude 2
B r e n t B l e n d , U.K. N o r t h S e a Pennington, Nigeria
Whole crude Range, ~ IBP-I50 Aniline point, ~ 70.6 Kinematic viscosity at 80 ~C, cSt: 31.87 CALM with export from floating Freeze point, ~ - 6 9
Density at 15~ kg/L: 0.8334 Yield, vol%: 23.8 Cetane index: 51 Kinematic viscosity at 100 ~C, cSt: 16.80 storage in offshore producing Aniline point, ~ 135
Gravity, ~ 38.3 Yield, wt%: 20.8 Color stability: Stable Kinematic viscosity at 120~ cSt: 9.5* area Cloud point, ~ < - 7 6
Sulfur, wt%: 0.40 Density at 15~ kg/L: 0.7279 Kinematic viscosity at 150~ cSt: 5.5*
Range, ~ 350-375 Crude Gas oil
Kinematic viscosity at 20~ cSt: 6.07 Sulfur, wt%: 0.0007 Pour point, ~ 36
Yield. vd%: 3.8 Gravity, ~ 36.6 Range, ~ 470-650
Kinematic viscosity at 30 ~C, cSt: 14.673 Mercaptan S, ppm: 21 Wax content, wt%: 14.6
Yield, wt%: 4.0 Sulfur, wt%:0.07 Yield, vol%: 31.7
Pour point, ~ - 4 2 Paraffins, wt%: 65.3 Total N, mg/kg: 2667
Density at 15~ kg/L: 0.8795 Viscosity cSt at 104~ 3.06 Gravity, ~ 33.7
Acidity, mg KOI-I/g: 0.10 Naphthenes, wt%: 23.8 Acidity, mg KOH/g: 0.18
Sulfur, wt%: 0.49 Pour point, ~ 43 Sulfur, wt%: 0.05
Micro carbon res., wt%: 2.13 Aromatics, wt%: 10.9 Micro carbon res., wt%: 5.0
Kinematic viscosity at 50~ cSt: 7.807 C4 & lighter, vol%: 1.5 Viscosity, cSt at 104~ 3.72
Asphaltenes, wt%: 0.45 n-Paraffins, wt%: 25.9 Asphaltenes, wt%: 0.95
Kinematic viscosity at 100~C, cSt: 2.714 RVP at 100~ psi: 5.1 Pour point, ~ - 10
V/Ni, ppm: 6/1 V/Ni, ppm: 12/1
Range, ~ 150-230 Cloud point, ~ 19 Con. carbon residue, wt%: 0.7 Aniline point, ~ 158
H 2 S, wt%: <0.0001 Xylene equivalent, vol%: 5-10+
Yield, vol%: 13.9 Pour point, ~ 15 V/Ni, ppm: 2/1 Cloud point, ~ 6.0
Salt (as NaCI), wt%: 0.015
Yield, wt%: 13.3 Wax content, wt%: 17.3 Range, ~ 550+
Water content, wt%: 0.38 Light naphtha Vacuum gas oil
Density at 15~ kg/L: 0.8001 Total N, mg/kg: 532 Yield, vol%: 12.2
Range, ~ C5-200 Range, ~ 650-1000
Light ends (CI-C5) Sulfur, wt%: 0.0048 Aniline point, ~ 80.5 Yield, wt%: 14.4
Yield, vol%: 7.4 Yield vol: 234
Yield, wt%: 5.87 Mercaptan S, ppm: 42 Cetane index: 49.5 Density at 15~ kg/L: 0.9876 Sulfur, wt%: 0.001 Gravity, API 23.5
Range, ~ IBP-95 Kinematic viscosity at 40 ~ cSt: 1.135 Sulfur, wt%: 1.21
Range, ~ 375-550 Paraffins, vol%: 70.5 Sulfur, wt%: 0.15
Yield, vd%: 12.3 Kinematic viscosity at 60~ eSt: 0.9408 Kinematic viscosity at 80~ cSt: 975.4
,q Yield, vd%: 20.7 Naphthenes, vol%: 25.5 Viscosity cSt at 122~ 28.9
Yield, wt%: 10.3 Acidity, mg KOH/g: 0.036 Kinematic viscosity at 100~C, cSt: 303.1
Yield, wt%: 22.5 Aromatics, vol%: 3.5 Pour point, ~ 102
Density at 15~ kg/L: 0.6924 Smoke point, rnm: 22 Kinematic viscosity at 120~ cSt: 120"
Density at 15~ kg/L: 0.9059 RON, clear: 74.9 Aniline point, ~ 184
Sulfur, wt%: 0.0006 Freeze point, • -61.5 Kinematic viscosity at 150 ~C, cSt: 40*
Sulfur, wt%: 0.61 Heavy naphtha Con. Carbon residue, wt%: 0.12
Mercaptan S, ppm: 51 Aniline point, ~ 55.0 Penetration at 25~ >400
Kinematic viscosity at 60~ cSt: 26.23 V/Ni, ppm: 2/<1
Paraffins, wt%: 82.8 Cetane index: 36.6 Softening point, ~ 31.8 Range, ~ 200--300
Kinematic viscosity at 80~ cSt: 12.69
Naphthenes, wt%: 11.9 Hydrogen, wt%: 13.45 Total N, mg/kg: 5242 Yield, vol%: 14.7 Residue
Kinematic viscosity at 100~C, cSt: 7.711
Aromatics, wt%: 5.3 Color stability: Stable Micro carbon res., wt%: 2.05 Sulfur, wt%: 0.001 Range, ~ 650+
Wax content, wt%: 20.7
n-Paraffins, wt%: 30.9 Naphthalenes, vol%: 2.30 Asphaltenes, wt%: 2.05 P/N/A, vol~ 35.0/52.5/12.0 Yield, vol%: 26.7
Total N, mg/kg: 1447
V/Ni, ppm: 35/3 RON, clear: 55.5 Gravity, ~ 21.8
Range, ~ 95-175 Range, ~ 230-350 Basic N, mg/kg: 354
Xylene equivalent, vol%: 5-10 Aniline point, ~ 118 Sulfur, wt%: 0.17
Yield, vol%: 16.7 Yield, vd%: 22.2 Acidity, mg KOH/g: 0.052
Kero~ne Viscosity cSt at 122~ 48.5
Yield, wt%: 15.4 Yield, wt% 22.5 Aniline point, ~ 90.5 Dewaxed oil
Range, ~ 34~470 Pour point, ~ 91
Density at 15~ kg/L; 0.7693 Density at 15~ kg/L: 0.8461 Refractive index at 60~ 1.4884 Density at 15~ kg/L: 0.9244
Yield, vol%: 18.0 Aniline point, ~ 179
Sulfur, wt%: 0.0012 Sulfur, wt%: 0.24 Kinematic viscosity at 40~ cSt: 104.2
Range, ~ 350+ Con. Carbon residue, wt%: 2.3
Mercaptan S, ppm: 14 Kinematic viscosity at 50~ cSt: 2.831 Kinematic viscosity at 60 ~C, cSt: 37.17 Gravity, ~ 40.6
Yield, vol%: 36.7
Paraffins, wt%: 44.9 Kinematic viscosity at 100~C, cSt: 1.339 Kinematic viscosity at 100~C, cSt: 9.539 Sulfur, wt%: 0.02
Yield, wt%: 40.9
Naphthenes, wt%: 36.7 Cloud point, ~ - 9 Viscosity index: 54 Viscosity cSt at 104~ 1.45
Density at 15~ kg/L: 0.9285
Aromatics, wt%: 18.4 Pour point, ~ - 9 Pour point, ~C: - 18 Smoke point, ram: 19.5
Sulfur, wt%: 0.86
n- Paraffins, wt%: 20.1 Wax content, wt%: 6.2 *Extrapolated
Kinematic viscosity at 50~ cSt: 126.8
Total N, mg/kg: 107 +Sample contained sediment
Kinematic viscosity at 60~C, cSt: 77.38
Acidity, mg KOI-Ug:0.028
TABLE 4.3--(Continued)
Crude 3 Crude 4 Crude 5
Arabian Light, Saudi Arabia Kuwait Export, Kuwait Fereidoon,/ran
Ras Tanura and Yanbu Range, ~ 650-1049 Crude Kerosine Kharg Island Diesel index, ~ 64.3
Yield, vol%: 29.74 Gravity, ~ 31.4 Range, ~ 310-520 Cetane index: 45.4
Whole crude Crude
SG, 60/60:0.9231 Viscosity, SUV, sec at 100~ 58.1 Yield, vol%: 18.43 Gravity, ~ 31.3 Freeze point, ~ - 4 1
SG, 60/60:0.8581 (33.4 ~ gravity) Sulfur, wt%: 2.46 Pour point, ~ +5 Gravity, ~ 45.9 Cloud point, ~ - 5 2
Sulfur, wt%: 1.77 Sulfur, wt%: 2.50
Nitrogen, wt%: 0.080 Sulfur, wt%: 2.52 Pour point, ~ - 4 5 RVP, psi: 5.7 Pour point, ~ - 5 2
Viscosity, cSt at 60~ 12.19 V/Ni, ppm: 0.07/0.05 Ni/V (calc.), ppm: 8/30 Sulfur, wt%: 0.39 Smoke point, m m : 22
Neut. n o , m u g : 0.00 H 2 S (evolved), wt%: 0.015
RVP, psi: 6.7 Aniline point, ~ 142.9 Viscosity, SUS at 68~ 32.4
Nitrogen, wt%: 0.09 Range, ~ 650-1500 Pour point, ~ - 3 5
Yield, vol%: 44,71 H2S , wt%: < 0.001 Cetane index: 53.8 Neut. no., m g KOFUg: 0.70 Gas Oil
Pour point, ~ - 6 5 C1-C4 (GC analysis), vol%: 3.03 Smoke point, ram: 26
SG, 60/60:0.9561 Rams. carbon, wt%: 5.4 Range, ~ 520-650
Range, ~ 68-347 Sulfur, wt%: 3.07 Heating oil Kinematic viscosity, cSt at Yield, volvo: 10.8
Naphtha
Yield, vol%: 22.73 Viscosi~ cSt at 140~ 102.8 Range, ~ 520-680 68~ 15.33 Gr~tvity, ~ 33.4
Range, ~ Over point-140
RON, clear: 48.7 Pour point, ~ 48 Yield, vol%: 14.89 V/Ni, ppm: 36/11 Sulfur, wt%: 1.56
Yield, vol%: 3.80
Sulfur, wt%: 0.0370 V/Ni, ppm: 27.1/6.7 Gravity, ~ 32.9 C I ~ 5, vol%: 6.35 Charac. factor: 11.74
Gravity, ~ 89.7
Iron, ppm: 5.5 Sulfur, wt%: 0.04 Viscosity, SUV;, sec. at 100~ 43.3 Naphtha Aniline point, ~ 161.2
Range, ~ 347-563
RON (micro), clear: 71.8 Pour point, ~ +20 Diesel index: 53.8
Yield, vol%: 22,72 Range, ~ 761-1500 Range, ~ C5-200
Yield, vol%: 35,50 RON + 3 mL TEL: 89.3 Total N, ppm: 100 ~eld, vol%: 9.9 Cetane index: 49.4
SG, 60/60:0.8131
SG, 60/60:0.9714 Sulfur, wt%: 1.98 Gravity, ~ 82.6 Cloud point, ~ 20
Sulfur, wt%: 0.440
Naphtha Aniline point, ~ 160.2 Sulfur, wt%: 0.017 Pour point, ~ 19
Aniline point, ~ 146.0 Abs. Viscosity, poise at 60~ 2.79
00 Range, ~ 140-170 Cetane index: 54.3 Charac. Factor: 12.72 Viscosit34, SUS at 68~ 48.0
Viscosity, cSt at 100~ 1.83 Viscosity, cSt at 212~ 45.93
Yield, vol%: 2.26 Gas oil Paraffins, vol%: 76.8 Residuum
Freeze point, ~ -21 Range, ~ 878-1500
Pour point, ~ - 2 4 Gravity, ~ 78.6 Range, ~ 680-1000 Naphthenes, vol%: 21.8 Range, ~ 650+
Yield, vol%: 26.80 Sulfur, wt%: 0.02
Smoke point, mm: 21.3 SG, 60/60:0.9889 Yield, vol%: 24.09 Aromatics, vol%: 1.4 Yield, vol%: 48.8
Aniline point, ~ 147.7 Gravity, ~ 21.4 RON/MON, clear: 64.3/62.6 Gravity, ~ 13.6
Range, ~ 563-650 Abs. Viscosity, poise at 600C: 11.37 RON, clear/+3 m L TEL: 58.2/80.40 Pour point, ~ +100 Naphtha Sulfur, wt%: 4.13
Yield, vol%: 8.32 Viscosity, cSt at 212~ 117.9 Paraffins, vol%: 88.5 Sulfur, wt%: 3.12 Range, ~ 200-320 Pour point, ~ 34
SG, 60/60:0.8599 Range, ~ 1049-1500 Monocycloparaffms, vol%: 9.9 Total N, ppm: 900 Rams. Carbon, wt%: 9.86
Sulfur wt%: 1.56 Aromatics, total, vol%: 1.6 Yield, vol%: 10.5
Yield, vol%: 14.97 Ni/V, ppm: <0.1/<0.1 Viscosity, SUS at 122~ 3616.0
Aniline point, ~ 165.4 Gravity, ~ 58.1
SG, 60/60:1.0217 Viscosit3;, SUV, sec at 130~ 175.2 Sulfur, wt%: 0.056 V/Ni, ppm: 66/20
Viscosity, cSt at 100~ 5.57 Heavy Naphtha
Abs. viscosity, poise at 60~ 330 Rams. carbon residue, wt%: 0.47 Charac. Factor: 12.02 Residuum
Freeze poinL ~ 39 Range, ~ 170-310
Viscosity, cSt at 212~ 1,222 Yield, vol%: 11.91 Residuum P/N/A, vol%: 71.6/17.2/11.3 Range, ~ 1050+
Pour point, ~ 36 Mina AI Ahmadi Gravity, ~ 62.1 Range (VEFV), ~ 1000+ RON/MON, clear: 30.2/32.0 Yield, vol%: 20.4
Sulfur, wt%: 0.03 Yield, volvo: 21.78 Kerosine Gravity, ~ 3.8
Aniline point, ~ 136.2 Gravity, ~ 5.8 Range, ~ 320-520 Sulfur, wt%: 5.6
Paraffins vo]%: 70.1 Sulfur, wt%: 5.11 Pour point, ~ 158
Yield, vol%: 17.7
Monocycloparaffins, vol%: 20.4 V/Ni, ppm: 116/34 V/Ni, ppm: 143/44
Gravity, ~ 44.3
Dicycloparaffirts, vol%: 0.4 Viscosity, SUV, sec at 250~ 3394 Sulfur, wt%: 0.435
Aromatics, total, vol%: 9.1 Rams. carbon residue, wt%: 21.9 Charac. factor: 11.84
Aniline point, ~ 145.2
TABLE 4.3--(Continued)
Crude 6 Crude 7
H o n d o Monterey, C a l i f o r n i a , U S A Boscan, Venenzuela
OST, Production Area Range, ~ 650-1049 Bajo Grande Gas o i l
Whole crude Yield, vol%: 25.02 Lake Maracaibo Range, ~ 400-500
Gravity, ~ 19.4 SG, 60/60:0.9453 Crude Yield, vol%: 3.96
SG, 60/60:0.9377 Sulfur, wt%: 4.61 Gravity, ~ 10.10 Gravity, ~ 34.0
Sulfur, wt%: 4.70 Nitrogen, wt%: 0.460 Sulfur, wt%: 5.5 Sulfur, wt%: 3.14
Viscosity, cSt at 60~ 1034 V/Ni, ppm: 0.05/0.07* Kinematic viscosity at 122~ 1.81
Viscosity, SUS at IO0~ 90000
Neut. no., mg/g: 0.43 Range, ~ 650-1500 Pour point, ~ +50 Pour point, ~ -77.53
Nitrogen, wt%: 0.65 Yield, vol%: 63.70 Smoke point, ~ 16.11
Light naphtha Aniline point, ~ 113.83
Pour point, ~ - 1 0 SG, 60/60:1.0082 Range, ~ 82-200 Cetane index: 37.24
Light ends streams (Cl-C5) Sulfur, wt%: 6.22
Yield, vol%: 0.55
Yield, vol%: 0.62 Viscosity, eSt at 140~ 17121 Gas Oil
Pour point, ~ 87 Heavy Naphtha Range, ~ 500-550
Range, ~ 68-347 V/Ni, ppm: 440/210 Range, ~ 200-300 Yield, vol%: 2.66
Yield, vol%: 15.12 iron, ppm: 48.0 Yield, vol%: 1.32 Gravity, ~ 28.62
RON, clear: 59.4 Sulfur, wt%:0.68 Sulfur, wt%: 4.02
Sulfur, wt%: 0.66 Range, ~ 761-1500 Paraffins, %: 53.28
Yield, vol%: 55.92 Gas Oil
Range, ~ 347-563 SG, 60/60:1.0198 Naphtha Range, ~ 550-650
Yield, vol%: 14.99 Penetration at 25~ mm: 238.2* Range, ~ 300-350 Yield, vol%: 6.46
SG, 60/60:0.8494 Abs. Viscosity, poise at 60~ 743 Yield, vol%: 0.88 Gravity, ~ 25.54
Sulfur, wt%: 3.14 "~rmcosity, cSt at 212~ 2028 Sulfur, wt%: 1.28 Sulfur, wt%: 4.47
Aniline point, ~ 120.0 Paraffins, %: 34.9 Pour point, ~ -10.9
Viscosity, cSt at 100~ 2.26 Range, ~ 878-1500
Yield, vo]%: 48.29 Kerosine Cetane no.: 38.0
Freeze point, ~F: - 3 0 Aniline point, ~ 117.3
SG, 60/60:1.0317 Range, ~ 350-400
Pour point, ~ - 3 8 Cetane index: 42.27
Penetration at 25~ ram: 35.6 Yield, vol%: 1.28
Smoke point, mm: 18.2
Abs. Viscosity, poise at 60~ 8029 Sulfur, wt%: 1.88 Residue
Range, ~ 563-650 Viscosity, cSt at 212~ 10145 Aromatics, %: 24.04 Range, ~ 650+
Yield, vo]%: 6.16 Pour point, ~ -141.0 Yield, vol%: 82.9
SG, 60/60:0.8911 Range, ~ 1049-1500 Smoke point, mm: 16.0 Sulfur, wt%: 5.86
Sulfur, wt%: 3.89 Yield, vol%: 38.68 Aniline point, oF: 112.69
SG, 60/60:1.0489 Gravity, ~ 6.10
Aniline point, ~ 136.8 Sulfur, wt%: 5.86
Viscosity, cSt at 100~ 7.83 Penetration at 25~ ram: 3.2*
Ni/V, ppm: 175/1407
Freeze point, ~ 37* Abs. Viscosity, poise at 60~ 650524
Kinematic Viscosity at 122~ 22,502
Pour point, ~ 29 Viscosity, cSt at 212~ 190077
* Extrapolated values.
160 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
4.2 GENERALIZED CORRELATIONS natural gas. However, for gas condensate systems simple cor-
FOR PSEUDOCRITICAL PROPERTIES relations in terms of specific gravity have been proposed in
OF NATURAL G A S E S A N D GAS the following forms similar to the above correlations and are
CONDENSATE SYSTEMS usually used by reservoir engineers [ 10]:
(4.3) Tpc -- 103.9 + 183.3SGg - 39.7SGg2
Natural gas is a mixture of light hydrocarbon gases rich in
methane. Methane content of natural gases is usually above (4.4) Ppc = 48.66 + 3.56SGg - 0.77SG~
75 % with C7+ fraction less than 1%. If mole fraction of H2 S in
These equations give higher critical temperature and lower
a natural gas is less than 4 x I0 -6 (4 ppm on gas volume basis)
critical pressure than do the equations for natural gases since
it is called "sweet" gas (Section 1.7.15). A sample composition
gas condensate samples contain heavier compounds. Because
of a natural gas is given in Table 1.2. Dry gases contain no
of the greater variation in carbon number, the equations for
C7+ and have more than 90 mol% methane. The main differ-
gas condensate systems are much less accurate than those
ence between natural gas and other reservoir fluids is that the
for natural gas systems. For this reason properties of gas con-
amount of C7+ or even C6+ in the mixture is quite low and the
densate systems may be estimated more accurately from the
main components are light paraffinic hydrocarbons. Proper-
distribution models presented in Section 4.5.
ties of pure light hydrocarbons are given in Tables 2.1 and 2.2.
Equations (4.1)-(4.4) proposed for pseudocritical proper-
The C6+ or C7+ fraction of a mixture should be treated as an
ties of natural gas and gas condensate systems are based
undefined fraction and its properties may be determined from
on the assumption that mixtures contain only hydrocarbon
the correlations given in Chapter 2. If the detailed composi-
compounds. However, these reservoir fluids generally contain
tion of a natural gas is known the best method of characteriza-
components such as carbon dioxide (CO2), hydrogen sulfide
tion is through Eq. (3.44) with composition in terms of mole
(H2S), or nitrogen (N2). Presence of such compounds affects
fraction (xi) for calculation of pseudocritical properties, acen-
the properties of the gas mixture. For such cases, correc-
tric factor, and molecular weight of the mixture. Although
tions are added to the calculated pseudocritical properties
the Key's mixing rule is not the most accurate mixing rule
from Eqs. (4.1)-(4.4). Corrections proposed by Wichert and
for pseudocritical properties of mixtures, but for natural gas
Aziz [12] and Carr et al. [13] are recommended for the effects
systems that mainly contain methane it can be used with rea-
of nonhydrocarbons on properties of natural gases [8]. The
sonable accuracy. More advanced mixing rules are discussed
method of Carr et al. for adjustment of calculated Tpc and Pvc
in Chapter 5. Once the basic characterization parameters for
is given as follows:
the mixture are determined various physical properties can
be estimated from appropriate methods. (4.5) Tpc = Tpc -44.44yco2 + 72.22yH2S -- 138.89yN2
The second approach is to consider the mixture as a single
pseudocomponent with known specific gravity. This method (4.6) Ppc = Ppc + 30.3369yco2 + 41.368yH2S -- 11.721yN2
is particularly useful when the exact composition of the mix- where T~c and Ppc are the adjusted (corrected) pseudocriti-
ture is not known. There are a number of empirical correla- cal temperature and pressure in kelvin and bar, respectively.
tions in the literature to estimate basic properties of natural Yco2, YH2Sand YN2 are the mole fractions of CO2, H2S, and N2,
gases from their specific gravity. Some of these methods are respectively. Tpc and Ppc are unadjusted pseudocritical tem-
summarized below. perature and pressure in kelvin and bar, respectively. These
In cases that the composition of a natural gas is unknown unadjusted properties may be calculated from Eqs. (4.1) and
Brown presented a simple graphical method to estimate pseu- (4.2) for a natural gas. The following example shows calcula-
docritical temperature and pressure from gas specific gravity tion of pseudocritical properties for a natural gas sample.
(SGg) as shown by Ahmed [10]. Standing [11] converted the
graphical methods into the following correlations for estima- Example 4.1--A natural gas has the following composition in
tion of Tpc and Ppc of natural gases free of CO2 and H2S: mol%: H2S 1.2%, N2 0.2%; CO2 1%, C1 90%, C2 4.8%, Ca 1.7%,
iC4 0.4, F/C40.5%, iC5 0.1, nC5 0.1%.
(4.1) Tpc = 93.3 § 180.6SGg - 6.94SG~
a. Calculate Tc, Pc, o9, and M using properties of pure com-
(4.2) Ppc = 46.66 § 1.03SGg - 2.58SG~ pounds.
b. Calculate the gas specific gravity.
where Too and Ppc are the pseudocritical temperature and
c. Calculate Tc and Pc using Eqs. (4.1) and (4.2) and SG cal-
pressure in kelvin and bar, respectively. SGg is defined in
culated from Part (b).
Eq. (2.6). This method is particularly useful when the exact
d. Adjust Tc and Pc for the effects of nonhydrocarbon com-
composition of the mixture is not available. This method pro-
pounds present in the gas.
vides acceptable results since nearly 90% of the mixture is
methane and the mixture is close to a pure component. There-
fore, assumption of a single pseudocomponent is quite rea-
Solution--Values of M, Tc, Pc, and co for pure components
present in the gas mixture can be obtained from Table 2.1.
sonable without significant difference with detailed composi-
These values as well as calculated values of M, Tc, Pc, and co
tional analysis. Application of these equations to wet gases is
for the mixture based on Eq. (3.44) are given in Table 4.4. The
less accurate.
calculated values of M, To, Pc, and w as shown in Table 4.4 are:
Another type of reservoir fluids that are in gaseous phase
M = 18.17, Tpc = -68.24~ Ppc = 46.74 bar, and w = 0.0234.
under reservoir conditions are gas condensate systems. Com-
This method should be used for gases with SGg > 0.75 [10].
position of a gas condensate sample is given in Table 1.2. Its
C7+ content is more than that of natural gases and it is about a. Equation (2.6) can be used to calculate gas specific gravity:
few percent, while its methane content is less than that of SGg = 18.17/28.96 = 0.6274.
4. C H A R A C T E R I Z A T I O N OF RESERVOIR FLUIDS AND CRUDE OILS 161
b. The system is a n a t u r a l gas so Eqs. (4.1) a n d (4.2) are u s e d 02for SCN groups f r o m C 6 to C45. W h i t s o n used R i a z i - D a u b e r t
to e s t i m a t e pseudocritical p r o p e r t i e s using SGg as the only correlations (Eqs. (2.38), (2.50), (2.63), and (2.64)) to generate
available i n p u t data. critical p r o p e r t i e s a n d m o l e c u l a r weight from Tb a n d SG. He
also u s e d E d m i s t e r m e t h o d (Eq. (2.108)) to generate values of
Tpc = 93.3 + 180.6 • 0.6274 - 6.9 x 0.62742
acentric factor. As discussed in C h a p t e r 2 (see Section 2.10),
= 203.9 K = -69.26~ these are not the best m e t h o d s for calculation o f p r o p e r t i e s of
Ppc = 46.66 + 1.03 x 0.6274 2.58 x 0.62742 = 46.46 bar. h y d r o c a r b o n s heavier t h a n C22 (M > 300). However, physical
p r o p e r t i e s r e p o r t e d b y Katz a n d F i r o o z a b a d i have b e e n u s e d
c. To calculate the effects of n o n h y d r o c a r b o n s p r e s e n t in in reservoir engineering calculations a n d b a s e d o n their tab-
the system Eqs. (4.5) a n d (4.6) are used to calculate ad- u l a t e d data, analytical correlations for calculation of M, To
j u s t e d p s e u d o c r i t i c a l properties. These equations r e q u i r e Pc, Vc, Tb, a n d SG of SCN groups from C6 to C45 in t e r m s of
m o l e fractions of H2S, CO2, a n d N2, w h i c h are given in Nc have b e e n developed [I0].
Table 4.4 as: Yrhs = 0.012, yco2 = 0.01, a n d YN2 = 0.002. Un- Riazi a n d M s a h h a f [16] r e p o r t e d a new set of d a t a on prop-
a d j u s t e d Tpc a n d Ppc are given in Part c as Tpc = -69.26~ erties of SCN from C6 to C50. They used boiling p o i n t a n d
a n d Ppc = - 4 6 . 4 6 bar. specific gravity d a t a for SCN groups p r o p o s e d b y Katz a n d
F i r o o z a b a d i from C6 to C22 to e s t i m a t e PNA c o m p o s i t i o n of
Tpc = - 6 9 . 2 6 - 44.44 • 0.01 + 72.22 x 0.012 each g r o u p using the m e t h o d s discussed in Section 3.5.1.2.
- 1 3 8 . 8 9 • 0.002 = -69.12~ Then Eq. (2.42) was used to generate physical p r o p e r t i e s of
P ~ = 46.46 + 30.3369 x 0.01 + 41.368 x 0.012 paraffinic, naphthenic, a n d a r o m a t i c groups. Properties of
SCN f r o m C6 to C22 have b e e n calculated t h r o u g h Eq. (3.39)
- 11.721 x 0.002 = 47.24 bar.
using the p s e u d o c o m p o n e n t a p p r o a c h . These d a t a have b e e n
Although use of Eqs. (4.1) a n d (4.2) t o g e t h e r with Eqs. (4.5) used to o b t a i n coefficients of Eq. (2.42) for p r o p e r t i e s of
a n d (4.6) gives reliable results for n a t u r a l gas systems, use SCN groups. The p s e u d o c o m p o n e n t m e t h o d p r o d u c e d boil-
of Eqs. (4.3) a n d (4.4) for gas c o n d e n s a t e systems or gases ing p o i n t s for SCN groups from C6 to C22 n e a r l y identical to
t h a t c o n t a i n C7+ fractions is not reliable. F o r such systems those r e p o r t e d b y Katz a n d F i r o o z a b a d i [ 14]. Development of
p r o p e r t i e s of C7+ fractions should be e s t i m a t e d a c c o r d i n g to Eq. (2.42) was discussed in C h a p t e r 2 a n d it is given as
the m e t h o d s discussed later in this chapter. r
(4.7) 0 = 0o~ - exp(a - b M c)
Values of critical properties and acentric factor for SCN correlations and dc is calculated from Eq. (2.98). Actual val-
groups greater than C30 estimated by different methods vary ues of w are probably greater than those given in this table.
significantly especially for higher SCN groups. Reliability of Values of Zc calculated from its definition by Eq. (2.8) for C60,
these values is subject to further research and no concrete C70, and Cs0 are 0.141, 0.132, and 0.125, respectively. ,
r e c o m m e n d a t i o n is given in the literature.
Values of M given in Table 4.6 are more consistent with
values of M r e c o m m e n d e d by Pedersen et al. [6] than those 4.4 CHARACTERIZATION APPROACHES
suggested by Whitson [15] as discussed by Riazi et al. [16]. FOR C7+ FRACTIONS
Molecular weights of SCN groups r e c o m m e n d e d by Whitson
[15] are based on the Katz-Firoozabadi values. However, In description of composition of a reservoir fluid, C6 is a very
more recently Whitson [ 17] r e c o m m e n d s values of M for SCN narrow boiling range fraction and characterization methods
groups, which are very close to those suggested by Riazi [16]. discussed in Chapter 2 (i.e., Eq. 2.38 [17]) can be used to esti-
Although he refers to Katz-Firrozabadi molecular weights, mate various properties of this group. Contrary to C6 group,
his reported values are m u c h higher than those given in Ref. the C7+ fraction has a very wide boiling range especially for
[14]. In addition the Watson characterization factor (Kw) re- crude oils. Therefore, methods of Chapter 2 or 3 cannot be di-
ported in Ref. [15] from C6 to C4s should be considered with rectly applied to a C7+ fraction. However, for a natural gas that
caution as they are nearly constant at 12 while toward heavier its C7+ fraction has a narrow boiling range and the a m o u n t
h y d r o c a r b o n factions as the a m o u n t of aromatics increases, of C7+ is quite small, equations such as Eq. (2.40) in terms
the Kw values should decrease. Values of properties of SCN of M and SG m a y be used to estimate various properties. In
groups given in Table 4.6 or those given in Ref. [15] are ap- some references there are specific correlations for properties
proximate as these properties m a y vary for one fluid mixture of C7+ fractions. For example, Pedersen et al. [6] suggested
to another. As shown in Table 4.1, values of M for the SCN use of the following relation for calculation of critical volume
of C7 in four different reservoir fluids vary from 92 to 100. of C7+ fractions in terms of MT+ and SG7+:
In Table 4.2, molecular weight of each SCN fraction is deter-
Vc7+ = 0.3456 + 2.4224 x 10-4M7+ - 0.443SG7+
mined by adding 14 to molecular weight of preceding SCN
group [7]. Thus M for C6 is specified as 86, which is deter- (4.11) + 1.131 x 10-3MT+SG7+
mined by adding 14 to molecular weight nC5 that is 72. The where Vc7+ is the critical volume of C7+ fraction in cma/gmol.
Pedersen molecular weight of SCN groups is given in terms Standing represented the graphical correlation of Katz for the
of Nc by the following relation [6]: pseudocritical temperature and pressure of C7+ fractions into
the following analytical correlations [18]:
(4.10) M = 14Nc - 4
To7+ = 338 + 202 x log(MT+ - 71.2)
In very few references the value of 2 is used instead of 4
in Eq. (4.10). To obtain properties of heavier SCN groups (4.12) +(1361 x log M7§ - 2111)logSG7+
(Nc > 50), Eq. (4.7) should be used with coefficients given in Po7+ = 81.91 - 29.7 x log(M7+ - 61.1) + (SG7§ - 0.8)
Table 4.5. However, to use this equation, it is necessary to
calculate the molecular weight from boiling point through (4.13) x [159.9 - 58.7 x log(MT+ - 53.7)]
reversed form of Eq. (4.7). Tb is calculated by Eq. (4.8) using where To7+ and Pc7+ are in K and bar, respectively. The origi-
the carbon number. The calculation method is demonstrated nal development of these correlations goes back to the early
in Example 4.2. 1940s and there is no information on the reliability of these
equations. Use of such relations to a C7+ fraction as a sin-
Example 4.2--Calculate Tb, SG, d20, n2o, To, Pc, Vc, a, and 8 gle p s e u d o c o m p o n e n t leads to serious errors especially for
for C60, C70, and C80 SCN groups. mixtures with considerable a m o u n t of C7+. Properties of C7+
fractions have significant effects on estimated properties of
S o l u t i o n - - T h e only data needed for calculation of physical the reservoir fluid even when they are present in small quan-
properties of SCN groups is the carbon number. For Nc = 60, tities [20, 21].
from Eq. (4.8): Tb = 1 0 9 0 - exp(6.9955 --0.11193 x 602/3) = Choru and Mansoori [21 ] have documented various meth-
894 K. Equation (4.7) in a reversed form can be used to ods of characterization of C7+ fractions. Generally there are
estimate M from Tb with coefficients given in Table 4.5: two techniques to characterize a h y d r o c a r b o n plus fraction:
M = [ ~ (6.97996 - ln(I080 - Tb)] 3/2. For Tb = 894 K we (i) The p s e u d o c o m p o n e n t and (ii) the continuous mixture ap-
get M = 844. This value of M should be used to calculate other proaches. In the p s e u d o c o m p o n e n t approach the C7+ is split
properties. For example SG is calculated as: SG = 1 . 0 7 - into a n u m b e r of subfractions with k n o w n mole fraction, Tb,
exp(3.56073 - 2 . 9 3 8 8 6 x 844 ~ -- 0.96. Similarly other prop- SG, and M [22-26]. In this m e t h o d the TBP curve can also
erties can be estimated and the results are given in Table 4.7. be used to split the mixture into a n u m b e r of pseudocompo-
Values of To Pc, and o~ are calculated from Lee-Kesler nents. Moreover, each subfraction m a y further be split into
TABLE 4.7--Calculated values for physical properties of C60, C70, and Ca0 of Example 4.2.
NC M Tb SG n2o d2o Tc Pc dc o~ a ?J
60 844 894.0 0.960 1.532 0.952 1010.5 5.7 0.410 1.699 30.1 8.4
70 989 927,0 0.969 1.538 0.961 1035.6 5.1 0.447 1.817 30.1 8.5
80 1134 953.0 0.977 1.542 0.969 1055.7 4.7 0.487 1.909 30.2 8.5
Units: Tb, Tc in K; Pc in bar; d2o and dc in g/cm3;a in dyne/cm;* in (cal/crn3)1/2.
164 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
three pseudocomponents from paraffinic, naphthenic, and plus fraction, it is important to know distribution of carbon
aromatic groups. Although a higher number of pseudocom- numbers to describe the mixture properly. A mathematical
ponents leads to more accurate results, the increase in the function that describes intensity of amount of a carbon num-
number of components complicates the calculations as the ber, or value of molecular weight, or boiling point for com-
number of input data required increases significantly. For pounds with Nc > 6 is referred as probability density function
example, the application of a two-parameter equation of state (PDF). The PDF can be obtained from a distribution function
(such as Peng-Robinson EOS) requires four input parame- that describes how various components or their properties
ters for each component: Tc, Pc, co, and a binary interaction are distributed in a mixture. In this section, general char-
coefficient (kii), which is a correction factor for a mixture of acteristics of density functions are discussed and then three
dissimilar components. The number of variables needed for different distribution models used to describe properties of
a 20-component mixture in two-parameter EOS calculations hydrocarbon-plus fractions are presented.
is 290! [27].
The second approach is the continuous mixture character-
ization method. In this method instead of mole fractions, a 4.5.1 G e n e r a l C h a r a c t e r i s t i c s
distribution function is introduced to describe the composi-
tion of many component mixtures [24, 25, 28-32]. Since com- Distribution functions can be applied to determine distribu-
position of a reservoir fluid up to C5 is given in terms of dis- tion of compounds from hexane or heavier in a reservoir fluid.
crete mole fractions, application of this approach to reservoir However, since the mole fraction of C6 fraction in reservoir
fluids is also referred as semicontinuous approach in which fluid is usually known and heavier hydrocarbons are grouped
the distribution function is applied to C6+ part of the mixture. in a C7+ group, distribution functions are generally used to de-
Distribution of components in mixtures that consist of many scribe properties of C7+ fractions. Mole fraction versus molec-
species is presented by a distribution function F(P) whose ular weight for SCN groups heavier than C6 in the West Texas
independent property P is defined in terms of a measurable gas condensate sample in Table 4.1 and the waxy and Kuwaiti
property such as molecular weight (M), boiling point (Tb), or crude oils of Table 4.2 are shown in Fig. 4.4. Such graphs are
carbon number (Nc) and varies from a value for the light- known as molar distribution for the hydrocarbon plus (in this
est component (Po) to the value for the heaviest component case C6+) fraction of reservoir fluids. As can be seen from this
(P0o) present in the mixture. Generally the value of Poo for M figure the molar distribution of gas condensates is usually
or Tb for a plus fraction is assumed as infinity (oo). Classical exponential while for the black oil or crude oil samples it is
thermodynamics for vapor-liquid equilibrium (VLE) calcula- left-skewed distribution.
tions of multicomponent systems require equality of temper- For the same three samples shown in Fig. 4.4, the proba-
ature, pressure, as well as equality of chemical potential of bility density functions (PDF) in terms of molecular weight
each component in both phases: are shown in Fig. 4.5. Functionality of molecular weight ver-
sus cumulative mole fraction, M(x), for the three samples is
(4.14) #L=/zV i=1,2 ..... N
shown in Fig. 4.6.
where #/L a n d / z v are chemical potential of component i in
liquid and vapor phase, respectively. Equation (4.14) should
be valid for all N components in the mixture. For VLE calcu-
Carbon Number, N c
lations of continuous mixtures Eq. (4.12) becomes
10 20 30 40
(4.15) / z L ( p ) = #V(p) Po < P < oo
15 i I r I i I r
fF(P)dP=fdx=l
oo 1
0 Po
where P is a property such as M, Tb, Nc, SG, or I (defined
where P (x) is the distribution function for property P in terms
by Eq. 2.36) and F is the probability density function. If the
of cumulative mole, weight, or volume fraction, Xc. For all
original distribution of P is in terms of cumulative mole frac-
the components whose parameters varies from/~ to P2 the
tion (Xcm),then xc in Eq. (4.16) is the cumulative mole fraction.
average value of property P, Pav(t,l__,e2), is determined as
As mentioned before, parameter P for a continuous mixture
varies from the initial value of Po to infinity. Therefore, for the f~2 PF(P)dP
whole continuous mixture (i.e., C7+), integration of Eq. (4.16) (4.21) Pav(Pl-~P2)- fle2 F(P)dP
This is shown in Fig. 4.7 where the total area under the curve
600 45
from Po to c~ is equal to unity (Eq. 4.17) and the area under
. . . . . W. TexasGas Condensate curve from P1 to Pa represents the fraction of components
- - Kuwaiti Crude / whose property P is greater than P1 but less than P2- Fur-
ther properties of distribution functions are discussed when
. 400 ~2
different models are introduced in the following sections.
///////
///////
I I / l l l l
/ / / / / / /
I / i i l / J
/ / / / / / J
F(P) ,MHH
/ / f f / f /
/ / / / / / /
/ / / / / / /
i i i i i r l
/ i / / l i /
/ / / f / / /
/ / / / / / /
I / / / / / /
l / r / / r /
l i t / i l l
f i l l / l /
I / / / / / /
l / r i l l 1
PO P1 ~ ~
P
h y d r o c a r b o n plus fraction into SCN groups by Eq. (4.22), the t h e n based on mole fraction of C7+ they are converted into
last h y d r o c a r b o n group is s h o w n by CN+ fraction (i.e., 4 0 + mole fractions in the original mixture. For this problem, N§ =
i n the waxy oil sample of Table 4.2). Molecular weight a n d i0+; therefore, mole fractions of C7, C8, a n d C9 should be esti-
specific gravity of the last fraction can be d e t e r m i n e d from mated. The results are x7 = 0.225, x8 = 0.174, a n d x9 = 0.134.
the following equations Mole fraction of Ct0+ is calculated from material balance
N as xl0+ = 1.0 - (0.225 + 0.174 + 0.134) = 0.467. The tool% of
(4.23) E x , Mn -----/1//7+ these c o m p o n e n t s i n the original mixture can be o b t a i n e d
n=7 t h r o u g h multiplying n o r m a l i z e d mole fractions by mol% of
C7+, 3.06; that is mol% of C7 = 0.225 x 3.06 = 0.69. Sum-
~ x.M~ mT+ m a r y of results is given i n Table 4.8. The M10+ a n d SG10+
(4.24) are calculated t h r o u g h Eqs. (4.23) a n d (4.24) using M, a n d
,=7 SG, SG7+
SG, from Table 4.5 as M10+ = [124 - (0.225 x 95 + 0.174 x
where MT+ a n d SG7+ are k n o w n i n f o r m a t i o n for the C7+ frac- 107 + 0.134 x 121)]/0.467 = 145.2. Similarly from Eq. (4.24),
tion. M. a n d SG, are molecular weight a n d specific gravity of M10+/SG10+ = 181.5. Therefore, SGl0+ = 145.2/181.5 = 0.80.
SCN group that m a y be taken from Table 4.6 or estimated C o m p a r i s o n with the actual values from Table 4.1 is presented
from Eq. (4.7) a n d coefficients given i n Table 4.5. Equa- also in Table 4.8. As is s h o w n in this table this m e t h o d does
tion (4.24) is i n fact equivalent to Eq. (3.45) w h e n it is ap- not provide a good estimate of SCN distribution. The errors
plied to SG. The following example shows application of this will be m u c h larger for oil systems. r
m e t h o d to generate SCN groups.
Yarborough [34] a n d Pedersen et al. [35] have suggested for
Example 4 . 3 - - U s e the M7+ a n d SG7+ for the North Sea gas gas condensate systems to assume a logarithmic d i s t r i b u t i o n
condensate mixture of Table 4.1 to generate the composition of the mole fraction x~ versus the c a r b o n n u m b e r CN as
of SCN groups up to C10+ a n d c o m p a r e with actual data.
(4.25) lnx,~ = A + B x CN
Solution--The C7+ has/147+ = 124, SG7+ = 0.7745, andx7+ = where A a n d B are constants specific for each mixture. This
0.0306. At first the mole fractions of SCN groups based o n e q u a t i o n is in fact similar to Eq. (4.22), except i n this case
n o r m a l i z e d c o m p o s i t i o n are calculated from Eq. (4.22) a n d the coefficients A a n d B are deterrnined for each mixture. If
TABLE 4.8--Prediction of SCN groups from Eq. (4.22) for a gas condensate system
of Example 4.3.
Actual data from Table 4.1 Predicted valuesfrom Eqs. (4.22)-(4.24)a
mol%b tool%b
CN Nor.b M SG Nor? M SG
7 0.80 26 95 0.7243 0.69 23 95 0.727
8 0.76 25 103 0.7476 0.53 17 107 0.749
9 0.47 15 116 0.7764 0.41 13 121 0.768
10+ 1.03 34 167 0.8120 1.43 47 145.2 0.800
~Valuesof M and SG for SCN groups are taken from Table4.6.
bValuesof mol%in the first column represent compositionin the whole originalfluid whilein the second
column under Nor. represent normalizedcomposition(~ = 100) for the C7+ fraction.
4. C H A R A C T E R I Z A T I O N OF R E S E R V O I R F L U I D S A N D C R U D E O I L S 167
TABLE 4.9~Prediction of SCN groups from Eq. (4.27) for a gas condensate system
of Example 4.4.
Actual data from Table 4.1 Predicted values from Eqs. (4.27)-(4.30) a
mol%b mol%b
CN Nor. b M SG NorP M SG
7 0.80 26 95 0.7243 0.95 31 95 0.727
8 0.76 25 103 0.7476 0.69 22 107 0.749
9 0.47 15 116 0.7764 0.45 15 121 0.768
10+ 1.03 34 167 0.8120 0.97 32 166 0.819
~Values of M and SG for SCN groups up to C50 are taken from Table 4.6. C10+ fraction represents SCN
~roups from 10 to 50.
Values of mola~ in the first column represent composition in the whole original fluid while in the second
column under Nor. represent normalized composition for the C7+ fraction.
Eq. (4.10) is c o m b i n e d with Eq. (4.25), an e q u a t i o n for m o l a r use values of Mn f r o m Table 4.6 w h i c h yields B = -0.0276.
d i s t r i b u t i o n of h y d r o c a r b o n - p l u s fractions can be o b t a i n e d as P a r a m e t e r A is calculated from Eq. (4.29) as: A = -4.2943.
Using p a r a m e t e r s A a n d B, n o r m a l i z e d m o l e fractions for
(4.26) lnx~ = A1 + Bi x M~ SCN from 7 to 50 are calculated from Eq. (4.27). Mole fraction
w h i c h m a y also be written in the following exponential form. of C10+ can be e s t i m a t e d from s u m of mole fractions of C10
to C50. M10+ a n d SG10+ are c a l c u l a t e d as in E x a m p l e 4.3 a n d
(4.27) x~ = a exp(B x M~) the s u m m a r y of results are given in Table 4.9. In this m e t h o d
w h e r e B = B1 a n d A = exp(A:). It should be n o t e d that A a n d M10+ a n d SG10+ are calculated as 166 a n d 0.819, w h i c h are
close to the actual values of 167 a n d 0.812. r
B in Eq. (2.27) are different f r o m A a n d B in Eq. (2.25). Para-
m e t e r s A1 a n d Bi can be o b t a i n e d by regression of d a t a be-
tween ln(x~) a n d MR such as those given in Table 4.2 for the As s h o w n in E x a m p l e 4.4, the exponential m o d e l w o r k s well
waxy oil. Most c o m m o n case is t h a t the detailed composi- for p r e d i c t i o n of SCN d i s t r i b u t i o n of s o m e gas c o n d e n s a t e
tional analysis of the m i x t u r e is n o t available a n d only M7+ systems, b u t generally shows w e a k p e r f o r m a n c e for crude oils
is known. F o r such cases coefficients A1 a n d BI (or A a n d B) a n d heavy reservoir fluids. As a n example for the N o r t h Sea
can be d e t e r m i n e d b y applying Eqs. (4.23) a n d (4.24) a s s u m - oil d e s c r i b e d in Table 4. i, b a s e d on the p r o c e d u r e d e s c r i b e d
ing t h a t the mixture contains h y d r o c a r b o n s to a certain g r o u p in E x a m p l e 4.4, M10+ is calculated as 248 versus actual value
(i.e., 45, 50, 60, 80, o r even higher) w i t h o u t a plus fraction. of 259. In this m e t h o d we have used u p to C50 a n d the coef-
This is d e m o n s t r a t e d in E x a m p l e 4.4. ficients of Eq. (4.27) are A = 0.2215, B = -0.0079. If we use
SCN groups up to C40, we get A = 0.1989 a n d B = - 0 . 0 0 7 3
Example 4 . 4 - - R e p e a t E x a m p l e 4.3 using the P e d e r s e n expo- with M10+ = 246.5, b u t if we include SCN h i g h e r t h a n C50
nential d i s t r i b u t i o n model, Eq. (4.27). slight i m p r o v e m e n t will be observed. The exponential distri-
b u t i o n m o d e l as expressed in t e r m s of Eq. (4.27) is in fact a
Solution--For this case the only i n f o r m a t i o n n e e d e d are discrete function, w h i c h gives m o l e fraction of SCN groups.
/147+ = 124 a n d x7+ = 0.0306. We calculate n o r m a l i z e d m o l e The c o n t i n u o u s f o r m of the exponential m o d e l will be s h o w n
fractions f r o m Eq. (4.27) after o b t a i n i n g coefficients A a n d B. later in this section.
Since the mixture is a gas c o n d e n s a t e we a s s u m e m a x i m u m
SCN g r o u p in the mixture is C50. Molecular weights of CN
f r o m 7 to 50 are given in Table 4.6. If Eq. (4.26) is a p p l i e d to 4.5.3 G a m m a Distribution Model
all SCN groups f r o m 7 to 50, since it is a s s u m e d there is no The g a m m a d i s t r i b u t i o n m o d e l has been used to express
N50+, for the whole C7+ we have m o l a r d i s t r i b u t i o n of w i d e r range of reservoir fluids including
50 50 black oils. Characteristics, specifications, a n d a p p l i c a t i o n of
(4.28) Z xn ----Z a exp(BM~) = 1 this d i s t r i b u t i o n m o d e l to m o l e c u l a r weight a n d boiling p o i n t
n=7 n=7 have been discussed by W h i t s o n in details [15, 17, 22, 23,
w h e r e x, is n o r m a l i z e d m o l e fraction of SCN g r o u p n in the 36, 37]. The PDF in t e r m s of m o l e c u l a r weight for this dis-
C7+ fraction. F r o m this equation A is f o u n d as t r i b u t i o n m o d e l as suggested b y W h i t s o n has the following
form:
(4.34) (
t / = 110 1 - 1 + 4.0zi3cr ~ ' ) 150
o
{
By substituting Eq. (4.31) into Eq. (4.18), cumulative mole
fraction, Xcm versus molecular weight, M can be obtained 100 t
which in terms of an infinite series is given as:
Carbon Number, CN following relation for the average molecular weight of the
0 20 40 60 80 SCN group n:
0.01 / t 1
M,v~, = - ( M
) T +X
nm
,- - ~ )\ e x p ( E _ 0
[aexp(bM)aM=
P
where for 0 < x < 1, A1 = -0.577191652, A2 = 0.988205891, (4.51) Xcm : (b)[exp(bM)- exp(br/)]
A3 - -0.897056937, A4 = 0.918206857, As = -0.756704078,
A6 = 0.482199394, A7 --- -0.193527818, and As =
For a SCN group n with lower and upper molecular weights
0.035868343. And for x > 1, the recurrence formula may be
of M~- and M~ and use of Eq. (4.19) we get mole fraction of
used:
the group, x,:
(4.45) F(x + 1) = xF(x)
(4.52) x~ = ( ; ) [ e x p ( b M ~ ) - exp(b/~-)]
where from Eq. (4.44), F(1) = 1 and thus from the above equa-
tion P(2) = 1. From Eqs. (4.39) and (4.40) we have
Equation (4.31) with ~ = 1 reduces to an exponential dis-
tribution form. From Eq. (4.33) with ot = 1, fl = M7+ - ~ and M, + M,,+I Mn-i + Mn
substituting these coefficients into Eq. (4.31) the following ~- 2 ~-= 2
density function can be obtained:
Now if we assume the difference between M, and M,-1 is a
1 M-~ constant number such as h we have Mn+l = M, + h and M,_I =
(4.46) F(M)=(M7+_O)exp(-~77+--) M ~ - h, thus M~+ = M~ + h/2 and M~- = M~ - h/2. A typical
value for h is usually 14. Substituting for M~ and M~- in
For a SCN group n, with molecular weight boundaries of Mff
Eq. (4.52) gives
and M+, substitution of Eq. (4.46) into Eq. (4.19) will result:
Xn = (;){exp[b(Mn + h / Z ) ] - exp[b(M~ - h/2)]}
xm,~=-exP(M7;_ 0)
(4.53) = ( ; ) [exp(~ff) - e x p ( - ~ ) ] exp(bM~)
(4.47) x [exp ( - M7+M+
e xrT)
p ( ---- - - M7+
M2- 7 ) ]
This equation can be written as
where x,~n is the mole fraction of SCN group n. Substitut-
ing Eq. (4.46) in Eq. (4.21) for molecular weight gives the (4.54) x~ = A exp(BMn)
170 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 4. I O~Prediction of molecular weightof SCN groups to SG can be estimated for each subfraction using their cor-
from Eq. (4.48)for Example 4.6. responding boiling point. As it will be shown, this approach
SCN, n M~ Mn- M~ x.... Mn, cMc. Mn.calc-Mn dos not provide an accurate distribution of specific gravity in
7 95 88 101 0.0314 94.5 --0.5 a wide and heavy hydrocarbon-plus fraction. As was shown in
8 107 101 114 0.0304 107.5 0.5
9 121 114 128.5 0.0328 121.2 0.2 Chapter 2, specific gravity is an important parameter in char-
10 136 128.5 142.5 0.0306 135.5 --0.5 acterization of petroleum fractions and errors in its value
11 149 142.5 156 0.0285 149.2 0.2 cause errors in estimation of physical properties of the sys-
12 163 156 169.5 0.0276 162.7 --0.3 tem. However, when these models are applied to very heavy
13 176 169.5 184 0.0286 176.7 0.7
14 191 184 199 0.0286 191.5 0.5 fractions especially for mixtures in which the density func-
15 207 199 214 0.0275 206.5 --0.5 tion F(M) sharply decreases for the heaviest components,
16 221 214 229 0.0265 221.5 0.5 their performance decreases [24, 25]. In fact these distribu-
17 237 229 243 0.0239 236.0 --1.0 tion functions are a m o n g m a n y standard PDF models that
18 249 243 255 0.0199 249.0 0.0 has been selected for application to petroleum mixtures for
19 261 255 268 0.0209 261.5 0.5
20 275 268 282 0.0217 275.0 0.0 expression of their molar distributions because of its math-
ematical convenience. For these reasons, Riazi attempted to
develop a general distribution model for various properties
and applicable to different types of petroleum mixtures espe-
Equation (4.54) is identical to Eq. (4.27) with parameters A
cially heavy oils and residues [24, 25].
and B defined in terms of parameters a and b in the exponen-
tial distribution model (Eq. (4.49)) as following: 4.5.4.1 Versatile Correlation
An extensive analysis was made on basic characterization pa-
(4.55) A = ( ; ) Iexp ( ~ ) - exp ( - ~ ) 1 rameters for C7+ fractions of wide range of gas condensate
systems and crude oils, light and heavy as well as n a r r o w and
B=b
wide petroleum fractions. Based on such analysis the follow-
where a and b are defined in terms of distribution parameters ing versatile equation was found to be the most suitable fit
by Eq. (4.50). , for various properties of more than 100 mixtures [24]:
m i x t u r e c o n t a i n s all c o m p o u n d s including extremely heavy for the value Po is always needed. F o r a C7+ fraction value of
c o m p o u n d s u p to M --~ oo. However, w h a t differs from one p r o p e r t y P for C7 o r C6 h y d r o c a r b o n group from Table 4.6
m i x t u r e to a n o t h e r is the a m o u n t of individual c o m p o n e n t s . m a y be u s e d as the initial guess. Although l i n e a r regression
F o r low a n d m e d i u m m o l e c u l a r weight range fractions t h a t can be p e r f o r m e d w i t h s p r e a d s h e e t s such as Excel o r Lotus,
do n o t c o n t a i n high m o l e c u l a r weight c o m p o u n d s , the m o d e l coefficients C1 a n d C2 in Eq. (4.57) c a n be d e t e r m i n e d by h a n d
expressed b y Eq. (4.56) a s s u m e s t h a t e x t r e m e l y heavy com- calculators using the following relation derived f r o m the least
p o u n d s do exist in the m i x t u r e b u t their a m o u n t is infinitely squares linear regression m e t h o d :
small, w h i c h in m a t h e m a t i c a l calculations do n o t affect mix- Zx, ~ -N~(Xi~)
ture properties. C2 =
W h e n sufficient d a t a on p r o p e r t y P versus c u m u l a t i v e mole, (4.60) ( ~ X~)z - N Z (X~/)
weight, o r v o l u m e fraction, Xo are available constants in ZY~ - c 2 z x~
Eq. (4.56) can be easily d e t e r m i n e d by converting the equa- C1 =
N
tion into the following l i n e a r form:
w h e r e e a c h s u m applies to all d a t a p o i n t s u s e d in the regres-
(4.57) Y = Ca + C z X sion a n d N is the total n u m b e r of points used. The least
squares l i n e a r regression m e t h o d is a s t a n d a r d m e t h o d for
where Y=lnP* and X=ln[ln(1/x*)]. By combining o b t a i n i n g the e q u a t i o n of a straight line, such as Eq. (4.57),
Eqs. (4.56) a n d (4.57) we have f r o m a set of d a t a on Xi a n d Y/.
1
B=-- E x a m p l e 4 . 7 - - T h e n o r m a l i z e d c o m p o s i t i o n of a C7+ fraction
(4.58) C2
derived f r o m a N o r t h Sea gas c o n d e n s a t e s a m p l e (GC) in
A = B exp(C1B) t e r m s of weight fractions of SCN groups u p to C17 is given
It is r e c o m m e n d e d t h a t for s a m p l e s with a m o u n t of residues in Table 4.11. M a n d SG of Cas+ fraction are 264 a n d 0.857,
(last h y d r o c a r b o n group) greater t h a n 30%, the residue d a t a respectively. F o r the whole C7+ fraction the M7+ a n d SG7+ are
should not be i n c l u d e d in the regression analysis to o b t a i n 118.9 a n d 0.7597, respectively. Obtain p a r a m e t e r s Po, A, a n d
the coefficients in Eq. (4.57). If a fixed value of B is u s e d for B in Eq. (4.56) for M, Tb, a n d SG a n d c o m p a r e calculated
a certain property, then only p a r a m e t e r Ca should be u s e d to values of these p r o p e r t i e s with d a t a s h o w n in Table 4.6.
o b t a i n coefficient A from Eq. (4.58).
To e s t i m a t e Po in Eq. (4.56), a t r i a l - a n d - e r r o r p r o c e d u r e S o l u t i o n - - F o r SCN groups f r o m C 7 to C17 values of M, Tb, a n d
can be used. By choosing a value for Po, w h i c h m u s t be lower SG can be taken from Table 4.6 a n d are given in Table 4.11.
t h a n the first d a t a p o i n t in the dataset, p a r a m e t e r s A a n d B An alternative to this table w o u l d be values r e c o m m e n d e d by
can be d e t e r m i n e d from liner regression of data. P a r a m e t e r W h i t s o n [ 16] for SCN g r o u p s less t h a n Cz5. Discrete m o l e frac-
Po can be d e t e r m i n e d b y m i n i m i z i n g the e r r o r function E(Po) tions, Xr~ can be calculated from discrete weight fractions, Xwi
equivalent to the r o o t m e a n squares (RMS) defined as a n d Mi b y a reversed f o r m of Eq. (1.15) as follows:
Xwi/ Mi
(4.61) Xmi -- N
(4.59) E(Po) = (p21 c _ pi~xp)2 Ei=mxwd Mi
i=1
w h e r e N is the total n u m b e r of c o m p o n e n t s (including the
w h e r e N is the total n u m b e r of d a t a p o i n t used in the regres- last plus fraction) a n d for this e x a m p l e it is 12. Discrete vol-
sion process a n d p/~al~is the c a l c u l a t e d value of p r o p e r t y P for u m e fractions x~/can be c a l c u l a t e d f r o m Xwi a n d SGi t h r o u g h
the s u b f r a c t i o n i f r o m Eq. (4.56) using e s t i m a t e d p a r a m e t e r s Eq. (1.16). Values of x ~ a n d x,~ are given in Table 4.11. To ob-
Po, A, a n d B. As a n alternative objective function, best value tain p a r a m e t e r s in Eq. (4.56), c u m u l a t i v e m o l e (xcm), weight
of Po can be o b t a i n e d b y m a x i m i z i n g the value of R 2 defined (Xcw), o r volume (XCv) fractions are needed. A s a m p l e calcu-
b y Eq. (2.136). W i t h s p r e a d s h e e t s such as Microsoft Excel, lation for the e s t i m a t i o n of m o l e c u l a r weight versus Xcm is
p a r a m e t e r Po c a n be directly e s t i m a t e d f r o m the Solver tool s h o w n here. A s i m i l a r a p p r o a c h c a n b e t a k e n to e s t i m a t e cu-
w i t h o u t t r i a l - a n d - e r r o r procedure. However, a n initial guess mulative weight o r v o l u m e fractions.
TABLE 4.11---Sample data on characteristics of a C7+ fraction for a gas condensate system in Example 4. 7.
Fraction Carbon
No. No. Xw M Tb, K SG Xm xv Xcm Xcw Xcv
1 7 0.261 95 365 0.727 0.321 0.273 0.161 0.130 0.137
2 8 0.254 107 390 0.749 0.278 0.259 0.460 0.388 0.403
3 9 0.183 121 416 0.768 0.176 0.181 0.687 0.607 0.622
4 10 0.140 136 440 0.782 0.121 0.137 0.836 0.768 0.781
5 11 0.010 149 461 0.793 0.008 0.009 0.900 0.843 0.854
6 12 0.046 163 482 0.804 0.033 0.043 0.920 0.871 0.880
7 13 0.042 176 500 0.815 0.028 0.040 0.951 0.915 0.922
8 14 0.024 191 520 0.826 0.015 0.022 0.972 0.948 0.953
9 15 0.015 207 539 0.836 0.009 0.014 0.984 0.967 0.971
10 16 0.009 221 556 0.843 0.005 0.008 0.990 0.979 0.982
11 17 0.007 237 573 0.851 0.003 0.006 0.994 0.987 0.988
12 18+ 0.010 264 -- 0.857 0.004 0.009 0.998 0.995 0.996
xw, xm, andxvare w~ght, m~e, and volume ~actions, respectivet~ V~uesofM, Tb, and SGaretaken ~omTab~ 4.6. xcm,Xcw,and
x~ arecumulative mole, weight, andvolume ~a~ionsc~culated from Eq.(4.62).
172 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 4.12--Determination of coefficients of Eq. (4.56) for molecular weight from data of Table 4.11.
Mo = 90, C1 = -1.1809, C2 = 1.0069, A = 0.3074, Mo = 91, C1 = -1.2674, C2 = 1,0606, A = 0.2854,
B = 0.9932, R2 = 0.998, RMS = 2.70, AAD= 1.32% B = 0.9429, Rz = 0.999, RMS = 2.139, AAD= 0.99%
95 0.839 -1.743 0.056 -2.89 94.8 0.0 0.2 0.044 -3.125 95.0 0.0 0.0
107 0.54 -0.484 0.189 -1.667 107.0 0.0 0.0 0.176 -1.738 106.3 0.4 0.6
121 0.313 0.15 0.344 -1.066 122.1 1.3 0.9 0.330 -1.110 121.1 0.0 0.0
136 0.164 0.591 0.511 -0.671 140.1 16.9 3.0 0.495 -0.704 139.0 8.8 2.2
149 0.1 0.833 0.656 -0.422 153.9 24.5 3.3 0.637 -0.450 153.0 16.1 2.7
163 0,08 0.927 0.811 -0.209 160.3 7.5 1.7 0.791 -0.234 159.5 12.4 2.2
176 0.049 1.101 0.956 -0.045 173.7 5.3 1.3 0.934 -0.068 173.3 7.1 1.5
191 0.028 1.273 1.122 0.115 189.6 2.0 0.7 1.099 0.094 189.9 1.2 0.6
207 0.016 1.413 1.3 0.262 204.6 5.8 1.2 1.275 0.243 205.6 1.8 0.7
221 0.01 1.53 1.456 0.375 219.0 4.0 0.9 1.429 0.357 220.9 0.0 0.1
237 0.006 1.634 1.633 0.491 233.2 14.3 1.6 1.604 0.473 236.0 1.0 0.4
264 0.002 1.814 1.933 0.659 261.6 5.6 0.9 1.901 0.642 266.4 5.9 0.9
AMi2 = (Mcalc- M~)2, %AD= Percent absoluterelativedeviation.
4. CHARACTERIZATION OF RESERVOIR FLUIDS AND CRUDE OILS 173
..= C.)
f z I T T Z r r
0 [ r I T T r l r P r 0.7
0 0.2 0.4 0.6 0.8
0.2 0.4 0.6 0.8 1
Cumulative Mole Fraction, Xcm Cumulative Weight Fraction, Xcm
FIG. 4.11--Prediction of molar distribution from FIG. 4.12--Prediction of boiling point and specific gravity dis-
Eq. (4.56) for the GC system of Example 4.7. tributions from Eq. (4.56) for the GC system of Example 4.7.
TABLE 4.13---Coefficients of Eq. (4.56) for M, Tb, and SG for data of Table 4.11.
Property Type of Xc Po A B RMS %AAD R2
M Mole 91 0.2854 0.9429 2.139 0.99 0.999
Tb Weight 350 (K) 0.1679 1.2586 3.794 0.62 0.998
SG Volume 0.705 0.0232 1.8110 0.004 0.32 0.997
SG Weight 0.705 0.0235 1.8248 0.004 0.33 0.997
Coefficients of Po and A with fixed value of B for each property
M Mole 89.86 0.3105 1 2.83 1.39 0.998
Tb Weight 340 (K) 0.1875 1.5 5.834 1.15 0.993
SG Volume 0.665 0.0132 3 0.005 0.54 0.984
SG Weight 0.6661 0.0132 3 0.005 0.53 0.985
174 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
4 10
X__J
r,r 6
2
a~
i
0.2 0.4 0.6 0.8
Tb* 0.1 0.2 0.3 0.4
SG*
0.012
14
12
0.009
10
0.006 --- 8
0.003
2
0 I A
F(p)=(~)x(~_~2)• [p_po.~S
) lexp[_B(P_po,~
Po .] J "~ 0.006
.=
(4.70)
in which F(1 + 1/B) can be evaluated directly from parame- where J is just an integration parameter defined in Eq. (4.79).
ter B. This equation was developed empirically for mathemat- As6 is the coefficient in Eq. (4.56) when SG is expressed
ical convenience. Values estimated from this equation vary by in terms of cumulative weight fraction, Xcw. For most sam-
a m a x i m u m of 0.02% (at B = 1) with those from Eq. (4.43). ples evaluated, parameter Asa is between 0.05 and 0.4; how-
Therefore, for simplicity we use Eq. (4.73) for calculation of ever, for no system a value greater than 0.4 was observed.
average values through Eq. (4.72). By combining Eqs. (4.78) and (4.79) with definition of SG* by
As mentioned earlier for m a n y systems fixed values of B Eq. (4.56), SG~ can be calculated from the following relation:
for different properties may be used. These values are BM = 1
for M, BT = 1.5 for Tb, and BsG = t31 = Ba = 3 for SG,/20 or SOa ( )SOo
d20. For these values of B, F(1 § 1/B) has been evaluated by
Eq. (4.73) and substituted in Eq. (4.72), which yields the fol- this equation should be used when SG is expressed in terms
lowing simplified relations for calculation of average proper- of Xcw by Eq. (4.56). For analytical integration of Eq. (4.78)
ties of whole C7+ fraction in terms of coefficient A for each see Problem 4.4.
property: In general, once P*~ is determined from Eq. (4.72), P~ can
be determined from the definition of P* by the following
(4.74) M~*v= AM relation:
(4.75) T~,~ = 0.689A~ 3 (4.81) Pay = Po(1 + Pay)
(4.76) SG*v = 0.619A~ Average properties determined by Eqs. (4.74)-(4.76) can be
converted to M~v, Tbav,and SGa~ by Eq. (4.81). Equation (4.76)
It should be noted that Eq. (4.76) can be used when SG derived for SG~, can also be used for refractive index para-
is expressed in terms of cumulative volume fraction. Equa- meter I or absolute density (d) when they are expressed in
tion (4.76) is based on Eq. (4.72), which has been derived terms ofx~. Similarly Eqs. (4.78)-(4.80) can be applied to/2o
from Eq. (4.71). As it was discussed in Chapter 3 (Section 3.4), or d20 when they are expressed in terms of Xcw. The following
for SG, d (absolute density), and I (defined by Eq. 2.36) two example shows application of these equations.
types of mixing rules may be used to calculate mixture prop-
erties. Linear Kay mixing rule in the form of Eq. (3.45) can Example 4.8--For the gas condensate system of Example 4.7
be used if composition of the mixture is expressed in volume calculate mixture molecular weight, boiling point, and spe-
fractions, but when composition if given in terms of weight cific gravity using the coefficients given in Table 4.13. The ex-
fractions, Eq. (4.46) must be used. Both equations give sim- perimental values are MT+ = 118.9 and SG7+ --0.7569 [24].
ilar accuracy; however, for mixtures defined in terms of very Also calculate the boiling point of the residue (component
few compounds that have SG values with great differences, no. 12 in Table 4.11).
Eq. (3.46) is superior to Eq. (3.45). Equation (4.46) can be
applied to SG in a continuous form as follows: Solution--For molecular weight the coefficients of PDF in
1 terms Of Xcmfor Eq. (4.66) as given in Table 4.13 are: Mo = 91,
1 f dxcw AM ----0.2854, and BM = 0.9429. From Eq. (4.73), F(I + 1/B) =
(4.77) SG~ - SG(xcw) 1.02733 and from Eq. (4.72), M*v = 0.2892. Finally May is cal-
0 culated from Eq. (4.81) as 117.3. For this system BM is very
where SG~ is the average specific gravity of C7+ and SG(x~w) is close to unity and we can use the coefficients in Table 4.13 for
the continuous distribution function for SG in terms of cumu- Mo = 89.86, AM = 0.3105, and BM = 1. From Eqs. (4.74) and
lative weight fraction. SG(xcw) can be expressed by Eq. (4.56). (4.81) we get May -- 89.86 x (1 + 0.3105) = 117.8. Comparing
Equation (4.77) in a dimensionless form in terms of SG* with the experimental value of 118.9, the relative deviation
becomes is - 1 % .
For specific gravity, the coefficients in terms of Xcv are:
oo
SGo -- 0.705, ASG = 0.0232, and Bsa -- 1.811. From Eq. (4.76),
(4.78) - - 1 - f F(SG*) dSG* SG~v = 0.0801 and from Eq. (4.81), SGav = 0.7615. Compar-
SG*v + 1 SG* + 1
0 ing with experimental value of 0.7597, the relative devia-
tion is 0.24%. If the coefficients in terms of Xcw are used,
In this equation integration is carried on the variable SG* SGo = 0.6661, AsG = 0.0132, and from Eq. (4.79) we get 1/J =
and F(SG*) is the PDF for SG* in terms of Xcw. Integration 1.1439 using appropriate range for As6. From Eq. (4.80),
in Eq. (4.78) has been evaluated numerically and has been SG~v = 0.7619 which is nearly the same as using cumulative
correlated to parameter Aso in the following form [24]: volume fraction.
oo For Tb the coefficients in terms Xcwwith fixed value of BT are
r F G* dSG* To = 340 K, AT = 0.1875, and Br = 1.5. From Eq. (4.75) and
J=J (S)SG*+I (4.81) we get: Ta~ = 416.7 K. To calculate Tb for the residue we
0
1 use the following relation:
= 1.3818 + 0.3503AsG - 0.1932A~G for AsG > 0.05
T a v __ N- 1
XwiTbi
~-~i= 1
1 (4.82) TbN =
= 1.25355 + 1.44886AsG -- 5.9777A2G + 0.02951 lnAsc XwN
where TbN is the boiling point of the residue. For this exam-
(4.79) for As6 < 0.05 ple from Table 4.11, N - - 1 2 and XwN = 0.01. Using values
4. C H A R A C T E R I Z A T I O N OF R E S E R V O I R FLUIDS A N D CRUDE OILS 177
1 = j F(SG*)SGodSG*
qi
SGoSG~v + SGo SGoSG* + SGo FIG. 4.18~lncomplete gamma function
0
I'(1 + lIB, @) for different values of B. Taken
which after simplification reduces to Eq. (4.78). r with permission from Ref. [40].
(4.85) * - - -zi
Pi,,v-
1/ P*F(P*)dP*
( A ) [(1+ ~Pi-1 { Pi-l'~) - (1 + ~-)
- ) "~ex pt--~-- p*'x
i ex / P*'x']
i
eL~ (4.91)
by substituting F(P*) from Eq. (4.66) and carrying the inte- where zi is obtained from Eq. (4.84) which for the case of
gration we get B = 1 becomes:
where In these relations, Pi* and P/*-I are the upper and lower bound-
aries of the subfraction i. One can see that if we set P/* = M~*
(4.87) qi = B pi,B and Pi*-i = M~-*, then Eq. (4.91) is equivalent to Eq. (4.48)
A for estimated molecular weight of SCN groups through the
zi should he calculated from Eq. (4.84). Pi,avis calculated from exponential model.
P,*
t , a v through Eq. (4.81) as
(4.88) Pi,~v = Po(1 + Pi*av) Example 4.10--For the C7+ fraction of Example 4.7, com-
position and molecular weight of SCN groups are given in
In Eq. (4.86), F(1 + I/B, qi) is the incomplete g a m m a function Table 4.11. Coefficients of Eq. (4.56) for the molar distribu-
defined as [38] tion of this system are given in Table 4.13 as Mo = 89.86,
oo
A --- 0.3105, and B = 1. Calculate average molecular weights
(4.89) F(a, q) = f t~-le-tdt of C~2-C13 group and its mole fraction. Compare calculated
q values from those given in Table 4.11.
1 7 8 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
Solution--In Table 4.11 for Ct2 and C13 the mole fractions are 15
89.9 = 0.7353 and P~* = 1047. Substituting in Eq. (4.84) we 500 1000 1500
get: z~ = 0.059.
For this system, A = 0.3501 and B = 1; therefore, from Molecular Weight, M
Eq. (4.87), qi-1 = 2.37048 and qi = 3.37401, which gives [39]
F(1 + 1/B, q~_~) = 0.3149 and F(1 + 1/B, q~) = 0.1498. Substi- FIG. 4.1~ Comparison of various distribution
tuting these values in Eq. (4.86) gives M~ = 0.8662 which models for molecular weight of a heavy petroleum
mixture. Taken with permission from Ref. [25].
yields Mn = 167.7. Therefore the predicted values for zi and
Mi,av for group of Clz-C13 are 0.059 and 167.7, respectively,
versus actual values of 0.061 and 169. r I~ Applying the g a m m a distribution model by Eq. (4.31)
gives Or = 349.9 K, ar = 1.6, and fir = 112.4 K. Use of these
coefficients in Eq. (4.31) for prediction of Tb distribution gives
4.5.4,5 Model Evaluations
average error of 1.6~ The exponential model (Eq. (4.56) with
The distribution model expressed by Eq. (4.56) can be used for B = 1) gives an average error of 4~ For this mixture with
M, Tb, SG, d, and refractive index parameter I. The exponen- intermediate molecular weight, the generalized and g a m m a
tia] model expressed by Eq. (4.27) or Eq. (4.31) with a = 1 can distribution models both are predicting boiling point with a
only be used for molecular weight of light otis and gas conden- good accuracy. However, the exponential model is the least
sate systems. The g a m m a distribution model can be applied accurate model for the boiling point distribution since Br in
to both M and Tb, but for SG, the method of constant Watson Eq. (4.56) is greater than unity.
K is recommended by Whitson [20]. In this method Kw for Distribution of specific gravity for the C7+ fraction of a black
the whole C7+ is calculated from its M7+ and SG7+ (Eq. 2.133) oil system from Ekofisk field of North Sea fields is shown in
and it is assumed to be constant for all components. For each Fig. 4.21. The generalized model, exponential model, and the
component, SG is calculated from Eq. (2.133) using the Kw
of the mixture and M for the component.
As mentioned earlier the main advantage of generalized 700
model is its capability to predict distribution of properties
of heavy oils. This is demonstrated in Fig. 4.19, for molecu- o Experimental
lar weight distribution of a heavy residue [25]. Experimental - - GeneralizedModel
data on the molar distribution are taken from Rodgers et al. 600 . . . . . . . GammaDist. Model /
[41 ]. The experimentally determined mixture weight averaged
molecular weight is 630 [41 ]. For this sample, parameters Mo,
Am, and Bra for Eq. (4.56) in terms of cumulative weight frac-
tion are calculated as 144, 71.64, and 2.5, respectively. For 500
O
this heavy oil sample both Mo and parameter B are higher
than their typical values for oil mixtures. Predicted mixture ~0
TABLE 4.16---Prediction of distribution of refractive index of a C6+ fraction from Eq. (4.56).
Nc Vol% FiE0 Xcv I I, pred. n2o,pred. %AD
6 2.50 1.3866 0.013 0.2352 0.2357 1.3875 0.06
7 5.47 1.4102 0.056 0.2479 0.2467 1.4080 0.16
8 4.53 1.4191 0.109 0.2526 0.2542 1.4222 0,22
9 5.06 1.4327 0.161 0.2597 0.2596 1.4324 0.02
10 2.55 1.4407 0.201 0.2639 0.2632 1.4393 0.09
11 3.62 1.4389 0.234 0.2630 0.2659 t.4445 0.39
I2 3.70 1.4472 0.274 0.2673 0.2688 1.4502 0.20
13 4.19 1.4556 0.316 0.2716 0.2718 1.4560 0.03
14 3.73 1,4615 0.358 0.2747 0.2747 1.4615 0.00
15 3.96 1.4694 0.399 0.2787 0.2774 1.4668 0.18
16 3.03 1.4737 0.437 0.2809 0.2798 1.4715 0.15
17 3.40 1.4745 0.471 0.2813 0.2819 1.4758 0.09
18 3.13 1.4755 0.506 0.2818 0.2842 1.4802 0.31
19 2.94 1.4808 0.538 0.2845 0.2862 1.4842 0.23
20+ 41.70 1.5224 0.777 0.3052 0.3031 1.5182 0.28
Mixture 93.51 0.14
tExperimentaldata on n20 are taken from Berge [42].
m e t h o d can be used to predict d i s t r i b u t i o n of refractive index predicted with B -- 3 with a n accuracy of 0.2% as s h o w n i n
of a C6+ fraction. Table 4.14. As discussed in Chapter 2, p a r a m e t e r I is a size pa-
r a m e t e r similar to density or specific gravity a n d therefore the
Example 4.11--For a C6+ of a n oil sample experimental data average for a mixture should be calculated t h r o u g h Eq. (4.76)
o n refractive index at 20~ are given versus vol% of SCN or (4.80). #
groups from C6 to C20+ i n Table 4.16. Refractive index of the
whole fraction is 1.483. Use Eq. (4.56) to predict refractive
1.6
index d i s t r i b u t i o n a n d o b t a i n the AAD% for the model pre-
diction. Also graphically c o m p a r e the model prediction with o Experimental
the experimental data a n d calculate the mixture refractive
-- Predicted
index.
O
4.5.4.6 Prediction o f Property Distributions This equation can be used for narrow boiling range fractions
Using Bulk Properties with M between 70 and 350. In this molecular weight range
As discussed above, Eq. (4.56) can be used as a two-parameter this equation is slightly more accurate than Eq. (2.117). Once
relation with fixed values of B for each property (BM = 1, distribution of I is determined from these equations, if the
Br = 1.5, and Bsa = Bz = 3). In this case Eq. (4.56) is referred initial values of Mn and SGo are correct then 1"v calculated
as a two-parameter distribution model. In such cases only from the distribution coefficients and Eq. (4.72) should be
parameters Po and A must be known for each property to ex- close to the experimental value obtained from n7+. For cases
press its distribution in a hydrocarbon plus fraction. The two- that experimental data on n7+ is not available it can be esti-
parameter model is sufficient to express property distribu- mated from M7+ and SG7+ using Eq. (4.95) or (2.117). Equa-
tion of light oils and gas condensate systems. For very heavy tion (2.117) estimates values of nT+ for 48 systems [23] with
oils two-parameter model can be used as the initial guess to an average error of 0.4%. Steps to predict M, Tb, SG, and I
begin calculations for determination of the three parameters (or n) distributions can be summarized as follows [23]:
in Eq. (4.56). In some cases detailed composition of a C7+
fraction in a reservoir fluid is not available and the only infor-
mation known are MT+ and SG7+, while in some other cases
in addition to these properties, a third parameter such as re- 1. Read values of M7+, SG7+, and 17+ for a given C7+ sample.
fractive index of the mixture or the true boiling point (TBP) If 17+ is not available Eq. (2.117) may be used to estimate
curve are also known. For these two scenarios we show how this parameter.
parameters Po and A can be determined for M, Tb, SG, and 2. Guess an initial value for Mo (assume Mo = 72) and cal-
/20. culate M*v from Eq. (4.93).
Method A: MT+, SGT+, and nT+ are known--Three bulk 3. Calculate AM from Eq. (4.72) or Eq. (4.74) when B = 1.
properties are the minimum data that are required to predict 4. Choose 20 (or more) arbitrary cuts for the mixture with
complete distribution of various properties [24, 43]. In ad- equal mole fractions (Xmi)of 0.05 (or less). Then calculate
dition to/147+ and SGT+, refractive index, n7+, can be easily Mi for each cut from Eq. (4.56).
measured and they are known for some 48 C7+ fractions [24]. 5. Convert mole fractions (x~/) to weight fractions (x~)
As shown by Eqs. (4.74)-(4.76), if P*v is known, parameter A through Eq. (1.15) using Mi from step 4.
can be determined for each property. For example, if Mo is 6. Guess an initial value for So (assume So = 0.59 as a starting
known, Ma*v can be determined from definition of M* as: value).
7. Calculate 1/J from Eq. (4.80) using SGo and SG7+. Then
May- Mo calculate As6 from Eq. (4.79) using Newton's method.
(4.93) M~*v- /14o 8. Using Eq. (4.56) with AsG and So from steps 6 and 7 and
B = 3, SG distribution in terms of Xcw is determined and
where Mav is the mixture molecular weight of the C7+ fraction, for each cut SGi is calculated.
which is known from experimental measurement. Similarly, 9. Convert Xwi to Xvi using Eq. (1.16) and SG/ values from
SG*v and I~*~can be determined from $7+ and n7+ (or I7+). step 8.
Parameters AM, AsG, and AI are then calculated from Eqs. i0. For each cut calculate Tbi from M/ and SGi through
(4.72) and (4.81). For fixed values of B, Eqs. (4.74)-(4.76) Eq. (2.56) or (2.57).
and (4.79) and (4.80) may be used. One should realize that 11. For each cut calculate // from Mi and SGi through
Eq. (4.74) was developed based on cumulative mole fraction, Eq. (2.95).
while Eqs. (4.79) and (4.80) are based on cumulative weight 12. From distribution of I versus Xcv find parameters Io, At
fraction. Once distribution of M and SG are known, distribu- and BI through Eqs. (4.56)-(4.57). Then calculate Iav from
tion of Tb can be determined using equations given in Chapter Eq. (4.72) and (4.81).
2, such as Eqs. (2.56) or (2.57), for estimation of Tb from M 13. Calculate el = [(Iav,calc.- I7+)/I7+[.
and SG. Based on data for 48 C7+ samples, the following re- 14. If el < 0.005, continue from step 15, otherwise go back
lation has been developed to estimate Io from Mo and SGo to step 6 with SGo,new = SGo,old+ 0.005 and repeat steps
[23]: 7-13.
15. Calculate Io from Eq. (4.94).
/o ~ 0.7454exp(-0.01151Mo - 2.37842SGo 16. Calculate 82 = [(lo,step15 -- Io,step12)/Io.step15[9
(4.94) + 0.01225 MoSGo)M~ SG~"53147 17. GO back to step 2 with a new guess for Mo (higher than the
previous guess). Repeat steps 2-16 until either e2 < 0.005
This equation can reproduce values of Io with an average or 82 becomes minimum.
deviation of 0.3%. Furthermore, methods of estimation of 18. For heavy oils large value of 82 may be obtained, because
parameter I from either Tb and SG or M and SG are given in value of BM is greater than 1. For such cases values of
Section 2.6.2 by Eqs. (2.115)-(2.117). Equation (2.117) may BM = 1.5, 2.0, and 2.5 should be tried successively and
be applied to the molecular weight range of 70-700. However, calculations from step 2 to 17 should be repeated to min-
a more accurate relation for prediction of parameter I from imize 82.
M and SG is Eq. (2.40) with coefficients from Table 2.5 as 19. Using data for Tb versus Xcw,determine parameters To, At,
follows: and Br from Eqs. (4.56) and (4.57).
20. Print Mo, AM, BM, SGo, AsG, To, At, Br, Io, AI, and Bz.
I = 0.12399 exp(3.4622 x 10-*M + 0.90389SG 21. Generate distributions for M, Tb, SG, and n20 from
(4.95) -6.0955 • 10-4MSG)M~176 0'22423 Eq. (4.56).
182 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 4.17--Sample calculations for prediction of distribution of properties of the C7+ fraction
in Example 4.12.
Xm Xcm M Xw Xcw SG x~ Xc~ Tb,K 1
No. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1 0.05 0.025 90.7 0.038 0.019 0.719 0.002 0.001 353.9 0.242
2 0.05 0.075 92.3 0.039 0.058 0.727 0.007 0.006 357.9 0.245
3 0.05 0.125 93.9 0.040 0.097 0.732 0.012 0.015 361.6 0.246
I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3. Choose 20 (or m o r e ) a r b i t r a r y cuts for the m i x t u r e with To = 350 K, Ar = 0.1679, a n d Br = 1.2586. An initial guess
equal weight (or volume) fractions of 0.05 (or less). Then value of SGo = 0.7 is u s e d to calculate ASG a n d SG distribu-
d e t e r m i n e Tbi for each cut f r o m Eq. (4.56) a n d coefficients tion coefficients. N o w we divide the whole fraction into 20
from step 2. cuts w i t h equal weight fractions as x ~ = 0.05. S i m i l a r to cal-
4. Guess a n initial value for SGo (lowest value is 0.59). culations s h o w n in Table 4.17, Xcw is calculated a n d t h e n for
5. Calculate 1/J f r o m Eq. (4.80) using SGo a n d SG7+. Then each cut values of Tbi a n d SGi are calculated. F r o m these two
calculate AsG f r o m Eq. (4.79) using Newton's m e t h o d o r p a r a m e t e r s Mi a n d Ii are c a l c u l a t e d b y Eqs. (2.56) a n d (2.115),
o t h e r a p p r o p r i a t e procedures. If original TBP is in t e r m s respectively. F r o m Xwi a n d Mi m o l e fractions ( X m i ) a r e calcu-
of X~v, then Eq. (4.76) should be u s e d to d e t e r m i n e ASG in lated. S a m p l e calculation for the first few points is given in
t e r m s of x~. Table 4.19 w h e r e calculation continues u p to i = 20. The co-
6. If original TBP is in t e r m s of xov, find SG d i s t r i b u t i o n f r o m efficients o f Eq. (4.56) d e t e r m i n e d from d a t a in Table 4.19
Eq. (4.56) in t e r m s ofx~v. Then use SG to convert xv to xw a r e given in Table 4.20. In this m e t h o d p a r a m e t e r el = 0.0018
t h r o u g h Eq. (1.16). (step 13), w h i c h is less t h a n 0.005 a n d there is no n e e d to
7. Using values of SG a n d Tb for each cut d e t e r m i n e values re-guess SGo. In this set of calculations since initial guess for
of M from Eq. (2.56). SGo is the s a m e as the actual value only one r o u n d of cal-
8. Use values of M f r o m step 7 to convert Xw into Xr, t h r o u g h culations was needed. Coefficients given in Table 4.20 have
Eq. (1.15). b e e n u s e d to generate d i s t r i b u t i o n for various p r o p e r t i e s a n d
9. F r o m d a t a c a l c u l a t e d in step 8, find m o l a r d i s t r i b u t i o n b y they are c o m p a r e d with p r e d i c t e d values f r o m M e t h o d A as
e s t i m a t i n g coefficients Mo, AM, a n d BM in Eq. (2.56). well as actual values given in Table 4.11. Results are s h o w n in
10. Calculate value of I for each cut from Tb a n d SG t h r o u g h Figs. 4.23 a n d 4.24 for p r e d i c t i o n of M a n d Tb distributions.
Eq. (2.115) o r (2.116). M e t h o d s A a n d B predict s i m i l a r d i s t r i b u t i o n curves m a i n l y
11. F i n d coefficients Io, AI, a n d BI (set Bt = 3) f r o m d a t a ob-
t a i n e d in step 10. S e t / 1 = Io.
TABLE 4.20---Estimated coefficients of Eq. (4.56)
12. F r o m SGo a s s u m e d in step 4 a n d Mo d e t e r m i n e d in step 9, for the C7+ of Example 4.13.
e s t i m a t e Io t h r o u g h Eq. (4.94). Set 12 = Io. Property Po A B Type of Xc
13. Calculate el = 1(12 - 11)/111. M 90 0.3324 1.096 Xcm
14. If el _> 0.005 go b a c k to step 4 b y guessing a new value for Tb 350 0.1679 1.2586 Xcw
SGo. If el < 0.005 o r it is m i n i m u m go to step 15. SG 0.7 0.0029 3.0 Xcw
15. Print Mo, AM, BM, SGo, As6, To, At, By, Io, At, a n d Bz. I 0.236 7.4 x 10-4 3.6035 Xcv
16. G e n e r a t e d i s t r i b u t i o n s for M, Tb, SG, a n d n20 from
Eq. (4.56). 300
i
Example 4.13--For the C7+ of E x a m p l e 4.7, a s s u m e t h a t ~0
.,..4
0)
200
Tb distillation curve is available, as given in c o l u m n s 3 a n d
5 in Table 4.11. I n a d d i t i o n a s s u m e t h a t for this s a m p l e
M7+ = 118.9 a n d SG7+ -- 0.7597 are also available. Using the
m e t h o d d e s c r i b e d above (Method B) generate M, Tb, SG, a n d
I distributions. G r a p h i c a l l y c o m p a r e p r e d i c t i o n of various
d i s t r i b u t i o n s by Methods A a n d B with actual d a t a given in
Table 4.11.
(9
o
150
100 o--------"
J
f
Solution--Similar to E x a m p l e 4.12, step-by-step p r o c e d u r e 50 i i i i i i I i i
TABLE 4.22---Generation of pseudocomponents from Gaussain Eq. (4.92) and molecular weights from Eq. (4.91). Step-by-
quadrature method for the C7+ sample in Example 4.14. step calculations for both of these methods with an example
Yi wi Zi Mi zi A/~ are given in the next section.
1 0.416 0.711 0.711 103.6 73.7
2 2.294 0.279 0.279 154.6 43.1
3 6.290 0.010 0.010 252.2 2.6 4.6.2 Lumping Scheme
Mixture ... 1.000 1.000 ... 119.4
The lumping scheme is applied when composition of a reser-
voir fluid or crude oil is given in terms of SCN groups such
as those given in Table 4.2. Whitson [15, 17] suggests that the
weighting factors according to Eq. (4.102). Summary of cal- C7+ fraction can be grouped into Alppseudocomponents given
culations and mole fractions and molecular weights of the by
components are given in Table 4.22. As shown in this table
average molecular weight of C7+ calculated from the 3 pseu- (4.108) ATe = 1 + 3.3 log~0(N+ - 7)
docomponents is 119.4, which varies with experimental value where Np is the number of pseudocomponents and N+ is the
of 118.9 by 0.4%. To apply the Gaussian method to the gamma carbon number of heaviest fraction in the original fluid de-
distribution model first we must determine coeff• or, r, scription. Obviously Np is the nearest integer number calcu-
and 0. Since detailed compositional data are not available as lated from the above equation. The groups are separated by
discussed in Section 4.5.3, we assume ~ = 90 and ~ = 1. Then molecular weight M i given by
fl is calculated from Eq. (4.33) as fl = (i 18.9 - 90)/1.0 = 28.9.
Substituting ~ = 1 in Eq. (4.105), considering that F(a) = 1 (4.109) Mi = M7+(MN+/M7+) 1/Nv
we get zi ----w/. Values of M/ are calculated from Eq. (4.106) where j = 1. . . . . Ale. SCN groups in the original fluid descrip-
with yi taken from Table 4.21. Three components have molec- tion that have molecular weights between boundaries Mi_l
ular weights of 102, 156.3, and 271.6, respectively. M7+ calcu- and Mi are included in the group j. This method can be ap-
lated from these values and mole fractions given in Table 4.22 plied only to those C7+ fractions that are originally separated
is exactly 118.9 the same as the experimental data. The rea- by SCN groups and Ne >__20 [17].
son is that this value was used to obtain parameter ft. For The lumping scheme is very similar to the pseudoization
this sample zi calculated from the gamma PDF is the same as method, except the distribution coefficients are determined
those obtained from Eq. (4.102) since o~= 1 and Eq. (2.105) for data on distribution of carbon number. For this reason
reduces to Eq. (2.102). But for values of u different from unity, the lumping scheme generates better and more accurate pseu-
the two models generate different mole fractions and dif- docomponents than does the splitting method when distribu-
ferent Mi values, t tion coefficients are determined from only two bulk properties
such as M7+ and SG7+. Method of lumping is very similar to
the calculations made in Example 4.10 in which SCN groups
4.6.1.2 Carbon Number Range Approach of C12 and C13 for the C7+ sample in Table 4.11 were lumped
In this approach we divide the whole C7+ fraction into a num- together and the mole fraction and molecular weight of the
ber of groups with known carbon number boundaries. As an group were estimated. Here the two methods that can be used
example if five pseudocomponents are chosen to describe the for lumping and splitting schemes are summarized to show
mixture, then five carbon number ranges must be specified. It the calculations [24]. In these methods the generalized distri-
was found that for gas condensate systems and light oils the bution model is used; however, other models (i.e., gamma or
carbon number ranges of C7-C10, CH-C15, C16-C2s, C25-C36, exponential) can be used in a similar way.
and C36§ well describe the mixture [23]. It should be noted
M e t h o d I: G a u s s i a n Q u a d r a t u r e A p p r o a c h
that the heaviest component in the first group is C+0, which
is the same as the lightest component in the second group is 1. Read properties of SCN groups and properties of plus frac-
C~-1. Values of the lower and upper limit molecular weights tions (e.g., M30+ and SG30+). Normalize the mole fractions
for each SCN group can be calculated from Eqs. (4.39) and ( ~ x m i = 1).
(4.40) and SCN up to C20 were calculated and are given in 2. If M and SG for each SCN group are not available, obtain
Table 4.10. For example for the C7-C10, the molecular weight these properties from Table 4.6.
range is from Mo (initial molecular weight of a C7+ fraction) to 3. Determine distribution parameters for molecular weight
M~0 or Mo - 142.5 and for the CH-CI5, the molecular weight (Mo, AM, and BM) in terms of cumulative mole fraction and
range is M~-M~5 or 142.5-214. Similarly molecular weight for specific gravity (SGo, AsG, and BsG) in terms of cumu-
range of other groups can be determined as: 214-352 for C16- lative weight fraction.
C25,352--492 for C26-C35, and 492-oo for C36+-cx~. For the last 4. Choose the number of pseudocomponents (i.e., 5) and cal-
group; i.e., C36+the molecular weight range is from M36 to culate their mole fractions (zi) and molecular weight (M/)
to. Once the lower and upper values of M are known mole from Eqs. (4.102) and (4.103) or from Eqs. (4.105) and
fraction and molecular weight for each group can be deter- (4.106) for the case of gamma distribution model.
mined from appropriate equations developed for each dis- 5. Using Mi and zi in step 4, determine discrete weight frac-
tribution model. Mole fraction and molecular weight of each tions, zwi, through Eq. (1.15).
group for the gamma distribution model are determined from 6. Calculate cumulative weight fraction, Z~w,from z~ and es-
Eqs. (4.37) and (4.38), respectively. For the generating distri- timate SGi for each pseudocomponent through Eq. (4.56)
bution model these equations are Eq. (4.84) and (4.86) and with coefficients determined for SG in step 3. For example,
for the exponential model mole fractions are calculated from SG1 can be determined from Eq. (4.56) at z~1 = Zwl/2.
4. C H A R A C T E R I Z A T I O N OF R E S E R V O I R F L U I D S A N D C R U D E O I L S 187
TABLE 4.23--Lumping of SCN groups by two methods for the C7+ sample in Example 4.15 [24].
Method I: Gaussian quadrature approach Method II: Carbon number range approach
Component Mole Weight Mole Weight
i fraction fraction Mi SGi fraction fraction Mi SGi
1 0.5218 0.3493 102.1 0.7436 0.532 0.372 106.7 0.7457
2 0.3987 0.4726 180.8 0.8023 0.302 0.328 165.5 0.7957
3 0.0759 0.1645 330.4 0.8591 0.144 0.240 254.4 0.8389
4 0.0036 0.0134 569.5 0.9174 0.019 0.049 392.7 0.8847
5 2.3 x 10-5 1.4 x 10-4 950.1 0.9809 0.003 0.011 553.5 0.9214
Mixture 1.0000 1.0000 152.5 0.7905 1.000 1.0000 1 5 2 . 5 0.7908
Taken with permission from Ref. [24].
7. Obtain May and SGav for the mixture from May = ~,~/P=IZi Mi in step 4 of the above methods. For carbon n u m b e r s from C7
and 1/SGav = ~7~1Zwi/SGj. to C19, values of M and SG are taken from Table 4.6 and mole
fractions are converted into weight fraction (x~-). Distribu-
Method II: Carbon Number Range Approach tion coefficients for M in terms of Xcm and SG in terms of Xcw
1. Same as Method I. are then determined from Eqs. (4.56) and (4.57). The results
2. Same as Method I. for M are Mo = 84, AM = 0.7157, and BM = 1 and for SG the
3. Same as Method I. coefficients are SGo = 0.655, Asc = 0.038, and Bso = 3. For
4. Choose n u m b e r of p s e u d o c o m p o n e n t s (i.e., 5) and corres- the 5 pseudocomponents, Methods I and II have been applied
ponding carbon n u m b e r ranges, e.g., group i: C7-C10, step by step and for each group j values of Zr~, Z~, Mi, and
group 2: Cll-C15, group 3: C16-C25, group 4:C25-C36 and SGi are given in Table 4.23. Specific gravity and molecular
C36+.
weight of C7+ calculated from p s e u d o c o m p o n e n t s generated
5. Obtain molecular weight boundaries from Eqs. (4.39) and by Method I are 0.7905 and 152.5, which are very close to ex-
(4.40). For example for the groups suggested in step 4 perimental values of 0.7917 and 151.6. Method II gives similar
the molecular weight ranges are: (Mo-142.5), (142.5-214), results as shown in Table 4.23. Specific gravity differs from
(214-352), (352-492) and (492-oo). The n u m b e r of pseu- the experimental value by 0.1%. Obviously components
d o c o m p o n e n t s (Np) and molecular weight boundaries m a y 1, 2 . . . . generated in Method I are not the same compo-
also be determined by Eqs. (4.108) and (4.109). nents generated by Method II, but combination of all 5 com-
6. Using the molecular weight boundaries determined in ponents by two methods represent the same mixture. That is
step 5, calculate mole fractions (zi) and molecular weight why Mi and SGi for the 5 pseudocomponents generated by
(Mi) of these p s e u d o c o m p o n e n t s from Eqs. (4.84) and Methods I and II are not the same.
(4.86) or from Eqs. (4.92) and (4.91) when BM in Eq. (4.56)
is equal to unity.
7. Same as step 5 in Method I. 4.7 CONTINUOUS M I X T U R E
8. Same as step 6 in Method I. CHARACTERIZATION APPROACH
9. Same as step 7 in Method I.
A more complicated but more accurate treatment of a C7+
In this m e t h o d if the calculated mole fraction for a pseudo- fraction is to consider it as a continuous mixture. In this ap-
component in step 6 is too high or too low, we m a y reduce or proach the mixture is not expressed in terms of a finite num-
increase the corresponding carbon n u m b e r range chosen for ber of p s e u d o c o m p o n e n t but its properties are given by a con-
that p s e u d o c o m p o n e n t in step 4. Application of these meth- tinuous function such as Eq. (4.56). This m e t h o d is in fact
ods is shown in the following example. equivalent to the p s e u d o c o m p o n e n t approach but with infi-
nite n u m b e r of c o m p o n e n t s (Np = oo). Mansoori and Chorn
Example 4.15--Fluid description of a C7+ from North Sea [27] discussed a general approach toward characterization of
fields (sample 42 in Ref. [24]) is given in terms of mole frac- continuous mixtures. In this approach instead of specifying
tions of SCN groups from C7 to C20+ as a c o m p o n e n t by i, it is expressed by one of its characteris-
tic parameters such as Tb or M. Formulation of continuous
Nc 7 8 9 10 11 12 13 14 15 16 17 18 19 20+
mixtures for phase equilibrium calculations is best expressed
xi 0.178 0.210 0.160 0.111 0.076 0,032 0.035 0.029 0.022 0.020 0.020 0.016 0,013 0.078
by Eq. (4.15), while for the p s e u d o c o m p o n e n t approach for a
where Nc represents carbon n u m b e r group and x~ is its cor- defined discrete mixture it is formulated through Eq. (4.14).
responding normalized mole fraction. For this mixture the To show application of a PDF in characterization of a crude
/147+ = 151.6 and SG7+ = 0.7917. L u m p these components oil by the continuous mixture approach, we use Eq. (4.15)
into an appropriate n u m b e r of p s e u d o c o m p o n e n t s and give to formulate vapor-liquid equilibrium (VLE) and to obtain
their molecular weight and specific gravity using the above species distribution of vapor and liquid products once such
two methods. distribution is k n o w n for the feed during a flash distillation
process. Theory of VLE is discussed in Chapter 6 and its appli-
cation is shown in Chapter 9. In Eq. (4.15), if we take boiling
Solution--For this sample N+ = 20 and we m a y use
point as the characterization parameter for P the equilibrium
Eq. (4.108) to determine the n u m b e r of pseudocomponents.
relation in terms of fugacity is (see Eq. 6.173)
Np = 1 + 3.31 log(20 - 7) = 4.7. The nearest integer n u m b e r
is 5, therefore Np = 5, which is the same n u m b e r as suggested (4.110) fV(T) = fL(T) To < T < oo
188 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
where fV(T) is the fugacity of a specie in the vapor phase gas the left-hand side of Eq. (4.114) should be multiplied by
whose boiling point is T. When T = To the above equation is fugacity coefficient ~0(T, T s, p) for components with boiling
applied to the lightest component in the mixture and when point T at temperature T s and pressure p. These thermody-
T = o0 it is applied to the last and heaviest component whose namic properties are defined in Chapter 6 and can be obtained
boiling point may be considered as infinity. For simplicity it is from an equation of state for hydrocarbon systems. A more
assumed that the vapor phase is ideal gas and the liquid phase general form of Eq. (4.114) for high-pressure VLE calcula-
is an ideal solution. Under such conditions Eq. (4.110) for a tions is in terms equilibrium ratio can be written as
component with boiling point T in a mixture can be written
as (4.115)
(4.111) dyrp = dxrpl's
where Kr is the equilibrium ratio for a component with boil-
where dyr is the mole fraction of a component having boil-
ing point T at temperature T s and pressure p. As it will be
ing point T in the vapor phase and dxr is the mole fraction
shown in Chapter 6, Kr depends on vapor pressure pS.
of the same component in the liquid phase, prs is the satura-
Now we apply the above equations for design and operation
tion pressure (or vapor pressure) of components with boiling
of a separation unit for flash distillation of reservoir fluids
point T at temperature T s and p is the total pressure at which
and crude oils. As shown in Fig. 4.25 we assume 1 mol of
vapor and liquid are in equilibrium. T s is in fact the tempera-
feed enters the unit that is operating at temperature T s and
ture at which separation occurs and Prs is a function of T s and
pressure p. The products are r moles of vapor and 1 - ~bmoles
type of component that is characterized by boiling point (see
of liquid in which ~ is the fraction of the feed vaporized in
Problem 4.16). This relation is known as the Raoult's law and
a single-stage flash distillation unit. Material balance on the
its derivation will be discussed in Chapter 6. In Eq. (4.111),
distillation unit for a component whose boiling point is T can
dyTp is the fugacity of components with boiling point T in an
be written as
ideal gas vapor phase while dxTprs is the fugacity of compo-
nents with boiling point T in an ideal liquid solution. To apply
Eq. (4.111) for a continuous mixture, we can use Eq. (4.16) (4.116) dzr x 1 = dxr x (I - ~ ) + d y r x
to express dxr and dyr by a PDF in each phase:
where dzT is the mole fraction of all components having boil-
(4.112) dxr = FLdT
ing point T and can be expressed in terms of a PDF similar to
(4.113) dyr = F~dr Eq. (4.112). Substituting Eqs. (4.112) and (4.113) for dxr and
dyT and similarly for dzr into the above equation gives
where FL and Fv are the PDF in terms of boiling point T for
the liquid and vapor phases, respectively. Equations (4.70)
or (4.31) may be used to express F v or F L. Substituting (4.117) FTF = (1 - ~)FTv + 4)FL4)
Eqs. (4.112) and (4.113) into Eq. (4.111) we get
where FF is the density function for the feed in terms of boil-
(4.114) FVp = F~.p
ing point T. For all three probability density functions, F v,
Equation (4.114) is the Raouh's law in terms of a PDF appli- F v, and F L we have
cable to a continuous mixture. If the liquid phase is nonideal,
the right-hand side of above equation should be multiplied by
activity coefficient V(T, T s) for those components with boiling
point T at temperature T ~. And if the vapor phase is nonideal
(4.118) /
TO
FFdT= /
To
FVdT= /
To
FLdT = 1
Vapor
Cmole
| Tb
~ Feed ~ Ts,P
1mole
Tb
~Liquid
e~
I-~mole
Tb
From Eqs. (4.114), (4.117), and (4.118) one can derive the 0.005
following relation for calculation of parameter r ""-. -- Crude Feed
0.004 , ' Liquid Product
P P~ . ...... Vapor Product
(4.119) f (1 -- ~-)-p-~r FFdT = 0
To "-~ 0.003 "
where the integration should be carried numerically and
may be determined by trial-and-error procedure. As will be ~ .002
shown in Chapter 7, combination of Trouton's rule for the heat
of vaporization and the Clasius-Clapeyron equation leads to
0.001
the following relation for the vapor pressure:
(4.122) P~ + Cp~ F v
FV = (1 - ~b)p where Eqs. (4.123)-(4.125) are equivalent to Eqs. (4.119),
(4.121), and (4.122) for ideal systems, respectively. Calcula-
After finding r from Eq. (4.119), it can be substituted in the tion of equilibrium ratios from equations of state will be dis-
above equations to find density functions for the vapor and cussed in Chapters 5 and 10. Probability density functions in
liquid products. these equations may be expressed in terms of other character-
For evaluation and application of these equations, data on ization parameters such as molecular weight or carbon num-
boiling point distribution of a Russian crude oil as given by ber. However, as discussed in Chapter 2, boiling point is the
Ratzch et al. [31] were used. In this case TBP distributions for most powerful characterization parameter and it is preferable
feed, vapor, and liquid streams during flash distillation of the to be used once it is available. Similarly the same approach
crude are available. Molecular weight, specific gravity, and can be used to obtain distribution of any other property (see
refractive index of the mixture are 200, 0.8334 and 1.4626, Problem 4.16).
respectively. Applying Method A discussed in Section 4.5.4.6, In treatment of a reservoir fluid, the mixture may be pre-
we obtain distribution coefficients for boiling point of feed sented in terms of composition of pure hydrocarbon com-
as: To = 241.7 K, Ar = 1.96, and Br = 1.5 and FrF was deter- pounds from Ct to C5 and nonhydrocarbon compounds such
mined from Eq. (4.70). Fraction of feed vaporized, q~, was as HES and CO2 as well as grouped C6+ or a SCN group of C6
determined from Eq. (4.119) as 0.7766. Boiling point distri- and C7+. For these mixtures the continuous mixture approach
butions for the liquid (F L) and vapor (Fv) products were de- discussed in this section can be applied to the hydrocarbon-
termined from Eqs. (4.121) and (4.122), respectively. Results plus portion, while the discrete approach can be applied
of calculations for F~, Fv, and F L for this crude are shown to the lower portion of the mixture containing compounds
in Fig. 4.26 and compared with the experimental values pro- with known composition. This approach is known as semi-
vided in Ref. [31]. Since heavier components appear in the continuous approach and calculation of different properties
liquid product, therefore, the curve for F L is in the right side of reservoir fluids by this approach has been discussed by
of both F F and F v corresponding to higher values of boiling various researchers [27, 28, 43].
points.
Part of errors for predicted distributions of F L and Frv is
due to assumption of an ideal solution for VLE calculations
as well as an approximate relation for the estimation of vapor 4 . 8 CALCULATION OF P R O P E R T I E S
pressures. For more accurate calculations Eq. (4.115) can be OF C R U D E OILS A N D R E S E R V O I R F L U I D S
used which would result in the following relations:
As discussed in Chapter 2, properties of a hydrocarbon com-
pound depend on its carbon number and molecular type.
(4.123) f (i - -12u,,
77 -r
'K---r-r'
~,Kr~ FrFdT = 0 Accurate calculation of properties of a petroleum mixture
rely on accurate knowledge of the composition of the mixture
To
by individual constituents, their properties, and an appropri-
1 ate mixing rule to estimate the mixture properties. In this
(4.124) FL = ( 1 - ~ b ) + ~ K r F rv
part based on the methods outlined in this chapter a crude
Kr FF oil or a reservoir fluid is presented by a number of pseudo-
(4.125/ FrY = (1 - r + r r components and a general approach is outlined to estimate
190 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
p r o p e r t i e s of such mixtures. Application of this a p p r o a c h is TABLE 4.25--Pseudoization of the C7+ for the Kuwaiti crude
s h o w n t h r o u g h e s t i m a t i o n of sulfur c o n t e n t of c r u d e oils. in Example 4.16.
Pseudocomponent 1 2 3 4 5
Weight fraction 0.097 0.162 0.281 0.197 0.264
Mole fraction 0.230 0.255 0.280 0,129 0,106
4.8.1 General Approach Molecular weight 112.0 169.1 267.1 405,8 660.9
Specific gravity 0.753 0.810 0.864 0,904 0.943
F o r a reservoir fluid a c c u r a t e p r o p e r t i e s can be calculated
t h r o u g h detailed c o m p o s i t i o n a l analysis o f p u r e c o m p o u n d s
f r o m C1 to C5 a n d SCN g r o u p s from C6 a n d heavier g r o u p s by an a d e q u a t e n u m b e r of p s e u d o c o m p o n e n t s with k n o w n
u p to at least C50. The r e m a i n i n g p a r t can be g r o u p e d as C50+. parameters.
F o r very heavy oils, SCN g r o u p s e p a r a t i o n m a y be e x t e n d e d
u p to C80 a n d the residue g r o u p e d as C80+. To e s t i m a t e various E x a m p l e 4 . 1 6 - - C o m p o s i t i o n a l d a t a on a K u w a i t i c r u d e oil is
p r o p e r t i e s of these SCN groups at least two c h a r a c t e r i z i n g pa- given as follows:
r a m e t e r s such as Tb a n d SG o r M a n d SG should be known.
This is s h o w n in Table 4.24, w h e r e k n o w n d a t a are i n d i c a t e d Component C2 C3 iC4 nC4 iC5 nC5 C6 C7+
b y + sign. Methods o u t l i n e d in Sections 4.5 a n d 4.6 can l e a d Wt% 0.03 0.39 0.62 1 . 0 8 0.77 1 . 3 1 1.93 93.87
to generate such i n f o r m a t i o n for a reservoir fluid. F o r SCN
g r o u p s of C6 a n d heavier, m e t h o d s in C h a p t e r 2 c a n be used The c h a r a c t e r i s t i c s of the C7+ fraction a r e M7+ = 266.6 a n d
to e s t i m a t e various p r o p e r t i e s (Tb, To Pc. . . . ) using M a n d SG7+ =-- 0.891 [44]. Divide the C7§ fraction into 5 p s e u d o c o m -
SG as available input p a r a m e t e r s . F o r p u r e c o m p o u n d s u p p o n e n t s a n d p r e s e n t the crude in t e r m s of m o l e a n d weight
to C5, all b a s i c p r o p e r t i e s are given in Tables 2.1 a n d 2.2 a n d fractions of representative constituents with k n o w n M, SG,
no e s t i m a t i o n m e t h o d is required. F o r m o r e a c c u r a t e pre- a n d Tb. E s t i m a t e M a n d SG for the whole crude.
diction of p r o p e r t i e s of a reservoir fluid, each SCN g r o u p s
f r o m C6 up to C50+ m a y be divided into f u r t h e r three pseudo-
c o m p o n e n t s as paraffinic, naphthenic, a n d aromatic. Meth- S o l u t i o n - - F o r the C 6 g r o u p from Table 4.6 we have M6 = 82,
ods of Section 3.5 can be u s e d to d e t e r m i n e PNA c o m p o s i t i o n SG6 --- 0.69, a n d Tb6 :- 337 K. F o r p u r e c o m p o n e n t s f r o m C2
of each SCN group. In this w a y n u m b e r of c o m p o n e n t s in to C5, M a n d SG can be taken from Table 2.1. Using M a n d
Table 4.24 increases to 152. F o r heavy oils the n u m b e r of c o m - x~, m o l e fraction Xm c a n be e s t i m a t e d t h r o u g h Eq. (4.61).
p o n e n t s w o u l d be even higher. F o r each h o m o l o g o u s group, Using M e t h o d A in Section 4.5.4.6 d i s t r i b u t i o n coefficients
different p r o p e r t i e s m a y be e s t i m a t e d f r o m m o l e c u l a r weight in Eq. (4.56) for the C7+ fraction are f o u n d as Mo = 90,
of individual SCN g r o u p t h r o u g h the relations given in Sec- AM = 1.957, a n d BM = 1.0. F r o m M e t h o d II outlined in Sec-
tion 2.3.3. Obviously calculation of m i x t u r e p r o p e r t i e s w h e n tion 4.6.1.2 a n d specifying 5 c a r b o n n u m b e r ranges the C7+
it is expressed in t e r m s of large n u m b e r of c o m p o n e n t s is n o t can be split into 5 p s e u d o c o m p o n e n t s with k n o w n m o l e frac-
an easy task. F o r this r e a s o n the n u m b e r of c o m p o n e n t s in tion (normalized), M a n d SG as given in Table 4.25. I n this
Table 4.24 m a y be r e d u c e d b y g r o u p i n g to SCN c o m p o n e n t s table the weight fractions are calculated t h r o u g h Eq. (1.15)
o r splitting the C7 fraction into j u s t 3 o r 5 p s e u d o c o m p o n e n t s . using mole fraction a n d m o l e c u l a r weight. Values of weight
F u r t h e r m o r e , iC4 a n d nC4 m a y b e g r o u p e d as C4 a n d iC5 a n d fractions in Table 4.25 s h o u l d be m u l t i p l i e d b y w t % of C7+ in
nC5 could be g r o u p e d as C5. I n this w a y the mixture can be pre- the whole c r u d e to e s t i m a t e wt% of each p s e u d o c o m p o n e n t
sented b y 10-15 c o m p o n e n t s with k n o w n specifications. The in the crude. Values of m o l % in the original fluid are cal-
following e x a m p l e shows h o w a c r u d e oil can be p r e s e n t e d culated from wt% a n d m o l e c u l a r weight of all c o m p o n e n t s
p r e s e n t in the m i x t u r e as s h o w n in Table 4.26. F o r the 5 pseu-
d o c o m p o n e n t s g e n e r a t e d by splitting the C7+, boiling p o i n t s
TABLE 4.24--Matrix table of components for estimation are calculated from M a n d SG using Eq. (2.56). F r o m Tb a n d
of properties of reservoir fluids. SG of p s e u d o c o m p o n e n t s given in Table 4.26, one m a y esti-
No. Compound Mole fraction M SG m a t e basic c h a r a c t e r i z a t i o n p a r a m e t e r s to e s t i m a t e various
1 H2S q-
2 CO2 +
3 N2 + TABLE4.26--Characterization of the Kuwait crude oil
4 H20 + in Example 4.16.
5 C1 + Component Wt% Mol% M SG Tb,; C
6 C2 + C2 0.03 0.22 30.1 0.356
7 C3 + C3 0.39 1.99 44.1 0.507
8 iC4 q- iC4 0.62 2.40 58.1 0.563
9 nC4 q- nC4 1.08 4.18 58.1 0.584
10 iC5 + iC5 0.77 2.40 72.2 0.625
11 nC5 + nC5 1.31 4,08 72.2 0.631
12 C6 + -I- + C6 1.93 5.29 82 0.690 64
13 C7 + + + C7+(1) 9.1 18,28 112,0 0.753 123
14 C8 + + + C7+(2) 15.2 20.22 169.1 0.810 216
15-5 5a C9-C49 q- + + C7+(3) 26.4 22.23 267.1 0.864 333
56 C50 + + + C7+(4) 18.5 10.26 405.8 0.904 438
57 C5o+ + + + C7+(5) 24.8 8.44 660.9 0.943 527
aCompounds from 15 to 55 represent SCN groups from C9 to C49.
For compounds 1-11, properties are given in Tables 2.1 and 2.2. Total 100 100 225.2 0.8469
4. C H A R A C T E R I Z A T I O N OF R E S E R V O I R F L U I D S A N D C R U D E O I L S 191
properties, For example, for the whole crude the molecular 4.8.2 E s t i m a t i o n o f Sulfur Content o f a Crude Oil
weight is calculated from Xmi and Mi using Eq. (3.1) and spe-
Estimation of sulfur content of crude oils is based on the gen-
cific gravity of the whole crude is calculated from Xwi and SGi
eral a p p r o a c h f o r e s t i m a t i o n o f v a r i o u s p r o p e r t i e s o f c r u d e
using Eq. 3.44. Calculated M and SG for the crude are 225
oils a n d r e s e r v o i r fluids d e s c r i b e d i n S e c t i o n 4.8.1. O n c e a
and 0.85, respectively. These values are lower than those for
c r u d e is p r e s e n t e d b y a n u m b e r o f p u r e c o m p o n e n t s a n d s o m e
the C7+ as the crude contains components lighter than C7. In
narrow boiling range pseudocomponents with known mole
the next example estimation of sulfur content of this crude is
f r a c t i o n , Tb, a n d SG, a n y p h y s i c a l p r o p e r t y m a y b e e s t i m a t e d
demonstrated.
t h r o u g h m e t h o d s d i s c u s s e d i n C h a p t e r s 2 a n d 3. P r o p e r t i e s
Solution--The crude is presented in terms of 12 compo- FIG. 4.27--Distribution of sulfur content in the
nents (6 pure compounds and 6 pseudocompounds) in Ta- crude oil of Example 4.18. Taken with permission
ble 4.26. Sulfur content of the crude should be estimated from Ref. [45],
through Eq. (4.126). For pure hydrocarbons from C2 to nC5
the sulfur content is zero; however, sulfur content of pseu-
docomponents from C6 to C7§ should be calculated from crude is calculated from Eq. (4.126) as shown in the last col-
Eqs. (3.96) and (3.97). Parameters n20 and d20 needed in these umn of Table 4.27. The estimated sulfur content of the crude
equations have been estimated from methods discussed in is 2.34 wt%, which is near the experimental value of 2.4%.
Sections 2.6.1 and 2.6.2. Estimated sulfur content of C6 and A graphical comparison between predicted and experimen-
the 5 C7+ pseudocomponents are 0.2, 0.1, 0.7, 1.9, 2.9, and 3.8, tal sulfur distribution along distillation curve is presented in
respectively. Substituting these values into Eq. (4.126) would Fig. 4.27.
estimate sulfur content of the whole crude as 2.1% versus
experimental value of 2.4% with -0.3 wt% error. Calculations made in Examples 4.17 and 4.18 show that as
more characterization data for a crude are available better
Example 4.18--For the crude oil of Example 4.17 a complete property prediction is possible. In many cases characteriza-
TBP, SG and sulfur wt% curves versus weight fraction are tion data on a crude contain only the TBP curve without SG
available as given in the first five columns of Table 4.27. Esti- distribution. In such cases M and SG distributions can be de-
mate the sulfur content curve and graphically compare with termined from Eq. (4.7) and coefficients given in Table 4.5.
the experimental values. Also estimate the sulfur content of Equation (4.7) can be used in its reversed form using Tb as in-
whole crude from predicted sulfur content curve and com- put instead of M. Once M is determined it can be used to esti-
pare with the experimental value of 2.4 wt%. mate SG, n20, and d20 from Eq. (4.7) with corresponding coef-
ficients in Table 4.5. This approach has been used to estimate
Solution--A complete characterization dataset on a crude oil sulfur content of 7 crudes with API gravity in the range of 31-
include two suitable characterization parameters such as Tb 40. An average deviation of about 0.3 wt% was observed [45].
and SG versus cumulative weight or volume fraction with low
residue. When such data are available, properties of the crude
may be estimated quite accurately. In Table 4.27, Tb, d2o, and 4.9 CONCLUSIONS AND
sulfur wt% of 55 cuts are given with known wt%. The boil- RECOMMENDATIONS
ing point of last cut (residue) was not originally known from
experimental data. Based on the fractions with Tb greater than In this chapter methods of characterization of reservoir fluids,
100~ (cuts 21-54), weight fractions were normalized and cu- crude oils, natural gases and wide boiling range fractions have
mulative weight fractions were calculated. Temperature of been presented. Crude assay data for seven different crudes
100~ is near the boiling point of nC7. For the C7+ portion from around the world are given in Section 4.1.2. Charac-
of the crude distribution, coefficients in Eq. (4.56) were de- terization of reservoir fluids mainly depends on the charac-
termined as To = 360 K, Ar = 1.6578, and Br = 1.485. Using terization of their C7+ fractions. For natural gases and gas
these values, boiling point of the residue (cut 55) was deter- condensate samples with little C7+ content, correlations de-
mined from Eq. (4.56) as 678.3~ Specific gravity of cuts were veloped directly for C7+, such as Eqs. (4.11)-(4.13), or the
determined from d20 and parameters M, t~, and m were de- correlations suggested in Chapters 2 and 3 for narrow-boiling
termined using the methods discussed in Chapters 2 and 3. range fractions may be used. However, this approach is not
Sulfur content of each cut were determined from Eq. (3.96) applicable to reservoir fluids with considerable amount of C7+
for cuts with M < 200 and from Eq. (3.97) for cuts 44-55 with such as volatile or black oil samples. The best way of charac-
M > 200. For cuts 1-7 calculated values of S% from Eq. (3.96) terizing a reservoir fluid or a crude oil is to apply a distribu-
were slightly less than zero and they are set as zero as dis- tion model to its C6+ or C7+ portion and generate a distribu-
cussed in Section 3.5.2.2. Finally sulfur content of the whole tion of SCN groups or a number of pseudocomponents that
4. C H A R A C T E R I Z A T I O N OF R E S E R V O I R FLUIDS A N D CRUDE O I L S 193
represent the C7+ fraction. Various characterization param- is presented in terms weight, volume, or mole fraction of pure
eters and basic properties of SCN groups from C6 to C50 are compounds, the basic characterization parameters and prop-
given in Table 4.6 and in the form of Eq. (4.7) for computer erties may be taken from Tables 2.1 and 2.2. Once a crude
applications. is expressed in term of a number of components with known
Characterization of C7+ fraction is presented through appli- properties, a mixture property can be determined through ap-
cation of a distribution model and its parameters may be de- plication of an appropriate mixing rule for the property as it
termined from bulk properties with minimum required data will be shown in the next chapter.
on M7+ and SG7+. Three types of distribution models have
been presented in this chapter: exponential, gamma, and a
generalized model. The exponential model can be used only to 4.10 PROBLEMS
molecular weight and is suitable for light reservoir fluids such
as gas condensate systems and wet natural gases. The gamma 4.1. Consider the dry natural gas, wet natural gas, and gas
distribution model can be applied to both molecular weight condensate systems in Table 1.2. For each reservoir fluid
and boiling point of gas condensate systems. However, the estimate the following properties:
model does not accurately predict molar distribution of very a. SGg for and the API gravity.
heavy oils and residues. This model also cannot be applied b. Estimate Tp~ and/'pc from methods of Section 4.2.
to other properties such as specific gravity or refractive in- c. Estimate Tpc, Ppo and Vp~ from Eq. (3.44) using pure
dex. The third model is the most versatile distribution model components properties from Table 2.1 and C7+ prop-
that can be applied to all major characterization parameters erties from Eqs. (4.12) and (4.13).
of M, Tb, SG, and refractive index parameter I. Furthermore, d. Compare the calculated values for Tp~ and Pp~ in parts
the generalized distribution model predicts molar distribu- b and c and comment on the results.
tion of heavy oils and residues with reasonable accuracy. Ap- 4.2. Calculate Tb, SG, d20, n20, Tc, Pc, Vo a, and ~ for C55, C65,
plication of the generalized distribution model (Eq. 4.56) to and C75 SCN groups.
phase behavior prediction of complex petroleum fluids has 4.3. Predict SCN distribution for the West Texas oil sample
been reported in the literature [46]. Both the gamma and the in Table 4.1, using Eq. (4.27) and M7+ and x7+ (mole
generalized distribution models can be reduced to exponen- fraction of C7+) as the available data.
tial in the form of a two-parameter model. 4.4. Derive an analytical expression for Eq. (4.78), and show
Once a distribution model is known for a C7+ fraction, the that when SG is presented in terms of X~wwe have
mixture can be considered as a continuous mixture or it could
be split into a number of pseudocomponents. Examples for
both cases are presented in this chapter. The method of con- SG.v+~-~]- - J0 = ~ ( - 1)k+l !~)F 1 + ---~-
tinuous distribution approach has been applied to flash dis- k=0
tillation of a crude oil and the method of pseudocomponent 4.5. Basic characterization data, including M, Tb, and SG,
approach has been applied to predict sulfur content of an oil. versus weight fraction for seven subfractions of a C7+
Several characterization schemes have been outlined for dif- fluid are given in Table 4.28. Available experimental bulk
ferent cases when different types of data are available. Meth- properties are MT+ = 142.79, and SG7+ -- 0.7717 [47].
ods of splitting and grouping have been presented to represent Make the following calculations:
a crude by a number of representative pseudocomponents. a. Calculate Xr~ and x~.
A good characterization of a crude oil or a reservoir fluid is b. Estimate distribution parameter I from Tb and SG
possible when TBP distillation curve is available in addition using methods of Chapter 2.
to M7+ and SG7+. The most complete and best characteriza- c. Using experimental data on M, Tb, SG and I distribu-
tion data on a crude oil or a C7+ fraction would be TBP and SG tions calculate distribution coefficients Po, A and B in
distribution in terms of cumulative weight or volume fraction Eq. (4.56) for these properties. Present M in terms of
such as those shown in Table 4.27. Knowledge of carbon num- x ~ and Tb, SG and I in terms of Xcw.
ber distribution up to C40 and specification of residue as C40+ d. Calculate PDF from Eq. (4.66) and show graphical
fraction is quite useful and would result in accurate property presentation of F(M), F(T), F(SG), and F(I).
prediction provided the amount of the residue (hydrocarbon e. Find refractive index distribution
plus) is not more than a few percent. For heavy oils separation f. Calculate mixture M, Tb, SG, and n20 based on the
up to C60+ or C80+ may be needed. When the boiling point of coefficients obtained in part c.
the residue in a crude or a C7+ fraction is not known, a method g. For parts b and f calculate errors for M, Tb, and SG in
is proposed to predict this boiling point from the generalized terms of AAD.
distribution model. When data on characterization of a crude
are available in terms of distribution of carbon number such TABLE 4.28---Characterization parameters for the C7+
as those shown in Table 4.2, the method of grouping should be fraction of the oil.system in Problem 4.5 [47].
used to characterize the mixture in terms of a number of sub- x~ M~ Tb~,K SG
fractions with known mole fraction, M, Tb and SG. Further in- 0.1269 98 366.5 0.7181
formation on options available for crude oil characterization 0.0884 110 394.3 0.7403
0.0673 121 422.1 0.7542
from minimum data is given by Riazi et al. [40]. Properties 0.1216 131 449.8 0.7628
of subfractions or pseudocomponents can be estimated from 0.1335 144 477.6 0.7749
Tb and SG using methods presented in Chapters 2 and 3. For 0.2466 165 505.4 0.7859
light portion of a crude or a reservoir fluid whose composition 0.2157 216 519.3 0.8140
194 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
4.6. Repeat Problem 4.5 for the gamma and exponential dis- (crude number 6). In constructing this figure the mid-
tribution models to volume points may be used for the specific gravity. De-
a. find the coefficients of Eq. (4.31): 0, a, and fl in termine the distribution coefficients in Eq. (4.56) for Tb
Eq. (4.31) for M and Tb. and SG in terms Of Xcvand compare with the experimen-
b. estimate SG distribution based on exponential model tal values. Also estimate crude sulfur content.
and constant Kw approach. 4.15. Show how Eqs. (4.104), (4.105), and (4.106) have been
c. calculate mixture M, Tb, and SG and compare with derived.
experimental data. 4.16. As it will be shown in Chapter 7, Lee and Kesler have
d. make a graphical comparison between predicted dis- proposed the following relation for estimation of vapor
tributions for M, Tb, and SG from Eq. (4.56) as ob- pressure (Pvap) of pure compounds, which may be ap-
tained in Problem 4.5, gamma and exponential mod- plied to narrow boiling range fractions (Eq. 7.18).
els with each other and experimental data.
4.7. For the C7+ of Problem 4.5 find distributions of M, Tb, lnP~ap = 5.92714 - 6.09648/Tbr -- 1.28862 In Tbr
and SG assuming: + 0.169347T6r + o)(15.2518 -- 15.6875/Tbr
a. Only information available are MT+ = 142.79 and
SG7+ = 0.7717. -- 13.4721 In Tbr + 0.43577T~r)
b. Only information available is MT+ = 142.79.
where p~ap = pv,p/pc and Tb~ = Tb/Tc in which both Tb
c. Only information available is SGT+ = 0.7717.
and Tc must be in K. Use the continuous mixture ap-
d. Graphically compare predicted distributions from
proach (Section 4.7) to predict distribution of vapor
parts a, b and c with data given in Table 4.28.
pressure at 311 and 600 K for the waxy crude oil in
e. Estimate MT+ and SG7+ form distribution parameters
Table 4.2 and graphically show the vapor pressure dis-
obtained in parts a, b and c and compare with the
tribution versus cumulative mol% and carbon number.
experimental data.
4.17. Minimum information that can be available for a crude
4.8. Using the Guassian Quadrature approach, split the C7+
oil is its API or specific gravity. A Saudi light crude has
fraction of Problem 4.7 into three pseudocomponents.
API gravity of 33.4 (SG = 0.8581), and experimental data
Determine, xm, M, Tb, and SG for each component. Cal-
on boiling point and specific gravity of its various cuts
culate the mixture M and SG from the three pseudocom-
are given in the following table as given in the Oil and
ponents. Repeat using carbon number range approach
Gas Journal Data Book (2000) (p. 318 in Ref. [8]).
with t 5 pseudocomponents and appropriate boundary
values of Mi.
4.9. For the C7+ fraction of Problem 4.5 estimate total sulfur Vol% SG Tb, K SG (calc) Tb, K (calc)
content in wt%. 23.1 _ 370.8 ? ?
4.10. For the waxy oil in Table 4.2 present the oil in six 23.1 0.8"1"31 508.3 ? ?
groups as C2-C3, C4-C6, C7-Cl0, Cll-C20, C21-C30, and 8.5 0.8599 592.5 ? ?
30.2 0.9231 727.5 ? ?
C31+. Determine M and SG for each group and calculate 15.1 1.0217 ... ?
M and the API gravity of the oil. Compare estimated
M from the six groups with M calculated for the crude
based on the detailed data given in Table 4.2. a. Using the minimum available data (API gravity), es-
4.11. Use the crude assay data for crude number 7 in Table 4.3 timate values of Tb and SG in the above table and
to compare with given experimental data graphically.
a. determine Tb and SG distributions. b. Similar data exist for a Saharan crude oil from Algeria
b. estimate Tb for the residue based on the distribution (page 320 in Ref. [8]) with API gravity of 43.7. Con-
found in Part a. struct Tb and SG distribution diagram in terms of cu-
c. estimate M for the residue from Tb in Part b and mulative volume fraction.
SG. 4.18. Similar to the continuous mixture approach introduced
d. estimate M for the residue from viscosity and SG and in Section 4.7, calculate vapor and liquid product distri-
compare with value from c. butions for flash distillation of the same crude at 1 atm
e. Determine distribution of sulfur for the crude and and 400~ Present the results in a fashion similar to
graphically evaluate variation of S% versus cumula- Fig. 4.26 and calculate the vapor to feed ratio (~b).
tive wt%. 4.19. Repeat Problem 4.18 but instead of Eq. (4.120) for the
f. Estimate sulfur content of the crude based on the pre- vapor pressure, use the Lee-Kesler correlation given in
dicted S% distribution. Problem 4.16. Compare the results with those obtained
4.12. For the crude sample in Problem 4.8 find distribution of in Problem 4.17 and discuss the results.
melting point and estimate average melting point of the
whole crude.
4.13. Estimate molecular weights of SCN groups from 7 to REFERENCES
20 using Eqs. (4.91) and (4.92) and compare your re-
sults with those calculated in Example 4.6 as given in [1] Pedersen, K. S., Thomassen, R, and Fredenslund, Aa.,
Table 4.10. "Characterization of Gas Condensate Mixtures," C7+ Fraction
4.14. Construct the boiling point and specific gravity curves Characterization, L. G. Chorn and G. A. Mansoori, Eds., Taylor
for the California crude based on data given in Table 4.3 & Francis, New York, 1989, pp. 137-152.
4. C H A R A C T E R I Z A T I O N OF R E S E R V O I R F L U I D S A N D C R U D E O I L S 195
[2] Hoffmann, A. E., Crump, J. S., and Hocott, C. R. E, [25] Riazi, M. R., "Distribution Model for Properties of
"Equilibrium Constants for a Gas-Condensate System," Hydrocarbon-Plus Fractions," Industrial and Engineering
Petroleum Transactions AIME, Vol. 198, 1953, pp. 1-10. Chemistry Research, Vol. 28, 1989, pp. 1731-1735.
[3] Osjord, E. H., Ronnisgsen, H. E, and Tau, L., "Distribution of [26] Ahmed, T., Hydrocarbon Phase Behavior, Gulf Publishing,
Weight, Density, and Molecular Weight in Crude Oil Derived Houston, TX, 1989.
from Computerized Capillary GC Analysis," Journal of High [27] Mansoori, G. A. and Chorn, L. G., "Multicomponent Fractions
Resolution Chromatography, Vol. 8, 1985, pp. 683-690. Characterization: Principles and Theories,"
[4] ASTM, Annual Book of Standards, Section Five, Petroleum C7+Characterization, L. G. Chorn and G. A. Mansoori,
Products, Lubricants, and Fossil Fuels (in 5 Vol.), ASTM Eds.,Taylor & Francis, New York, 1989, pp. 1-10.
International, West Conshohocken, PA, 2002. [28] Cotterman, R. L. and Prausnitz, J. M., "Flash Calculation for
[5] KISR, Private Communication, Petroleum Research Studies Continuous or Semicontinuous Mixtures Using An Equation of
Center, Kuwait Institute for Scientific Research (KISR), State," Industrial Engineering Chemistry, Process Design and
Ahmadi, Kuwait, May 2002. Development, Vol. 24, 1985, pp. 434-443.
[6] Pedersen, K. S., Fredenslund, Aa., and Thomassen, E, Properties [29] Kehlen, H., Ratsch, M. T., and Berhmann, J., "Continuous
of Oils and Natural Gases, Gulf Publishing, Houston, TX, 1989. Thermodynamics of Multicomponent Systems," AIChE
[7] Pan, H., Firoozabadi, A., and Fotland, E, "Pressure and Journal, Vol. 31, 1985, pp. 1136-1148.
Composition Effect on Wax Precipitation: Experimental Data [30] Ratsch, M. T. and Kehlen, H., "Continuous Thermodynamics of
and Model Results," Society of Petroleum Engineers Production Complex Mixtures," Fluid Phase Equilibria, Vol. 14, 1983,
and Facilities, Vol. 12, No. 4, 1997, pp, 250-258. pp. 225-234.
[8] Oil and Gas Journal Data Book, 2000 edition, PennWell, Tulsa, [31] Ratzch, M. T., Kehlen, H., and Schumann, J., "Flash
OK, 2000, pp. 295-365. Calculations for a Crude Oil by Continuous Thermodynamics,"
[9] Gary, J. H. and Handwerk, G. E., Petroleum Refining, Technology Chemical Engineering Communications, Vol. 71, 1988,
and Economic, 3rd ed., Marcel Dekker, New York, 1994. pp. 113-125.
[10] Ahmed, T., Reservoir Engineering Handbook, Gulf Publishing, [32] Wang, S. H. and Whiting, W. B., "A Comparison of Distribution
Houston, TX, 2000. Functions for Calculation of Phase Equilibria of Continuous
[11] Standing, M. B., Volumetric and Phase Behavior of Oil Field Mixtures," Chemical Engineering Communications, Vol. 71,
Hydrocarbon Systems, Society of Petroleum Engineers, Dallas, 1988, pp. 127-143.
TX, 1977. [33] Katz, D. L., "Overview of Phase Behavior of Oil and Gas
[12] Wichert, E. and Aziz, K., "Calculation of Z's for Sour Gases," Production," Journal of Petroleum Technology, June 1983,
Hydrocarbon Processing, Vol. 51, No. 5, 1972, pp. 119-122. pp. 1205-1214.
[13] Carr, N., Kobayashi, R., and Burrows, D.,"Viscosity of [34] Yarborough, L., "Application of a Generalized Equation of
Hydrocarbon Gases Under Pressure," Transactions of AIME, State to Petroleum Reservoir Fluids," Paper presented at the
Vol. 201, 1954, pp. 23-25. 176th National Meeting of the American Chemical Society,
[14] Katz, D. L. and Firoozabadi, A. "Predicting Phase Behavior of Miami Beach, FL, 1978.
Condensate Crude/Oil Systems Using Methane Interaction [35] Pedersen, K. S., Thornassen, P., and Fredenslund, Aa.,
Coefficients," Journal of Petroleum Technology, November 1978, "Thermodynamics of Petroleum Mixtures Containing Heavy
pp. 1649-1655. Hydrocarbons, 1: Phase Envelope Calculations by Use of the
[15] Whitson, C. H., "Characterizing Hydrocarbon Plus Fractions," Soave-Redlich-Kwong Equation of State," Industrial
Society of Petroleum Engineers Journal, August 1983, Engineering Chemisty, Process Design and Development, Vol. 23,
pp. 683-694. 1984, pp. 163-170.
[16] Riazi, M. R. and A1-Sahhaf, T. A., "Physical Properties of Heavy [36] Whitson, C. H., Anderson, T. E, and Soreide, I., "C7+
Petroleum Fractions and Crude Oils," Fluid Phase Equilibria, Characterization of Related Equilibrium Fluids Using the
Vol. 117, 1996, pp. 217-224. Gamma Distribution," C7+ Fraction Characterization, L. G.
[17] Whitson, C. H. and Brule, M. R., Phase Behavior, Monograph Chorn and G. A. Mansoori, Eds., Taylor & Francis, New York,
Volume 20, SPE, Richardson, TX, 2000. 1989, pp. 35-56.
[18] Danesh, A., PVT and Phase Behavior of Petroleum Reservoir [37] Whitson, C. H., Anderson, T. E, and Soreide, I., "Application
Fluids, Elsevier, Amsterdam, 1998. of the Gamma Distribution Model to Molecular Weight
[19] Riazi, M. R. and Daubert, T. E., "Characterization Parameters and Boiling Point Data for Petroleum Fractions,"
for Petroleum Fractions," Industrial and Engineering Chemistry Chemical Engineering Communications, Vol. 96, 1990,
Research, VoL 26, 1987, pp. 755-759. Corrections, p. 1268. pp. 259-278.
[20] Whitson, C. H., "Effect of C7+ Properties on Equation-of-State [38] Abramowitz, M. and Stegun, I. A., Handbook of Mathematical
Predictions," Society of Petroleum Engineers Journal, December Functions, Dover Publication, New York, 1970.
1984, pp. 685-696. [39] Press, W. H., Flannery, B. P., Teukolsky, S. A., and Vetterling,
[21] Chorn, L. G. and Mansoori, G. A., C7+ Fraction W. T., Numerical Recipes, The Art of Scientific Computing,
Characterization, Taylor & Francis, New York, 1989, 235 p. Cambridge University Press, Cambridge, London, 1986,
[22] Benmekki, H. and Mansoori, G. A., "Pseudization techniques pp. 160-161.
and heavy fraction characterization with equation of state [40] Riazi, M. R., A1-Adwani,H. A., and Bishara, A., "The Impact of
models, Advances in Thermodynamics, Vol. 1: C7+ Fraction Characterization Methods on Properties of Reservoir Fluids
Characterization, L. G. Chorn and G. A. Mansoori, Eds,, Taylor and Crude Oils: Options and Restrictions," Journal of Petroleum
& Francis, New York, 1989, pp. 57-78. Science and Engineering, Vol. 42., No. 2-4, 2004, pp. 195-207.
[23] Soreide, I., Improved Phase Behavior Predictions of Petroleum [41] Rodgers, P. A., Creagh, A. L., Prauge, M. M., and Prausnitz,
Reservoir Fluids from a Cubic Equation of State, Doctoral J. M., "Molecular Weight Distribution for Heavy Fossil Fuels
Dissertation, Norwegian Institute of Technology (NTH), from Gel-Permeation Chromatograph,," Industrial and
Trondheim, Norway, 1989. Engineering Chemistry Research, Vol. 26, 1987, pp. 2312-
[24] Riazi, M. R. "A Distribution Model for C7+ Fractions 2321.
Characterization of Petroleum Fluids," Industrial and [42] Berge, O., Damp/Vaeske-Likevekteri Raoljer: Karakterisering av
Engineering Chemistry Research, Vol. 36, 1997, pp. 4299-4307. Hydrokarbonfraksjon, M.Sc. Thesis, Department of Chemical
196 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
Engineering, Norwegian Institute of Technology, Trondheim, Engineering Chemistry Research, Vol. 38, No. 11, 1999,
Norway, 1981. pp. 4507-4512.
[43] Manafi, H., Mansoori, G. A., and Ghotbi, S., "Phase Bahavior [46] Fazlali, A., Modarress, H., and Mansoori, G. A., "Phase
Prediction of Petroleum Fluids with Minimum Behavior Prediction of Complex Petroleum Fluids," Fluid Phase
Characterization Data," Journal of Petroleum Science and Equilibra, Vol. 179, 2001, pp. 297-317.
Engineering, Vol. 22, 1999, pp. 67-93. [47] Jacoby, R. H., Koeller, R. C., and Berry Jr., V. J., "Effect of
[44] Stroud, A. H. and Secrest, D., Gaussian Quadrature Formulas, Composition and Temperature on Phase Behavior and
Prentice-Hall, Englewood Cliffs, NJ, 1966. Depletion Performance of Rich Gas-Condensate
[45] Riazi, M. R., Nasimi, N., and Roomi, Y., "Estimating Sulfur Systems," Transactions ofAIME, Vol. 216, 1959,
Content of Petroleum Products and Crude Oils," Industrial and pp. 406-411.
MNL50-EB/Jan. 2005
197
property. All molar properties are intensive properties and are T and P three phases of a pure c o m p o n e n t can coexist all the
related to total property as time. This temperature and pressure are k n o w n as triple point
0t
temperature and triple point pressure and are characteristics
(5.1) 0 = -- of any pure c o m p o u n d and their values are given for m a n y
n c o m p o u n d s [2, 3]. For example, for water the triple point tem-
where n is the n u m b e r of moles, 0 t is a total property such as perature and pressure are 0.01~ and 0.6117 kPa (~0.006 bar),
volume, V t, and 0 is a molar property such as molar volume, respectively [3]. The most recent tabulation and formulation
V. The n u m b e r of moles is related to the mass of the system, of properties of water r e c o m m e n d e d by International Asso-
m, through molecular weight by Eq. (1.6) as ciation for the Properties of Water and Steam (IAPWS) are
m given by Wagner and Pruss [4].
(5.2) n = -- A t h e r m o d y n a m i c property that is defined to formulate the
M
first law of thermodynamics is called internal energy shown
If total property is divided by mass of the system (m), instead by U and has the unit of energy per mass or energy per mole
of n, then 0 is called specific property. Both molar and spe- (i.e., J/mol). Internal energy represents both kinetic and po-
cific properties are intensive properties and they are related tential energies that are associated with the molecules and for
to each other t h r o u g h molecular weight. any pure substance it depends on two properties such as T and
(5.3) Molar Property = Specific Property x M V. W h e n T increases the kinetic energy increases and when
V increases the potential energy of molecules also increases
Generally t h e r m o d y n a m i c relations are developed a m o n g and as a result U increases. Another useful t h e r m o d y n a m i c
molar properties or intensive properties. However, once a mo- property that includes PV energy in addition to the internal
lar property is calculated, the total property can be calculated energy is enthalpy and is defined as
from Eq. (5.1).
The phase rule gives the m i n i m u m n u m b e r of independent (5.5) H = U + PV
variables that must be specified in order to determine ther-
m o d y n a m i c state of a system and various t h e r m o d y n a m i c where H is the molar enthalpy and has the same unit as U.
properties. This n u m b e r is called degrees of freedom and is Further definition of t h e r m o d y n a m i c properties a n d basic re-
shown by F. The phase rule was stated and formulated by the lations are presented in Chapter 6.
American physicist J. Willard Gibbs in 1875 in the following
form [1]:
5.2 PVT RELATIONS
(5.4) F = 2 + N - Fl
where rI is the n u m b e r of phases and N is the n u m b e r of For a pure c o m p o n e n t system after temperature and pressure,
components in the system. For example for a pure compo- a property that can be easily determined is the volume or
nent (N = 1) and a single phase (Fl = 1) system the degrees molar volume. According to the phase rule for single phase
of freedom is calculated as 2. This means when two intensive and pure c o m p o n e n t systems V can be determined from T
properties are fixed, the state of the systems is fixed and its and P:
properties can be determined from the two k n o w n parame-
ters. Equation (5.4) is valid for nonreactive systems. If there (5.6) V = f~(T, P)
are some reactions a m o n g the components of the systems, de-
grees of freedom is reduced by the n u m b e r of reactions within where V is the molar volume and fl represents functional
the system. If we consider a pure gas such as methane, at least relation between V, T, and P for a given system. This equation
two intensive properties are needed to determine its thermo- can be rearranged to find P as
dynamic properties. The m o s t easily measurable properties
(5.7) P = f2(T, V)
are temperature (T) and pressure (P). N o w consider a mix-
ture of two gases such as methane and ethane with mole frac- where the forms of functions fl and f2 in the above two rela-
tions xl and x2 (x2 = 1 - xl). According to the phase rule three tions are different. Equation (5.6) for a mixture of N compo-
properties must be known to fix the state of the system. In ad- nents with k n o w n composition is written as
dition to T and P, the third variable could be mole fraction of
one of the components (xl or x2). Similarly, for a mixture with (5.8) P = f3(T, V,.~l,X2 ..... XN_I)
single phase and N components the n u m b e r of properties that
m u s t be k n o w n is N + 1 (i.e., T, P, xl, x2 ..... XN-1). When the where x4 is the mole fraction of c o m p o n e n t i. Any mathemat-
n u m b e r of phases is increased the degrees of freedom is de- ical relation between P, V, and T is called an equation of state
creased. For example, for a mixture of certain a m o u n t of ice (EOS). As will be seen in the next chapter, once the PVT rela-
and liquid water (H = 2, N = 1) from Eq. (5.4) we have F = 1. tion is k n o w n for a system all t h e r m o d y n a m i c properties can
This means when only a single variable such as temperature be calculated. This indicates the importance of such relations.
is k n o w n the state of the system is fixed and its properties In general the PVT relations or any other t h e r m o d y n a m i c re-
can be determined. M i n i m u m value of F is zero. A system of lation m a y be expressed in three forms of (1) mathematical
pure c o m p o n e n t with three phases in equilibrium with each equations, (2) graphs, and (3) tables. The graphical a p p r o a c h
other, such as liquid water, solid ice, and vapor, has zero de- is tedious and requires sufficient data on each substance to
grees of freedom. This means the temperature and pressure construct the graph. Mathematical or analytical forms are
of the system are fixed and only under unique conditions of the most important and convenient relations as they can be
200 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
Gas
Fusion / C
Curve',,.~/ Liquid Region / i
........ 1- .....
P SolidRegion ~ / :
%
S~ t~ O~ 0%/ 'Liquid/vapor'lTOO)o' ~s : ~ : ~ ~ <Tc ,atm ............ F .... -f :
Trip,e Point i
Vapor ',
I I
I I
vc I I
V Tf Tb Tc
FIG. 5.1--Typical PV diagram for a pure substance. (a) T
(5.9) ~ n,Pc \b-~/In,po high the pressure is and it is usually referred as a gas. When
T and P of a substance are greater than its Tc and Pc the sub-
The first and second partial derivatives of P with respect to stance is neither liquid nor vapor and it is called supercriti-
V (at constant T) may be applied to any EOS in the form cal fluid or simply fluid. However, the word fluid is generally
of Eq. (5.7) and at the critical point they should be equal to used for either a liquid or a vapor because of m a n y similari-
zero. Simultaneous solution of resulting two equations will ties that exist between these two phases to distinguish them
give relations for calculation of EOS parameters in terms of from solids.
critical constants as will be seen later in Section 5.5.1. As is seen in Fig. 5.1, lines of saturated liquid and vapor
The two-phase region in the PV diagram of Fig. 5.1 is are identical in the PT diagram. This line is also called vapor
under the envelope. As is seen from this figure the slope of an pressure (or vaporization) curve where it begins from the
isotherm in the liquid region is much greater than its slope in triple point and ends at the critical point. The saturation line
the vapor phase. This is due to the greater change of volume of between solid and liquid phase is called fusion curve while
a gas with pressure in comparison with liquids that show less between solid and vapor is called sublimation curve. In Fig.
dependency of volume change with pressure under constant 5.2 typical PT diagrams for pure substances (a) and mixtures
temperature condition. The dotted lines inside the envelope (b) are shown.
indicate percentage of vapor in a mixture of liquid and vapor, In Fig. 5.2a the freezing point temperature is almost the
which is called quality of vapor. On the saturated vapor curve same as triple point temperature but they have different
(right side) this percentage is 100% and on the saturated liq- corresponding pressures. The normal boiling point and crit-
uid curve (left side) this percentage is zero. Vapor region is ical point both are on the vaporization line. A comparison
part of a greater region called gas phase. Vapor is usually re- between PV and PT diagrams for pure substances (Figs. 5.1
ferred to a gas that can be liquefied under pressure. A vapor and 5.2a) shows that the two-phase region, which is an area in
at a temperature above Tc cannot be liquefied no matter how the PV diagram, becomes a line in the PT diagram. Similarly
5. PVT RELATIONS AND EQUATIONS OF STATE 2 0 1
and dew points curves meet each other at the critical point.
The m a i n application of PT diagram is to determine the phase
of a system u n d e r certain conditions of temperature and pres-
sure as will be discussed in Chapter 9 (Section 9.2.3). Figure / ! ' t ,<
5.1 shows that as temperature of a pure substance increases,
~++ t t t t ~ i '~ifi
at constant pressure, the following phase changes occur:
0
o
o
o
o o o
O O
By taking the second terms to the right-hand side in each Equation (5.37) indicates that value of Zc is the same for all
equation and dividing Eq. (5.24) by Eq. (5.25) we get compounds. Values of Zc given in Table 2.1 varies from 0.28
to 0.21 for most hydrocarbons. Therefore, vdW EOS signifi-
(5.26) b = Vc cantly overpredicts values of Zc (or Vc) and its performance
3
in the critical region is quite weak. Similar approaches can
By substituting Eq. (5.26) into Eq. (5.24) we get be used to determine EOS parameters and Zc for any other
9 Ec EOS. r
(5.27) a = ~Tc
Since Tc and Pc are usually available, it is common to express Since the introduction of the vdW EOS as the first cubic
parameters a and b in terms of Tc and Pc rather than Tc and equation 130 years ago, dozens of cubic EOSs have been pro-
Vc. For this reason Vc can be found from Eq. (5.21) in terms posed, many of them developed in recent decades. The math-
of Tc and Pc and replaced in the above equations. Similar ematical simplicity of a cubic EOS in calculation of thermo-
results can be obtained by a more straightforward approach. dynamic properties has made it the most attractive type of
At the critical point we have V = Vc or V - Vc = 0, which can EOS. When van der Waals introduced Eq. (5.21) he indicated
be written as follows: that parameter a is temperature-dependent. It was in 1949
when Redlich and Kwong (RK) made the first modification
(5.28) ( v - vc) 3 = 0 to vdW EOS as [11]
Application of Eq. (5.23) at Tc and Pc gives RT a
(5.38) P -- - -
V - b V ( V + b)
(5.29) V3 - b+ Pc ] V2 + Pc V - ~ = 0
where parameter a depends on temperature as ac/T ~ in
Expansion of Eq. (5.28) gives which ac is related to Tc and Pc. Parameters a and b in Eq.
(5.38) are different from those in Eq. (5.22) but they can be
(5.30) (V - Vc)3 = V3 - 3VcV2 + 3Vc2V - V3 = 0 obtained in a similar fashion as in Example 5.1 (as shown
Equations (5.29) and (5.30) are equivalent and the corre- later). The repulsive terms in Eqs. (5.38) and (5.22) are iden-
sponding coefficients for V3, V2, V 1, and V~ must be equal tical. Performance of RK EOS is much better than vdW EOS;
in two equations. This gives the following set of equations for however, it is mainly applicable to simple fluids and rare gases
the coefficients: s u c h as Kr, CH4, o r O2, but for heavier and complex com-
pounds it is not a suitable PVT relation.
(5.31) - ( b + RTc~ = -3Vc coefficients of V2 The RK EOS is a source of many modifications that began
\ Pc/ in 1972 by Soave [12]. The Soave modification of Redlich-
(5.32) ~a = 3Vc2 coefficients of V Kwong equation known as SRK EOS is actually a modifica-
tion of parameter a in terms of temperature. Soave obtained
ab parameter a in Eq. (5.38) for a number of pure compounds us-
(5.33) - -V3 coefficients of V~
Pc ing saturated liquid density and vapor pressure data. Then he
By dividing Eq. (5.33) by (5.32), Eq. (5.26) can be obtained. By correlated parameter a to reduced temperature and acentric
substituting Vc = 3b (Eq. 5.26) to the right-hand side (RHS) factor. Acentric factor, co, defined by Eq. (2.10) is a parameter
of Eq. (5.31) the following relation for b is found: that characterizes complexity of a molecule. For more com-
plex and heavy compounds value of o~ is higher than simple
RTc molecules as given in Table 2.1. SRK EOS has been widely
(5.34) b = --
sic used in the petroleum industry especially by reservoir engi-
Combining Eqs. (5.26) and (5.34) gives neers for phase equilibria calculations and by process engi-
neers for design calculations. While RK EOS requires Tc and
3 RTc
(5.35) Vc - Pc to estimate its parameters, SRK EOS requires an additional
8Pc
parameter, namely a third parameter, which in this case is oJ.
Substituting Eq. (5.35) into Eq. (5.27) gives As it will be seen later that while SRK EOS is well capable of
9 (3RTc'~ _ 27R2Tc2 calculating vapor-liquid equilibrium properties, it seriously
/2 underestimates liquid densities.
R T c \ 8Pc ] 64Pc
Another popular EOS for estimation of phase behavior
Therefore, the final relation for parameter a in terms of Tc and and properties of reservoir fluids and hydrocarbon systems
Pc is as follows: is Peng-Robinson (PR) proposed in the following form [13]:
27R2T 2 RT a
(5.36) a - - - (5.39) P - --
64Pc V - b V ( V + b) + b(V - b)
In calculation of parameters a and b unit of R should be con- where a and b are the two parameters for PR EOS and are
sistent with the units chosen for Tc and Pc. Another useful calculated similar to SRK parameters. Parameter a was cor-
result from this analysis is estimation of critical compress- related in terms of temperature and acentric factor and later it
ibility factor through Eq. (5.35). Rearranging this equation was modified for properties of heavy hydrocarbons [14]. The
and using definition of Zc from Eq. (2.8) gives original idea behind development of PR EOS was to improve
PcVc 3 liquid density predictions. The repulsive term in all four cubic
(5.37) Zc - - - 0.375 equations introduced here is the same. In all these equations
RTc 8
206 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
T A B L E 5.1--Constants in Eq. (5.40) for four common cubic EOS (with permission from Ref. [15]).
Equation /A1 b/2 ac ~ b Zc
vdW 0 0 27 RzT~
Pe 1 ~RZ 0.375
RI~ 1 0
Pc
0"42748R2
T2 T r 1/2 ~
c
0.333
, 0 04274s.2pc 0.333
f,0 = 0.48 + 1.574c0 - 0.176w 2
Z2 = 2 ~ / - ~ Cos 7 0 + 240 ~
)al 3 bsRx = 0.08664 x 83.14 • 568.7 = 164.52 c m 3 / m o l .
24.9
(5.47)
P a r a m e t e r s A a n d B are c a l c u l a t e d f r o m Eq. (5.42):
Z3 = 2 - , r Cos (31-0) a13
3.957 x 107 x 19.9
L A= = 0.373
w h e r e Cos 0 - (83.14) 2 • (552.65) 2
and
w h e r e 0 is in degrees. To c h e c k v a l i d i t y of t h e s o l u t i o n , t h e
164.52 x 19.9
t h r e e r o o t s m u s t satisfy t h e f o l l o w i n g r e l a t i o n s B = = 0.07126
83.14 • 552.65
Z 1 + Z2 + Z3 = - a l
Coefficients al, a2, a n d a3 are c a l c u l a t e d f r o m Eq. (5.49) as
(5.48) Z1 • Z2 + 22 • Z3 + Z3 • Zl = a2
al = - ( 1 + 0.07126 - 1 x 0.07126) -- - 1
Z 1 • Z2 • Z3 = -a 3
a2 = 0.373 + 0 • 0.071262 - 1 • 0.07126 - 1 x 0.071262
A c o m p a r i s o n of Eq. (5.42) a n d (5.43) i n d i c a t e s t h a t t h e
f o l l o w i n g r e l a t i o n s exist b e t w e e n coefficients ai(s) a n d E O S = 0.29664
parameters a3 = - 0 . 3 7 3 0 5 • 0.07126 - 0 x 0.071262 - 0 x 0.071263
al = - ( I + B - ulB)
---- - 0 . 0 2 6 5 8 4
(5.49) a2 = A + u z B 2 - u l B - u a B 2
F r o m Eq. (5.44), Q = - 0 . 0 1 2 2 3 , L --- 8.84 • 10 -4, a n d D =
a3 = - A B - u2 B2 - u2 B3 - 1.048 x t 0 -6. S i n c e D < 0, t h e s o l u t i o n is g i v e n b y Eq. (5.47).
F o r t h e c a s e t h a t t h e r e are t h r e e d i f f e r e n t real r o o t s (D < 0), 0 = Cos-1(8.84 x 1 0 - 4 / x / - ( - 0 . 0 1 2 2 3 ) 3) = 492 ~ a n d t h e r o o t s
Z liq is e q u a l to t h e l o w e s t r o o t (Z1) w h i l e Z Vap is e q u a l to t h e a r e Z1 = 0.17314, Z2 = 0.28128, a n d Z3 = 0.54553. Accept-
h i g h e s t r o o t (Z3). T h e m i d d l e r o o t (Z2) is d i s r e g a r d e d as phys- a b l e r e s u l t s a r e t h e l o w e s t a n d h i g h e s t r o o t s w h i l e t h e in-
ically m e a n i n g l e s s . E q u a t i o n (5.42) m a y also be solved b y suc- t e r m e d i a t e r o o t is n o t useful: Z L = Z1 = 0.17314 a n d Z v =
cessive s u b s t i t u t i o n m e t h o d s ; however, a p p r o p r i a t e f o r m s of Z2 = 0.54553. M o l a r v o l u m e , V , c a n b e c a l c u l a t e d f r o m Eq.
t h e e q u a t i o n a n d initial v a l u e s a r e d i f f e r e n t for v a p o r a n d liq- (5.15): V = ZRT/P in w h i c h T = 552.65 K, P = 19.9 bar, a n d
u i d cases. F o r e x a m p l e , for gases t h e b e s t initial v a l u e for Z is R = 83.14 c m 3 - b a r / m o l . K ; t h e r e f o r e , V L = 399.9 c m 3 / m o l
1 w h i l e for l i q u i d s a g o o d initial g u e s s is b P / R T [ 1 ]. S o l u t i o n of a n d V v ---- 1259.6 c m 3 / m o l . F r o m Table 5.1, Zc = 0.333 a n d
c u b i c e q u a t i o n s t h r o u g h Eq. (5.42) is s h o w n in t h e f o l l o w i n g Vc is c a l c u l a t e d f r o m Eq. (2.8) as Vc = (0.333 x 83.14 x
example. 568.7)/24.9 = 632.3 c m 3 / m o l . E r r o r s for V L, V v, a n d Vc a r e
31.5, 3.6, a n d 30%, respectively. It s h o u l d b e n o t e d t h a t Zc
E x a m p l e 5 . 2 - - E s t i m a t e m o l a r v o l u m e of s a t u r a t e d l i q u i d a n d c a n also b e f o u n d f r o m t h e s o l u t i o n of c u b i c e q u a t i o n w i t h
v a p o r f o r n - o c t a n e at 279.5~ a n d p r e s s u r e of 19.9 b a r f r o m T = Tc a n d P = Pc. H o w e v e r , for this case D > 0 a n d t h e r e is
t h e RK, S R K , a n d P R c u b i c E O S . Values of V L a n d V v ex- o n l y o n e s o l u t i o n w h i c h is o b t a i n e d by Eq. (5.45) w i t h s i m i l a r
t r a c t e d f r o m t h e e x p e r i m e n t a l d a t a are 304 a n d 1216 c m 3 / m o l , answer. As is s e e n in this e x a m p l e , l i q u i d a n d c r i t i c a l v o l u m e s
r e s p e c t i v e l y [18]. Also e s t i m a t e t h e critical v o l u m e . a r e g r e a t l y o v e r e s t i m a t e d . S u m m a r y of results for all f o u r cu-
bic e q u a t i o n s a r e g i v e n in Table 5.2. t
Solution--To use SRK and PR EOS pure component data
for n-C8 a r e t a k e n f r o m Table 2.1 as Tc = 295.55~ (568.7 K), 5.5.3 Volume Translation
Pc = 24.9 bar, co = 0.3996, a n d Vc = 486.35 c m 3 / m o l . W h e n T is
in K, P is in bar, a n d V is i n c m 3 / m o l , v a l u e of R f r o m S e c t i o n In practice the SRK and PR equations are widely used for VLE
1.7.24 is 83.14 c m 3 9b a r / m o l . K. S a m p l e c a l c u l a t i o n is s h o w n c a l c u l a t i o n s in i n d u s t r i a l a p p l i c a t i o n s [ 19-21 ]. H o w e v e r , t h e i r
h e r e for S R K E O S . Tr = (279.5 + 2 7 3 . 1 5 ) / 5 6 8 . 7 = 0.972. ability to p r e d i c t v o l u m e t r i c d a t a e s p e c i a l l y for l i q u i d s y s t e m s
TABLE 5.2--Prediction o f saturated liquid, vapor and critical molar volumes for n-octane in
Example 5.2.
Equation VL, cma/mol %D Vv, cma/mol %D Vc, cm3/mol %D
Data* 304.0 ... 1216.0 ... 486.3
RK 465.9 53.2 1319.4 8.5 632.3 i6"
SRK 399.9 31.5 1259.6 3.6 632.3 30
PR 356.2 17.2 1196.2 -1.6 583.0 19.9
Source: V L and Vv from Ref. [18]; Vc from Table 2.1.
208 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
is weak. Usually the SRK equation predicts densities more ac- 5.5.4 Other Types of Cubic Equations of State
curately for compounds with low acentric values, while the
In 1972 Saove for the first time correlated parameter a in a
PR predicts better densities for compounds with acentric fac-
cubic EOS to both Tr and co as given in Table 5.1. Since then
tors near 0.33 [21]. For this reason a correction term, known
this approach has been used by many researchers who tried
as volume translation, has been proposed for improving vol-
to improve performance of cubic equations. Many modifica-
umetric prediction of these equations [8, 15, 19, 22]:
tions have been made on the form of f~ for either SRK or
(5.50) V = V E~ - c PR equations. Graboski and Daubert modified the constants
in the f~ relation for the SRK to improve prediction of va-
where c is the volume translation parameter and has the same por pressure of hydrocarbons [24]. Robinson and Peng [14]
unit as the molar volume. Equation (5.50) can be applied to also proposed a modification to their fo~ equation given in
both vapor and liquid volumes. Parameter c mainly improves Table 5.1 to improve performance of their equation for heav-
liquid volume predictions and it has no effect on vapor pres- ier compounds. They suggested that for the PR EOS and for
sure and VLE calculations. Its effect on vapor volume is negli- compounds with ~o > 0.49 the following relation should be
gible since Vv is very large in comparison with c, but it greatly used to calculate f~:
improves prediction of liquid phase molar volumes. Values of
(5.53) fo, = 0.3796 + 1.485co - 0.1644o92 + 0.01667co 3
c have been determined for a number of pure components up
to C10 for both SRK and PR equations and have been included Some other modifications give different functions for param-
in references in the petroleum industry [19]. Peneloux et al. eter ot in Eq. (5.41). For example, Twu et al. [25] developed
[22] originally obtained values of c for some compounds for the following relation for the PR equation.
use with the SRK equation. They also suggested the following
ct = Tr 0"171813exp [0.125283 (I - Tr1"77634)]
correlation for estimation of c for SRK equation:
(5.54) + (.0 {Tr-0"607352exp [0.511614 (I - V2~
Rrc
(5.51) c = 0.40768 (0.29441 - ZRA)- -
Pc - Tr ~ exp [0.125283 (1 - T)77634)]}
where ZRA is the Rackett parameter, which will be discussed Other modifications of cubic equations have been derived by
in Section 5.8.1. Similarly Jhaveri and Yougren [23] obtained suggesting different integer values for parameters ui and uz in
parameter c for a number of pure substances for use with PR Eq. (5.40). One can imagine that by changing values of ul and
EOS and for hydrocarbon systems have been correlated to u2 in Eq. (5.40) various forms of cubic equations can be ob-
molecular weight for different families as follows: tained. For example, most recently a modified two-parameter
cubic equation has been proposed by Moshfeghian that cor-
Cp = bpR (1 -- 2.258M~-~ responds to ut = 2 and u2 = - 2 and considers both parame-
(5.52) CN = bva (1 -- 3.004M~ ~ ters a and b as temperature-dependent [26]. Poling et al. [8]
have summarized more than two dozens types of cubic equa-
CA= bpR (1 -- 2.516M~ ~176176
tions into a generalized equation similar to Eq. (5.42). Some
where bpR refers to parameter b for the PR equation as given of these modifications have been proposed for special sys-
in Table 5.1. Subscripts P, N, and A refer to paraffinic, naph- tems. However, for hydrocarbons systems the original forms
thenic, and aromatic hydrocarbon groups. The ratio of c/b of SRK and PR are still being used in the petroleum industry.
is also called shift parameter. The following example shows The most successful modification was proposed by Zudke-
application of this method. fitch and Joffe [27] to improve volumetric prediction of RK
EOS without sacrificing VLE capabilities. They suggested that
parameter b in the RK EOS may be modified similar to Eq.
Example 5.3--For the system of Example 5.2, estimate VL
(5.41) for parameter a as following:
and Vv for the PR EOS using the volume translation method.
(5.55) b = bRKfl
Solution--For n-Cs, from Table 2.1, M = 114 and beR are where fl is a dimensionless correction factor for parameter
calculated from Table 5.1 as 147.73 cm3/mol. Since the hy- b and it is a function of temperature. Later Joffe et al. [28]
drocarbon is paraffinic Eq. (5.51) for Cp should be used, determined parameter 0~ in Eq. (5.41) and fl in Eq. (5.55) by
c = 7.1 cma/mol. From Table 5.2, V L(PR) = 356.2 and Vv(vR) = matching saturated liquid density and vapor pressure data
1196.2 cma/mol. From Eq. (5.50) the corrected molar volumes over a range of temperature for various pure compounds. In
are VL = 356.2 - 7.1 = 3491 and Vv = 1196.2 - 7.1 = 1189.1 this approach for every case parameters a and/~ should be
cm3/mol. By the volume translation correction, error for VL determined and a single dataset is not suitable for use in all
decreases from 17.2 to 14.8% while for Vv it has lesser effect cases. SRK and ZJRK are perhaps the most widespread cu-
and it increases error from -1.6% to -2.2%. r bic equations being used in the petroleum industry, especially
for phase behavior studies of reservoir fluids [19]. Other re-
As is seen in this example improvement of liquid volume searchers have also tried to correlate parameters oe and fl in
by volume translation method is limited. Moreover, estima- Eqs. (5.41) and (5.55) with temperature. Feyzi et al. [29] cor-
tion of c by Eq. (5.51) is limited to those compounds whose related a 1/2 and fl1/2 for PR EOS in terms Tr and co for heavy
ZRAis known. With this modification at least four parameters reservoir fluids and near the critical region. Their correla-
namely Tc, Pc, co, and c must be known for a compound to tions particularly improve liquid density prediction in com-
determine its volumetric properties. parison with SRK and PR equations while it has similar VLE
5. P V T R E L A T I O N S A N D E Q U A T I O N S O F S T A T E 209
N
0.8~ O Z~ ,exp ~
& Zv,exp ~&.&
o Zc, e)q:, ~
0.6
Q z~.~ ~,
Z1, calc
0.4' ---"Z% Ca|C
0.2"
0.0 '
A~ A A ~
L4 o.s E6 0:8 0.9
R e d u c e d T e m p e r a t u r e , Tr
FIG. 5.10--Prediction of saturation curves for ethane using a modified PR EOS [29].
prediction capabilities. Another improvement in their corre- where V/t is the total volume occupied by i at T and P of the
lation was prediction of saturation curves near the critical mixture. For the whole mixture this equation becomes
region. This is shown for prediction of compressibility fac-
(5.57) P V t = ntRT
tor of saturated liquid and vapor curves as well as the critical
point for methane and ethane in Fig. 5.10. where Vt is the total volume of mixture (V t = EV/t) and nt is
Prediction of isotherms by a cubic EOS is shown on PV di- the total number of moles (nt -- En/). By dividing Eq. (5.56)
agram in Fig. 5.9. As shown in this figure in the two-phase by Eq. (5.57) we get
region the prediction of isotherm is not consistent with true
r~ Viit
behavior of the isotherm. In addition, performance of these (5.58) Yi -- --
n Vt
cubic equations in calculation of liquid densities and derived
thermodynamic properties such as heat capacity is weak. This where yi is the mole fraction of i in the gas mixture. The
indicates the need for development of other EOS. Further in- above equation indicates that in an ideal gas mixture the mole
formation on various types of cubic EOS and their character- fractions and volume fractions are the same (or mol% of i =
istics are available in different sources [30-34]. vol% of i). This is an assumption that is usually used for gas
mixtures even when they are not ideal.
For nonideal gas mixtures, various types of mixing rules
5.5.5 A p p l i c a t i o n t o Mixtures for determining EOS parameters have been developed and
Generally when a PVT relation is available for a pure sub- presented in different sources [6, 8]. The mixing rule that is
stance, the mixture property may be calculated in three ways commonly used for hydrocarbon and petroleum mixtures is
when the mixture composition (mole fraction, x~) is known. called quadratic mixing rule. For mixtures (vapor or liquid)
The first approach is to use the same equation developed for with composition xi and total of N components the following
pure substances but the input parameters (To, Pc, and to) are equations are used to calculate a and b for various types of
estimated for the mixture. Estimation of these pseudocritical cubic EOS:
properties for petroleum fractions and defined hydrocarbon N N
mixtures were discussed in Chapter 3. The second approach (5.59) amix = ~ ] ~ xixjai j
i=1 j = l
is to estimate desired physical property (i.e., molar volume
or density) for all pure compounds using the above equations N
and then to calculate the mixture property using the mixture (5.60) bmix = ~ x q b i
i=1
composition through an appropriate mixing rule for the prop-
erty (i.e., Eq. (3.44) for density). This approach in some cases where aij is given by the following equation:
gives good estimate of the property but requires large calcu-
lation time especially for mixtures containing many compo- (5.61) aq = (aiai)l/2(1 - kq)
nents. The third and most widely used approach is to calculate For the volume translation c, the mixing rule is the same as
EOS parameters (parameters a and b) for the mixture using for parameter b:
their values for pure components and mixture composition. N
The simplest EOS for gases is the ideal gas law given by Eq. (5.62) Cmi~ = y~xici
(5.14). When this is applied to component i with n~ moles in i=1
a mixture we have
In Eq. (5.61), kq is a dimensionless parameter called binary in-
(5.56) PV//t = n~RT teractionparameter (BIP), where kqi = 0 andkq = kji. For most
210 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
hydrocarbon systems, k/l = 0; however, for the key hydrocar- of Eq. (5.59) for a ternary gas mixture (N = 3) becomes
bon compounds in a mixture where they differ in size value
of k/i is nonzero. For example, for a reservoir fluid that con- amix "---y2all + y2a22 + y2a33 + 2ylY2a12 + 2ylx3a13 + 2y2Yaa23
tains a considerable amount of methane and C7+ the BIP for (5.64)
C1 and C7 fractions cannot he ignored. For nonhydrocarbon-
hydrocarbon pairs k/i values are nonzero and have a signif- where all = al, a22 -----a2, and a33 -----a3. Interaction coefficients
icant impact on VLE calculations [20, 35]. Values of k/i for such as a12 can be found from Eq. (5.61): a12 = a~d]-~(1 - k12)
a particular pair may be determined from matching exper- w h e r e k12 may be taken from Table 5.3 or estimated from Eq.
imental data with predicted data on a property such as va- (5.63). a13 and a23 can be calculated in a similar way.
por pressure. Values of k/i are specific to the particular EOS
being used. Some researchers have determined k/j for SRK
or PR equations. Values of BIP for N2, CO2, and methane 5.6 N O N C U B I C EQUATIONS OF STATE
with components in reservoir fluids from C1 to C6 and three
subfractions of C7+ for PR and SRK are tabulated by Whit- The main reason for wide range application of cubic EOS
son [19]. Values that he has recommended for use with SRK is their application to both phases of liquids and vapors,
and PR equations are given in Table 5.3. There are some gen- mathematical simplicity and convenience, as well as pos-
eral correlations to estimate BIPs for any equation [36, 37]. sibility of calculation of their parameters through critical
The most commonly used correlation for estimating BIPs constants and acentric factor. However, these equations are
of hydrocarbon-hydrocarbon (HC-HC) systems is given by mainly useful for density and phase equilibrium calculations.
Chueh and Prausnitz [37]: For other thermodynamic properties such as heat capacity
and enthalpy, noncubic equations such as those based on
statistical associating fluid theory (SAFT) or perturbed hard
(5.63) kq -----A - (V~i)l/3 + (V)/3)J /
chain theory (PHCT). Some of these equations have been
particularly developed for special mixtures, polar molecules,
where Vci and V~j are critical molar volume of components hard sphere molecules, and near critical regions. Summary
i and j in cm3/mol. Originally A = 1 and B = 3; however, in of these equations is given by Poling et al. [8]. In this sec-
practical cases B is set equal to 6 and A is adjusted to match tion three important types of noncubic EOS are presented:
saturation pressure and other variable V I E data [20, 38]. For (I) virial, (2) Carnahan-Starling, and (3) modified Benedict-
most reservoir fluids, A is within 0.2-0.25; however, as is seen Webb-Rubin.
in Chapter 9 for a Kuwaiti oil value of A was found as 0.18.
As discussed by Poling et al. [8], Tsonopoulos recommends
the original Chueh-Prausnitz relation (A = 1 and B = 3) for 5.6.1 Virial Equation o f State
nonpolar compounds. Pedersen et al. [39] proposed another
The most widely used noncubic EOS is the virial equation or
relation for calculation of BIPs for HC-HC systems. Their cor-
relation is based on data obtained from North Sea reservoir its modifications. The original virial equation was proposed
in 1901 by Kammerlingh-Onnes and it may be written either
fluids and it is related to molecular weights of components
in the form of polynomial series in inverse volume (pressure
i and ] as k4i ~ O.O01MilMi where Mi > Mj. Another corre-
lation was proposed by Whitson [40] for estimation of BIPs explicit) or pressure expanded (volume explicit) as follows:
of methane and C7+ fraction components based on the data B C D
presented by Katz and Firoozabadi [36] for use with PR EOS. (5.65) Z = I+ V + ~ + V~ + "
His correlation is as: k U = 0.14 SGi - 0.0688, where 1 refers
to methane and j refers to the CT+(j) fraction, respectively. Z--l+
Equations (5.59)-(5.62) can be applied to either liquid or
vapor mixtures. However, for the case of vapor mixtures with
N components, mole fraction yi should be used. Expansion
(5.66) + ( D - 3 B~C- ~ + 2B3\
\
)m3+ .
5. P V T R E L A T I O N S A N D E Q U A T I O N S O F S T A T E 211
TABLE 5.4--Second virial coefficients for several gases [41~ A n o t h e r form of Eq. (5.59) for calculation of Bm~ c a n be writ-
Temperature, K ten as following:
Compound 200 300 400 500
N INN
N2 -35.2 -4.2 9 16.9
CO2 - 122.7 60.5 -29.8 Bmix = E y i B i i + "~ Z EyiYjSi/ w h e r e 8i/~-2Bij -Bii - B i i
CH4 -i05 -42 -15 -0.5 i=l i=l j=l
C2H6 -410 -182 -96 -52 (5.70)
C3H8 ... -382 -208 - 124
Note: Valuesof B are given in cma/mol. There are several correlations developed b a s e d on the the-
ory of c o r r e s p o n d i n g state principles to e s t i m a t e the s e c o n d
virial coefficients in t e r m s of t e m p e r a t u r e . S o m e of these rela-
w h e r e B, C, D .... are called second, third, a n d fourth virial co-
tions correlate B/Vc to Tr a n d o9. P r a u s n i t z et al. [6] reviewed
efficients a n d they are all t e m p e r a t u r e - d e p e n d e n t . The above
s o m e of these relations for e s t i m a t i o n of the s e c o n d virial
two forms of virial e q u a t i o n are the s a m e a n d the s e c o n d
coefficients. The r e l a t i o n developed b y Tsonopoulos [42] is
e q u a t i o n can b e derived f r o m the first e q u a t i o n (see P r o b l e m
useful to e s t i m a t e B f r o m To, Pc, a n d o9.
5.7). The s e c o n d f o r m is m o r e p r a c t i c a l to use since usually T
a n d P are available a n d V s h o u l d be estimated. The n u m b e r BPc = B(0) + OgB(1)
of t e r m s in a virial EOS can be e x t e n d e d to infinite t e r m s RTc
b u t c o n t r i b u t i o n of h i g h e r t e r m s reduces w i t h increase in 0.330 0.1385 0.0121 0.000607
B (~ = 0.1445
p o w e r of P. Virial e q u a t i o n is p e r h a p s the m o s t a c c u r a t e PVT r/ r3 T/
r e l a t i o n for gases. However, the difficulty w i t h use of virial 0.331 0.423 0.008
e q u a t i o n is availability of its coefficients especially for h i g h e r B (1) = 0.0637 + - -
Tr2 Tr3 Trs
terms. A large n u m b e r of d a t a are available for the s e c o n d
virial coefficient B, b u t less d a t a are available for coefficient (5.71)
C a n d very few d a t a are r e p o r t e d for the fourth coefficient w h e r e Tr = T/Tc. There are s i m p l e r relations t h a t c a n be u s e d
D. Data on values of virial coefficients for several c o m p o u n d s for n o r m a l fluids [ 1].
a r e given in Tables 5.4 a n d 5.5. The virial coefficient has firm
basis in t h e o r y a n d the m e t h o d s of statistical m e c h a n i c s allow BPc = B(O) + ogBO)
d e r i v a t i o n of its coefficients. RTc
0.422
B r e p r e s e n t s t w o - b o d y interactions a n d C r e p r e s e n t e d (5.72) B (~ = 0.083 -
t h r e e - b o d y interactions. Since the c h a n c e of t h r e e - b o d y in- Tr1-6
t e r a c t i o n is less t h a n t w o - b o d y interaction, therefore, the im- 0.172
B (1) = 0.139 - - -
p o r t a n c e a n d c o n t r i b u t i o n of B is m u c h greater t h a n C. F r o m Tr4.2
q u a n t u m m e c h a n i c s it can be s h o w n that the s e c o n d virial
A n o t h e r relation for p r e d i c t i o n of s e c o n d virial coefficients of
coefficient c a n be calculated from the k n o w l e d g e of p o t e n t i a l
simple fluids is given b y M c G l a s h a n [43]:
function (F) for i n t e r m o l e c u l a r forces [6]:
oo (5.73) BPc = 0.597 - 0.462e ~176176
(5.67) S = 2zrNA ] (1 -- e-r(r)/kr)r2dr RTc
0 A g r a p h i c a l c o m p a r i s o n of Eqs. (5.71)-(5.73) for p r e d i c t i o n
of s e c o n d virial coefficient of ethane is s h o w n in Fig. 5.11.
w h e r e NA is the Avogadro's n u m b e r (6.022 x 1023 mo1-1) a n d Coefficient B at low a n d m o d e r a t e t e m p e r a t u r e s is negative
k is the Boltzman's c o n s t a n t (k = R/NA). Once the relation for a n d increases with increase in t e m p e r a t u r e ; however, as is
F is known, B c a n be estimated. F o r example, if the fluid seen from the above correlations as T --+ ~ , B a p p r o a c h e s a
follows h a r d sphere p o t e n t i a l function, one by substituting positive number.
Eq. (5.13) for F into the above e q u a t i o n gives B = (2/3)zr NA33. To p r e d i c t Bm~, for a m i x t u r e of k n o w n c o m p o s i t i o n , the in-
Vice versa, constants in a p o t e n t i a l r e l a t i o n (e a n d a ) m a y t e r a c t i o n coefficient Bi/is needed. This coefficient can be cal-
be e s t i m a t e d f r o m the knowledge of virial coefficients. F o r culated f r o m / ~ a n d Bi/using the following relations [1, 15]:
mixtures, Bmi• c a n b e calculated from Eq. (5.59) with a being
r e p l a c e d b y B. F o r a t e r n a r y system, B can be calculated f r o m B~/ = ~RTcii (B(O) + ogiiB(1))
Eq. (5.64). Bii is calculated from Eq. (5.67) using Fii with r
a n d eli given as [6] B (~ a n d B (1) are calculated t h r o u g h T~i = T/Tcij
1
(5.68) a# = ~(ai + a/) o9~+ w/
o9i/-- 2
(5.69) eli = (eiSi) 1/2 Tcij = (TdTcj)l/2(1 - kii)
(5.74)
TABLE 5.5--Sample values of different virial coefficients for Pii - Zij RTc~i
several compounds [1]. Vc#
Compound T, ~ B, cma/mol C, cm6/mol2
z~ + zc/
Methane a 0 -53.4 2620 Zci/ - 2
Ethane 50 - 156.7 9650
Steam (H20) 250 -152.5 -5800
Sulfur dioxide (SO2) 157.5 -159 9000
aFor m e t h a n e at 0~ the fourth virial coefficient D is 5000 cmg/mol3, 2
212 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
o3
-450
100
f
? ~
300
' ~
500 700
' P ~
900
'
1100
(5.78)
0.02432
C ~~ = 0.01407 + - -
0.0177
r/8
0.040
0.00313
Ty5
0.003 0.00228
C (1) = -0.02676 + ~ + T~ Tr6 Trl~
Temperature, K
where C (~ and C (~ are dimensionless parameters for simple
FIG. 5.11--Prediction of second virial coefficient for
ethane from different methods. Experimental data and correction terms in the generalized correlation. Estima-
from Table 5.4: McGlashan, Eq. (5.73); Normal fluids, tion of the third virial coefficients for mixtures is quite diffi-
Eq. (5.72); Tsopoulos, Eq. (5.71). cult as there are three-way interactions for C and it should be
calculated from [6]:
where k/i is the interaction coefficient and for hydrocarbons of (5.79) Cmix- E E E y i y j y k C q k
similar size it is zero. B ~~ and B ~x)must be calculated from the
Methods of estimation of cross coefficients Ciik are not re-
same relations used to calculate Bu and Bii. Another simpler
liable [6]. For simplicity, generally it is assumed that Ciii =
m e t h o d that is fairly accurate for light, nonpolar gases is the
Ciii = Ciii but still for a binary system at least two cross coef-
geometric mean:
ficients of Cl12 and C~22 must be estimated. In a binary system,
Bij = (ninj) 1/2 Cl12 expresses interaction of two molecules of c o m p o n e n t 1
with one molecule of c o m p o n e n t 2. Orbey and Vera [44] sug-
gest the following relation for calculation of Ciik as
Bmixm(i~=lYiB1/2)2
(5.80) Cij k : (CijCikCjk) 1/3
The importance of these relations is that at moderate pres-
sures, Eq. (5.66) m a y be truncated after the second term as where C# is evaluated from Eq. (5.78) using Tr Pcij and ogii
follows: obtained from Eq. (5.74). This approach gives satisfactory
estimates for binary systems.
BP
(5.75) z = 1 + -- There are certain specific correlations for the virial coeffi-
RT cients of some specific gases. For example, for hydrogen the
This equation is usually referred to as the truncated virial following correlations for B and C are suggested [6]:
equation and m a y be used with a reasonable degree of accu- g
racy in certain ranges of reduced temperature and pressure: B = ~ bi X(2i-1)/4
Vr > 2.0 [i.e., (Pr < 0.5, Tr > 1), (0.5 < Pr < 1, Tr > 1.2), (1 < 1
Pr < 1.7, Tr > 1.5)]. At low-pressure range (Pr < 0.3), Eq. C = 1310.5x 1/2 (1 + 2.1486x 3) • [i - exp (1 - x-3)]
(5.72) provides good prediction for the second virial coeffi- (5.81)
cients for use in Eq. (5.75) [1]. 109.83 b
where x - ~- , a = 42.464, bz = -37.1172,
A more accurate form of virial equation for gases is ob-
tained when Eq. (5.65) or (5.66) are truncated after the third b3 = -2.2982, and b4 = - 3 . 0 4 8 4
term:
where T is in K, B is in cm3/mol, and C is in cm6/mol 2. The
B C
(5.76) Z = 1 + ff + V-5 range of temperature is 15-423 K and the average deviations
for B and C are 0.07 cm3/mol and 17.4 cm6/mo] 2, respectively
An equivalent form of this equation in terms of P can be ob- [6].
tained by Eq. (5.66) with three terms excluding fourth virial As determination of higher virial coefficients is difficult,
coefficient and higher terms. Because of lack of sufficient application of truncated virial EOS is mainly limited to gases
data, a generalized correlation to predict the third virial co- and for this reason they have little application in reservoir
efficient, C, is less accurate and is based on fewer data. The fluid studies where a single equation is needed for both liquid
generalized correlation has the following form [6]: and vapor phases. However, they have wide applications in
estimation of properties of gases at low and moderate pres-
C _ (0.232Tr_0.25 + 0.468Tr5 ) x [1 - e(1-1"S9T2)]
sures. In addition, special modifications of virial equation has
industrial applications, as discussed in the next section. F r o m
(5.77) + d e - ( 2"49-2'30Tr § )
mathematical relations it can be shown that any EOS can be
5. P V T R E L A T I O N S A N D E Q U A T I O N S O F S T A T E 213
converted into a virial form. This is s h o w n by the following Considering the fact that a is a t e m p e r a t u r e - d e p e n d e n t
example. p a r a m e t e r one c a n see that the virial coefficients are all
t e m p e r a t u r e - d e p e n d e n t parameters. With use of SRK EOS,
E x a m p l e 5 . 4 - - C o n v e r t RK EOS into the virial form a n d ob- similar coefficients are o b t a i n e d b u t p a r a m e t e r a also depends
tain coefficients B a n d C i n terms of EOS parameters. on the acentric factor as given in Table 5.1. This gives bet-
ter estimation of the second a n d third virial coefficients (see
S o l u t i o n - - T h e R K EOS is given by Eq. (5.38). If both sides Problem 5.10) r
of this equation are multiplied by V/RT we get
PV V a The following example shows application of t r u n c a t e d
(5.82) Z - virial equation for calculation of vapor m o l a r volumes.
RT V - b R T ( V + b)
Assume x = b/V a n d A = alRT, t h e n the above e q u a t i o n c a n E x a m p l e 5 . 5 - - P r o p a n e has vapor pressure of 9.974 b a r at
be written as 300 K. Saturated vapor m o l a r volume is Vv = 2036.5 cma/mol
1 1 1 [Ref. 8, p. 4.24]. Calculate (a) second virial coefficient from
(5.83) Z---l_x A~x l+x Eqs. (5.71)-(5.73), (b) third virial coefficient from Eq. (5.78),
(c) Vv from virial EOS t r u n c a t e d after second term u s i n g Eqs.
Since b < V, therefore, x < 1 a n d the terms in the RHS of
(5.65) a n d (5.66), (d) Vv from virial EOS t r u n c a t e d after third
the above e q u a t i o n can be expanded t h r o u g h Taylor series
term u s i n g Eqs. (5.65) a n d (5.66), a n d (e) Vv from ideal gas
[16, 17]:
law.
(5.84) f ( x ) = s f(~)(x~
n!
(x - x~
n=0 S o l u t i o n - - ( a ) a n d (b): For p r o p a n e from Table 2.1 we
get Tc = 96.7~ (369.83 K), Pc 42.48 bar, a n d c0 = 0.1523.
where f(')(Xo) is the nth order derivative d" f(x)/dx" evaluated Tr = 0.811, Pr = 0.23, a n d R = 83.14 c m 3 9bar/mol - K. Second
at x = Xo. The zeroth derivative of f is defined to be f itself virial coefficient, B, can be estimated from Eqs. (5.71) or
a n d both 0! a n d 1! are equal to 1. Applying this expansion rule (5.72) or (5.73) a n d the third virial coefficient from Eq. (5.78).
at Xo = 0 we get: Results are given i n Table 5.6. (c) Truncated virial e q u a t i o n
i after second t e r m from Eq. (5.65) is Z = 1 + B/V, which is
-- 1 +x +x 2 ~-X 3 +X 4 -~- . . .
1--x referred to as V expansion form, a n d from Eq. (5.66) is Z =
(5.85) 1 + B P / R T , which is the same as Eq. (5.75) a n d it is referred
1
- 1 --XArX2--xa+x 4 .... to as P expansion form. For the V expansion (Eq. 5.65), V
l+x should be calculated t h r o u g h successive s u b s t i t u t i o n m e t h o d
It should be noted that the above relations are valid w h e n or from m a t h e m a t i c a l solution of the equation, while i n P
Ixl < 1. S u b s t i t u t i n g the above two relations in Eq. (5.83) we expansion form (Eq. 5.66) Z can be directly calculated from
get T a n d P. Once Z is determined, V is calculated from Eq.
(5.15): V = ZRT/P. I n part (d) virial e q u a t i o n is t r u n c a t e d
Z = (1 + x + x 2 " ~ - X 3 "~- - 9 ") - - A 1- x (1 - x +x 2 --X 3 -~-'" ') after the third term. The V expansion form reduces to Eq.
V
(5.86) (5.76). S u m m a r y of calculations for m o l a r volume is given
in Table 5.6. The results from V expansion (Eq. 5.65) a n d P
If x is replaced by its definition b/V a n d A by a/RT we have expansion (5.66) do n o t agree with each other; however, the
b - a/RT b2 + a b / R T + b 3 - ab2/RT difference between these two forms of virial e q u a t i o n reduces
Z=I+ - - + + + as the n u m b e r of terms increases. W h e n the n u m b e r of terms
V V2 V3
(5.87) becomes infinity (complete equation), t h e n the two forms
of virial e q u a t i o n give identical results for V. Obviously for
A c o m p a r i s o n with Eq. (5.65) we get the virial coefficients in t r u n c a t e d virial equation, the V expansion form, Eq. (5.65),
terms of RK EOS p a r a m e t e r s as follows: gives more accurate result for V as the virial coefficients
a ab ab 2 are originally d e t e r m i n e d from this equation. As can be seen
(5.88) B=b-R~ C=b 2+~ D=b 3--RT from Table 5.6, w h e n B is calculated from Eq. (5.71) better
TABLE 5.6--Prediction of molar volume of propane at 300 K and 9. 974 bar from virial equation with different methods for second virial
coefficient (Example 5.5).
Virialequation with two terms Virialequation with three termsa
Method of estimation of P expansionb V expansionc P expansiond V expansione
second virial coefficient(B) B, cma/mol V, cma/mol %D V, cma/mol %D V, cm3/mol %D V, cm3/mol %D
Tsonopoulos (Eq. 5.71) -390.623 2110.1 3.6 2016.2 -1.0 2056.8 1.0 2031.6 -0.2
Normal fluids (Eq. 5.72) -397.254 2103.5 3.3 2005.3 -1.5 2048.1 0.6 2021.0 -0.7
McGlashan (Eq. 5.73) -360.705 2140.0 5.1 2077.8 2.0 2095.7 2.9 2063.6 1.3
The experimentalvalue of vapor molar volumeis: V = 2036.5 cm3/mol(Ref. [8], p. 4.24).
aIn all calculationswith three terms, the third virial coefficientC is calculatedfrom Eq. (5.78) as C = 19406.21 cm6/mol2.
bTruncatedtwo terms (P expansion)refers to pressure expansionvirial equation (Eq. 5.66) truncated after secondterm (Eq. 5.75): Z = 1 + BP/RT.
CTruncatedtwo terms (V expansion)refers to volumeexpansionvirial equation (Eq. 5.65) truncated after secondterm: Z = 1 + B/V.
aTruncatedthree terms (P expansion)refers to pressure expansionvirial equation (Eq. 5.66) truncated after third term: Z = 1 + BP/RT+ (C - B2)P2/(RT)2.
eTruncatedthree terms (V expansion)refers to volumeexpansionvirial equation (Eq. 5.65)truncated after third term (Eq. 5.76): Z = 1 + B/V + C/V2.
214 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
predictions are obtained. Equation (5.72) also gives reason- volume that be can occupied by N molecules of diameter a is
able results but Eq. (5.73) gives a less accurate estimate of B.
The best result is obtained from Eq. (5.76) with Eqs. (5.71) V o N = N ( V~
and (5.78), which give a deviation of 0.2%. (e) The ideal gas \NAJ
(5.90)
law (Z - i) gives V v -- 2500.7 cm3/mol with a deviation of
+22.8%. #
Vo-~ (-~2 or3) NA
where NA is the Avogadro's n u m b e r and Vo is the volume of
5.6.2 Modified Benedict-Webb--Rubin 1 mol (NA molecules) of hard spheres as packed molecules
E q u a t i o n o f State without empty space between the molecules. VoN is the total
volume of packed N molecules. If the molar volume of fluid is
Another important EOS that has industrial application is the
V, then a dimensionless reduced density, ~, is defined in the
Benedict-Webb-Rubin (BWR) EOS [45]. This equation is in
following form:
fact an empirical expansion of virial equation. A modification
of this equation by Starling [46] has found successful applica-
tions in petroleum and natural gas industries for properties
of light hydrocarbons and it is given as
P = RT
( co Oo
+ B o R T - A o - ~-~ + T3 ~-~ V2
Parameter g is also known as packing fraction and indicates
fraction of total volume occupied by hard molecules. Substi-
tuting Vo from Eq. (5.90) into Eq. (5.91) gives the following
1 (a + 1 relation for packing fraction:
(5.89) +(bRT-a-:)~-~+a_ :)~6
+T-T-V-~v
3 I+~-~ exp
where the 11 constants Ao, Bo, ..., a, b . . . . . a and y are given The Carnahan-Starling EOS is then given as [6]
in Table 5.7 in terms of Vc, Tc, and w as reported in Ref. [21]. PV 1+~+~2-~ 3
This equation is k n o w n as BWRS EOS and m a y be used for (5.93) Z Hs = - - =
RT (1 _~)3
calculation of density of light hydrocarbons and reservoir flu-
ids. In the original BWR EOS, constants Do, Eo, and d were all where Z us is the compressibility factor for hard sphere
zero and the other constants were determined for each spe- molecules. For this EOS there is no binary constant and
cific c o m p o u n d separately. Although better volumetric data the only parameter needed is molecular diameter a for each
can be obtained from BWRS than from cubic-type equations, molecule. It is clear that as V -+ oo (P --~ 0) from Eq. (5.93)
but prediction of phase equilibrium from cubic equations are ( ~ 0 and Z ns ~ 1, which is in fact identical to the ideal
quite comparable in some cases (depending on the mixing gas law. Carnahan and Starling extended the HS equation to
rules used) or better than this equation in some other cases. fluids whose spherical molecules exert attractive forces and
Another problem with the BWRS equation is large computa- suggested two equations based on two different attractive
tion time and mathematical inconvenience to predict various terms [6]:
physical properties. To find m o l a r volume V from Eq. (5.89),
(5.94) Z = Z Hs a
a successive substitutive m e t h o d is required. However, as it RTV
will be discussed in the next section, this type of equations can or
be used to develop generalized correlations in the graphical
or tabulated forms for prediction of various thermophysical (5.95) Z = Z ~s - ~a (V - b) -1 T -t/2
properties.
where Z ns is the hard sphere contribution given by Eq.
(5.93). Obviously Eq. (5.94) is a two-parameter EOS (a, a)
5.6.3 Carnahan-Starling E q u a t i o n o f State and Eq. (5.95) is a three-parameter EOS (a, b, a). Both Eqs.
a n d Its Modifications (5.94) and (5.95) reduce to ideal gas law (Z -+ Z Hs --* 1) as
Equations of state are mainly developed based on the un- V -~ cc (or P ~ 0), which satisfies Eq. (5.18). For mixtures,
derstanding of intermolecular forces and potential energy the quadratic mixing rule can be used for parameter a while
functions that certain fluids follow. For example, for hard a linear rule can be applied to parameter b. Application of
sphere fluids where the potential energy function is given by these equations for mixtures has been discussed in recent ref-
Eq. (5.13) it is assumed that there are no attractive forces. For erences [8, 47]. Another modification of CS EOS is through
such fluids, Carnahan and Starling proposed an EOS that has LJ EOS in the following form [48, 49]:
been used extensively by researchers for development of m o r e
(5.96) Z --- Z Hs 32e~
accurate EOS [6]. For hard sphere fluids, the smallest possible 3kBT
5. PVT R E L A T I O N S A N D E Q U A T I O N S OF STATE 215
where e is the molecular energy parameter and ( (see Eq. 5.92) pressure variations from the critical point. By substituting pa-
is related to a the size parameter, e and a are two parameters rameter h into the first equation in Eq. (5.99) one can see that
in the LJ potential (Eq. 5.11) and ks is the Boltzman constant.
(5.101) Z = f(Tr, Pr)
One advanced noncubic EOS, which has received significant
attention for property calculations specially derived proper- This equation indicates that for all fluids that obey a two-
ties (i.e., heat capacity, sonic velocity, etc.), is that of SAFT parameter EOS, such as RK, the compressibility factor, Z,
originally proposed by Chapman et al. [50] and it is given in is the only function of Tr and Pr. This means that at the
the following form [47]: critical point where Tr = Pr = 1, the critical compressibility
(5.97) Z sAFT = 1 + Z Hs + Z cHAIN+ Z DIsc + Z Ass~ factor, Zc, is constant and same for all fluids (0.333 for RK
EOS). As can be seen from Table 2.1, Zc is constant only for
where HS, CHAIN, DISE and ASSOC refer to contributions simple fluids such as N2, CH4, O2, or Ar, which have Zc of
from hard sphere, chain formation molecule, dispersion, and 0.29, 0.286, 0.288, and 0.289, respectively. For this reason RK
association terms. The relations for Z Hs and Z cH~aNare simple EOS is relatively accurate for such fluids. Equation (5.101)
and are given in the following form [47]: is the fundamental of corresponding states principle (CSP) in
classical thermodynamics. A correlation such as Eq. (5.101)
zSAFr= l + r ~_~)33+(1--r) is also called generalized correlation. In this equation only
two parameters (To and Pc) for a substance are needed to
(5.98) + Z DISP -[- Z AssOC determine its PVT relation. These types of relations are usu-
ally called two-parameter corresponding states correlations
where r is a specific parameter characteristic of the substance (CSC). The functionality of function f in Eq. (5.101) can be
of interest. ( in the above relation is segment packing fraction determined from experimental data on PVT and is usually
and is equal to ( from Eq. (5.92) multiplied by r. The relations expressed in graphical forms rather than mathematical
for Z DIsP and Z gss~ are more complex and are in terms of equations. The most widely used two-parameter CSC in a
summations with adjusting parameters for the effects of asso- graphical form is the Standing-Katz generalized chart that
ciation. There are other forms of SAFT EOS. A more practical, is developed for natural gases [52]. This chart is shown in
but much more complex, form of SAFT equation is given by Li Fig. 5.12 and is widely used in the petroleum industry [19, 21,
and Englezos [51]. They show application of SAFT EOS to cal- 53, 54]. Obviously this chart is valid for light hydrocarbons
culate phase behavior of systems containing associating fluids whose acentric factor is very small such as methane and
such as alcohol and water. SAFT EOS does not require criti- ethane, which are the main components of natural gases.
cal constants and is particularly useful for complex molecules Hall and Yarborough [55] presented an EOS that was based
such as very heavy hydrocarbons, complex petroleum fluids, on data obtained from the Standing and Katz Z-factor chart.
water, alcohol, ionic, and polymeric systems. Parameters can The equation was based on the Carnahan-Stafling equation
be determined by use of vapor pressure and liquid density (Eq. 5.93), and it is useful only for calculation of Z-factor of
data. Further characteristics and application of these equa- light hydrocarbons and natural gases. The equation is in the
tions are given by Prausnitz et al. [8, 47]. In the next chapter, following form:
the CS EOS will be used to develop an EOS based on the
velocity of sound. (5.102) Z=O,O6125PrT~-ay-'exp[-1.2(1- Trl)2]
where Tr and Pr are reduced temperature and pressure and y
5.7 C O R R E S P O N D I N G STATE is a dimensionless parameter similar to parameter ~ defined
CORRELATIONS in Eq. (5.91). Parameter y should be obtained from solution
of the following equation:
One of the simplest forms of an EOS is the two-parameter RK F(y) = - 0.06125PrTr 1 exp [-1.2 (1 - T r ' ) 2]
EOS expressed by Eq. (5.38). This equation can be used for
fluids that obey a two-parameter potential energy relation. In + y + y2 + y3 _ y4 (14"76Tfl-9"76Tr-E +4"58T~-3)Y 2
fact this equation is quite accurate for simple fluids such as (1 - y)3
methane. Rearrangement of Eq. (5.38) through multiplying + (90.7T~-1 - 242.2T~-2 + 42.4Tr 3) y(2aS+2"82r~-~)= 0
both sides of the equation by V/RT and substituting param-
eters a and b from Table 5.1 gives the following relation in (5.103)
terms of dimensionless variables [ 1]: The above equation can be solved by the Newton-Raphson
1 4.934( h ) 0.08664Pr method. To find y an initial guess is required. An approximate
(5.99) Z -- 1 - ~ TrL~ ~ where h =- ZTr relation to find the initial guess is obtained at Z = 1 in Eq.
(5.102):
where T~ and P~ are called reduced temperature and reduced
pressure and are defined as: (5.104) y(k) = 0.06125PrTr-1 exp [-1.2 (1 - T~-t)2]
T P Substituting y(k) in Eq. (5.103) gives F (k), which must be used
(5.100) Tr ~ Pr = in the following relation to obtain a new value of y:
Tc Pc
where T and Tc must be in absolute degrees (K), similarly P F(k)
and Pc must be in absolute pressure (bar). Both T~ and Pr are (5.105) y(k+l) = y(k) dF (k)
dimensionless and can be used to express temperature and dy
216 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
Reduced Pressure, Pr
s 6
!,! 'L1
!.0 I.e
*t7
0.95
o-~ l
O.| ~.7
0,7 1.6
N N
O O
It. LL
a ~
"~
t~ ~ - - ~.i ":'' ~. . . . . . : 1.4
E
~ . . . . . . . . . . . . . . . ~ E
-.+- }
i ~ "+"
Reduced Pressure, Pr
FIG. 5 . 1 2 ~ S t a n d i n g - K a t z generalized chart for compressibility factor of natural gases
(courtesy of GPSA and GPA [53]).
defined by Eq. (2.10). Acentric factor, co, is defined in a w a y TABLE 5.8----Constants for the Lee-Kesler modification of BWR
that for simple fluids it is zero or very small. For example, N2, EOS---Eq. (5.109) [581.
Constant Simple fluid Re~rencefluid
CH4, 02, or Ar have acentric factors of 0.025, 0.011, 0.022,
bl 0.1181193 0.2026579
and 0.03, respectively. Values of w increase with complexity b2 0.265728 0.331511
of molecules. In fact as shown in Section 2.5.3, Z~ can be cor- b3 0.154790 0.027655
related to w and both indicate deviation from simple fluids. b4 0.030323 0.203488
Acentric factor was originally introduced by Pitzer [56, 57] to cl 0.0236744 0.0313385
extend application of two-parameter CSC to more complex c2 0.0186984 0.0503618
c3 0.0 0.016901
fluids. Pitzer and his coworkers realized the linear relation c4 0.042724 0.041577
between Zc and w (i.e., see Eq. (2.103)) and assumed that dl x 104 0.155488 0.48736
such linearity exists between w and Z at temperatures other d2 x 104 0.623689 0.0740336
than To. They introduced the concept of three-parameter cor- fl 0.65392 1.226
responding states correlations in the following form: y 0.060167 0.03754
(5.107) Z = Z (~ + wZ O)
In determining the constants in these equations the con-
where both Z (~ and Z O) are functions of Tr and Pr- For straints by Eq. (5.9) and equality of chemical potentials or
simple fluids (w ~ 0), this equation reduces to Eq. (5.101). fugacity (Eq. 6.104) between vapor and liquid at saturated
Z (~ is the contribution of simple fluids and Z (1) is the cor- conditions were imposed. These coefficients for both simple
rection term for complex fluids, tt can be shown that as and reference fluids are given in Table 5.8.
P --~ 0, Z (~ --+ 1 while Z O) --+ 0, therefore, Z --+ 1. The origi- In using Eq. (5.108), both Z (~ and Z (r) should be calculated
nal three-parameter CSC developed by Pitzer was in the form from Eq. (5.109). Lee and Kesler also tabulated values of Z (~
of two graphs similar to Fig. (5.12): one for Z (~ and the other and Z 0) versus Tr and Pr for use in Eq. (5.107). The original
for Z (~), both in terms of Tr and Pr. Pitzer correlations found Lee-Kesler (LK) tables cover reduced pressure from 0.01 to
wide application and were extended to other t h e r m o d y n a m i c 10. These tables have been widely used in m a j o r texts and ref-
properties. They were in use for more than two decades; how- erences [1, 8, 59]. However, the API-TDB [59] gives extended
ever, they were found to be inaccurate in the critical region tables for Z (~ and Z O) for the Pr range up to 14. Lee-Kesler
and for liquids at low temperatures [58]. tables and their extension by the API-TDB are perhaps the
The most advanced and accurate three-parameter corre- most accurate m e t h o d of estimating PVT relation for gases
sponding states correlations were developed by Lee and and liquids. Values of Z (~ and Z (I) as given by LK and their
Kesler [58] in 1975. They expressed Z in terms of values of Z extension by API-TDB are given in Tables 5.9-5.11. Table 5.11
for two fluids: simple and a reference fluid assuming linear give values of Z (~ and Z (1) for Pr > 10 as provided in the API-
relation between Z and w as follows: TDB [59]. In Tables 5.9 and 5.10 the dotted lines separate liq-
uid and vapor phases from each other up to the critical point.
(5.108) Z = Z (~ + -~r~(Z (r) - Z (~ Values above and to the right are for liquids and below and to
the left are gases. The values for liquid phase are highlighted
where Z (r) and w(r) represent compressibility factor and acen- with bold numbers. Graphical representations of these tables
tric factor of the reference fluid. A comparison between Eqs. are given in the API-TDB [59]. For c o m p u t e r applications,
(5.107) and (5.108) indicates that [Z (r) -Z(~ or) is equiva- Eqs. (5.108)-(5.111) should be used with coefficients given in
lent to Z (1). The simple fluid has acentric factor of zero, but Table 5.8. Graphical presentation of Z (~ and Z (1) versus Pr
the reference fluid should have the highest value of o) to cover and Tr with specified liquid and vapor regions is shown in
a wider range for application of the correlation. However, Fig. 5.13. The two-phase region as well as saturated curves
the choice of reference fluid is also limited by availability of are also shown in this figure. For gases, as Pr -* 0, Z (~ ~ 1
PVT and other t h e r m o d y n a m i c data. Lee and Kesler chose and Z (1) ~ 0. It is interesting to note that at the critical point
n-octane with ~o of 0.3978 (this n u m b e r is slightly different (T~ = Pr = 1), Z (~ = 0.2901, and Z (I) = -0.0879, which after
from the most recent value of 0.3996 given in Table 2.1) as substitution into Eq. (5.107) gives the following relation for
the reference fluid. The same EOS was used for both the sim- Zc:
ple and reference fluid, which is a modified version of BWR
EOS as given in the following reduced form: (5.112) Zr -- 0.2901 - 0.0879w
w h e r e BPc/RTc c a n b e e s t i m a t e d f r o m E q . (5.71) o r (5.72) These equations are equivalent to the following equations as
t h r o u g h T~ a n d 0). E q u a t i o n (5.114) m a y b e u s e d a t l o w Pr p r o v i d e d b y t h e A P I - T D B [59].
a n d Vr > 2 o r Tr > 0,686 + 0 . 4 3 9 P r [60] i n s t e a d o f c o m p l e x
E q s . ( 5 . 1 0 8 ) - ( 5 . 1 1 1 ) . T h e A P I - T D B [59] a l s o r e c o m m e n d s Vu = ZaRTcg/P,~
t h e f o l l o w i n g r e l a t i o n , p r o p o s e d b y P i t z e r e t al. [56], f o r Z a = 0.2905 - 0.085w~
c a l c u l a t i o n o f Z f o r g a s e s a t Pr --< 0.2.
TABLE 5.11--Values of Z (0) and Z (1) for use in Eq. (5.107) from the Lee-Kesler modification of BWR EOS---Eq. (5.109) [59].
Pr ~ Tr Z (~ Z (1)
~, 10 11 12 14 10 11 12 14
0.30 2.851 3.131 3.411 3.967 -0.792 -0.869 -0.946 - 1.100
0.35 2.554 2.804 3.053 3.548 -0.886 -0.791 - 1.056 - 1.223
0.40 2.321 2.547 2.772 3.219 -0.894 -0.978 - 1.061 - 1.225
0.45 2.134 2.340 2.546 2.954 -0.861 -0.940 - 1.019 - 1.173
0.50 1.980 2.171 2.360 2.735 -0.810 -0.883 -0.955 -1.097
0.55 1.852 2.029 2.205 2.553 -0.752 -0.819 -0.885 - 1.013
0.60 1.744 1.909 2.073 2.398 -0.693 -0.753 -0.812 -0.928
0.65 1.652 1.807 1.961 2.266 -0.635 -0.689 -0.742 -0.845
0.70 1.573 1.720 1.865 2.152 -0.579 -0.627 -0.674 -0.766
0.75 1.505 1.644 1.781 2.053 -0.525 -0.568 -0.610 -0.691
0.80 1.446 1.578 1.708 1.966 -0.474 -0.512 -0.549 -0.621
0.85 1.394 1.520 1.645 1.890 -0.425 -0.459 -0.491 -0.555
0.90 1.350 1.470 1.589 1.823 -0.379 -0.408 -0.437 -0.493
0.95 1.311 1.426 1.540 1.763 -0.334 -0.360 -0.385 -0.434
0.98 1.290 1.402 1.513 1.731 -0.308 -0.331 -0.355 -0.40I
0.99 1.284 1.395 1.504 1.721 -0.299 -0.322 -0.345 -0.390
1.00 1.277 1.387 1.496 1.710 -0.290 -0.313 -0.335 -0.379
1.01 1.271 1.380 1.488 1.701 -0.282 -0.304 -0.326 -0.368
1.02 1.265 1.373 1.480 1.691 -0.273 -0.295 -0.316 -0.357
1.03 1.259 1.367 1.473 1.682 -0.265 -0.286 -0.307 -0.347
1.04 1.254 1.360 1.465 1.672 -0.256 -0.277 -0.297 -0.337
1.05 1.248 1.354 1.458 1.664 -0.248 -0.268 -0.288 -0.326
1.06 1.243 1.348 1.451 1.655 -0.239 -0.259 -0.278 -0.316
1.07 1.238 1.342 1.444 1.646 -0.231 -0.250 -0.269 -0.306
1.08 1.233 1.336 1.438 1.638 -0.222 -0.241 -0.260 -0.296
1.09 1.228 1.330 1.431 1.630 -0.214 -0.233 -0.251 -0.286
1.10 1.223 1.325 1.425 1.622 -0.206 -0.224 -0.242 -0.276
1.11 1.219 t.319 1.419 1.614 -0.197 -0.215 -0.233 -0.267
1.12 1.214 1.314 1.413 1.606 -0.189 -0.207 -0.224 -0.257
1.13 1.210 1.309 1.407 1.599 -0.181 -0.198 -0.215 -0.247
1.15 1.202 1.299 1.395 1.585 -0.164 -0.181 -0.197 -0.228
1.20 1.184 1.278 1.370 1.552 -0.123 -0.139 -0.154 -0.183
1.25 1.170 1.259 1.348 1.522 -0.082 -0,098 -0.112 -0.139
1.30 1.158 1.244 1.328 1.496 -0.042 -0.058 -0.072 -0.097
1.40 1.142 1.220 1.298 1.453 0,035 0.019 0.005 -0.019
1.50 1.134 1.205 1.276 1.419 O. 106 0.090 0.076 0.052
1.60 1.132 1.197 1.262 1.394 0.167 0.152 0.138 0.116
1.70 1.134 1.193 1,253 1.374 0.218 0.204 0.192 0.171
1.80 1.139 1.192 1,247 1.359 0.258 0.247 0.237 0.218
2.00 1.152 1.196 1.243 1.339 0.310 0.305 0.300 0.290
2.50 1.176 1.210 1.244 1.316 0.348 0.356 0.362 0.37I
3.00 1.185 1.213 1.241 1.300 0.338 0.353 0.365 0.385
3.50 1.183 1.208 1.233 1.284 0.319 0.336 0.350 0.376
4.00 1.177 1.200 1.222 1.268 0.299 0.316 0.332 0.360
High Pressure Range: Value of Z (~ a n d Z (1) for 10 < Pr < 14.
0.0
0.01 0.1 1 10
04 i i :' :
i i~ iil i !! i i~!i
-0.4 .................
:............
i-i ': : ! : :.......................
:..............
?:: : ,: .............
i............
,
-u.o ..............
: .....
?-?:-1 ............
i i ''~ i i ~ i ! i } i:
0.01 0.1 1 10
(b) R e d u c e d P r e s s u r e , Pr
FIG. 5.13~Compressibility factor (a) Z 1~ and (b) ~1) from Tables 5.9 and 5.10.
Eq. (5.121) can be rearranged to get ZRA: This method is also included in the API-TDB [59]. Another
( .~l/n
MP~ approach to estimate density of defined liquid mixtures at its
(5.123) ZRA = \ RTcdy: bubble point pressure is through the following mixing rule:
Saturation pressure for a mixture is also called bubble point p~ = 0.999SG - 10 -3 X (2.34 -- 1.898SG) • (T - 288.7)
pressure and saturation molar volume is shown by Vbp. Liquid (5.127)
density at the bubble point is shown by pbp, which is related
to Vbp by the following relation: where SG is the specific gravity at 15.5~ (60~176 and T
is absolute temperature in K. p~ is liquid density in g/cmaat
M
(5.124) p b p = V bp temperature T and atmospheric pressure. If instead of SG at
15.5~ (288.7 K), density at another temperature is available
where pbp is absolute density in g/cm3 and M is the molec- a similar equation can be derived from Eq. (2.115). Equation
ular weight. V bp c a n be calculated from the following set of (5.127) is not accurate if T is very far from the reference tem-
equations recommended by Spencer and Danner [65]: perature of 288.7 K.
The effect of pressure on liquid density or volume becomes
g bp = R x/ ZRAm important when the pressure is significantly higher than 1
atm. For instance, volume of methanol at 1000 bar and 100~
n = 1 + (1 - Tr) 2/7 is about 12% less than it is at atmospheric pressure. In gen-
N eral, when pressure exceeds 50 bar, the effect of pressure on
ZRA~ = ~--~X4ZR~ liquid volume cannot be ignored. Knowledge of the effect of
i=1 pressure on liquid volume is particularly important in the de-
Tr = T/Tcm sign of high-pressure pumps in the process industries. The
following relation is recommended by the API-TDB [59] to
N N calculate density of liquid petroleum fractions at high pres-
(5.125) Tcm= E E ~)i4)jTcij sures:
i=1 j=l pO p
xi V~i (5.128) -- = 1.0- --
p By
~i - E~=I xi vci
where pO is the liquid density at low pressures (atmospheric
pressure) and p is density at high pressure P (in bar). Br
is called isothermal secant bulk modulus and is defined
as -(1/p~ Parameter By indicates the slope of
~j 1.O- change of pressure with unit volume and has the unit of pres-
L (vl,, + d sure. Steps to calculate By are summarized in the following
224 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
Br = r e X + Bz , [ B + P / P c e "~]
v-- v j
m = 1492.1 + 0.0734P + 2.0983 x 10-6p 2
K --
X = (/320 - 105)/23170 vs.t 5-~ r (BP~e+ P)
(5.129)
logB2o = - 1 . 0 9 8 • 10-3T + 5.2351 + 0.7133p ~
(5.132) B --- - I - 9.070217r I/3 + 62.45326r 2/3
/3i = 1.0478 x 1 0 3 + 4.704 P - 3.744 • IO-4p 2
- 135.1102r + e r 4/3
+2.2331 x lO-Sp 3
C = 0.0861488 + 0.034448309
w h e r e Br is in b a r a n d po is the liquid density at a t m o s p h e r i c
e = exp (4.79594 + 0.25004709 + 1.14188o) z)
p r e s s u r e in g/cm 3. In the above e q u a t i o n T is absolute tem-
p e r a t u r e in kelvin a n d P is the p r e s s u r e in bar. The average
r = 1 -T/Tc
e r r o r f r o m this m e t h o d is a b o u t 1.7% except n e a r the critical
p o i n t w h e r e e r r o r increases to 5% [59]. This m e t h o d is n o t w h e r e V s~t is the s a t u r a t i o n m o l a r volume a n d psat is the sat-
r e c o m m e n d e d for liquids at Tr > 0.95. I n cases t h a t po is not u r a t i o n p r e s s u r e at T. V is liquid m o l a r v o l u m e at T a n d P
available it m a y be e s t i m a t e d from Eq. (5.121) o r (5.127). Al- a n d x is the i s o t h e r m a l bulk c o m p r e s s i b i l i t y defined in the
t h o u g h this m e t h o d is r e c o m m e n d e d for p e t r o l e u m fractions above e q u a t i o n (also see Eq. 6.24). T~ is the critical temper-
b u t it gives r e a s o n a b l e results for p u r e h y d r o c a r b o n s (>C5) a t u r e a n d o) is the acentric factor. Poe is equivalent critical
as well. pressure, w h i c h for all alcohols was n e a r the m e a n value of
F o r light a n d m e d i u m h y d r o c a r b o n s as well as light 27.0 bar. This value for diols is a b o u t 8.4 bar. F o r o t h e r series
p e t r o l e u m fractions the Tait-COSTALD ( c o r r e s p o n d i n g s t a t e
of c o m p o u n d s Pce w o u l d be different. Garvin f o u n d that use of
liquid density) correlation originally p r o p o s e d b y H a n k i n s o n
P c e significantly i m p r o v e s p r e d i c t i o n of V a n d x for alcohols.
a n d T h o m s o n m a y be used for the effect of p r e s s u r e on liquid F o r example, for e s t i m a t i o n of r of m e t h a n o l at 1000 b a r a n d
density [66]: 100~ Eq. (5.132) predicts x value of 7.1 x 10 -7 b a r -1, w h i c h
gives an e r r o r of 4.7% versus e x p e r i m e n t a l value of 6.8 x 10 -7
(5.130) [
Pe = peo 1 - C l n [\ f fB- ~
+-P~)]Ij- ' b a r -1, while using Pc the e r r o r increases to 36.6%. However,
one s h o u l d note t h a t the n u m e r i c a l coefficients for B, C, a n d e
in Eq. (5.132) m a y vary for o t h e r types of p o l a r liquids such
w h e r e pe is density at p r e s s u r e P and Ovo is liquid d e n s i t y at
as coal liquids.
reference p r e s s u r e of po at w h i c h density is known. W h e n peo
A n o t h e r c o r r e l a t i o n for calculation of effect of p r e s s u r e on
is calculated from the Rackett equation, po = psat w h e r e p~at
liquid density was p r o p o s e d b y Chueh a n d P r a u s n i t z [69] a n d
is the s a t u r a t i o n (vapor) pressure, w h i c h m a y be e s t i m a t e d
is b a s e d on the e s t i m a t i o n of i s o t h e r m a l compressibility:
from m e t h o d s of C h a p t e r 7. P a r a m e t e r C is a d i m e n s i o n l e s s
c o n s t a n t a n d B is a p a r a m e t e r t h a t has the same unit as pres-
Pe = peo[I + 9fl(P - p o ) ] U 9
sure. These constants can be c a l c u l a t e d f r o m the following
equations: fl = a (1 - 0.894'~ ) exp (6.9547 - 76.2853Tr + 191.306T2
B
- - = - 1 - 9.0702 (1 - Tr) ~/3 + 62.45326 (1 - T~)2/3 - 203.5472T~ + 82.7631T 4)
Pc
- 135.1102 (1 - Tr) + e (1 - T~)4/3 V~ Zc
(5.131) RTc Pc
e = exp (4.79594 + 0.250047o) + 1.14188o) 2) (5.133)
C = 0.0861488 + 0.0344483o9
The p a r a m e t e r s are defined the s a m e as were defined in Eqs.
(5.132) a n d (5.133). V~ is the m o l a r critical v o l u m e a n d the
w h e r e Tr is the r e d u c e d t e m p e r a t u r e a n d o) is the acentric
units of P, V~, R, a n d Tr m u s t be consistent in a w a y that
factor. All the above relations are in d i m e n s i o n l e s s forms.
PVc/RT~ b e c o m e s dimensionless. This e q u a t i o n is a p p l i c a b l e
Obviously Eq. (5.130) gives very a c c u r a t e result w h e n P is
for Tr r a n g i n g from 0.4 to 0.98 a n d a c c u r a c y of Eq. (5.134) is
close to P~ however, it s h o u l d n o t be used at Tr > 0.95. The
j u s t m a r g i n a l l y less a c c u r a t e t h a n the COSTALD c o r r e l a t i o n
COSTALD correlation has been r e c o m m e n d e d for i n d u s t r i a l
[67].
a p p l i c a t i o n s [59, 67]. However, in the API-TDB [59] it is rec-
o m m e n d e d t h a t special values of acentric factor o b t a i n e d
from v a p o r p r e s s u r e d a t a s h o u l d be used for o). These values E x a m p l e 5 . 8 m P r o p a n e has v a p o r p r e s s u r e of 9.974 b a r at 300
for s o m e h y d r o c a r b o n s are given b y the API-TDB [59]. The K. S a t u r a t e d liquid a n d v a p o r volumes are V L = 90.077 a n d
following e x a m p l e d e m o n s t r a t e s a p p l i c a t i o n of these m e t h - V v = 2036.5 cma/mol [Ref. 8, p. 4.24]. Calculate s a t u r a t e d liq-
ods. The m o s t r e c e n t modification of the T h o m s o n m e t h o d uid m o l a r volume using (a) Rackett equation, (b) Eqs. (5.127)-
for p o l a r a n d associating fluids was p r o p o s e d b y Garvin in (5.129), (c) Eqs. (5.127) a n d (5.130), a n d (d) Eq. (5.133).
5. P V T R E L A T I O N S A N D E Q U A T I O N S OF S T A T E 225
$olution--(a) Obviously the most accurate method to esti- Generally constants of cubic equations are determined
mate VL is through Eq. (5.121). From Table 2.1, M = 44.I, based on data for hydrocarbons up to C8 or C9. As an ex-
SG = 0.507, T~ = 96.7~ (369.83 K), Pc 42.48 bar, and w = ample, the LK generalized correlations is based on the data
0.1523. From Table 5.12, ZgA ----0.2763. Tr ----0.811 so from for the reference fluid of n-C8. The parameter that indicates
Eq. (5.121), Vsat = 89.961 cma/mol (-0.1% error). (b) Use of complexity of a compound is acentric factor. In SRK and PR
Eqs. (5.127)-(5.129) is not suitable for this case that Rack- EOS parameter a is related to w in a polynomial form of at
ett equation can be directly applied. However, to show the least second order (see f~ in Table 5.1). This indicates that
application of method Vsat is calculated to see their perfor- extrapolation of such equations for compounds having acen-
mance. From Eq. (5.127) and use of SG = 0.507 gives pO __ tric factors greater than those used in development of EOS
0.491 g/cm 3. From Eq. (5.129), m = 1492.832, B20 = 180250.6, parameters is not accurate. And it is for this reason that most
X = 3.46356, BI -- 1094.68, and Br = 6265.188 bar. Using Eq. cubic equations such as SRK and PR equations break down
(5.128), 0.491/p = 1-9.974/6265.188. This equation gives when they are applied for calculation of liquid densities for
density at T (300 K) and P (9.974 bar) as p = 0.492 g/cm 3. C10 and heavier hydrocarbons. For this reason Riazi and Man-
V sat = M/p = 44.1/0.492 -- 89.69 cma/mol (error of -0.4%). soori [70] attempted to improve capability of cubic equations
(c) Use of Eqs. (5.127) and (5.130) is not a suitable method for liquid density prediction, especially for heavy hydrocar-
for density of propane, but to show its performance, satu- bons.
rated liquid volume is calculated in a way similar to part Most modifications on cubic equations is on parameter a
(b): From Eq. (5.131), B = 161.5154 bar and C = 0.091395. and its functionality with temperature and w. However, a pa-
For Eq. (5.130) we have ppo --- 0.491 g/cm 3, po = 1.01325 bar, rameter that is inherent to volume is the co-volume parameter
P -- 9.974 bar, and calculated density is pp = 0.4934 g/cm 3. b. RK EOS presented by Eq. (5.38) is the simplest and most
Calculated Vsat is 89.4 cm3/mol, which gives a deviation of widely used cubic equation that predicts reasonably well for
-0.8% from experimental value of 90.077 cma/mol. (d) Us- prediction of density of gases. In fact as shown in Table 5.13
ing the Chueh-Prausnitz correlation (Eq. 5.133) we have for simple fluids such as oxygen or methane (with small oJ)
Z~ = 0.276, 0t = 0.006497, fl = 0.000381, pp = 0.49266 g]cm 3, RK EOS works better than both SRK and PR regarding liquid
and VSa]tc= 89.5149 cm3/mol, which gives an error of -0.62% densities.
from the actual value. ) For liquid systems in which the free space between
molecules reduces, the role of parameter b becomes more
important than that of parameter a. For low-pressure gases,
5.9 R E F R A C T I V E I N D E X B A S E D E Q U A T I O N however, the role of parameter b becomes less important than
OF STATE a because the spacing between molecules increases and as
a result the attraction energy prevails. Molar refraction was
From the various PVT relations and EOS discussed in this defined by Eq. (2.34) as
chapter, cubic equations are the most convenient equations
that can be used for volumetric and phase equilibrium cal- M(n2-1~
culations. The main deficiency of cubic equations is their (5.134) Rm = V I = ~ \ n 2 + 2 /
inability to predict liquid density accurately. Use of volume
translation improves accuracy of SRK and PR equations for where Rm is the molar refraction and V is the molar volume
liquid density but a fourth parameter specific of each equa- both in cma/mol. Rm is nearly independent of temperature
tion is required. The shift parameter is not known for heavy but is normally calculated from density and refractive index
compounds and petroleum mixtures. For this reason some at 20~ (d20 and n20). Rm represents the actual molar volume
specific equations for liquid density calculations are used. As of molecules and since b is also proportional to molar volume
an example Alani-Kennedy EOS is specifically developed for of molecules (excluding the free space); therefore, one can
calculation of liquid density of oils and reservoir fluids and conclude that parameter b must be proportional to Rm. In
is used by some reservoir engineers [19, 21]. The equation fact the polarizability is related to Rm in the following form:
is in van der Waals cubic EOS form but it requires four nu-
merical constants for each pure compound, which are given 3
from Ca to Ca0. For the C7+ fractions the constants should be (5.135) a' = "7"'77-~.Rm -- #(T)
~Tr/VA
estimated from M7+ and SG7+. The method performs well for
light reservoir fluids and gas condensate samples. However, where NA is the Avogadro's number and/z(T) is the dipole
as discussed in Chapter 4, for oils with significant amount of moment, which for light hydrocarbons is zero [7]. Values of
heavy hydrocarbons, which requires splitting of C7+ fraction, Rm calculated from Eq. (5.134) are reported by Riazi et al.
the method cannot be applied to C7+ subfractions. In addition [70, 71] for a number of hydrocarbons and are given in Table
the method is not applicable to undefined petroleum fractions 5.14. Since the original RK EOS is satisfactory for methane we
with a limited boiling range. choose this compound as the reference substance. Parameter
TABLE 5.13--Evaluation of RK, SRK, and PR EOS for prediction of density of simple fluids.
%AAD
Compound No. of data points Temperaturerange, K Pressurerange, bar RK SRK PR Data source
Methane 135 90-500 0.7-700 0.88 1.0 4.5 Goodwin [72]
Oxygen 120 80-1000 1-500 1,1 1.4 4.0 TRC [73]
226 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
TABLE 5.14---Data source for development of Eq. (5.139), values of parameter r and predicted Zc from MRK EOS [70].
Critical compressibility, Zc
Rm, at 20~ No. of Temp. Pressure
No. Compound cm3/mol R data points range, K range, bar Ref. Table 2.1 Pred.MRI( %AD
1 Methane (C1) 6.987 1.000 135 90-500 0.5-700 Goodwin [72] 0.288 0.333 15.6
2 Ethane (C2) 11.319 1.620 157 90-700 0.1-700 Goodwin et al. [74] 0.284 0.300 5.6
3 Ethylene 10.508 1.504 90 100-500 1--400 McCarty and 0.276 0.295 6.9
Jacobsen [75]
5 Propane (Ca) 15.784 2.259 130 85-700 0.1-700 Goodwin and 0.280 0.282 0.7
Haynes [76]
6 Isobutane 20.647 2.955 115 110-700 0.1-700 Goodwin and 0.282 0.280 0.7
Haynes [76]
7 n-Butane (C4) 20.465 2.929 183 130-700 0.1-700 Haynes and 0.274 0.278 1,5
Goodwin [77]
8 n-Pentane (C5) 25.265 3.616 . . . . . . 0.269 0.271 0,7
9 n-Hexane (C6) 29.911 4.281 "1"()0 298-1000 1-500 TI~C Tables [73] 0.264 0.266 0.7
10 Cyclohexane 27.710 3.966 140 320-1000 t-500 TRC Tables [73] 0.273 0.269 1.5
11 Benzene 26.187 3.748 110 310-1000 1-500 TRC Tables [73] 0.271 0.270 0.4
12 Toluene 31.092 4.450 110 330-1000 1-500 TRC Tables [73] 0.264 0.265 0.4
13 n-Heptane (C7) 34.551 4.945 100 300-1000 1-500 TRC Tables [73] 0.263 0.262 0.4
14 n-Octane (C8) 39.183 5.608 80 320-1000 1-500 TRC Tables [73] 0.259 0.258 0.4
15 /-Octane 39.260 5.619 70 340-1000 1-500 TRC Tables [73] 0.266 0.256 3.8
16 n-Heptane (C7)a 34.551 4.945 35 303-373 50-500 Doolittle [78] . . . . . . . . .
17 n-Nonane (C9)a 43,836 6.274 35 303-373 50-500 Doolittle [78] 0.255 0.254 0.4
18 n-Decane (C10)a 48.497 6.941 . . . . . . . 0.249 0.250 0.4
19 n-Undecane (CH)a 53.136 7.605 "35 303-373 50-500 Doolittle [78] 0.243 0.247 1.6
20 n-Dodecane (Clz)a 57.803 8.273 . . . . . . . 0.238 0.245 2.9
21 n-Tridecane (C13) a 62.478 8.942 "30 303-373 50-500 Doolittle [78] 0.236 0.242 2.6
22 n-Tetradecane (C14)a 67.054 9.597 . . . . . . . . . .., 0.234 0.240 2.5
23 n-Pentadecane (C15) a 71.708 10.263 . . . . . . . . . ... 0.228 0.238 4.3
24 n-Hexadecane (C16)~ 76.389 10.933 . . . . . . . 0.225 0.235 4.2
25 n - H e p t a d e c a n e (C17) a 81.000 11.593 '30 323-573 50-500 l)oolittle [78] 0.217 0.233 7.4
26 n-Eicosane (C20)a 95.414 13.656 20 373-573 50-500 Doolitfle [78] 0.213 0.227 6.6
27 n-Triacosane (C30)a 141.30 20.223 20 373-573 50-500 Doolittle [78] ... 0.213 ...
28 n-Tetracontane (C40)a 187.69 26.862 20 423-573 50-500 Doolittle [78] . . . . . .
Overall . . . . . . 1745 90-1000 0.1-700 3.0
Density data for compounds 16-28 are all only for liquids [78]. Compounds specifiedby bold are used in developmentof Eq. (5.139). Calculatedvalues of Zc from
SRK and PR EOSs for all compounds are 0.333 and 0.307, respectively.These give average errors of 28.2 and 18.2%, respectively.
aPVT data for the followingcompounds were not used in developmentof Eq. (5.I39).
TABLE 5.15--Evaluation of various EOS for prediction of liquid TABLE 5.17nMixing rules for MRK EOS parameters (Eqs. (5.38)
density of heavy hydrocarbons [70]. and (5.137)-(5.140)).
%AAD Tcm ~i ~,j xixi Tc2ij/Pci]
No. of rc. = (rdrcj) '/~ (i - ~j)
Compound data points MRK RK SRK PR = ~,i ~j xixjTcij/Pcij
n-Heptane (n-C7) 35 0.6 12.1 10.5 1.4 8r~q
n-Nonane (n-C9) 35 0.6 15.5 13.4 3.4 Pcm (Ei v/;~jxixjTcij/Pcij) 2 P~q = [(r~i/p~i)l/3+(rc./pc.),/3]3
n-Undecane (n-C11) 35 1.7 18.0 15.5 5.4
n - T r i d e c a n e (n-C13) 30 2.8 20.3 17.7 7.9 1/3_ 1/3X3
n-Heptadecane (n-C17) 30 1.2 27.3 24.8 16.0 rii +rjj )
Rm = ~ i )-~-j ~ x j r i j rij - - 8
n-Eicosane (n-C20) 20 2.8 29.5 26.7 18.2
n-Triacontane (n-C30) 20 0.6 41.4 39.4 32.5
n-Tetracontane (n-C40) 20 4.1 50.9 49.4 44.4 Predicted liquid densities f r o m S R K a n d PR equations (Exam-
Total 225 1.6 24.3 22.1 13.3
MRK: Eqs. (5.38), (5.138), and (5.141). Note none of these data were used ple 5.2) deviate f r o m e x p e r i m e n t a l d a t a by +31.5 a n d 17.2%,
in development of Eq. (5.139). respectively. Advantage of M R K over o t h e r cubic equations
for liquid density is greater for heavier c o m p o u n d s as s h o w n
interesting p o i n t a b o u t this e q u a t i o n is t h a t it can be u s e d u p in Table 5.15. t
to C40 for density estimations. Obviously this e q u a t i o n is n o t
designed for VLE calculations as no VLE d a t a were u s e d to This modified version of R K EOS is developed only for den-
develop Eq. (5.139). P r e d i c t i o n of Zc f r o m M R K EOS is s h o w n sity calculation of h y d r o c a r b o n systems a n d t h e i r mixtures.
in Table 5.14. Evaluation of M R K with PR a n d S R K equations It can be u s e d directly to calculate density of p e t r o l e u m frac-
for p r e d i c t i o n of liquid density of heavy h y d r o c a r b o n s is given tions, once M, d20,/'/20, Tc, a n d Pc are calculated from m e t h -
in Table 5.15. Data sources for these c o m p o u n d s are given in ods discussed in Chapters 2 a n d 3. M o r e o v e r p a r a m e t e r r c a n
Table 5.14. Overall results for p r e d i c t i o n of density for b o t h be a c c u r a t e l y e s t i m a t e d for heavy fractions, while p r e d i c t i o n
liquid a n d gaseous h y d r o c a r b o n c o m p o u n d s from C1 to C40 of acentric factor for heavy c o m p o u n d s is n o t reliable (see
is s h o w n in Table 5.15. The overall e r r o r for the M R K EOS Figs. 2.20-2.22). The m a i n c h a r a c t e r i s t i c of this e q u a t i o n is its
for m o r e t h a n 1700 d a t a p o i n t s is a b o u t 1.3% in c o m p a r i s o n a p p l i c a t i o n to heavy h y d r o c a r b o n s a n d undefined p e t r o l e u m
with 4.6 for PR a n d 7.3 for S R K equations. fractions. The fact t h a t Eq. (5.139) was developed b a s e d on
To a p p l y this EOS to defined m i x t u r e s a set of mixing rules d a t a for h y d r o c a r b o n s f r o m C2 to C8 a n d it can well be u s e d u p
are given in Table 5.17 [70]. F o r p e t r o l e u m fractions p a r a m e - to C40 shows its e x t r a p o l a t i o n capability. The linear relation
ters can be directly calculated for the mixture. F o r b i n a r y a n d t h a t exists b e t w e e n 1/fl a n d p a r a m e t e r r m a k e s its extrapo-
t e r n a r y liquid m i x t u r e s c o n t a i n i n g c o m p o u n d s f r o m C1 to C20 lation to heavier h y d r o c a r b o n s possible. In fact it was f o u n d
a n average e r r o r of 1.8% was o b t a i n e d for 200 d a t a points [70]. that b y c h a n g i n g the functionality of 1/fi with r, b e t t e r pre-
F o r the s a m e d a t a s e t RK, SRK, a n d PR equations gave errors diction of density is possible b u t the relation w o u l d no l o n g e r
of 15, 13, a n d 6%, respectively. F u r t h e r characteristics a n d be linear a n d its e x t r a p o l a t i o n to heavier c o m p o u n d s w o u l d
evaluations of this modified RK EOS a r e discussed by Riazi be less accurate. F o r example, for C17 and C18, if the c o n s t a n t
a n d R o o m i [71]. Application of this m e t h o d in calculation of 0.02 in Eq. (5.139) is r e p l a c e d b y 0.018, the %AAD for these
d e n s i t y is s h o w n in the following example. c o m p o u n d s reduces f r o m 2 to 0.5%. The following e x a m p l e
shows a p p l i c a t i o n of this m e t h o d .
Example 5 . 9 - - R e p e a t E x a m p l e 5.2 for p r e d i c t i o n of liquid Analysis of various EOS shows t h a t use of refractive index
a n d v a p o r density of n-octane using M R K EOS. in o b t a i n i n g c o n s t a n t s of an EOS is a p r o m i s i n g a p p r o a c h .
F u r t h e r w o r k in this a r e a s h o u l d involve use of s a t u r a t i o n
Solution--The M R K EOS is to use Eq. (5.38) w i t h p a r a m e t e r s p r e s s u r e in a d d i t i o n to liquid density d a t a to o b t a i n relations
o b t a i n e d f r o m Eqs. (5.139)-(5.141). The i n p u t d a t a n e e d e d to for EOS p a r a m e t e r s t h a t w o u l d be suitable for b o t h liquid
use M R K EOS are Tc, Pc, a n d r . F r o m E x a m p l e 5.2, Tc = 568.7 density a n d VLE calculations.
K, Pc = 24.9 bar, a n d Tr = 0.9718 K. F r o m Table 5.14 for
n-Cs, r = 5.608. F r o m Eq. (5.139), fl = 1.5001 x 10 -4. F r o m
Eq. (5.139), b - - 150.01 cma/mol and from Eq. (5.141), a = 5.10 SUMMARY AND CONCLUSIONS
3.837982 x 107 cm6/mol 2. Solving Eq. (5.42) with ul = 1 a n d
u2 = 0 (Table 5.1) a n d in a w a y s i m i l a r to t h a t p e r f o r m e d in In this c h a p t e r the f u n d a m e n t a l of PVT relations a n d m a t h -
E x a m p l e 5.2 we get V L = 295.8 a n d V v -- 1151.7 cma/mol. De- e m a t i c a l EOS are presented. Once the PVT r e l a t i o n for a
viations of p r e d i c t e d values f r o m e x p e r i m e n t a l d a t a are -2.7% fluid is k n o w n various physical a n d t h e r m o d y n a m i c proper-
a n d - 5 . 3 % for liquid a n d v a p o r m o l a r volume, respectively. ties can be d e t e r m i n e d as d i s c u s s e d in Chapters 6 a n d 7. In-
t e r m o l e c u l a r forces a n d t h e i r i m p o r t a n c e in p r o p e r t y predic-
TABLE 5.16---Comparison of various EOSs for prediction of tions were discussed in this chapter. F o r light h y d r o c a r b o n s
density of liquid and gaseous hydrocarbons. t w o - p a r a m e t e r p o t e n t i a l energy relations such as LJ describes
No. of %AAD the i n t e r m o l e c u l a r forces a n d as a result t w o - p a r a m e t e r E O S
Compound data points MRK RK SRK PR are sufficient to describe the PVT relation for such fluids. It is
C1C~ 1520 1.3 4.9 5.1 3.3 s h o w n that EOS p a r a m e t e r s can be directly calculated f r o m
CT-Cb0 225 1.6 24.3 22.1 13.3 the potential energy relations. Criteria for correct EOS are
Total 1745 1.33 7.38 7.28 4.59 given so that validity of a n y EOS can be analyzed. Three cat-
aThese are the compounds that have been marked as bold in Table 5.14 and
are used in development of Eq. (5.139). egory o f EOSs a r e p r e s e n t e d in this chapter: ( I ) cubic type,
bThese are the same compounds as in Table 5.15. (2) n o n c u b i c type, a n d (3) generalized correlations.
228 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
Four types of cubic equations vdW, RK, SRK, and PR and calculation of density, enthalpy, entropy, and heat capacity of
their modifications have been reviewed. The main advantage hydrocarbons and petroleum fractions. Analytical form of LK
of cubic equations is simplicity, mathematical convenience, correlation is provided for computer applications, while the
and their application for both vapor and liquid phases. The tabulated form is given for hand calculations. Simpler two-
main application of cubic equations is in VLE calculations parameter empirical correlation for calculation of Z-factor of
as will be discussed in Chapters 6 and 9. However, their abil- gases, especially for light hydrocarbons and natural gases, is
ity to predict liquid phase density is limited and this is the given in a graphical form in Fig. 5.12 and Hall-Yarborough
main weakness of cubic equations. PR and SRK equations are equation can be used for computer applications..
widely used in the petroleum industry. PR equation gives bet- For calculation of liquid densities use of Rackett equa-
ter liquid density predictions, while SRK is used in VLE calcu- tion (Eq. 5.121) is recommended. For petroleum fractions
lations. Use of volume translation improves capability of liq- in which Racket parameter is not available it should be de-
uid density prediction for both PR and SRK equations; how- termined from specific gravity through Eq. (5.123). For the
ever, the method of calculation of this parameter for heavy effect of pressure on liquid density of light pure hydrocar-
petroleum fractions is not available and generally these equa- bons, defined hydrocarbon mixtures and light petroleum frac-
tions break down at about C10. Values of input parameters tions, the COSTALD correlation (Eq. 5.130) m a y be used. For
greatly affect EOS predictions. For heavy hydrocarbons, ac- petroleum fractions effect of pressure on liquid density can
curate prediction of acentric factor is difficult and for this rea- be calculated through Eq. (5.128).
son an alternative EOS based on modified RK equation is pre- For defined mixtures the simplest approach is to use Kay's
sented in Section 5.9. The MRK equation uses refractive index mixing rule (Eqs. 3.39 and 5.116) to calculate pseudocriti-
parameter instead of acentric factor and it is recommended cal properties and acentric factor of the mixture. However,
for density calculation of heavy hydrocarbons and undefined when molecules in a mixture are greatly different in size (i.e.,
petroleum fraction. This equation is not suitable for VLE and Cs and C20), more accurate results can be obtained by using
vapor pressure calculations. In Chapter 6, use of velocity of appropriate mixing rules given in this chapter for different
sound data to obtain EOS parameters is discussed [79]. EOS. For defined mixtures liquid density can be best calcu-
Among noncubic equations, virial equations provide more lated through Eq. (5.126) when pure component densities are
accurate PVT relations; however, prediction of fourth and known at a given temperature and pressure. For undefined
higher virial coefficients is not possible. Any EOS can be con- narrow boiling range petroleum fractions Tc, Pc, and coshould
verted into a virial form. For gases at moderate pressures, be estimated according to the methods described in Chapters
truncated virial equation after third term (Eq. 5.75) is recom- 2 and 3. Then the mixture may be treated as a single pseudo-
mended. Equation (5.71) is recommended for estimation of component and pure component EOS can be directly applied
the second virial coefficient and Eq. (5.78) is recommended to such systems. Some other graphical and empirical meth-
for prediction of the third virial coefficients. For specific com- ods for the effect of temperature and pressure on density and
pounds in which virial coefficients are available, these should specific gravity of hydrocarbons and petroleum fractions are
be used for more accurate prediction of PVT data at certain given in Chapter 7. Further application of methods presented
moderate conditions such as those provided by Gupta and in this chapter for calculation of density of gases and liquids
Eubank [80]. especially for wide boiling range fractions and reservoir fluids
Several other noncubic EOS such as BWRS, CS, LJ, SPHC, will be presented in Chapter 7. Theory of prediction of ther-
and SAFT are presented in this chapter. As will be discussed modynamic properties and their relation with PVT behavior
in the next chapter, recent studies show that cubic equations of a fluid are discussed in the next chapter.
are also weak in predicting derivative properties such as en-
thalpy, Joule Thomson coefficient, or heat capacity. For this
reason, noncubic equations such as simplified perturbed hard 5.11 P R O B L E M S
chain (SPHC) or statistical associating fluid theory (SAFT)
are being investigated for prediction of such derived prop- 5.1. Consider three phases of water, oil, and gas are in equi-
erties [81]. For heavy hydrocarbons in which two-parameter librium. Also assume the oil is expressed in terms of
potential energy functions are not sufficient to describe the in- 10 components (excluding water) with known specifica-
termolecular forces, three- and perhaps four-parameter EOS tions. The gas contains the same compounds as the oil.
must be used. The most recent reference on the theory and Based on the phase rule determine what is the m i n i m u m
application of EOSs for pure fluids and fluid mixtures is pro- information that must be known in order to determine
vided by Sengers et al. [82]. In addition, for a limited number oil and gas properties.
of fluids there are highly accurate EOS that generally take on a 5.2. Obtain coefficients a and b for the PR EOS as given in
modified MBWR form or a Helmholtz energy representation Table 5.1. Also obtain Zc = 0.307 for this EOS.
like the IAPWS water standard [4]. Some of these equations 5.3. Show that the Dieterici EOS exhibits the correct limiting
are even available free on the webs [83]. behavior at P ~ 0 (finite T) and T --~ o0 (finite P)
The theory of corresponding state provides a good PVT rela-
RTb e-a/n:rv
tion between Z-factor and reduced temperature and pressure. P-V-
The LK correlation presented by Eqs. (5.107)-(5.11 I) is based
on BWR EOS and gives the most accurate PVT relation if ac- where a and b are constants.
curate input data on To, Pc, and co are known. While the cubic 5.4. The Lorentz EOS is given as
equations are useful for phase behavior calculations, the a bV
LK corresponding states correlations are recommended for
5. PVT R E L A T I O N S AND EQUATIONS OF STATE 229
where a and b are the EOS constants. Is this a valid density at 68~ and pressure of 5400 psig using the fol-
EOS? lowing methods. The experimental value is 0.8838 g/cm 3
5.5. A graduate student has come up with a cubic EOS in the (Ref. [59] Ch 6)
following form: a. SRK EOS
b. SRK using volume translation
I aV2 ] (V - b) = RT c. MRK EOS
P + (V + b ) ( V - b ) d. Eq. (5.128)
e. COSTALD correlation (Eq. 5.130)
Is this equation a correct EOS? f. LK generalized correlation
5.6. Derive a relation for the second virial coefficient of a g. Compare errors from different methods
fluid that obeys the SWP relation. Use data on B for 5.15. Estimate liquid density of n-decane at 423 K and 506.6
methane in Table 5.4 to obtain the potential energy bar from the following methods:
parameters, a and s. Compare your calculated values a. PR EOS
with those obtained from LJ Potential as a = 4.01 A and b. PR EOS with volume translation
elk = 142.87 K [6, 79]. c. PR EOS with Twu correlation for parameter a (Eq.
5.7. Derive Eq. (5.66) from Eq. (5.65) and discuss about your 5.54)
derivation. d. MRK EOS
5.8. Show that for the second virial coefficient, Eq. (5.70) can e. Racket equation with COSTALD correlation
be reduced to a form similar to Eq. (5.59). Also show that f. Compare the values with the experimental value of
these two forms are identical for a binary system. 0.691 g/cm 3
5.9. Derive the virial form of PR EOS and obtain the virial 5.16. Estimate compressibility factor of saturated liquid and
coefficients B, C, and D in terms of PR EOS parameters. vapor (Z L and Z v) methane at 160 K (saturation pres-
5.10. With results obtained in Example 5.4 and Problem 5.9 sure of 15.9 bar) from the following methods:
for the virial coefficients derived from RK, SRK, and PR a. Z L from Racket equation and Z v from Standing-Katz
equations estimate the following: chart
a. The second virial coefficient for propane at temper- b . PR EOS
atures 300, 400 and 500 K and compare the results c. PR EOS with Twu correlation for parameter a (Eq.
with those given in Table 5.4. Also predict B from 5.54)
Eqs. (5.71)-(5.73). d. MRK EOS
b. The third virial coefficients for methane and ethane e. LK generalized correlation
and compare with those given in Table 5.5. f. Compare estimated values with the values from Fig.
c. Compare predicted third virial coefficients from (b) 6.12 in Chapter 6.
with those predicted from Eq. (5.78). 5.17. Estimate Z v of saturated methane in Problem 5.16 from
5.11. Specific volume of steam at 250~ and 3 bar is 796.44 virial EOS and evaluate the result.
cm3/g [1]. The virial coefficients (B and C) are given in 5.18. A liquid mixture of C1 and n-C5 exists in a PVT cell at
Table 5.5. Estimate specific volume of this gas from the 311.1 K and 69.5 bar. The volume of liquid is 36.64 cm 3.
following methods: Mole fraction of C1 is 0.33. Calculate mass of liquid in
a. RK, SRK, and PR equations. grams using the following methods:
b. Both virial forms by Eqs. (5.65) and (5.66). Explain a. PR EOS with and without volume translation
why the two results are not the same. b. Rackett equation and COSTALD correlation
c. Virial equation with coefficients estimated from Eqs. c. MRK EOS
(5.71), (5.72), and (5.78) 5.19. A natural gas has the following composition:
5.12. Estimate molar volume of n-decane at 373 K and 151.98
bar from LK generalized correlations. Also estimate the Component CO2 H2S N2 C1 C2 C3
critical compressibility factor. The actual molar volume mol% 8 16 4 65 4 3
is 206.5 cma/mol.
5.13. For several compounds liquid density at one tempera- Determine the density of the gas at 70 bar and 40~ in
ture is given in the table below. g/cm 3 using the following methods:
a. Standing-Katz chart
Componen# N2 H20 C1 C2 C3 n-C4 b. Hall-Yarborough EOS
T, K 78 293 112 183 231 293 c. LK generalized correlation
p, g/cm3 0.804 0.998 0.425 0.548 0.582 0.579 5.20. Estimate Z L and Z v of saturated liquid and vapor ethane
~Source:Reid et al. [15]. at Tr = 0.8 from MRK and virial EOSs. Compare calcu-
lated values with values obtained from Fig. 5.10.
For each compound calculate the Rackett parameter
from reference density and compare with those given
in Table 5.12. Use estimated Rackett parameter to cal- REFERENCES
culate specific gravity of Ca and n-Ca at 15.5~ and com-
pare with values of SG given in Table 2.1. [1] Smith, J. M., Van Ness, H. C., and Abbott, M. M., Introduction
5.14. For a petroleum fraction having API gravity of 31.4 and to Chemical Engineering Thermodynamics, 5th ed.,
Watson characterization factor of 12.28 estimate liquid McGraw-Hill, New York, 1996.
230 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
[2] Daubert, T. E., Danner, R. E, Sibul, H. M., and Stebbins, C. C., [24] Graboski, M. S. and Daubert, T. E., "A Modified Soave Equation
Physical and Thermodynamic Properties of Pure Compounds: of State for Phase Equilibrium Calculations, 1: Hydrocarbon
Data Compilation, DIPPR-AIChE, Taylor & Francis, Bristol, Systems," Industrial Engineering and Chemistry, Process Design
PA, 1994 (extant) (www.aiche.org/dippr). Updated reference: and Development, Vol. 17, No. 4, 1978, pp. 443-448.
Rowley, R. L., Wilding, W. V., Oscarson, J. L., Zundel, N. A., [25] Twu, C. H., Coon, J. E., and Cunningham, J. R., "A New
Marshall, 2". L., Daubert, 1". E., and Darmer, R. R, D1PPRData Alpha-Function for a Cubic Equation of State, 1:
Compilation of Pure Compound Properties, Design Institute for Peng-Robinson Equation," Fluid Phase Equilibria, Vol. 105,
Physical Properties (DIPPR), Taylor & Francis, New York, 2002 No. 1, 1995, pp. 49-59.
(http://dippr.byu.edu). [26] Nasrifar, K. and Moshfeghian, M., "A New Cubic Equation of
[3] AIChE DIPPR| Database, Design Institute for Physical Property State for Simple Fluids, Pure and Mixtures," Fluid Phase
Data (DIPPR), EPCON International, Houston, TX, 1996 (also Equilibria, Vol. 190, No. 1-2, 2001, pp. 73-88.
see http://dippr.byu.edu). [27] Zudkevitch, D. and Joffe, E., "Correlation and Prediction of
[4] Wagner, W. and Pruss, A., "The IAPWS Formulation for the Vapor-Liquid Equilibria with the Redlich-Kwong Equation
Thermodynamic Properties of Ordinary Water Substance for of State," American Institute of Chemical Engineers Journal,
General and Scientific Use," Journal of Physical and Chemical Vol. 16, No. 1, 1970, pp. 112-119.
Reference Data, Vol. 31, No. 2, 2002, pp. 387-535 (also see [28] Joffe, E., Schroeder, G. M., and Zudkevitch, D., "Vapor-Liquid
http://www.iapws.org/). Equilibrium with the Redlich-Kwong Equation of State,"
[5] KISR, Private Communication, Petroleum Research Studies American Institute of Chemical Engineers Journal, Vol. 16,
Center, Kuwait Institute for Scientific Research (KISR), No. 5, 1970, pp. 496-498.
Ahmadi, Kuwait, May 2002. [29] Feyzi, E, Riazi, M. R., Shaban, H. I., and Ghotbi, S.,
[6] Prausnitz, J. M., Lichtenthaler, R. N., and de Azevedo, E.G., "Improving Cubic Equations of State for Heavy Reservoir
Molecular Thermodynamics of Fluid-Phase Equilibria, Fluids and Critical Region," Chemical Engineering
Prentice-Hall, New Jersey, 1986. Communications, Vol. 167, 1998, pp. 147-166.
[7] I-Iirschfelder, O. J., Curtiss, C. E, and Bird, R. B., Molecular [30] Abbott, M. M., "Cubic Equations of State: An Interpretive
Theory of Gases and Liquids, Wiley, New York, 1964. Review," Chapter 3 in Equation of State in Engineering and
[8] Poling, B. E., Prausnitz, J. M., and O'Connell, J. E, Properties of Research, K. C. Chao and R. L. Robinson, Jr., Eds., Advances in
Gases & Liquids, 5th ed., Mc-Graw Hill, New York, 2000. Chemistry Series, American Chemical Society, Washington, DC,
[9] Alberty, R. A. and Silbey, R. J., Physical Chemistry, 2nd ed., Vol. 182, 1979, pp. 47-97.
Wiley, New York, 1999. [31] Martin, J. J., "Cubic Equation of State--Which?," Industrial
[10] Levine, I. N., Physical Chemistry, 4th ed., McGraw-Hill, New and Engineering Chemistry, Vol. 18, No. 2, 1979, pp. 81-97.
York, 1995, [32] Tsonopoulos, C. and Heidman, J. L., "From Redlich-Kwong to
[11] Redlich, O. and Kwong, J. N. S., "On the Thermodynamics of the Present," Fluid Phase Equilbria, Vol. 24, 1985, pp. 1-23.
Solutions," Chemical Review, Vol. 44, 1949, pp. 233-244. [33] Firoozabadi, A., "Reservoir Fluid Phase Behavior and
[12] Soave, G., "Equilibrium Constants from a Modified Volumetric Prediction with Equation of State," Journal of
Redlich-Kwong Equation of State," Chemical Engineering Petroleum Technology, Vol. 40, No. 4, 1988, pp. 397-406.
Science, Vol. 27, 1972, pp. 1197-1203. [34] Patel, N. C. and Teja, A. S., "A New Cubic Equation of State
[13] Peng, D. Y. and Robinson, D. B., "A New Two-Constant for Fluids and Fluid Mixtures," Chemical Engineering Science,
Equation of State," Industrial and Engineering Chemistry Vol. 77, No. 3, 1982, pp. 463-473.
Fundamentals, Vol. 15, No. 1, 1976, pp. 59-64. [35] Manafi, H., Mansoori, G. A., and Ghotbi, S., "Phase Behavior
[14] Robinson, D. B. and Peng, D. Y., The Characterization of the Prediction of Petroleum Fluids with Minimum
Heptanes and Heavier Fractions for the GPA Peng-Robinson Characterization Data," Journal of Petroleum Science and
Programs, Research Report 28, Gas Producers Association Engineering, Vol. 22, 1999, pp. 67-93.
(GPA), Tulsa, 1978. [36] Katz, D. L. and Firoozabadi, A., "Predicting Phase Behavior of
[15] Reid, R. C., Prausnitz, J. M., and Poling, B. E., Propertiesof Condensate/Crude Oil Systems Using Methane Interaction
Gases & Liquids, 4th ed., McGraw-Hill, New York, 1987. Coefficients," Journal of Petroleum Technology, Trans.,/LIME,
[16] Perry, R. H. and Green, D. W., Eds., Chemical Engineers' Vol. 265, November 1978, pp. 1649-1655.
Handbook, 7th ed., McGraw-Hill, New York, 1997. [37] Chueh, P. L. and Prausnitz, J. M., "Vapor-Liquid Equilibria at
[17] Abramowitz, M. and Stegun, I. A., Handbook of Mathematical High Pressures, Calculation of Partial Molar Volume in
Functions, Dover Publication, New York, 1970. Non-Polar Liquid Mixtures," American Institute of Chemical
[18] Guo, 1". M., Du, L., Pedersen, K. S., and Fredenslund, Aa., Engineers Journal, Vol. 13, No. 6, 1967, pp. 1099-1113.
"Application of the Du-Guo and the SRK Equations of State to [38] Whitson, C. H., Anderson, T. E, and Soreide, L, "C7+
Predict the Phase Behavior of Chinese Reservoir Fluids," SPE Characterization of Related Equilibrium Fluids Using the
Reservoir Engineering, Vol. 6, No. 3, 1991, pp. 379-388. Gamma Distribution," C7+ Fraction Characterization, L. G.
[19] Whitson, C. H. and Brule, M. R., Phase Behavior, Monograph Chorn and G. A. Mansoori, Eds., Taylor & Francis, New York,
Volume 20, SPE, Richardson, Texas, 2000. 1989, pp. 35-56.
[20] Soreide, I., Improved Phase Behavior Predictions of Petroleum [39] Pedersen, K. S., Thomassen, P., and Fredenslund, Aa., "On the
Reservoir Fluids from a Cubic Euation of State, Dr. Engineering Dangers of "Tuning" Equation of State Parameters," Chemical
Dissertation, Norwegian Institute of Technology, Trondheim, Engineering Science, Vol. 43, No. 2, pp. 269-278.
Norway, 1989. [40] Whitson, C. H., "Characterizing Hydrocarbon Plus Fractions,"
[21] Danesh, A., PVT and Phase Behavior of Petroleum Reservoir Society of Petroleum Engineers Journal, Trans., MME, Vol. 275,
Fluids, Elsevier, Amsterdam, 1998. 1983, pp. 683-694.
[22] Peneloux, A., Rauzy, E., and Freze, R., "A Consistent Correction [41] Dymond, J. H. and Smith, E. B., The Virial Coefficients of Pure
for Redlich-Kwong-Soave Volumes," Fluid Phase Equilibria, Gases and Mixtures, Clarendon Press, Oxford, UK, 1980.
I/ol. 8, 1982, pp. 7-23. [42] Tsonopoulos, C., "Empirical Correlation of Second Virial
[23] Jhaveri, B. S. and Youngren, G. K., "Three-Parameter Coefficients," American Institute of Chemical Engineers Journal,
Modification of the Peng-Robinson Equation of State to Vol. 20, 1974, pp. 263-272.
Improve Volumetric Predictions," SPE Reservoir Engineering, [43] McGlashml, M. L., Chemical Thermodynamics, Academic Press,
Trans., MME, Vol. 285, 1998, pp. 1033-1040. New York, 1980.
5. P V T R E L A T I O N S A N D E Q U A T I O N S O F S T A T E 231
[44] Orbey, H. and Vera, J. H., "Correlation for the Third Virial [65] Spencer, C. E and Danner, R. E, "Prediction of Bubble Point
Coefficient Using Tc, Pc and 0~as Parameters," American Pressure of Mixtures," Journal of Chemical and Engineering
Institute of Chemical Engineers Journal, Vol. 29, No. 1, 1983, Data, Vol. 18, No. 2, 1973, pp. 230-234.
pp. 107-113. [66] Thomson, G. H., Brobst, K. R., and Hankinson, R. W., "An
[45] Bendict, M., Webb, G. B., and Rubin, L. C., "An Empirical Improved Correlation for Densities of Compressed Liquids and
Equation for Thermodynamic Properties of Light Liquid Mixtures," American Institute of Chemical Engineers
Hydrocarbons and Their Mixtures, Methane, Ethane, Propane Journal, Vol. 28, No. 4, 1982, pp. 671-676.
and n-Butane," Journal of Chemical Physics, Vol. 8, 1940, [67] Tsonopoulos, C., Heidman, J. L., and Hwang, S.-C.,
pp. 334-345. Thermodynamic and Transport Properties of Coal Liquids, An
[46] Starling, K. E., Fluid Thermodynamic Properties for Light Exxon Monograph, Wiley, New York, 1986.
Petroleum Systems, Gulf Publishing, Houston, TX, 1973. [68] Garvin, J., "Estimating Compressed Liquid Volumes for
[47] Prausnitz, J. M., Lichtenthaler, R. N., and de Azevedo, E. G., Alcohols and Diols," Chemical Engineering Progress, 2004,
Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed., Vol. 100, No. 3, 2004, pp. 43--45.
Prentice-Hall, New Jersey, 1999. [69] Chueh, E L. and Prausnitz, J. M., "A Generalized Correlation
[48] Rao, R. V. G. and Dutta, S. K., "An Equation of State of L-J for the Compressibilities of Normal Liquids," American
Fluids and Evaluation of Some Thermodynamic Properties," Institute of Chemical Engineers Journal, Vol. 15, 1969, p. 471.
Zeitschrift f~r Physikalische Chemie, Vol. 264, 1983, pp. 771-783. [70] Riazi, M. R. and Mansoori, G. A., "Simple Equation of State
[49] Rao, R. V. G. and Gupta, B. D., "Visco-Elastic Properties of Accurately Predicts Hydrocarbon Densities," Oil and Gas
Ionic Liquids," Acoustics Letter, Vol. 8, No. 9, 1985, pp. 147-149. Journal, July 12, 1993, pp. 108-111.
[50] Chapman, W. G., Gubbins, K. E., Jackson, G., and Radosz, M., [71] Riazi, M. R. and Roomi, Y., "Use of the Refractive Index in the
"New Reference Equation of State for Associating Liquids," Estimation of Thermophysical Properties of Hydrocarbons and
Industrial and Engineering Chemistry Research, Vol. 29, 1990, Their Mixtures," Industrial and Engineering Chemistry
pp. 1709-1721. Research, Vol. 40, No. 8, 2001, pp. 1975-1984.
[51] Li, X.-S. and Englezos, E, "Vapor-Liquid Equilibrium of [72] Goodwin, R. D., The Thermophysical Properties of Methane from
Systems Containing Alcohols Using the Statistical Associating 90 to 500 K at Pressures to 700 Bar, National Bureau of
Fluid Theory Equation of State," Industrial and Engineering Standards, NBS Technical Note 653, April 1974.
Chemistry Research, Vol. 42, 2003, pp. 4953-4961. [73] Hall, K. R., Ed., TRC Thermodynamic Table--Hydrocarbons,
[52] Standing, M. B. and Katz, D. L., "Density of Natural Gases," Thermodynamic Research Center, Texas A&M University
AIME Transactions, Vol. 146, 1942, pp. 140-149. System, 1986.
[53] Engineering Data Book, Gas Processors (Suppliers) Association, [74] Goodwin, R. D., Roder, H. M., and Starty, G. C., The
10th ed., Tulsa, OK, 1987. Thermophysical Properties of Ethane from 90 to 600 K at
[54] Ahmed, T. Reservoir Engineering Handbook, Gulf Publishing, Pressures to 700 Bar, National Bureau of Standards, NBS
Houston, TX, 2000. technical note 684, August 1976.
[55] Hall, K. R. and Yarborough, L., "A New Equation of State for [75] McCarty, R. D. and Jaconsen, R. T., An Equation of State for
Z-Factor Calculations," Oil and Gas Journal, June 18, 1973, Fluid Ethylene, National Bureau of Standards, NBS Technical
pp. 82-92. Note 1045, July 1981.
[56] Pitzer, K. S., Lipmann, D. Z., Curl, R. E, Huggins, C. M., and [76] Goodwin, R. D. and Hayens, W. M., The Thermophysical
Petersen, D. E., "The Volumetric and Thermodynamic Properties of Propane from 85 to 700 K at Pressures to 70 MPa,
Properties of Fluids, II: Compressibility Factor, Vapor Pressure, National Bureau of Standards, NBS Monograph 170, April
and Entropy of Vaporization," Journal of American Chemical 1982.
Society, Vol. 77, 1955, pp. 3433-3440. [77] Haynes, W. M. and Goodwin, R. D., The Thermophysical
[57] Pitzer, K, S., Thermodynamics, 3rd ed., MacGraw-Hill, New Properties of Propane from 85 to 700 K at Pressures to 70 MPa,
York, 1995. National Bureau of Standards, NBS Monograph 170, April
[58] Lee, B. I. and Kesler, M. G., "A Generalized Thermodynamic 1982.
Correlation Based on Three-Parameter Corresponding States," [78] Doolittle, A. K., "Specific Volumes of Normal Alkanes,"
American Institute of Chemical Engineers Journal, Vol. 21, 1975, Journal of Chemical and Engineering Data, Vol. 9, 1964, pp.
pp. 510-527. 275-279.
[59] Daubert, T. E. and Danner, R. E, Eds., API Technical Data [79] Riazi, M. R. and Mansoori, G. A., "Use of the Velocity of Sound
Book--Petroleum Refining, 6th ed., American Petroleum in Predicting the PVT Relations," Fluid Phase Equilibria,
Institute (API), Washington, DC, 1997. Vol. 90, 1993, pp. 251-264.
[60] Elliott, J. R. and Lira, C. T., Introductory Chemical Engineering [80] Gupta, D. and Eubank, E T., "Density and Virial Coefficients of
Thermodynamics, Prentice Hall, New Jersey, 1999 Gaseous Butane from 265 to 450 K Pressures to 3.3 Mpa,"
(www.phptr.com). Journal of Chemical and Engineering Data, Vol. 42, No. 5,
[61] Rackett, H. G., "Equation of State for Saturated Liquids," Sept.-Oct. 1997, pp. 961-970.
Journal of Chemical and Engineering Data, Vol. 15, No. 4, 1970, [81] Konttorp, K., Peters, C. J., and O'Connell, J. E, "Derivative
pp. 514-517. Properties from Model Equations of State," Paper presented at
[62] Spencer, C. E and Danner, R. E, "Improved Equation for the Ninth International Conference on Properties and Phase
Prediction of Saturated Liquid Density," Journal of Chemical Equilibria for Product and Process Design (PPEPPD 2001),
and Engineering Data, Vol. 17, No. 2, 1972, pp. 236-241. Kurashiki, Japan, May 20-25, 2001.
[63] Spencer, C. E and Adler. S. B., "A Critical Review of Equations [82] Sengers, J. V., Kayser, R. E, Peters, C. J., and White, Jr., H. J.,
for Predicting Saturated Liquid Density," Journal of Chemical Eds., Experimental Thermodynamics, Vol. V: Equations of State
and Engineering Data, Vol. 23, No. 1, 1978, pp. 82-89. for Fluids and Fluid Mixtures, IUPAC, Elsevier, Amsterdam,
[64] Yamada, T. G., "Saturated Liquid Molar Volume: the Rackett 2000.
Equation," Journal of Chemical and Engineering Data, Vol. 18, [83] National Institute of Standards and Technology (NIST),
No. 2, 1973, pp. 234-236. Boulder, CO, 2003, http://webbook.nist.gov/chemistry/fluid/.
MNL50-EB/Jan. 2005
Thermodynamic Relations
for Property Estimations
232
SG g Specific gravity of gas fluid (pure or mixture) o" Molecular size parameter,/~ (10 -1~ m)
[= Mg/29], dimensionless (9 Acentric factor defined by Eq. (2.10)
SG Specific gravity of liquid substance at 15.5~ (60~ Packing fraction defined by Eq. (5.86),
defined by Eq. (2.2), dimensionless dimensionless
T Absolute temperature, K K Isothermal compressibility defined by Eq. (6.25),
Tc Critical temperature, K bar -I
Tr Reduced temperature defined by Eq. (5.100) Joule-Thomson coefficient defined by Eq. (6.27),
(= T/Tc), dimensionless K/bar
TB A parameter in the Standing correlation Y Heat capacity ratio (-- Ce/Cv), dimensionless
(Eq. 6.202), K Y/ Activity coefficient of component i in liquid solution
TM Freezing (melting) point for a pure component at defined by Eq. (6.112), dimensionless
1.013 bar, K Activity coefficient of a solid solute (component 1) in
Ttp Triple point temperature, K the liquid solution defined by Eq. (6.161), dimension-
V Molar volume, cm3/gmol less
Vt Saturated liquid molar volume, cm3/gmol Activity coefficient of component i in liquid solution
Vsat Saturation molar volume, cm3/gmol at infinite dilution (x4 ~ 0), dimensionless
Vv Saturated vapor molar volume, cm3/gmol Chemical potential of component i defined in
Vc Critical volume (molar), cm3/Inol (or critical specific Eq. (6.115)
volume, cm3/g) ACpi Difference between heat capacity of liquid and solid
Vr Reduced volume (= V/Vc) for pure component i (= cLi -- CSi), J/mol. K
V25 Liquid molar volume at 25~ cm3/mol Heat of fusion (or latent heat of melting) for pure
x4 Mole fraction of component i in a mixture (usually component i at the freezing point and 1.013 bar,
used for liquids), dimensionless J/tool
Xwi Weight fraction of component i in a mixture (usually A H yap Heat of vaporization (or latent heat of melting) at
used for liquids), dimensionless 1.013 bar defined by Eq. (6.98), J/mol
yi Mole fraction of i in mixture (usually used for gases), AHmix Heat of mixing. J/mol
dimensionless AM Property change for M due to mixing defined by
Z Compressibility factor defined by Eq. (5.15), Eq. (6.84)
dimensionless AS/ Entropy of fusion for pure component i at the freezing
Z L Compressibility factor of liquid phase, point and 1.013 bar, J/mol-K
dimensionless ASv~p Entropy of vaporization at 1.013 bar defined by
Zv Compressibility factor of vapor phase, dimensionless Eq. (6.97), J/mol
ATb2 Boiling point elevation for solvent 2 (Eq. 6.214), K
ATM 2 Freezing point depression for solvent 2
Greek Letters (Eq. 6.213), K
Parameter defined by Eq. (5.41), dimensionless AVm~x Volume change due to mixing defined by Eq. (6.86)
Ors, ]~S Parameters defined based on velocity of sound for
correction of EOS parameters a and b defined by
Eq. (6.242), dimensionless
Superscript
fl Coefficient of thermal expansion defined by E Excess property defined for mixtures (with respect to
Eq. (6.24), K -1. ideal solution)
A Difference between two values of a parameter exp Experimental value
~i Solubility parameter for i defined in Eq. (6.147), HS Value of a property for hard sphere molecules
(J/cm3)l/2 ig Value of a property for a component as ideal gas at
8i Parameter used in Eq. (6.126), dimensionless temperature T and P --~ 0
3ii Parameter defined in Eq. (5.70) id Value of a property for an ideal solution
e Energy parameter in a potential energy function L Value of a property for liquid phase
e Error parameter defined by Eq. (106), dimensionless R A residual property (with respect to ideal gas
qI)iVolume fraction of i in a liquid mixture defined by property)
Eq. (6.146) V Value of a property for vapor phase
Fugacity coefficient of pure i at T and P defined by vap Change in value of a property due to vaporization
Eq. (6.49), dimensionless S Value of a property for solid phase
q~i Fugacity coefficient of component i at T and P in an sat Value of a property at saturation pressure
ideal solution mixture, dimensionless t Value of a property for the whole (total) system
6i Fugacity coefficient of component i in a mixture at T [](0) A dimensionless term in a generalized correlation for
and P defined by Eq. (6.110) a property of simple fluids
0 A parameter defined in Eq. (6.203), dimensionless [](1) A dimensionless term in a generalized correlation for
p Density at a given temperature and pressure, g/cm 3 a property of acentric fluids
(molar density unit: cma/mol) [](r) A dimensionless term in a generalized correlation for
a Diameter of hard sphere molecule,/~ (10 -l~ rn) a property of reference fluids
234 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
or, fl Value of a property for phase a or phase fl IN CHAPTER 5 THE PVT relations and theory of intermolecu-
oo Value of a property for i in the liquid solution at lar forces were discussed. The PVT relations and equations of
infinite dilution as x~ --~ 0 states are the basis of property calculations as all physical and
o Value of a property at standard state, usually the thermodynamic properties can be related to PVT properties.
standard state is chosen at pure component at T In this chapter we review principles and theory of property
and P of the mixture according to the estimation methods and basic thermodynamic relations that
Lewis/Randall rule will be used to calculate physical and thermodynamic prop-
A Value of molar property of a component in the erties.
mixture The PVT relations and equations of state are perhaps the
most important thermodynamic relations for pure fluids and
their mixtures. Once the PVT relation is known, various phys-
Subscripts ical and thermodynamic properties needed for design and op-
eration of units in the petroleum and related industries can be
cValue of a property at the critical point
calculated. Density can be directly calculated from knowledge
iA component in a mixture
of molar volume or compressibility factor through Eq. (5.15).
jA component in a mixture
Various thermodynamic properties such heat capacity, en-
i, [Effect of binary interaction on a property
thalpy, vapor pressure, phase behavior and vapor liquid equi-
m Value of a property for a mixture
librium (VLE), equilibrium ratios, intermolecular parame-
mix Change in value of a property due to mixing at
ters, and transport properties all can be calculated through
constant T and P
accurate knowledge of PVT relation for the fluid. Some of
PR Value of a property determined from PR EOS
these relations are developed in this chapter through funda-
SRK Value of a property determined from SRK EOS
mental thermodynamic relations. Once a property is related
to PVT, using an appropriate EOS, it can be estimated at any
Acronyms temperature and pressure for pure fluids and fluid mixtures.
Development of such important relations is discussed in this
API-TDB American Petroleum Institute--Technical Data chapter, while their use to estimate thermophysical properties
Book for petroleum mixtures are discussed in the next chapter.
BIP Binary interaction parameter
bbl Barrel, unit of volume of liquid as given in Section
1.7.11 6.1 DEFINITIONS AND FUNDAMENTAL
CS Carnahan-Starling EOS (see Eq. 5.93) THERMODYNAMIC RELATIONS
DIPPR Design Institute for Physical Property Data
EOS Equation of state In this section, thermodynamic properties such as entropy,
GC Generalized correlation Gibbs energy, heat capacity, residual properties, and fugacity
GD Gibbs-Duhem equation (see Eq. 6.81) are defined. Thermodynamic relations that relate these prop-
HS Hard sphere erties to PVT relation of pure fluids are developed.
HSP Hard sphere potential given by Eq. (5.13)
IAPWS International Association for the Properties of
6.1.1 Thermodynamic Properties and
Water and Steam
Fundamental Relations
LJ Lennard-Jones potential given by Eq. (5.1 I)
LJ EOS Lennard-Jones EOS given by Eq. (5.96) Previously two thermodynamic properties, namely internal
LK EOS Lee-Kesler EOS given by Eq. (5.104) energy (U) and enthalpy (H), were defined in Section 5.1.
LLE Liquid-liquid equilibria The enthalpy is defined in terms of U and P V (Eq. 5.5) as
NIST National Institute of Standards and Technology
(6.1) H = U + PV
PVT Pressure-volume-temperature
PR Peng-Robinson EOS (see Eq. 5.39) Another thermodynamic property that is used to formulate
RHS Right-hand side of an equation the second law of thermodynamics is called entropy and it is
RK Redlich-Kwong EOS (see Eq. 5.38) defined as
SRK Soave-Redlich-Kwong EOS given by Eq. (5.38)
and parameters in Table 5.1 (6.2) dS = 8O~v
T
SAFT Statistical associating fluid theory (see
where S is the entropy and ~Q~v is the amount of heat trans-
Eq. 5.98)
ferred to the system at temperature T through a reversible
SLE Solid-liquid equilibrium
process. The symbol ~ is used for the differential heat Q to
SLVE Solid-liquid-vapor equilibrium
indicate that heat is not a thermodynamic property such as
VLE Vapor-liquid equilibrium
VLS Vapor-liquid-solid equilibrium H or S. The unit of entropy is energy per absolute degrees, e.g.
J/K, or on a molar basis it has the unit of J/tool. K in the SI
VS Vapor-solid equilibrium
%AAD Average absolute deviation percentage defined by unit system. The first law of thermodynamics is derived based
on the law of conservation of energy and for a closed system
Eq. (2.135)
%AD Absolute deviation percentage defined by (constant composition and mass) is given as follows [1, 2]:
Eq. (2.134) (6.3) dU = 8Q - P d V
6. T H E R M O D Y N A M I C R E L A T I O N S FOR P R O P E R T Y E S T I M A T I O N S 235
Combining Eqs. (6.2) and (6.3) gives the following relation: 6.1.2 Measurable Properties
(6.4) dU = T d S - PdV In this section some thermodynamic properties that are di-
This relation is one of the fundamental thermodynamic rela- rectly measurable are defined and introduced. Heat capacity
tions. Differentiating Eq. (6.1) and combining with Eq. (6.4) at constant pressure (Cp) and heat capacity at constant vol-
gives ume (Cv) are defined as:
There are two other measurable properties: coefficient of Values of C~ are known for many compounds and they are
thermal expansion, r, and the bulk isothermal compressibility, given in terms of temperature in various industrial handbooks
g. These are defined as [5]. Once C~ is known, C~, Uig, H ig, and Sig can also be de-
termined from thermodynamic relations discussed above. To
(6.24) fl=V ~ p calculate properties of a real gas an auxiliary function called
residual property is defined as the difference between prop-
erty of real gas and its ideal gas property (i.e., H - Hig). The
(6.25) x----V V r difference between property of a real fluid and ideal gas is
since OV/OP is negative, the minus sign in the definition of also called departure from ideal gas. All fundamental relations
x is used to make it a positive number. The units of fl and also apply to residual properties. By applying basic thermo-
x in SI system are K -1 and Pa -1, respectively. Values of fl dynamic and mathematical relations, a residual property can
and r can be calculated from these equations with use of an be calculated through a PVT relation of an equation of state.
equation of state. For example, with use of Lee-Kesler EOS If only two properties such as H and G or H and S are known
(Eq. 5.104), the value of ~ is 0.84 x 10 -9 Pa -1 for liquid ben- in addition to values of V at a given T and P, all other prop-
zene at temperature of 17~ and pressure of 6 bar, while the erties can be easily determined from basic relations given in
actual measured value is 0.89 x 10-9 Pa -1 [6]. Once fl and this section. For example from H and G, entropy can be calcu-
are known for a fluid, the PVT relation can be established lated from Eq. (6.6). Development of relations for calculation
for that fluid (see Problem 6.1). Through the above thermo- of enthalpy departure is shown here. Other properties may be
dynamic relations and definitions one can show that calculated through a similar approach.
Assume that we are interested to relate residual enthalpy
TVB 2 (H - HiE) into PVT at a given T and P. For a homogenous
(6.26) Ce - Cv = - - g
fluid of constant composition (or pure substance), H can be
Applying Eqs. (6.24) and (6.25) for ideal gases (Eq. 5.14) considered as a function of T and P:
gives Big = I / T and gig ___ 1/P. Substituting Big and gig into
(6.28) H = H(T, P)
Eq. (6.26) gives Eq. (6.23). From Eq. (6.26) it is clear that
Cp > Cv; however, for liquids the difference between Cp and Applying Eq. (6.10) gives
Cv is quite small and most thermodynamic texts neglect this
OH
difference and assume Cp ~- Cv. Most recently Garvin [6] has (6.29) dH=(-~-~)edT+(~p)rdP
reviewed values of constant volume specific heats for liquids
and concludes that in some cases Cp - Cv for liquids is signif- Dividing both sides of Eq. (6.5) to OP at constant T gives
icant and must not be neglected. For example, for saturated
liquid benzene when temperature varies from 300 to 450 K,
the calculated heat capacity ratio, Cp/Cv, varies from 1.58
(6.30) ()
OH
~ r
=V+T
r
to 1.41 [6]. Although these values are not yet confirmed as Substituting for (OS/OP)T from Eq. (6.15) into Eq. (6.30)
they have been calculated from Lee-Kesler equation of state, and substitute resulting (OH/OP)T into Eq. (6.29) with use of
but one should be careful that assumption of Cp ~- Cv for liq- Eq. (6.20) for (OH/OT)p,Eq. (6.29) becomes
uids in general may not be true in all cases. In fact for ideal
incompressible liquids, B ~ 0 and x - ~ 0 and according to (6.31) d H = C p d T + [ V-T(OV~kOT/Pj]dP
Eq. (6.26), (Cp - Cv) --~ O, which leads to y = Cp/Cv --~ 1.
There is an EOS with high accuracy for benzene [7]. It gives where the right-hand side (RHS) of this equation involves
Cp/C~ for saturated liquids having a calculated heat capacity measurable quantities of Cp and PVT, which can be deter-
ratio of 1.43-1.38 over a temperature range of 300-450 K. mined from an equation of state. Similarly it can be shown
Another useful property is Joule-Thomson coefficient that that
is defined as
(6.32) dS=C~l- -~ e
Equations (6.31) and (6.32) are the basis of calculation of en-
This property is useful in throttling processes where a fluid
thalpy and entropy and all other thermodynamic properties
passes through an expansion valve at which enthalpy is nearly
of a fluid from its PVT relation and knowledge of Cp. As an ex-
constant. Such devices are useful in reducing the fluid pres-
ample, integration of Eq. (6.31) from (T~, PL) to (T2, P2) gives
sure, such as gas flow in a pipeline, tj expresses the change of
change of enthalpy (AH) for the process. The same equation
temperature with pressure in a throttling process and can be
can be used to calculate departure functions or residual prop-
related to C~ and may be calculated from an equation of state
erties from PVT data or an equation of state at a given T and
(see Problem 6.10).
P. For an ideal gas the second term in the RHS of Eq. (6.32)
is zero. Since any gas as P --->0 behaves like an ideal gas, at a
6.1.3 Residual Properties and fixed temperature of T, integration of Eq. (6.31) from P --+ 0
Departure Functions to a desired pressure of P gives
Properties of ideal gases can be determined accurately P
0V
through kinetic theory. In fact all properties of ideal gases are (6.33) (H-H'g)r=f[V-T(~-f)p]dP (at constant T)
known or they can be estimated through the ideal gas law. 0
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 237
For practical applications the above equation is converted Solution--By substituting the Maxwelrs relation of Eq. (6.15)
into dimensionless form in terms of parameters Z defined by into Eq. (6.30) we get
Eq. (5.15). Differentiating Z with respect to T at constant P,
from Eq. (5.15) we get (6.40) ~ r \ST ]e
differentiating this equation with respect to T at constant P
(6.34) ~ e = ~ kS--T)e + ~ - gives
Dividing both sides of Eq. (6.33) by RT and combining with
Eq. (6.34) gives
P
(6.35) H - H ig f(OZ) dP (6.41) = -T \ 8T2,iv
R~ - T 8-T e --P- (at constant T)
0 From mathematical identity we have
0H 3H
It can be easily seen that for an ideal gas where Z - - 1 , (6.42) [ ~---~(-~-ff)r ] l = [ 8 (-ff~ )p]r
Eq. (6.35) gives the expected result of H - H i g = 0. Similarly
for any equation of state the residual enthalpy can be calcu- Using definition of Cv through Eq. (6.20) and combining the
lated. Using definitions of Tr and Pr by Eq. (5.100), the above above two equations we get
equation may be written as
(6.43) \ OP Jr =-T \ ~T~je
(6.36)
H-H
R~
ig
-- Tr2
f ( O~Z ) dPr
P~ ~ (at constant T) Upon integration from P = 0 to the desired pressure of P at
0 constant T we get
where the term in the left-hand side and all parameters in the P
V
(6.46) dG ig -- RTd In P
(6.39) (S-~g)r,v : f [ ( ~ T ) v - R ] d v For a real fluid a similar relation can be written in terms of
fugacity
Once H is known, U can be calculated from Eq. (6.1). Sim- (6.47) dG -- RTdln f
ilarly all other thermodynamic properties can be calculated where for an ideal gas fig = p. Subtracting Eq. (6.46) from
from basic relations and definitions. (6.47), the residual Gibbs energy, G R, can be determined
through fugacity:
Example 6.1--Derive a relation for calculation of Cp from GR G Gig
PVT relation of a real fluid at T and P. (6.48) RT -- g ~ - In f/-, = In ~b
238 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
1)3 1 In V
--/;a2 1 - In v+b--,/~
1a _ a~g 1/2 f ctI/2
--~ y rcTrl/2 I~
d2 r--~
3a
2r~r)/2 fo,( + f~)
a~ I o~ G ( t + f~)
2r2r)n
a = aca, where ac, a, f~, and b for RK SRK, and PR EOS are givenin Table 5.1.
The ratio of f / P is a dimensionless parameter called fugacity For the residual heat capacity (Ce - C~), the relation at a
coefficient and it is shown by r fixed T and P is given by Eq. (6.44). In general when fugac-
ity coefficient is calculated through Eq. (6.53), residual Gibbs
(6.49) ~b = --f energy can be calculated from Eq. (6.48). Properties of ideal
P
gases can be calculated accurately as will be discussed later
where for an ideal gas, r = 1. Once q~ is known, G can be in this chapter. Once H and G are known, S can be calcu-
calculated through Eq. (6.48) and from G and H, S m a y be lated from Eq. (6.6). Therefore, either H and S or H and r
calculated from Eq. (6.6). Vice versa when H and S are known, are needed to calculate various properties. In this chapter,
G and eventually f can be determined. methods of calculation of H, Ce, and r are presented.
When residual properties are related to PVT, any equation
6.1.5 General Approach for Property Estimation of state m a y be used to calculate properties of real fluids and
departure functions. Calculation of (H - Hig), (Cp - C~), and
Similar to the m e t h o d used in Example 6.1, every thermody- In r from RK, SRK, and PR equations of state are given in Ta-
namic property can be related to PVT relation either at a given ble 6.1. RK and SRK give similar results while the only differ-
T and P or at a given T and V. These relations for (H -/_/ig) ence is in parameter a, as given in Table 5. i. EOS parameters
are given by Eqs. (6.33) and (6.39). For the residual entropy an needed for use in Table 6.1 are given in Table 5.1. Relations
equivalent relation in terms of pressure is (see Problem 6.3) presented in Table 6.1 are applicable to both vapor and liq-
P uid phases whenever the EOS can be applied. However, when
(6.50) (S-Sig)r,, = f [ R - (~----f)e] dP they are used for the liquid phase, values of Z and V must
0
be obtained for the same phase as discussed in Chapter 5.
It should be noted that relations given in Table 6.1 for var-
This equation can be written in terms of Z as ious properties are based on assuming that parameter b in
P P the corresponding EOS is independent of temperature as for
(S- ~g)r,P:-Rf ~---~dP- RT f d_P_Pp RK, SRK, and PR equations. However, when parameter b is
considered temperature-dependent, then its derivative with
0 0
respect to temperature is not zero and derived relations for
(6.51) (at constant T)
residual properties are significantly m o r e complicated than
Once residual enthalpy and entropy are calculated, residual those given in Table 6. I. As will be discussed in the next sec-
Gibbs energy is calculated from the following relation based tion, cubic equations do not provide accurate values for en-
on Eq. (6.6): thalpy and heat capacity of fluids unless their constants are
adjusted for such calculations.
(6.52) (G -- Gig)T,p = (H - Hig)T,p -- T(S- ~g)T,P
Substituting Eqs. (6.33) and (6.50) into Eq. (6.52) and com-
bining with Eq. (6.48) give s the following equation which can
6.2 GENERALIZED CORRELATIONS
be used to calculate fugacity coefficient for a pure component:
F O R CALCULATION OF
THERMODYNAMIC PROPERTIES
P
(6.53) ln =f(z-1) 0
It is generally believed that cubic equations of state are not
suitable for calculation of heat capacity and enthalpy and
6. T H E R M O D Y N A M I C R E L A T I O N S FOR P R O P E R T Y E S T I M A T I O N S 239
E-2Tr3~,
c4 [ r ( f+l+~ ~/) exp ( _ ~r2
Y)1
~ 5
I
The coefficients in the above equations for the simple fluid
and reference fluid of n-octane are given in Table 5.8. Sim-
ilar equations for estimation of (Cv -Cipg), (Cv Cv), ig and -
(o)
TABLE 6.2--Values of --[ B~--T~] for use in Eq. (6.56).
L ~ J
For low and moderate pressures where truncated virial m a y b e u s e d f r o m Eq. (5.71), b u t c o r r e s p o n d i n g d e r i v a t i v e s
e q u a t i o n i n t h e f o r m o f E q . (5.113) is v a l i d , t h e r e l a t i o n f o r must be used, The above equation may be applied at the same
f u g a c i t y c o e f f i c i e n t c a n b e d e r i v e d f r o m Eq. (6.53) a s r e g i o n t h a t E q . (5.75) o r (5.114) w e r e a p p l i c a b l e , t h a t is, V~ >
2.0 o r T~ > 0 . 6 8 6 + 0 , 4 3 9 P r [2].
BP
(6.61) ln(~) - For real gases that follow truncated virial equation with
RT three terms (coefficients D and higher assumed zero in
T h i s r e l a t i o n m a y a l s o b e w r i t t e n as E q . 5.76), t h e r e l a t i o n s f o r Cp a n d Cv a r e g i v e n s a s
(6.62) ~ -~ e x p
[Pr(SPc]l
Z \R-~}J
Cv - C ~ _ T " (B - TB") z - C + TC' - T z C " / 2
w h e r e (BPc/RTc) c a n h e c a l c u l a t e d f r o m Eq. (5.71) o r (5.72).
Similarly enthalpy departure based on the truncated virial (6.64)
R
E: V2
]
e q u a t i o n is g i v e n a s [ 1 ] Cv - C ~ [ 2 T B ' + T2B" TC' + T 2 C ' / 2 ]
(6.65)
(6.63) H - H ig _
RTo Pr
[B(0 ) d B (~ (
Tr-d-~ + o~ \ B (1) -
d B (1) --W -= [ v v-~ "j
where B' and C' are the first-order derivatives of B and C with
w h e r e B (~ a n d B (1) a r e g i v e n b y Eq. (5.72) w i t h dB(~ = respect to temperature, while B" and C" are the second-order
0 . 6 7 5 / T r 26 a n d d B 0 ) / d T r = 0.722/Tr s2. O b v i o u s l y B (~ a n d B (1) derivatives of B and C with respect to temperature.
6. TH E RMO D YNA MIC R E LA TIO NS FOR P R O P E R T Y E S T I M A T I O N S 241
6.3 P R O P E R T I E S OF IDEAL G A S E S c h a n g e w i t h t e m p e r a t u r e a p p r e c i a b l y . As t h e n u m b e r o f
atoms in a molecule increases, dependency of ideal gas prop-
Calculation of thermodynamic properties through the meth- erties to temperature also increases. Data on properties of
o d s o u t l i n e d a b o v e r e q u i r e s p r o p e r t i e s o f i d e a l gases. B a s e d ideal gases for a large number of hydrocarbons have been
o n d e f i n i t i o n o f i d e a l gases, U ig, H ig, a n d C ~ a r e f u n c t i o n s r e p o r t e d b y t h e A P I - T D B [5]. T h e s e d a t a f o r i d e a l g a s h e a t ca-
of only temperature. Kinetic theory shows that the mo- pacity have been correlated to temperature in the following
l a r t r a n s l a t i o n a l e n e r g y o f a m o n o a t o m i c i d e a l g a s is 3 R T , form [5]:
w h e r e R is t h e u n i v e r s a l g a s c o n s t a n t a n d T is t h e a b s o -
l u t e t e m p e r a t u r e [ 10]. S i n c e f o r i d e a l g a s e s t h e i n t e r n a l e n - (6.66) C~Pg = A -b B T -k C T 2 + D T 3 q- E T 4
e r g y is i n d e p e n d e n t o f p r e s s u r e t h u s U ig = 23- RT. I This leads R
9 . . "g
t o H lg = ~ R 5
T , Cpl g = ~ R 5
, C vl g = ~3R , a n d F = 7 ~ -C-p - - - ~ =
5
1.667. w h e r e R is t h e g a s c o n s t a n t ( S e c t i o n 1.7.24), C ~ is t h e m o l a r
C
I d e a l g a s h e a t c a p a c i t y o f m o n o a t o m i c g a s e s s u c h a s ar- h e a t c a p a c i t y i n t h e s a m e u n i t a s R, a n d T is t h e a b s o l u t e
g o n , h e l i u m , etc. a r e c o n s t a n t w i t h r e s p e c t t o t e m p e r a t u r e temperature in kelvin. Values of the constants for a number
[10]. S i m i l a r l y f o r d i a t o m i c g a s e s s u c h as N2, 0 2 , air, etc., of nonhydrocarbon gases as well as some selected hydrocar-
H ig = ~R,7 w h i c h l e a d s t o C ~ -- ~R,7 CV 5
i g = 2R, a n d y -- 7 / 5 = b o n s a r e g i v e n i n T a b l e 6.8. T h e t e m p e r a t u r e r a n g e a t w h i c h
1.4. I n f a c t v a r i a t i o n o f h e a t c a p a c i t i e s o f i d e a l d i a t o m i c t h e s e c o n s t a n t s c a n b e u s e d is a l s o g i v e n f o r e a c h c o m p o u n d
g a s e s w i t h t e m p e r a t u r e is v e r y m o d e r a t e . F o r m u l t i a t o m i c i n T a b l e 6.8. F o r a c o m p o u n d w i t h k n o w n c h e m i c a l s t r u c -
molecules such as hydrocarbons, ideal gas properties do t u r e , i d e a l g a s h e a t c a p a c i t y is u s u a l l y p r e d i c t e d f r o m g r o u p
242 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
ofCff = A + B T + C(SG g) + D(SGg) 2 + E[T(SGg)], w h e r e T is range of 16-60. The average deviation for this e q u a t i o n for
t e m p e r a t u r e and SG g is gas specific gravity (Mg/29), Although these ranges is 5%; however, w h e n it is applied in the tem-
the e q u a t i o n is very useful for calculation of C~ of undefined p e r a t u r e range of 200-1000 K and m o l e c u l a r weight r a nge
na tur al gases but using the r e p o r t e d coefficients we could not of 16-50, the e r r o r reduces to 2.5%. Use of this e q u a t i o n is
obtain reliable values for C~. In a n o t h e r correlation, ideal r e c o m m e n d e d for undefined light h y d r o c a r b o n gas mixtures
heat capacities of h y d r o c a r b o n s (Nc > C5) were related to w h e n gas specific gravity (SG g) is k n o w n (M = 29 • SGg), F o r
boiling p o i n t an d specific gravity in the f o r m of Eq. (2.38) defined h y d r o c a r b o n m i x t u r e s of k n o w n c o m p o s i t i o n the fol-
at t h ree t e m p e r a t u r e s o f 0~ (~255 K), 600~ ( ~ 5 8 9 K), lowing relation m a y be used to calculate m i x t u r e ideal gas
a nd 1200~ (922 K) [18]. F o r light gases based on the data heat capacity:
g e n e r a t e d t h r o u g h Eq. (6.66) for c o m p o u n d s f r o m C1 to ig ig
C5 with H2S, CO2, a n d N2 the following relation has b e e n (6.74) Cp'mixR = EYi-RCpi
de t e r m i n ed : i
w h e r e Cpg is the m o l a r ideal heat capacity of c o m p o n e n t i
(6.73) -~-
C~ = ~2( A i + ~M)~ i (with m o l e fraction yi) an d m a y be calculated f r o m Eq. (6.66).
i=0
for natural and light gases with 16 < M < 60 E x a m p l e 6 . 2 - - C a l c u l a t e Cp for saturated liquid b e n z e n e at
where 450 K an d 9.69 b ar using generalized co r r el at i on and S R K
A0 = 3.3224 B0 = - 2 . 5 3 7 9 x 10 -2 E O S an d c o m p a r e with the value of 2.2 kJ/kg. K as given
A1 = - 7 . 3 3 0 8 x 10 -3 B1 = 7.5939 x 10 -4 in Ref. [6]. Also calculate heat capacity at c o n s t a n t volume,
A2 = 4.3235 x 10 -6 B2 = - 2 . 6 5 6 5 • 10 -7 heat capacity ratio, and residual enthalpy ( H - H ig) f r o m b o t h
S R K E O S an d generalized correlations of LK.
in w h i c h T is the absolute t e m p e r a t u r e in kelvin. This equa-
tion is based o n m o r e than 500 data points g e n e r a t e d in S o l u t i o n - - F r o m Table 2.1 for b e n z e n e w e have Tc = 562 K,
the t e m p e r a t u r e r an g e of 50-1500 K a n d m o l e c u l a r weight Pc = 49 bar, ~o = 0.21, M = 78.1, an d Kw = 9.72 a nd f r o m
6. T H E R M O D Y N A M I C RELATIONS FOR PROPERTY ESTIMATIONS 245
TABLE 6.8---Constants for Eqs. (6.66)-(6.69) for ideal gas heat capacity, enthaIpy, and entropy.
No. Compound name Formula M A B x 10 3 C x 10 6 D x 1010 E • 10 I4 Trrdn,K Tmax,K
Parat~lil$
1 Methane CH4 16.043 4.34610 -6.14488 26.62607 -219.2998 588.89965 50 1500
2 Ethane C2H6 30.070 4.00447 -1.33847 42.86416 -452.2446 1440.4853 50 1500
3 Propane C3H8 44.096 3.55751 10.07312 39.13602 -475.7220 1578.1656 50 1500
4 n-Butane C4Hlo 58.123 2.91601 28.06907 15.37435 -292.9255 1028.0462 200 1500
5 Isobutane C4H1o 58,123 2.89796 25.14031 26.04226 -405.3691 1396.6324 50 1500
6 n-Pentane C5H12 72,150 4.06063 29.87141 30.46993 -461.3523 1559.8971 200 1500
7 Isopentane C5H12 72,150 0.61533 49.99361 -9.72605 -121.1597 563.52870 200 1500
8 Neopentane C5H12 72.150 6.60029 24.43268 32.52759 -402.96336 1258.46299 220 1500
9 n-Hexane C6I-I14 86.177 3.89054 41.42970 24.35860 -457.5222 1599.4100 200 1500
10 n-Heptane C7H12 100.204 4.52739 47.36877 31.09932 -570.22085 1999.68224 200 1500
11 n-Octane C8Ht8 114.231 4.47277 57.81747 29.07465 -621.09106 2265.33690 200 1500
12 n-Nonane C9H20 128.258 3.96754 68.72207 31.85998 -758.47191 2875.17975 200 1000
13 n-Decane CIOH22 142.285 14.56771 -9.12133 283.5241 -3854.9259 16158.7933 200 1000
14 n-Undecane CllH24 156.312 15.72269 -8.39015 308.0195 -4205.1509 17634.1470 200 1000
15 n-Dodecane C12H26 170.338 16.87761 -7.65919 332.5144 -4555.3529 19109.3961 2O0 1000
16 n-Tridecane C13H28 184.365 30.63938 -107.2144 632.4036 -8053.8502 33377.9390 200 1000
17 n-Tetradecane C14H30 198.392 -2.95801 -6.19822 381.5291 -5256.0878 22061.2353 200 1000
18 n-Pentadecane C15H32 212.419 -2.65315 -5.09511 404.3408 -5576.7343 -23366.9827 200 1000
19 n-Hexadecane C16H34 226.446 -36.57941 -4.73820 430.5450 -5956.8328 25013.0768 200 1000
20 n-Heptadecane C17H36 240.473 23.25896 -4.00829 431.5163 -6307.0717 26488.4354 200 1000
21 n-Octadecane C18H38 254.500 -2.20866 -3.27840 479.5420 -6657.3161 27963.8136 200 1000
22 n-Nonadecane C19H40 268.527 25.68345 -2.54834 504.0402 -7007.5535 29439.1464 20O 1000
23 n-Eicosane C20H42 282.553 26.82718 -1.81886 528.5571 -7358.0352 30915.5067 200 1000
24 2-Methylpentane C6H14 86.177 0.44073 60.77573 -10.93570 -180.70573 833.40865 200 1500
25 3-Methylpentane C6H14 86.177 -0.07902 63.31181 -18.82562 -90.58759 510,28364 200 1500
26 2,2-Dimethylbutane C6H14 86.177 1.00342 56.10078 -1.05011 -237.84301 956.53142 200 1500
27 2-Meth~hexane C7H16 100.204 0.57808 70.71556 -15.00679 -187.48705 899.92106 200 1500
28 3-Methylhexane C7H16 100.204 -0.37490 75.26096 -22.63052 -131.65268 744.03635 200 1500
29 2,4-Dimeth~pentane C7H16 100.204 -3.20582 98.77224 -72.48550 293.10189 -500.62465 200 1500
30 2-Meth~heptane C8H18 114.231 0.92650 78.42561 -11,24742 -281.97592 1265.61161 200 1500
31 2,2,4-Trimethyl-pentane C8H18 114.231 -1.85230 96.08105 -47.77416 68.93159 137.05449 200 1500
Olefms
32 Ethylene C2H4 28.054 2.11112 8.32103 11.24746 -155.42099 516.14291 200 1500
33 Prop~ene C3H6 42.081 2.15234 17.76767 8.26700 -166.33525 582.37236 200 1500
34 1-Butene C4H8 56.107 4.25402 10.78298 47.84869 -597.94406 2039.55839 50 1500
35 1-Pentene C5HIo 70.135 2.04789 37.52066 5.51442 -254.11404 993.58809 300 1200
36 1-Hexene C6H12 84.162 0.00610 63.24725 -35.49665 88.16710 -57.32517 200 1500
37 1-Heptene C7H14 98.189 3.47887 48.09877 23.25712 -531.31261 1997.9072 200 1500
38 1-Octene C8H16 112.216 3.98703 54.62745 29.77494 -652.07494 2453.5780 200 1500
39 1-Nonene C9H18 126.243 4.54519 60.98631 36.26160 -769.45635 2891.5695 200 1500
40 1-Decene CloH20 140.270 4.95682 68.28457 40.89381 -873.20972 3296.6252 200 1500
41 1-Undecene CllH22 154.219 5.68918 73.39300 49.96821 -1012.1464 3796.2346 200 1500
42 1-Dodecene C12H24 168.310 5.94633 81.98315 51.37713 -1083.4987 4090.6826 200 1500
43 1-Tfidecene C13H26 182.337 -0.32099 129.0630 -25.20577 -505.45836 2487.2522 200 1500
44 1-Tetradecene C14H28 196.364 -0.29904 138.5541 -25.09881 -568.90191 2764.1344 200 1500
45 1-Pentadecene C15H3o 210.391 0.09974 145.9734 -20.68363 -670.89438 3163.1228 200 1500
46 1-Hexadecene C16H32 224.418 0.54495 152.863 -15.10253 -781.81088 3585.1609 200 1500
47 1-Heptadecene C17H34 238.445 0.41533 163.244 -16.44658 -838.48469 3862.2068 200 1500
48 1-Octadecene C18H36 252.472 31.69585 -73.95796 647.4299 -8000.0231 29069.9319 200 1500
49 1-Nonadecene CI9H38 266.490 0.77613 180.596 -13.04267 -993.43903 4506.8347 200 1500
50 1-Eicosene C20H40 280.517 -0.20146 196.7237 -27.69713 -923.83278 4366.8186 200 1500
Naphthenes
51 Cyclopentane C5H10 70.134 -7.43795 69.82174 -43.64337 122.59611 -92.59304 300 1500
52 Methylcyclopentane C6H12 84.161 -6.81073 80.58175 -50.42977 141.93915 -107.77270 300 1500
53 Ethylcyclopentane C7H14 98.188 -7.51027 93.72668 -58.81706 167.05996 -127.54803 300 1500
54 n-Propylcyclopentane C8H16 112.216 -7.61363 105.1051 -65.42900 183.75014 -138.61493 300 1500
55 n-Butylcyclopentane C9Hla 126.243 -7.58208 115.9123 -71.30847 197.13396 -146.30090 300 1500
56 n-Pentylcyclopentane C10H20 140.270 -8.03062 128.8771 -80.20325 225.20999 -169.73008 300 1500
57 n-Hexylcyclopentane CllH22 154.290 -5.33508 122.4990 -48.15565 -101.53331 871.78359 300 1500
58 n-Heptylcyclopentane C12H24 168.310 -8.17951 151.3485 -93.06028 256.93490 -190.08988 300 1500
59 n-Octylcyclopentane C13H26 182.340 -8.20466 162.3974 -99.23258 271.5150 -199.08614 300 1500
60 n-Nonylcyclopentane C14H28 196.360 -8.27104 173.6250 -105.7060 287.84321 -209.81363 300 1500
61 n-Decylcyclopentane C15H30 210.390 -8.70424 186.5104 -114.4444 314.93805 -232.04645 300 1500
62 n-Unoecylcyclopentane Ct6H32 224.420 -8.71319 197.4859 -120.5348 329.23444 -240.60241 300 1500
63 n-Dodecylcyclopentane C17H34 238.440 -8.81568 208.8487 -127.1364 345.90292 -251.65896 300 1500
(Con~nued)
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 247
TABLE 6,8--(Continued)
No. Compound name Formula M A B xl03 C • 106 D x l 0 l~ E • Tmln, K Tm~, K
64 n-Tridecylcyclopentane C18H36 252.470 -8.82057 219.8119 -133.2056 360.15605 -260.35782 300 1500
65 n-Tetradecylcyclopentane C19H38 266.490 -8.81992 227.8540 -139.2730 374.44617 -268.88152 300 1500
66 n-Pentadecylcyclopentane C20H40 280.520 -9.29147 243.8142 -148.2795 402.88398 -292.64837 300 1500
67 n-Hexadecylcyclopentane C21H42 294.550 -9.34807 254.9133 -154.7201 419.05364 -303.20007 300 1500
68 Cyclohexane C6H12 84.161 -7.66115 77.46123 -31.65303 -45.48807 456.29714 300 1500
69 Methylcyclohexane C7H14 98.188 -8.75751 100.2054 -62.47659 169.33320 -123.27361 300 1500
70 Ethylcyclohexane C8H16 112.215 -5.50074 91.59292 -26.04906 -192.84542 1021.80248 300 1500
71 n-Propylcyclohexane C9H18 126.243 -8.87526 124.6789 -76.99183 180.70008 20.22888 300 1500
72 n-Butylcyclohexane C10H20 140.270 -7.38694 127.4674 -67.63120 73.28814 355.51905 300 1500
73 n-Pentylcyclohexane CllH22 154.290 -10.16016 152.5757 -98.38009 265.14011 -106.39559 300 1500
74 n-Hexylcyclohexane C12H24 168.310 -9.58825 161.8750 -104.4133 302.29623 -236.75537 300 1500
75 n-Heptylcyclohexane C13H26 182.340 -12.53870 188.4588 -138.5801 523.83412 -881.68097 300 1500
76 n-Octylcyclohexane C14H28 196.360 -7.88711 178.2886 -112.8765 330.77533 -271.09270 300 1500
77 n-Nonylcycloi-iexane C15H30 210.390 -8.48961 187.0067 -105.2157 192.47573 192.71532 300 1500
78 n-Decylcyclohexane C16Ha2 224.420 -10.58196 209.8953 -134.2136 385.00443 -297.79779 300 1500
79 n-Undecylcyclohexane C17H34 238.440 -9.25980 214.8824 -131.9175 357.79740 -261.20128 300 1500
80 n-Dodecylcyclohexane C18H36 252.470 -9.94518 228.7293 -141.7915 389.80571 -289.05527 300 1500
81 n-Tridecylcyclohexane C19H38 266.490 -10.06895 240.3258 -148.9432 410.23509 -304.86051 300 1500
82 n-Tetradecylcyclohexane C20H40 280.520 -10.98687 255.5423 -161.2184 455.82197 -347.77976 300 1500
83 n-Hexadecylcyclohexane C22H44 308.570 -8.96825 268.1151 -159.9818 417.54247 -292.26560 300 1500
Aromatics
84 Benzene C6H6 78.114 -7.29786 75.33056 -69.66390 336.46848 -660.39655 300 1500
85 Toluene C7H8 92.141 -2.46286 57.69575 -19.66557 -106.61110 654.52596 200 1500
86 Ethylbenzene C8H10 106.167 4.72510 9,02760 141.1887 -1989.2347 8167.1805 50 1000
87 m-Xylene C8H10 106.167 -4.00149 76,37388 -44.21568 82.57499 90.13866 260 1500
88 o-Xylene C8H10 106.167 -1.51679 68.03181 -33.61164 24.37900 206.82729 260 1500
89 p-Xylene C8Hlo 106.167 -4.77265 80.94644 -51.89215 136.1966 -45.64845 260 1500
90 n-Propylbenzene C9H12 120.195 4.42447 33.21919 74.42459 -1045.5561 3656.7834 50 1500
91 n-Butylbenzene CloH14 134.222 -6.24190 110.6923 -74.17854 221.3160 -178.64701 300 1500
92 m-Cymene CloH14 134.222 -4.41825 103.1174 -65.46564 182.7512 -138.15307 300 1500
93 o-Cymene CloH14 134.222 -2.40242 96.87475 -58.63517 154.5568 -109.45170 300 1500
94 p-Cymene CloH14 134.222 -4.47668 102.5377 -64.61930 179.2371 -134.70678 300 1500
95 n-Pentylbeuzene CIIH16 148.240 -6.89760 124.5723 -84.11348 251.1513 -201.56517 300 1500
96 n-Hexylbenzene C12H18 162.260 -7.66975 139.1540 -95.04913 285.7856 -230.69678 300 1500
97 n-Heptylbenzene C13H2o 176.290 -8.36450 153.2807 -105.2641 316.7510 -255.85057 300 1500
98 n-Octylbenzene C14H22 190.320 -9.35221 168.8057 -117.4996 357.7099 -292.04881 300 1500
99 Styrene C8Hs 104.152 -6.20755 91.11255 -83.45606 411.3630 -842.07179 300 1500
Dienes and acetylenes
100 Propadiene C3H4 40.065 1.30128 23.37745 --13.57151 26.91489 26.81000 200 1500
101 1,2-Butadiene C4H6 54.092 3.43878 19.01555 11.36858 --212.98223 751.33700 50 1500
102 Acetylene C2H2 26.038 1.04693 21.20409 --29.08273 203.04028 -533.31364 50 1500
Diaromatics
103 Naphthalene CloH8 128.174 --5.74112 86.70543 --46.55922 --1.47621 531.58512 200 1500
Nonhydrocarbons
104 Water H20 18.015 4.05852 --0.71473 2.68748 --11.97480 13.19231 50 1500
105 Carbon dioxide CO2 44.01 3.51821 --2.68807 31.88523 --499.2285 2410.9439 50 1000
106 Hydrogen sulfide H2S 34.08 4.07259 --1.43459 6.47044 --45.32724 103.38528 50 1500
107 Nitrogen N2 28.014 3.58244 -0.84375 2.09697 --10.19404 11.22372 50 1500
108 Oxygen 02 32 3.57079 -1.18951 4.79615 --40.80219 110.40157 50 1500
109 A m m o n i a NH3 17.03 0.98882 --0.68636 3.61604 --32.60481 96.53173 50 1500
110 Carbon monoxide CO 28.01 3.56423 --0.78152 2.20313 --11.29291 13.00233 50 1500
111 Hydrogen H2 2.016 3.24631 1.43467 --2.89398 25.8003 --73.9095 160 1220
112 Nitrogen dioxide NO2 46.01 3.38418 3.13875 3.98534 --58.69776 197.35202 200 1500
113 Nitrous oxide NO 30.01 4.18495 --4.19791 9.45630 --72.74068 192.33738 50 1500
Tminand Tm~xare approximated to nearest 10. Data have been determined from Method 7A1.2 given in the API-TDB [5].
100 where H/and V/are molar enthalpy and volume of pure com-
ponents at T and P of the mixture. For ideal solutions both
0
H1 the heat of mixing and the volume change due to the mixing
are zero [19]. This means that in an ideal solution, partial
molar volume of component i in the mixture is the same as
-~ -100
pure component specific volume (~k~ri= V/) and 17i nor/)i vary
with composition. Figure 6.6 shows variation of molar vol-
-200
ume of binary mixture with mole fraction for both a real and
an ideal solution (dotted line) for two cases. In Fig. 6.6a the
-300 real solution shows positive deviation, while in Fig. 6.6b the
H2
solution shows negative deviation from ideal solution. Sys-
-400 I I I I I I I I I tems with positive deviation from ideality have an increase in
0 0.2 0.4 0.6 0.8 1 volume due to mixing, while systems with negative deviation
have decrease in volume upon mixing.
Weight Fraction H2SO4
Equations (6.85) and (6.86) are useful when pure compo-
FIG. 6.5--Specific enthalpy of sulfuric acid solution nents are mixed to form a solution. If two solutions are mixed
at 25~ (part of Example 6,3). then the volume change due to mixing can be calculated from
250 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
II
(6.91) G ia = ~ _ x i G i + R T ~ _ x i lnxi
i i
XI
(6.92) sid --- ~-~ xiSi - R E xi l n x i
Mole Fraction, xl i i
(b) Systems with decrease in volume due to mixing where H ia, V id, Gig, and ~g can be either molar or specific
enthalpy, volume, Gibbs energy, and entropy of mixture. In
FIG. 6.6--Variation of molar volume of a binary mixture with
case of specific property, x4 is weight fraction. For example, if
composition.
V is specific volume (= 1/p), Eq. (6.90) can be written in the
following form for density:
the following relation [ 17]:
/~Vtmixing = ~ F//,afterV/(r, P,/q-/,after)-- E Ft/'bef~ (T, P, t't/,before) (6.93) P
i i
(6.87) where x~/is the weight fraction of i and pi is the density of
where r~,beforeis the moles of i before mixing and r~,aner rep- pure i. This equation was previously introduced in Chapter 3
resents moles of i in the solution after the mixing. Obviously (Eq. 3.45). Although all hydrocarbon mixtures do not really
since the mixture composition before and after the mixing are behave like ideal solutions, mixtures that do not contain non-
not the same, 17i for i in the solution before the mixing and its hydrocarbons or very heavy hydrocarbons, may be assumed
value for i in the solution after the mixing are not the same. as ideal solutions. For simplicity, application of Eqs. (6.89)
The same equation may be applied to enthalpy by replacing and (6.90) is extended to m a n y thermodynamic properties as
H with V to calculate heat of mixing when two solutions are it was shown in Chapters 3 and 4. Mixture heat capacity, for
mixed at constant T and P. Partial molar volume and enthalpy example, is calculated similar to enthalpy as:
may be calculated from their definition, Eq. (6.78) through an
(6.94) Cv = E xiCpi
EOS. For example in deriving the relation for l?i, derivative i
[O(nV)/~Yli]r,p,nj~ i should be determined from the EOS. For
the PR EOS the partial molar volume is given as [20] where xi is either mole or mass fraction depending on the unit
of Cp. If Cp is the specific heat (i.e., J/g. ~ weight fraction
(6.88) 17,i _ X~ + X2 should be used for x~. Obviously the main application of these
X3+X4 equations is when values of properties of pure components
where are available. For cases that these properties are predicted
from equations of state or other correlations, the mixing rules
Xt = (RT + biV) x (V 2 + 2bV - b 2) are usually applied to critical properties and the input param-
X2 = [2bi R T - 2 ~ i xiaii - 2bi P (V - b)] x (V - b) + bia eters of an EOS rather than to calculated values of a thermo-
Xa = n (V2 + 2bV - b 2) + 2P (V - b) (V + b) dynamic property in order to reduce the time and complexity
of calculations. For hydrocarbon mixtures that contain very
X 4 = - 2 R T (V + b) + a
light and very heavy hydrocarbons the assumption of ideal so-
where V is the mixture molar volume calculated from PR lution and application of Eqs. (6.89)-(6.93) will not give accu-
EOS. For more accurate calculation of ffi, corrected V rate results. For such mixtures some correction terms to con-
through use of volume translation concept (Eq. 5.50) may sider the effects of nonideality of the system and the change in
be used. Similar relation for/}i can be obtained (see Prob- molecular behavior in presence of unlike molecules should be
lem 6.5). added to the RHS of such equations. The following empirical
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 251
Vapor (V)
Example 6.5--Calculate volume of a blend and its API gravity at T, psat
produced by addition of 10000 bbl of light n a p h t h a with API
gravity of 90 to 90000 bbl of a crude oil with API gravity of 30. fV(T psat)= fL(T psat)
freezing occurs. W h e n a system is in e q u i l i b r i u m its energy is Constant B is in fact s a m e as AI-FaO/R. Because of three m a -
in m i n i m u m level (dG = 0), w h i c h for a system with only va- j o r simplifying a s s u m p t i o n s m a d e above, Eq. (6.101) is very
p o r a n d liquid is d(G v - G L) = 0, w h i c h c a n be w r i t t e n as [1]: a p p r o x i m a t e a n d it m a y be u s e d over a n a r r o w t e m p e r a t u r e
range w h e n m i n i m u m d a t a are available. Constants A a n d
(6.96) dGV(T, psat) = dGL(T, psat)
B can be d e t e r m i n e d from m i n i m u m two d a t a p o i n t s on the
w h e r e psat indicates t h a t the relation is valid at the s a t u r a t i o n v a p o r p r e s s u r e curve. Usually the critical p o i n t (To Pc) a n d
t e m p e r a t u r e a n d pressure. S i m i l a r e q u a t i o n applies to solid- n o r m a l boiling p o i n t (1.01325 b a r a n d Tb) are used to o b t a i n
liquid o r s o l i d - v a p o r phases. the constants. If A/-/yap is known, then only one d a t a p o i n t (Tb)
During a p h a s e c h a n g e (i.e., v a p o r to liquid o r vice versa), w o u l d be sufficient to o b t a i n the v a p o r p r e s s u r e correlation.
t e m p e r a t u r e a n d p r e s s u r e of the system r e m a i n c o n s t a n t a n d A m o r e a c c u r a t e v a p o r p r e s s u r e c o r r e l a t i o n is the following
therefore f r o m Eq. (6.5) we have: t h r e e - c o n s t a n t c o r r e l a t i o n k n o w n as Antoine equation:
A/-yap B
(6.97) A S yap - - (6.102) In psat = A - - -
T T+C
w h e r e A H v~p is heat of vaporization a n d AS yap is the entropy
A, B, a n d C, k n o w n as Antoine constants, have b e e n deter-
o f vaporization. A/-/yap is defined as:
m i n e d for a large n u m b e r of c o m p o u n d s . Antoine p r o p o s e d
A H yap ----H(T, psat, s a t u r a t e d vapor) this simple modification of the C l a s i u s - C l a p e y r o n e q u a t i o n
(6.98) - H(T, psat, s a t u r a t e d liquid) in 1888. Various modifications of this e q u a t i o n a n d o t h e r cor-
relations for e s t i m a t i o n of v a p o r p r e s s u r e are discussed in the
S i m i l a r l y A S wp a n d AV v~p are defined. F o r a p h a s e change next chapter.
f r o m solid to liquid i n s t e a d of h e a t of v a p o r i z a t i o n A/-Fap,
h e a t of fusion o r melting A H fu~ is defined by the difference be- Example 6 . 6 - - F o r p u r e water, estimate v a p o r p r e s s u r e of wa-
t w e e n e n t h a l p y of s a t u r a t e d liquid a n d solid. Since ps~t is only
ter at 151.84~ W h a t is its h e a t of v a p o r i z a t i o n ? The actual
a function of t e m p e r a t u r e , A S yap a n d Avvap are also functions
values as given in the s t e a m tables are 5 b a r a n d 2101.6 kJ/kg,
of t e m p e r a t u r e only for a n y p u r e substance. A H yap a n d A W ap
respectively [1]. A s s u m e t h a t only Tb, To, a n d Pc a r e known.
decrease w i t h increase in t e m p e r a t u r e a n d at the critical p o i n t
they a p p r o a c h zero as v a p o r a n d liquid p h a s e s b e c o m e iden-
tical. While AV yap can be c a l c u l a t e d f r o m an e q u a t i o n of state Solution--From Table 2.1 for w a t e r we have Tc = 647.3 K,
as was discussed in C h a p t e r 5, m e t h o d s of calculation A/-/vap Pc = 220.55 bar, a n d Tb = I00~ Applying Eq. (6.101) at the
will be discussed in C h a p t e r 7. By applying Eq. (6.8) to b o t h critical p o i n t a n d n o r m a l boiling p o i n t gives lnPc = A - B~ Tc
d G v a n d d G L a n d use of Eqs. (6.97) a n d (6.98) the following a n d In (1.01325) ~- A - B/Tb. S i m u l t a n e o u s s o l u t i o n of these
r e l a t i o n k n o w n as Clapeyron equation can be derived: e q u a t i o n s gives the following relations to calculate A a n d B
from Tb, To, a n d Pc.
dpsat A/-/yaP
(6.99)
dT T A V yap l n ( J1.01325
~-)
B= 1 1
This equation is the basis of d e v e l o p m e n t of predictive meth- (6.103) rb r~
ods for v a p o r p r e s s u r e versus t e m p e r a t u r e . N o w t h r e e sim- B
A = 0.013163 +
plifying a s s u m p t i o n s are m a d e : (I) over a n a r r o w r a n g e of
t e m p e r a t u r e , A/-/vap is constant, (2) v o l u m e of liquid is small
in c o m p a r i s o n with v a p o r v o l u m e ( A V yap -~-- V v -- V L ~'~ V v ) , w h e r e Tc a n d Tb m u s t be in kelvin a n d Pc m u s t be in bar.
a n d (3) v o l u m e of v a p o r can be c a l c u l a t e d from ideal gas law The s a m e units m u s t be used in Eq. (6.101). In cases t h a t
(Eq. 5.14). These a s s u m p t i o n s are n o t true in general b u t at a value of v a p o r p r e s s u r e at one t e m p e r a t u r e is k n o w n it
a n a r r o w r a n g e of t e m p e r a t u r e a n d low p r e s s u r e c o n d i t i o n s s h o u l d be used instead of Tc a n d Pc so the resulting equa-
they can be u s e d for simplicity. Upon a p p l i c a t i o n of a s s u m p - tion will be m o r e a c c u r a t e between that p o i n t a n d the boiling
tions 2 a n d 3, Eq. (6.99) can be written in the following form point. As the difference between t e m p e r a t u r e s of two refer-
k n o w n as Clausius-Clapeyron equation: ence p o i n t s u s e d to o b t a i n constants in Eq. (6.101) reduces,
the a c c u r a c y of resulting e q u a t i o n for the v a p o r p r e s s u r e be-
d In psat aHvap
(6.100) tween two reference t e m p e r a t u r e s increases. F o r w a t e r from
d(1/T) R Eq. (6.103), A = 12.7276 b a r a n d B = 4745.66 b a r . K. Substi-
w h e r e R is the universal gas constant. This e q u a t i o n is the tuting A a n d B in Eq. (6.101) at T = 151.84 + 273.15 = 425 K
basis of d e v e l o p m e n t of s i m p l e correlations for e s t i m a t i o n gives In P = 1.5611 o r P = 4.764 bar. C o m p a r i n g p r e d i c t e d
of v a p o r p r e s s u r e versus t e m p e r a t u r e o r calculation of h e a t value with the actual value of 5 b a r gives a n e r r o r of - 4 . 9 % ,
of v a p o r i z a t i o n f r o m v a p o r p r e s s u r e data. F o r example, by w h i c h is acceptable considering simple r e l a t i o n a n d min-
using the first a s s u m p t i o n (constant A H v~p) a n d integrating i m u m d a t a used. H e a t of v a p o r i z a t i o n is calculated as
the above e q u a t i o n we get follows: A H yap = R B = 8.314 x 4745.66 = 39455.4 J/tool =
39455.4/18 = 2192 kJ/kg. This value gives a n e r r o r of +4.3%.
B
(6.101) In psat ~ A - -- Obviously m o r e a c c u r a t e m e t h o d of e s t i m a t i o n of h e a t of va-
T p o r i z a t i o n is t h r o u g h A/-Fap = H sat,yap - n sat,liq,w h e r e H sat,yap
w h e r e T is absolute t e m p e r a t u r e a n d A a n d B are two positive a n d H sat'liq can be calculated t h r o u g h g e n e r a l i z e d correla-
constants specific for each p u r e substance. This equation sug- tions. E m p i r i c a l m e t h o d s of calculation of h e a t of vaporiza-
gests that In pv~p versus 1/T is a straight line w i t h slope of - B . tion are given in C h a p t e r 7.
6. T H E R M O D Y N A M I C RELATIONS FOR PROPERTY ESTIMATIONS 253
where TM is the melting point temperature at pressure P. If equilibrium similar criteria must apply to every phase. In this
atmospheric pressure is shown by po (1.01325 bar) and the case every phase has different composition but all have the
melting point at P~ is shown by TMo (normal melting point), same T and P. We know for mechanical equilibrium, total
integration of the above equation from po to pressure P gives energy (i.e., kinetic and potential) of the system must be min-
[ A V M x ( P - Po)] imum. The best example is oscillation of hanging object that
(6.107) TM = TMoexp [_ AH M it comes to rest when its potential and kinetic energies are
minimum at the lowest level. For thermodynamic equilib-
where in deriving this equation it is assumed that both AV M rium the criterion is minimum Gibbs free energy. As shown
and AH M are constants with respect to temperature (melting by Eq. (6.73) a mixture molar property such as G varies with
point). This is a reasonable assumption since variation of TM T and P and composition. A mathematical function is mini-
with pressure is small (see Fig. 5.2a). Since this equation is mum when its total derivative is zero:
derived for pure substances, TM is the same as freezing point
(6.108) dG(T, P,x~) = 0
(Tf) and AH M is the same as heat of fusion (AHf).
(a) To calculate melting point of n-C18 at 300 bar, we have Schematic and criteria for VLE of multicomponent systems
P = 300 bar, Po = 1.01325 bar, TMo = 301.4 K, and AV M = are shown in Fig. 6.7b. Phase equilibria calculations lead to
1/p v - Up s = 0.I313 cm3/g. AH M = 242.4597 J/g, I/J = 10 determination of the conditions of T, P, and composition at
bar. cm 3, thus from Eq. (6.107) we have TM = 301.4 x exp which the above criteria are satisfied. In this section general
[0.1313 x (300 -- 1.013)/10 x 242459.7] = 301.4 x 1.0163 = formulas for phase equilibria calculations of mixtures are pre-
301.4 x 1.0163 = 306.3 K or TM = 33.2~ This indicates sented. These are required to define new parameters such as
that when pressure increases to 300 bar, the melting point activity, activity coefficient, and fugacity coefficient of a com-
of n-C18 increases only by 5~ In this temperature range as- ponent in a mixture. Two main references for thermodynam-
sumption of constant AV M and AH M is quite reasonable. ics of mixtures in relation with equilibrium are Denbigh [19]
(b) To calculate the triple point temperature, Eq. (6.107) and Prausnitz et al. [21 ].
must be applied at P = P t p = 3 . 3 9 x 1 0 -5 kPa = 3 . 3 9 x
10 -7 bar. This is a very low number in comparison
6.6.1 Definition of Fugacity, Fugacity
with Po = 1 bar, thus TM = 301.4 x exp(--0.1313 x 1.013/10 x
Coefficient, Activity, Activity Coefficient,
242459.7) = 301.4 x 0.99995 ~ 301.4 K. Thus, we get triple
and Chemical Potential
point temperature same as melting point. This is true for most
of pure substances as Ptp is very small. It should be noted that In this section important properties of fugacity, activity, and
Eq. (6.107) is not reliable to calcnlate pressure at which melt- chemical potential needed for formulation of solution ther-
ing point is known because a small change in temperature modynamics are defined and methods of their calculation are
causes significant change in pressure. This example explains presented. Consider a mixture of N components at T and P
why melting point of water decreases while for n-octadecane and composition yi. Fugacity of component i in the mixture
it increases with increase in pressure. As it is shown in Sec- is shown bye- and defined as
tion 7.2 density of ice is less than water, thus AVi for water
is negative and from Eq. (6.107), TM is less than TMo at high (6.109) lim ( f-~p ~ ~I
pressures. # \Yi }P-~O
where sign ^ indicates the fact that component i is in a mix-
ture. When yi --~ 1 we have f --~ f , where f is fugacity of
6.6 PHASE EQUILIBRIA OF pure i as defined in Eq. (6.45). The fugacity coefficient of i in
MIXTURES--CALCULATION a mixture is defined as
OF BASIC PROPERTIES
(6.110) q~i ~ f ~ /
Perhaps one of the biggest applications of equations of state YiP
and thermodynamics of mixtures in the petroleum science
is formulation of phase equilibrium problems. In petroleum where for an ideal gas, ~i = 1 or f = Yi P. In a gas mixture
production phase equilibria calculations lead to the determi- yi P is the same as partial pressure of component i. Activity of
nation of the composition and amount of oil and gas produced component i, di, is defined as
at the surface facilities in the production sites, PT diagrams to f
determine type of hydrocarbon phases in the reservoirs, solu- (6.111) ai = ~/o
bility of oil in water and water in oils, compositions of oil and
gas where they are in equilibrium, solubility of solids in oils, where ff is fugacity of i at a standard state. One common
and solid deposition (wax and asphaltene) or hydrate forma- standard state for fugacity is pure component i at the same
tion due to change in composition or T and P. In petroleum T and P of mixture, that is to assume f/~ = f , where f. is the
processing phase equilibria calculations lead to the determi- fugacity of pure i at T and P of mixture. This is usually known
nation of vapor pressure and equilibrium curves needed for as standard state base on Lewis rule. Choice of standard state
design and operation of distillation, absorption, and stripping for fugacity and chemical potential is best discussed by Den-
columns. bigh [I 9]. Activity is a parameter that indicates the degree of
A system is at equilibrium when there is no tendency to nonideality in the system. The activity coefficient of compo-
change. In fact for a multicomponent system of single phase nent i in a mixture is shown by yi and is defined as
to be in equilibrium, there must be no change in T, P, and ai
xl, x2. . . . . x~c-t. When several phases exist together while at (6.112) Fi = --
Yq
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 255
where x/is mole fraction of c o m p o n e n t i in the mixture. Both where G 7 is the molar Gibbs energy of pure i at T of the
di and yi are dimensionless parameters. With the above defini- system and pressure of 1 arm (ideal gas state). By replac-
tions one m a y calculate~ from one of the following relations: ing G = Y'~.Yi/2i, and V = ~yiVi in the above equation and
removing the s u m m a t i o n sign we get
(6.113) ~ = ~/yiP P
(6.117) d~/ = RTd In (Yi P) where V t is the total volume (V t = nV). In using these equa-
where in Eq. (6.117) if Yi = 1, it reduces to Eq, (6.116) for pure tions one should note that n is the sum of r~ and is not constant
c o m p o n e n t systems. For real gases these equations b e c o m e when derivative with respect to r~ is carried. These equations
are the basis of calculation of fugacity of a c o m p o n e n t in a
(6.118) dGi = d/zi = RTd In f~- mixture. Examples of such derivations are available in vari-
ous texts [I, 4, 11, 20-22]. One can use an EOS to obtain ITi
(6.119) dGi = d/2/= RTd In
and u p o n substitution in Eq. (6.124) a relation for calculation
Equation (6.118) is the same as Eq. (6.46) derived for pure of ~i can be obtained. For the general form of cubic equations
components. Equation (6.119) reduces to Eq. (6.118) at given by Eqs. (5.40)-(5.42), q~i is given as [11]
Yi -- I. Subtracting Eq. (6.118) f r o m Eq. (6.119) and using
Eq. (6.114) for]~ one can derive the following relation for/2i ln /= (z- 1)-In(Z-B)+
in a solution:
(6.126) x In 2 z + B ( . ~ + ~
(6.120) /2/-/z~ = RTln )'iN
where/z[ is the pure c o m p o n e n t chemical potential at T and b~ "&/P~
P of mixture and x4 is the mole fraction of i in liquid solu- where -
b Ejyjrcj/P~j
tion. For ideal solutions where Yi = 1, Eq. (6.120) reduces to
/2i - tz~ = RTln xi. In fact this is another way to define an ideal and - 2ay
solution. A solution that is ideal over the entire range of com- a i
position is called perfect solution and follows this relation.
ifallk/i = 0 then ~i = 2 ( - ~ ) 1/2
6.6.2 Calculation o f Fugacity Coefficients M1 parameters in the above equation for vdW, RK, SRK, and
from Equations o f State PR equations of state are defined in Tables 5.1 and 6. I. Pa-
Through t h e r m o d y n a m i c relations and definition of G one rameters a and b for the mixture should he calculated from
can derive the following relation for the mixture molar Gibbs Eqs. (5.59)-(5.61). Equation (6.126) can be used for calcula-
free energy [21]. tion of fugacity of i in both liquid and vapor phases provided
appropriate Z values are used as for the case of pure compo-
P
nent systems that was shown in Example 6.7. For calculation
(6.121) c= f (v-)de+RTyiln(r/P)+
y/C, of ~ / f r o m PR and SRK equations through the above relation,
0 use of volume translation is not required.
256 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
If truncated virial equation (Eq. 5.75) is used, ~i is calcu- - - G o o d approximation for gases at low pressure where the
lated from the following relation as derived from Eq. (6.124): gas phase is nearly ideal.
- - G o o d approximation at any pressure whenever i is present
in large excess (say, yi > 0.9). The Lewis rule becomes exact
(6.127) ln~)i = (2 E y i B i i - B) P-~
j RT in the limit of Yi ~ 1.
- - G o o d approximation over all range of pressure and com-
where B (for whole mixture) and Bii (interaction coefficients) position whenever physical properties of all components
should be calculated from Eqs. (5.70) and (5.74), respectively. present in the mixture are the same as (i.e., benzene and
As discussed earlier Eq. (5.70) is useful for gases at moderate toluene mixture).
pressures. Equation (6.127) is not valid for liquids. - - G o o d approximation for liquid mixtures whose behavior is
like an ideal solution.
Example 6.9--Suppose that fugacity coefficient of the whole - - A poor approximation at moderate and high pressures
mixture, CmJx, is defined similar to that of pure components. whenever the molecular properties of components in the
Through mixture Gibbs energy, derive a relation between f ~ mixture are significantly different from each other (i.e., a
and/~ for mixtures. mixture of methane and a heavy hydrocarbon).
Solution--Applying Eq. (6.80) to residual molar Gibbs free Lewis rule is attractive because of its simplicity and is usu-
energy ( G R = G - Gig) gives G R = ~-~yiGR and since t~i = Gi ally used when the limiting conditions are applied in certain
from Eq. (6.119) dGi =RTdln/~ and for ideal gases from situations. Therefore when the Lewis rule is used, fugacity of
Eq. (6.117) we have dGi g = RTdlnyiP. Subtracting these two i in a mixture is calculated directly from its fugacity as pure
relations from each other gives dG/R = RTd In ~i, which after component. When Lewis rule is applied to liquid solutions,
integration gives ~R = RTln ~i. Therefore for the whole mix- Eq. (6.114) can be combined with Eq. (6.131) to get Yi = 1
ture we have (for all components).
(6.128) G~ = RT E xi ln~,
6.6.4 Calculation of Fugacity of Pure Gases
where after comparing with Eq. (6.48) for the whole mixture and Liquids
we get
Calculation of fugacity of pure components using equations
(6.129) lnCmi~ = E x / l n S i of state was discussed in Section 6.5. Generally fugacity of
pure gases and liquids at moderate and high pressures may
or in terms of fugacity for the whole mixture, fmi~, it can be be estimated from equations given in Table 6.1 or through
written as generalized correlations of LK as given by EQ. (6.59). For pure
^
p~at is the saturation pressure or vapor pressure of pure i at T According to the definition of yi when x4 = 1 (pure i) then
a n d methods of its calculation are discussed in the next chap- Yi = 1. Generally for binary systems when a relation for ac-
ter. ~bsat is the vapor phase fugacity coefficient of pure compo- tivity coefficient of one c o m p o n e n t is known the relation for
nent i at p/sat and can be calculated from methods discussed activity coefficient of other components can be determined
in Section 6.2. The exponential term in the above equation is from the following relation:
called Poynting correction and is calculated from liquid molar
volume. Since variation of V/Lwith pressure is small, usually it dlnF1 dlnF2
is assumed constant versus pressure and the Poynting factor (6.141) xx dx 1 - x 2 dx 2
is simplified a s e x p [ V / L ( P -- P/sat)/RT]. In such cases V/L m a y
be taken as molar volume of saturated liquid at temperature which is derived from Gibbs-Duhem equation. One can ob-
T and it m a y be calculated from Racket equation (Section tain y2 from yl by applying the above equation with use of
5.8). At very low pressures or w h e n (P - psat) is very small, x2 = 1 - xl and dx2 = - d x l . Constant A in Eq. (6.140) can be
the Poynting factor approaches unity and it could be removed obtained from data on the activity coefficient at infinite di-
from Eq. (6.136). In addition, when p/sat is very small (~ 1 atm lution (Fi~ which is defined as limxi-~0(Yi). This will result
or less), ~b~~t m a y be considered as unity and fir is simply equal in A = R T l n y ~ = R T l n y ~ . This simple model applies well
to p/sat. Obviously this simplification can be used only in spe- to simple mixtures such as benzene-cyclohexane; however,
cial situations when the above assumptions can be justified. for more complex mixtures other activity coefficient models
For calculation of Poynting factor when V/L is in cm3/mol, P must be used. A more general form of activity coefficients for
in bar, and T in kelvin, then the value of R is 83.14. binary systems that follow Redlich-Kister model for G E a r e
g i v e n as
GE
0.8
(6.139) ~ = XlX2[A -}- B ( X l - x 2 ) -4- C ( X l - x2) 2 q-...]
based on the Wohl's model for the excess Gibbs energy [21 ]. c o m p o u n d s two parameters, namely energy parameter and
The G ~ relation for the van Laar model is given by size parameter describe the intermolecular forces. Energy of
vaporization is directly related to the energy required to over-
GE come forces between molecules in the liquid phase and molar
(6.143) - - = XlX2[A + B(x1 - x2)] -1
RT volume is proportional to the molecular size. Therefore, when
two components have similar values of 8 their molecular size
Upon application of Eq. (6.137), the activity coefficients are
and forces are very similar. Molecules with similar size and
obtained as
interrnolecular forces easily can dissolve in each other. The
A,2x, -2 importance of solubility parameter is that when two compo-
lnya --- A12 1 + A21x2] nents have 8 values close to each other they can dissolve in
(6.144), each other appreciably. It is possible to use an EOS to cal-
lnyz -~ A21( l + A21x2/ culate 8 from Eq. (6.147) (see Problem 6.20). According to
the theory of regular solutions, excess entropy is zero and it
can be shown that for such solutions RT lnF/ is constant at
where coefficients AlE and A21 are related to A and B in Eq. constant composition and does not change with temperature
(6.143) as A - B = 1/A12 and A + B = 1/A21. Coefficients A12 [11]. Values of V~L and 8i at a reference temperature of 298
and A21 can be determined from the activity coefficients at K is sufficient to calculate ~'i at other temperatures through
infinite dilutions (Alz = In y~, Azl = In y~). Once for a given Eq. (145). Values of solubility parameter for single carbon
system VLE data are available, they can be used to calculate n u m b e r components are given in Table 4.6. Values of V/L and
activity coefficients through Eqs. (6.179) or (6.181) and then ~i a t 25~ for a n u m b e r of pure substances are given in Ta-
GE/RT is calculated from Eq. (6.138). F r o m the knowledge of ble 6.10 as provided by DIPPR [13]. In this table values of 8
G~/RT versus (xl - x2) the best model for G ~ can be found. have the unit of (J/cm3) 1/2. In Table 6.10 values of freezing
Once the relation for G ~ has been determined the activity point and heat of fusion at the freezing point are also given.
coefficient model will be found. These values are needed in calculation of fugacity of solids as
For regular solutions where different components have the will be seen in the next section.
same intermolecular forces it is generally assumed that V ~ = Based on the data given in Table 6.10 the following rela-
S E = 0. Obviously systems containing polar c o m p o u n d s gen- tions are developed for estimation of liquid m o l a r volume of
erally do not fall into the category of regular solutions. Hy- n-alkanes (P), n-alkylcyclohexanes (N), and n-alkylhenzenes
drocarbon mixtures m a y be considered as regular solutions. (A) at 25~ V25 [23]: It gives Cp/Cv for saturated liquids hav-
The activity coefficient of c o m p o n e n t i in a binary liquid solu- ing a calculated heat capacity ratio of 1.43 to 1.38 over a tem-
tion according to the regular solution theory can be calculated perature range of 300-450 K.
from the Scatchard-Hfldebrand relation [21, 22]:
In V25 = -0.51589 + 2.75092M ~
(8, - h)
In Yl = RT for n-alkanes (C1 - C36)
(6.145) V2s = 10.969+ 1.1784M
(6.148)
In Yz = VL (81 - 82)2 ev~ for n-alkylcyclohexanes (C6 - C16)
RT
In V25 = -96.3437 + 96.54607M ~176
where V L is the liquid m o l a r volume of pure components (1 for n-alkylbenzenes (C6 - C24)
or 2) at T and P and 8 are the solubility parameter of pure
components 1 or 2. qb~ is the volume fraction of c o m p o n e n t where V25 is in cm3/mol. These correlations can reproduce
1 and for a binary system it is given by data in Table 6.10 with average deviations of 0.9, 0.4, and
0.2% for n-alkanes, n-alkylcyclohexanes, and n-alkylbenzens,
x~ V~ respectively. Similarly the following relations are developed
(6.146) Ol =
for estimation of solubility parameter at 25~ [23]:
where xl and xz are mole fractions of components 1 and 2. 6 = 16.22609 [I + exp (0.65263 - 0.02318M)] -~176176
The solubility parameter for c o m p o n e n t i can be calculated
from the following relation [21, 22]: for n-alkanes (C1 - C36)
8 = 16.7538 + 7.2535 • 10-5M
(6.149) for n-alkylcyclohexanes (C6 - C16)
(6.147) 8i = \ - - - ~ / L ] = \ v/L
8 = 2 6 . 8 5 5 7 - 0 . 1 8 6 6 7 M + 1.36926 x 10-3M 2
where AUyap and AH/yap a r e the molar internal energy and - 4 . 3 4 6 4 • 10-6M3 + 4.89667 • 10-9M 4
heat of vaporization of c o m p o n e n t i, respectively. The tradi- for n-alkylbenzenes (C6 - Cz4)
tional unit for ~ is (cal/cm3)l/z; however, in this chapter the
unit of (J/cm3) V2 is used and its conversion to other units is where 8 is in (J/cm3) U2. The conversion factor from this
given in Section 1.7.22. Solubility parameter originally pro- unit to the traditional units is given in Section 1.7.22:1
posed by Hildebrand has exact physical meaning. Two param- (cal/cm3) 1/2 = 2.0455 (J/cm3) V2. Values predicted from these
eters that are used to define 8 are energy of vaporization and equations give average deviation of 0.2% for n-alkanes, 0.5%
m o l a r volume. In Chapter 5 it was discussed that for n o n p o l a r for n-alkylcyclohexanes, and 1.4% for n-alkylbenzenes. It
6. T H E R M O D Y N A M I C RELATIONS FOR PROPERTY ESTIMATIONS 259
TABLE 6.10--Freezing point, heat of fusion, molar volume, and solubility parameters for some selected compounds [DIPPR].
No. Compound Formula Nc M TM, K AHf/RTuat TM V25, cm3/mol 82s, (J/cm3)l/2
n-ParmTms
1 Methane CH4 1 16.04 90.69 1.2484 37.969(52) a 11.6
2 Ethane C2H6 2 30.07 90.35 3.8059 55.229(68) a 12.4
3 Propane C3H8 3 44.09 85.47 4.9589 75.700(84) a 13.1
4 n-Butane C4H10 4 58.12 134.86 4.1568 96.48(99.5) a 13.7
5 n-Pentane C5H12 5 72.15 143.42 7.0455 116.05 14.4
6 n-Hexane C6H14 6 86.17 177.83 8.8464 131.362 14.9
7 n-Heptane C7HI6 7 100.20 182.57 9.2557 147.024 15.2
8 n-Octane C8H18 8 114.22 216.38 11.5280 163.374 15.4
9 n-Nonane C9H20 9 128.25 219.66 8.4704 179.559 15.6
10 n-Decane C10H22 10 142.28 243.51 14.1801 195.827 15.7
11 n-Undecane CllH24 11 156.30 247.57 10.7752 212.243 15.9
12 n-Dodecane C12H26 12 170.33 263.57 16.8109 228.605 16.0
13 n-Tridecane C13H28 13 184.35 267.76 12.8015 244.631 16.0
14 n-Tetradecane C14H30 14 198.38 279.01 19.4282 261.271 16.1
15 n-Pentadecane C15H32 15 212.41 283.07 14.6966 277.783 16.1
16 n-Hexadecane C16H34 16 226.43 291.31 22.0298 294.213 16.2
17 n-Heptadecane C17H36 17 240.46 295.13 16.3674 310.939 16.2
18 n-Octadecane C18H38 18 254.48 301.31 24.6307 328.233 16.2
19 n-Nonadecane C19H40 19 268.51 305.04 18.0620 345.621 16.2
20 n-Eicosane C20H42 20 282.54 309.58 27.1445 363.69 16.2
21 n-Heneicosane C21H44 21 296.56 313.35 18.3077 381.214 16.2
22 n-Docosane C22H46 22 310.59 317.15 18.5643 399.078 16.2
23 n-Triacosane C23H48 23 324.61 320.65 20.2449 416.872 16.3
24 n-Tetracosane C24H50 24 338.64 323.75 20.3929 434.942 16.3
25 n-Hexacosane C26H54 26 366.69 329.25 22.1731 469.975 16.3
26 n-Heptacosane C27H56 27 380.72 332.15 21.8770 488.150 16.2
27 n-Octacosane C28H58 28 394.74 334.35 23.2532 506.321 16.2
28 n-Nonacosane C29H60 29 408.77 336.85 23.6034 523.824 16.2
29 n-Triacontane C30H62 30 422.80 338.65 24.4439 540.500 16.2
30 n-Docontane C32H66 32 450.85 342.35 26.8989 576.606 16.2
31 n-Hexacontane C36H74 36 506.95 349.05 30.6066 648.426 16.2
Isoparaffins
32 Isobutane C4H10 4 58.12 113.54 4.8092 105.238 12.57
33 Isopentane C5H12 5 72.15 113.25 5.4702 117.098 13.86
34 Isooctane (2,2,4-trimethylpentane) C8H18 8 114.23 165.78 6.6720 165.452 14.08
n-Alkylcyelopentanes (naphthenes)
35 Cyclopentane C5H10 5 70.14 179.31 0.4084 94.6075 16.55
36 Methy]cyclopentane C6H12 6 84.16 146.58 4.7482 128.1920 16.06
37 Ethylcyclopentane C7H14 7 98.19 134.71 6.1339 128.7490 16.25
38 n-Propylcyclopentane C8H16 8 112.22 155.81 7.7431 145.1930 16.36
39 n-Butylcyclopentane C9H18 9 126.24 165.18 8.2355 161.5720 16.39
n.Alkylcyclohexanes (naphthenes)
40 Cyclohexane C6H12 6 84.16 279.69 1.1782 108.860 16.76
41 Methylcyclohexane C7H14 7 98.18 146.58 5.5393 128.192 16.06
42 Ethylcyclohexane C8H16 8 112.21 161.839 6.1935 143.036 16.34
43 n-Propylcyclohexane C9H18 9 126.23 178.25 6.9970 159.758 16.35
44 n-Butylcyclohexane CIoH20 10 140.26 198.42 8.5830 176.266 16.40
45 n-Decylcyclohexane C16H32 16 224.42 271.42 17.1044 275.287 16.65
n-Alkylbenzenes (aromatics)
46 Benzene C6H 6 6 78.11 278.65 4.2585 89.480 18.70
47 Methylbenzene (Toluene) C7H8 7 92.14 178.15 4.4803 106.650 18.25
48 Ethylbenzene Call10 8 106.17 178.15 6.1983 122.937 17.98
49 Propylbenzene C9H12 9 120.20 173.55 6.4235 139.969 17.67
50 n-Butylbenzene C10H14 10 134.22 185.25 7.2849 156.609 17.51
51 n-Pentylbenzene CllH16 11 148.25 198.15 9.2510 173.453 17.47
52 n-Hexylbenzene C12H18 12 162.28 211.95 10.4421 189.894 17.43
53 n-Heptylbenzene C13H20 t3 176.30 225.15 11.6458 206.428 17.37
54 n-Octylbenzene C14H22 14 190.33 237.15 13.1869 223.183 17.37
55 n-Nonylbenzene C15H24 15 204.36 248.95 13.9487 239.795 17.39
56 n-Decylbenzene C16H26 16 218.38 258.77 15.1527 256.413 17.28
57 n-Undecylbenzene C17H28 17 232.41 268.00 16.1570 272.961 17.21
58 n-Dodecylbenzene C18H30 18 246.44 275.93 17.5238 289.173 17.03
59 n-Tridecylbenzene C19H32 19 260.47 283.15 18.6487 306.009 16.87
6O n-Tetradecylbenzene C20H34 20 274.49 289.15 19.8420 322.197 16.64
61 n-Pentadecylbenzene C21H36 21 288.52 295.15 20.9874 339.135 16.49
(Continued)
260 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 6.10---(Continued)
No. Compound Formula Nc M TM, K AHf/RTM at T~ V2s, cm3/mol 82s, (J/cm3)1/2
62 n-Hexadecylbenzene C22H38 22 302.55 300.15 22.1207 356.160 16.39
63 n-Heptadecylbenzene C23H40 23 316.55 305.15 22.9782 373.731 16.30
64 n-Octadecylbenzene C24H42 24 330.58 309.00 23.7040 390.634 16.24
1-rvAlkylnaphthalenes (aromatics)
65 Naphthalene C10H8 10 128.16 353.43 6.4588 123.000 19.49
66 1-Methylnaphthalene CllHm 11 142.19 242.67 3.4420 139.899 19.89
67 1-Ethylnaphthalene C12H12 12 156.22 259.34 7.5592 155.579 19.85
68 1-n-Propylnaphthalene C13H14 13 170.24 264.55 7.9943 172.533 19.09
69 1-n-Butylnaphthalene C14H16 14 184.27 253.43 11.9117 189.358 19.10
70 1-n-Pentylnaphthalene ClsHt8 15 198.29 248.79 11.3121 205.950 18.85
71 1-n-Hexylnaphthalene C16H20 16 212.32 255.15 ... 224.155 18.72
72 1-n-Nonylnaphthalene C19H26 19 254.40 284,15 ... 272.495 17.41
73 l-n-Decylnaphthalene C20H2s 20 268.42 288.15 ... 289.211 17.20
Other organic compounds
74 Benzoic acid C7H602 7 122.12 395.52 5.4952 112.442 24.59
75 Diphenylmethane C13H12 13 168.24 298.39 7.3363 167.908 19.52
76 Antheracene C14H10 14 190.32 488.93 7.7150 182.900 17.75
Nonhydrocarbons
77 Water H20 ... 18.02 273.15 2.6428 18.0691 47.81
78 Methanol CH3OH 1 32.04 -97.68 0.2204 40.58 29.59
79 Ethanol C2HsOH 2 46.07 -114.1 0,3729 58.62 26.13
80 Isobutano] C4H9OH 4 74.12 -108.0 0.4634 ... 22.92
81 Carbon dioxide CO 2 1 44.01 216.58 5.0088 37.2744 14.56
82 Hydrogen sulfide H2S ... 34.08 187.68 1.5134 35.8600 18.00
83 Nitrogen N2 ... 28.01 63.15 1.3712 34.6723 9.082
84 Hydrogen H2 ... 2.02 13.95 1.0097 28.5681 6.648
85 Oxygen 02 ... 32.00 54.36 0.9824 28.0225 8.182
86 Ammonia NH3 17.03 195.41 3.4819 24.9800 29.22
87 Carbon monoxide CO "1' 28.01 68.15 1.4842 35.4400 6.402
aAPI-TDB [111 gives different values for V2s of light hydrocarbons. These values are given in parentheses and seem more accurate, as also given in Table 6.11.
Values in this table are obtained from a program in Ref. [13].
should be noted that the polynomial correlation given for n- Laar theory since by replacing A12 = (V~/RT)(SI - 82)2 and
alkylbenzenes cannot be used for compounds heavier than A21 = ( V ? / R T ) ( ~ 1 - ~2) 2 into Eq. (6.144), it becomes identical
C24. The other two equations may be extrapolated to heavier to Eq. (6.145). However, the main advantage of Eq. (6.145)
compounds. Equations (6.148) and (6.149) may be used to- over Eq. (6.144) is that parameters ViiL and ~i a r e calcula-
gether with the pseudocomponent method described in Chap- ble from thermodynamic relations. Riazi and Vera [23] have
ter 3 to estimate V2s and 8 for petroleum fractions whose shown that predicted values of solubility are sensitive to the
molecular weights are in the range of application of these values of ViiL and 8i and they have recommended some specific
equations. Values of V/L and 8 given in Table 6.10 are taken values for 8i of various light gases in petroleum fractions.
from Ref. [13] at temperature of 25~ It seems that for some Other commonly used activity coefficient models include
light gases (i.e., CH4), there are some discrepancies with re- Wilson and NRTL (nonrandom two-liquid) models, which
ported values in other references. Values of these properties are applicable to systems of heavy hydrocarbons, water, and
for some compounds as recommended by Pruasnitz et al. [21 ]
are given in Table 6.11. Obviously at 25~ for light gases such
as CH4 o r N2 values of liquid" properties represent extrapo- TABLE 6.1 l--Values of liquid molar volume and solubility
parameters for some pure compounds at 90 and 298 K.
lated values and for this reason they vary from one source to Compound V/L, (cm3/mol) 8i, (J/cm3)1/2
another. It seems that values given in Table 6.10 for light gases
N2 (at 90 K) 38.1 10.84
correspond to temperatures lower than 25~ For this reason N2 (at 298 K) 32.4 5.28
for compounds such as C1, C2, H2S, CO2, N2, and 02 values CO (at 90 K) 37.1 11.66
of V/L and 8 at 25~ as given in Table 6.11 are recommended CO (at 298 K) 32.1 6.40
to be used. 02 (at 90 K) 28.0 14.73
02 (at 298 K) 33.0 8.18
F o r m u l t i c o m p o n e n t s o l u t i o n s , E q s . (6.145) a n d (6.146) a r e CO2 (at 298 K) 55.0 12.27
replaced by the following relation: CH4 (at 90 K) 35.3 15.14
CH4 (at 298 K) 52.0 11.62
In Yi -- V/L (8i - 8mix)2 C2H6 (at 90 K) 45.7 19.43
RT C2H6 (at 298 K) 70.0 13.50
Taken from Ref. [21]. Components N2, CO, 02, CO2, CH4, and CzH6 at
(6.150) 6mlx = Y~ e;i&i 298 K are in gaseous phase (To < 298 K) and values of ~L and t~i are
i
x;vj r hypothetical liquid values which are recommended to be used. Values
given at 90 K are for real liquids. All other components are in liquid form
| - Ekx~V~ at 298 K. Values reported for hydrocarbons heavier than Cs are similar
to the values given in Table 6.10. For example, for n-Cl6 it provides
where the summation applies to all components in the mix- values of 294 and 16.34 for ~L and ~i, respectively. Similarly for benzene
values of 89 and 18.8 were provided in comparison with 89.48 and 18.7
ture. Regular solution theory is in fact equivalent to van given in Table 6.10.
6. T H E R M O D Y N A M I C R E L A T I O N S FOR P R O P E R T Y E S T I M A T I O N S 261
alcohol mixtures [21]. For hydrocarbon systems, UNIQUAC and 6.11 is shown by 8 L, then 8s may be calculated from the
(universal quasi chemical) model that is based on a group following relation [24]:
contribution model is often used for calculation of activity
coefficient of compounds with known structure. More details (6.154) (ss) 2 = (8/L)2 + An
on activity coefficient models and their applications are dis-
cussed in available references [4, 21]. The major application in which 8 is in (J/cm3) 1/2, A/~/ is in J/mol, and V/ is in
of activity coefficient models is in liquid-liquid and solid- cm3/mol.
liquid equilibria as well as low pressure VLE calculations Calculation of fugacity of solids through Eq. (6.151) re-
when cubic equations of state do not accurately estimate liq- quires calculation of f/~ For convenience the standard state
uid fugacity coefficients. for calculation of f~ is considered subcooled liquid at temper-
ature T and for this reason we show it by fiE. In the following
discussion solute component 1 is replaced by component i
6.6.6 Calculation of Fugacity of S o l i d s to generalize the equation for any component. Based on the
In the petroleum industry solid fugacity is used for SLE cal- SLE for pure i at temperature T it can be shown that [21, 25]
culations. Solids are generally heavy organics such as waxes ~s(r, p) = ~(ir, p)
and asphaltenes that are formed under certain conditions.
Solid-liquid equilibria follows the same principles as VLE. )<eXP LRTMi Q -- ~ - )
-----~ (1-- ~ ) - ACPi "
- ACvi
Generally fugacity of solids are calculated similar to the meth-
ods that fugacity of liquids are calculated. In the study of (6.155)
solubility of solids in liquid solvents usually solute (solid) is where f/S(T, P) is the fugacity of pure solid at T and P, A/ff is
shown by component 1 and solvent (liquid) is shown by com- the molar heat of fusion of solute, TMi is the melting or freezing
ponent 2. Mole fraction of solute in the solution is xl, which is point temperature, and A C p i = c L i -- CSi, which is the differ-
the main parameter that must be estimated in calculation of ence between heat capacity of liquid and solid solute at av-
solubility of solids in liquids. We assume that the solid phase erage temperature of (T + TMi)/2. Derivation of Eq. (6.155)
is pure component 1. In such a case fugacity of solid in the is similar to the derivation of Eq. (6.136) for calculation of
solution is shown b y f s, which is given by fugacity of pure liquids but in this case equilibrium between
solid and liquid is used to develop the above relation. Firooz-
(6.151) f~ (solid in liquid solution) = xl~,sf~ abadi [17] clearly describes calculation of fugacity of solids.
Since methods of calculation of f/L w e r e discussed in the pre-
where f~ is the fugacity of solute at a standard state but tem-
perature T of solution. )/s is the activity coefficient of solid vious section, f/s can be calculated from the above equation.
component in the solution. Obviously for ideal solutions yls Values of TM and A/~/for some selected compounds are given
is unity. Model to calculate ys is similar to liquid activity co- in Table 6.10 along with liquid molar volume and solubility
parameter. Estimation of freezing point TM for pure hydro-
efficients, such as two-suffix Margules equation:
carbons was discussed in Section 2.6.4. From Eq. (2.42) and
A coefficients given in Table 2.6 we have
(6.152) lnF s = ~ (1 --Xl) 2
TM = 397 -- exp (6.5096 -- 0.1487M ~
A more accurate activity coefficient model for nonpolar so- n-alkanes (C5 - C40)
lutes and solvents is given by the Scatchard-Hildebrand rela- TM ----370 -- exp (6.52504 -- 0.04945M 2/3)
tion (Eq. 6.145): (6.156)
n-alkylcyclopentanes (C7 - C40)
(6.153) In y1S : vL(81 -- 82)2(I)2 TM = 375 -- exp (6.53599 -- 0.04912M 2/3)
RT n-alkanes (C9 - C42)
where VIE is the liquid molar volume of pure component 1 at where TM is in kelvin. Average deviation for these equations
T and P, 82 is solubility of solvent, 81 is the solubility parame- are 1.5, 1.2, and 0.9%, for n-alkanes, n-alkylcyclopentanes,
ter of subcooled component 1, and ep2 is the volume fraction and n-alkylbenzenes, respectively. Similarly based on the data
of solvent and is given by Eq. (6.146). Methods of calculation given for A ~ in Table 6.10 the following relations are devel-
of 81 and apl have been discussed in Section 6.6.5.81 can be oped for estimation of heat of fusion of pure hydrocarbons
calculated from Eq. (6.147) from the knowledge of heat of for the PNA homologous families.
vaporization of solute, AH~ap. Values of the solubility param-
eter for heavy single carbon number components are given in In AHf _ -71.9215 + 70.7847M ~176
Table 4.6. When the liquid solvent is a mixture 82 is replaced RTM
for n-alkanes (C2 - Ca6)
by 8mu and ys is calculated through Eq. (6.150). It should be
noted that when Eq. (6.153) is used to calculate fugacity of a AHf
n R~M ----0.8325 + 0.009M
solid in a liquid solution value of 8 can be obtained from Table
for n-alkylcyclohexanes (C7 - C16)
6.10 from liquid solubility data. However, when this equation
is applied for calculation of fugacity of a solid component i AHf _ 1.1556 + 0.009M + 0.000396M 2 - 6.544 • 10-7M 3
in a homogeneous solid phase mixture (i.e., wax) then solid RTM
solubility, 8s, should be used for value of 8 as recommended for n-alkylbenzenes(C6 - C24)
by Won [24]. If a value of liquid solubility given in Tables 6.10 (6.157)
262 C H A R A C T E R I Z A T I O N AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
where TMi is the melting point in kelvin and R is the gas con- 30
n-Alkylbenzenes
stant. The ratio AI-Iif./RTta is dimensionless and represents en-
tropy of fusion. Unit of A/if/depends on the unit of R. A/ff o Data ^o~
may also be calculated from entropy change of fusion, AS/f, Proposed Equation
whenever it is available. 20 .......
(6.158) A / ~ / = TMiAS/f
The above equation may also be used to estimate AS/f from
A/~. calculated through Eq. (6.157). Equation (6.157) can ,0
reproduce data with average deviations of 12.5, 5.4, and
3.8% for n-alkanes, n-alkylcyclohexanes, and n-alkylbenzens,
respectively. Firoozabadi and co-workers [17, 24-26] have
provided the following equations for calculation of A/-~: 0 I i i I I
0 5 10 15 20 25 30
A/~f =0.07177M~ for paraffins CarbonNumber,Nc
RT~ai
FIG. 6.10--Prediction of entropy of fusion of
(6.159) A / ~ = 0.02652M/ for naphthenes and isoparaffins n-alkylbenzens. Proposed equation: Eq. (6.157);
RTM~
Won method: Eq. (6.159); data from DIPPR [13],
A/-ff = 5.63664 for aromatics
RTM~
n-alkanes at two different temperature of 298 K and freez-
where AI-~/RTM is dimensionless. The relation given for cal- ing point is shown in Fig. 6.11. As is seen from this figure,
culation of A/if/of aromatics (Eq. 6.159) suggests that the en- Eq. (6.160) gives values higher than actual values of ACpi.
tropy of fusion is constant for all aromatics. While this may Generally, actual values of ACr~ are small and as will be seen
be true for some multiring aromatics, it certainly is not true later they may be neglected in the calculation of f/s from
for n-alkylbenzenes. Graphical comparisons of Eqs. (6.157) Eq. (6.155) with good approximation.
and (6.159) for calculation of entropy of fusion of n-alkanes Another type of SLE that is important in the petroleum in-
and n-alkylbenzenes and evaluation with data given in Ta- dustry is precipitation of heavy organics, such as asphaltenes
ble 6.10 are shown in Figs. 6.9 and 6.10. As is seen from Fig. and waxes, that occurs under certain conditions. Wax and
6.10, the entropy of n-alkylbenzenes does change with carbon asphaltene precipitation can plug the well bore formations
number. and it can restrict or plug the tubing and facilities, such
Calculation of fugacity of solids also requires ACre. The fol- as flowlines and production handling facilities, which can
lowing relation developed for all types of hydrocarbons (P, N, lead to major economic problems. For this reason, knowl-
and A) by Pedersen et al. [26] is recommended by Firoozabadi edge of the conditions at which precipitation occurs is impor-
for calculation of ACpi [17]: tant. In formulation of this phase transition, the solid phase
is considered as a solution of mixtures of components that
(6.160) ACe~ = R(0.1526M~ - 2.3327 • 10-4M~T) fugacity of i is shown by ~ s and can be calculated from the
where T is the absolute temperature in kelvin and Mi is molec- following relation:
ular weight of i. The unit of ACr~ is the same as the unit of R. ^S
Evaluation of this equation with data from DIPPR [13] for (6.161) fi (solid i in solid mixture) = xSyis fis
40
n-Alkanes
Won Method at 25 C # ,
40 9 Data from DIPPR at 25 C J O
o Data .-'" /
30 - - - Won Method at Freezing Point / o
Proposed Equation . - " " ' f o Data from " " . *
. . . . . . . WonMethod o .:*/o 3O
o ..~ Ooo o
20 o .." oo
o ~176
20
"9 lO ~ ~176
./ t 9 -
9
10 9 o
@
o
82 o ,
0 l0 20 30 40
0
0 10 20 30 40
CarbonNumber,Nc
CarbonNumber,Nc
FIG. 6.9--Prediction of entropy of fusion
of n-alkanes. Proposed equation: Eq. (6.157); FIG. 6.11--Values of ACp~(= c L j - CS) for
Won method: Eq. (6.159); data from DIPPR n-alkanes. Won method: Eq. (6.160); data from
[13]. DIPPR [13].
6. T H E R M O D Y N A M I C R E L A T I O N S F O R P R O P E R T Y E S T I M A T I O N S 263
where f/s is the fugacity of pure i at T and P of the system. Eq. (6.138). Similarly VE and H E can be calculated from yi
In wax precipitation usually the solid solution is considered and H and V of the solution may be calculated from the fol-
ideal and yis is assumed as unity [17]. xs is the mole fraction lowing relations:
of solid i in the solid phase solution. Here the term solution ( Oln yi "]
means homogeneous mixture of solid phase. As it will be seen H = H id - - RT 2 Y'~xi \ ~ T ~ j l , ~ i
in the next chapter these relations can also be used to deter- (6.166)
mine the conditions at which hydrates are formed. (01n),i ~
V = V ia + RT ~_,xi \ - - ~ - - , / T m
Calculation of fugacity of pure solids through Eq. (6.155)
is useful for SLE calculations where the temperature is above
the triple-point temperature (Ttp). When temperature is less Once G, H, and V are known, all other properties can be cal-
than Ttp we have solid-vapor equilibrium as shown in Fig. culated from appropriate thermodynamic relations discussed
5.2a. For such cases the relation for calculation of fugacity of in Section 6.1.
Another common way of determining thermophysical pro-
pure solids can be derived from fugacity of pure vapor and
perties is through thermodynamic diagrams. In these dia-
effect of pressure on vapor phase fugacity similar to deriva-
tion of Eq. 6.136, where f/L, plat, and V/L should be replaced grams various properties such as H, S, V, T, and P for both
by f/s, p/sub, and V/s, respectively. However at T < Ttp, Pisub or liquid and vapor phases of a pure substance are graphically
solid-vapor pressure is very low and &at is unity. Furthermore shown. One type of these diagrams is the P - H diagram that
molar volume of solid, V/s is constant with respect to pressure is shown in Fig. 6.12 for methane as given by the GPA [28].
(see Problem 6.15). Such diagrams are available for many industrially important
pure compounds [28]. Most of these thermodynamic charts
and computer programs were developed by NIST [29]. Val-
6.7 G E N E R A L M E T H O D F O R CALCULATION ues used to construct such diagrams are calculated through
OF P R O P E R T I E S OF R E A L M I X T U R E S thermodynamic models discussed in this chapter. While these
diagrams are easy to use, but it is hard to determine an accu-
Two parameters have been defined to express nonideality of rate value from the graph because of difficulty in reading the
a system, fugacity coefficient and activity coefficient. Fugac- values. In addition they are not suitable for computer appli-
ity coefficient indicates deviation from ideal gas behavior and cations. However, these figures are useful for the purpose of
activity coefficient indicates deviation from ideal solution be- evaluation of an estimated property from a thermodynamic
havior for liquid solutions. Once residual properties (devia- model. Other types of these diagrams are also available. The
tion from ideal gas behavior) and excess properties (deviation H - S diagram known as Mollier diagram is usually used to
from ideal solution behavior) are known, properties of real graphically correlate properties of refrigerant fluids.
mixtures can be calculated from properties of ideal gases or
real solutions. Properties of real gas mixtures can be calcu-
lated through residual properties. For example, applying the 6.8 F O R M U L A T I O N OF P H A S E E Q U I L I B R I A
definition of residual property to G we get PROBLEMS FOR MIXTURES
(6.162) G = G ig + G R In this section equations needed for various phase equi-
where GR is the residual Gibbs energy (defined as G - Gig). librium calculations for mixtures are presented. Two cases
G R is related to q~i by Eq. (6.128), which when combined with of vapor-liquid equilibria (VLE) and liquid-solid equilibria
the above equation gives (LSE) are considered due to their wide application in the
petroleum industry, as will be seen in Chapter 9.
(6.163) G = G ig + RT E y i ln~i
Furthermore from thermodynamic relations one can show 6.8.1 Criteria for Mixture Phase Equilibria
that [ 1]
The criteria for phase equilibrium is set by minimum Gibbs
H=Y~.yiHiig-RT2~yi(Oln~)i~ free energy, which requires derivative of G to be zero at the
\or],,, conditions where the system is in thermodynamic equilib-
(6.164) rium as shown by Eq. (6.108). Gibbs energy varies with T, P,
V = y~yiv/ig + RT~]yi I~ and x4. At fixed T and P, one can determine x~ that is when G is
\ - S g - / T,,' minimized or at a fixed T (or P) and x~, equilibrium pressure
Calculation of properties of ideal gases have been discussed (or temperature) can be found by minimizing G. At different
in Section 6.3, therefore, from the knowledge of fugacity co- pressures functionality of G with x~ at a fixed temperature
efficients one can calculate properties of real gases. varies. Baker et al. [29] have discussed variation of Gibbs en-
Similarly for real liquid solutions a property can be cal- ergy with composition. A typical curve is shown in Fig. 6.13.
culated from the knowledge of excess property. Properties of To avoid a false solution to find equilibrium conditions, there
ideal solutions are given by Eqs. (6.89)-(6.92). Property of is a second constraint set by the second derivative of G as
a real solution can be calculated from knowledge of excess [17, 20, 30]
property and ideal solution property using Eq. (6.83): (0G)r,p = 0
(6.165) M ~--- M ig "4- M E (6.167) (02G)r,e > 0
where M E is the excess property and can be calculated from This discussion is known as stability criteria and it has re-
activity coefficients. For example, G ~ can be calculated from ceived significant attention by reservoir engineers in analysis
264 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
-~ll0 -I]1~0 -I~00 -]tl70 -1840 -III10 -171~ -0~0 -ff~0 -~80 -NS0 -1420 -15~0 -1560 - I ~ -/,~00 -I]PT0 -a40
100OO - - - 10OO0
100E ~O00
lOC I00
1t11
-1580 -IS30 -lg00 -1870 -181,0 -1810 -1780 -1"/~0 -172(I -l!10 -i6t)0
........
-1130 -1100 -1S70 -1~140 ~1510 -14~ -1450 -14ZO -1350 -1360 -1330 -1300 -1270 -1240
of fluid phase equilibrium of petroleum mixtures. Further dis- components with two phases of 0t and ft. Applying Eq. (6.79)
cussion regarding phase stability is given in a number of re- to total Gibbs energy, G t, and taking the derivative of G t with
cent references [17, 20, 31]. respect to r~ at constant T and P and combing with the Gibbs-
Derivation of the general formula for equilibrium condi- Duhem equation (Eq. 6.81) gives the following relation:
tions in terms of chemical potential and fugacity for multi-
component systems is shown here. Consider a mixture of N (6.168) dG t = ~/2idn/
i
0 1.0 This relation must apply to all components when the system
Composition,xi
is in equilibrium. If there are more than two phases (i.e., ~, r,
FIG. 6.13--A sample variation of Gibbs en- y .... ) the same approach leads to the following conclusion:
ergy versus composition for a binary system
at constant T and P. (6.17 I) /2~ =/~/~ =/21 . . . . for every i at constant T and P
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 265
Using the relation between fii and~ given by Eq. (6.119) we x~ ~ 0, will result in the following relation for calculation of
get activity coefficient at infinite dilution for component 1 (F~)
in a binary system of components 1 and 2 at T and P [17]:
(6.172) /~ --/~ --/~Y. . . . for e v e r y / a t constant T and P
bl , /' Zl -- B1 ~l
Equations (6.171) or (6.172) are the basis for formulation ln),~ = ~2(Z2 - 1) - (Z1 - 1) + m ~ )
of mixture phase equilibrium calculations. Application of
Eq. (6.172) to VLE gives + 2 _A1
_ ~ i in ( Z ~ + 2.414B1
- 0.414B1 )
(6.173) /~V(T, P, Yi) =~L(T, P, xi) (6.178) _ (al2P~ 1 ln(z~+2.414B2 ~
for all i components at constant T and P
blA2 ln(Z2 +2.414Bz']
For SLE, Eq. (6.172) becomes
-~ 2~B2b~2 \ Z2 ~ 1
(6.174) ~S(T, P,x s) =/~ L(T, P,x L)
where a12 a~/2a~/2(I -- k12) in which k12 is the binary inter-
=
for all i components at constant T and P action parameter. Parameters a, b, A, and B for PR EOS are
where xs is the mole fraction of i in the solid phase. Simi- given in Table 5.1. Z1 and Z2 are the compressibility factor
larly Eq. (6.172) can be used in liquid-liquid equilibria (LLE), for components 1 and 2 calculated from the PR EOS.
solid-liquid-vapor equilibria (SLVE), or vapor-liquid-liquid The main difference between Eqs. (6.175) and (6.176) for
equilibria (VLLE). VLE calculations is in their applications. Equation (6.176)
is particularly useful when both ~v and q~L are calculated
from equations of state. Cubic EOSs generally work well in
6.8.2 Vapor-Liquid Equilibria---Gas Solubility the VLE calculation of petroleum systems at high pressures
i n Liquids through this equation. ~v and ~/L may be calculated through
Eq. (6.126) with use of appropriate composition and Z; that
In this section general relations for VLE and specific relations is, xL and Z L must be used in calculation of ~ , while yV
developed for certain systems such as Raoult's and Henry's and ZVare used in calculation of ~v. Binary interaction coeffi-
laws are presented. For high pressure VLE calculations equi- cients (BIPs) given in Table 5.3 must be used when dissimilar
librium ratio (Ki) is defined and its methods of estimation for (very light and very heavy or nonhydrocarhon and hydrocar-
hydrocarbon systems are presented. bon) molecules exist in a mixture. However, as mentioned
earlier there is no need for use of volume translation or shift
6.8.2.1 Formulation of Vapor-Liquid parameter in calculation of ~v and ~/L for use in Eq. (6.176).
Equilibria Relations At low and moderate pressures use of Eq. (6.175) with
Formulation of VLE calculations requires substitution of rela- activity coefficient models is more accurate than use of
tions for/~ v and/~ L from Eqs. (6.113) and (6.114). Combining Eq. (6.176) with an EOS. Assuming constant V/L and substi-
Eqs. (6.114) and (6.173) gives the following relation: tuting Eq. (6.136) into Eq. (6.175) we have
Eq. (6.180) reduces to of bubble and dew point pressures and generation of such
diagrams will be discussed in Chapter 9.
(6.181) yiP -~- X./yi P/sat
6.8.2.2 Solubility o f Gases in Liquids--Henry's Law
This relation also k n o w n as modified Raouh's law is valid for
nonideal systems but at pressures of I atm or less where the Another important VLE relation is the relation for gas solubil-
gas phase is considered ideal gas. We know that as x/--~ 1 (to- ity in liquids. Many years ago it has been observed that solu-
ward a pure component) thus yi --~ 1 and therefore Eq. (6.181) bility of gases in liquids (x/) is proportional to partial pressure
reduces to Raoult's law even for a real solution. Nonideal sys- of c o m p o n e n t in the gas phase (Yi P), which can be formulated
tems with ~/i > 1 show positive deviation while with ~'i < 1 as [21]
show negative deviation from the Raoult's law. One direct ap-
plication of modified Raouh's law is to calculate composition (6.184) yi P = k~x~
of a c o m p o u n d in the air when it is vaporized from its pure
liquid phase (xi = 1, Yi = I). This relation is k n o w n as Henry's law and the proportion-
ality constant k~ is called Henry's constant. ]q-solvent has the
(6.182) YiP = psat unit of pressure per mole (or weight) fraction and for any
given solute and solvent system is a function of temperature.
Since for ideal gas mixtures volume and mole fractions are Henry's law is a good approximation when pressure is low
the same therefore we have (not exceeding 5-10 bar) and the solute concentration in the
solvent, x/, is low (not exceeding 0~03) and the temperature
(6.183) vol% o f i in air = p/sat is well below the critical temperature of solvent [21]. Henry's
Pa law is exact as x/--~ 0. In fact through application of Gibbs-
(for vaporization of pure liquid i) D u h e m equation in terms of Yi (Eq. 6.141), it can be shown
that for a binary system when Henry's law is valid for one
where Pa is atmospheric pressure. This is the same as Eq. c o m p o n e n t the Raoult's law is valid for the other c o m p o n e n t
(2.11) that was used to calculate a m o u n t of a gas in the air for (see Problem 6.32). Equation (6.184) m a y be applied to gases
flammability test. Behavior of ideal and nonideal systems is at higher pressures by multiplying the left side of equation
shown in Fig. 6.14 through Txy and Pxy diagrams. Calculation b y ~ v.
P-Const. T-Const.
T
W-Xl ~
T~ t p~at
L Bubble point Dew point V
0 1. 0
xbyl Xl,Yl
(a)Txydiagramfor an idealbinarysystem (b)Pxydiagramfor an idealbinarysystem
azeotropi {
L I
I az az
~,xl = yl x =yl J
0 1.0 0 1.
xj~yl Xl,Yl
(c)Txydiagramfor a real binarysystem (d)Pxydiagramfor a real binarysystem
FIG. 6.14--Txy and Pxy diagrams for ideal and nonideal systems.
6. T H E R M O D Y N A M I C R E L A T I O N S F O R P R O P E R T Y E S T I M A T I O N S 267
1300
/" kl
Constant T and P , ,' 1200
t t
i
t1 1100
t t
t
s
f 1000
Henry'sLaw / o
AL t"
t t
900
t'1 = klx] /
i t m 800
~L / f~ (Pure l) 700
600
0 50 100 150 200
S 9 sss sj. SS~S~SJS'~
Temperature,~
FIG. 6.16--Henry's constant for hydrogen in n-
hexadecane(~C16H~).
/ .- ~L =x,fL
1.0
XI where yi~ is the activity coefficient at infinite dilution and f/L
is the fugacity of pure liquid i at T and P of the system. Where
FIG. 6.15~Variation o f ~ with x~ in a binary liquid solu- if yi~ is calculated through Eq. (6.178) and the PR EOS is used
tion and comparison with its values from Henry's law and to calculate liquid fugacity coefficient ()rE = ~bLp), Henry's
Lewis rule.
constant can be calculated from the PR EOS.
The general mixing rule for calculation of Henry's constant
The RHS of Eq. (6.184) is~ L and in fact the exact definition for a solute in a mixed solvent is given by Prausnitz [21]. For
of Henry's constant is [1, 21] ternary systems, Henry's law constant for component 1 into
a mixed solvent (2 and 3) is given by the following relation:
(6.188) - +
Therefore, k/ is in fact the slope of ]~L versus x~ at x~ = 0. O/23
This is demonstrated in Fig. 6.15 for a binary system. The 2RT
Henry's law is valid at low values of x l ( ~ < 0.03) while as
Xl ~ 1, the system follows Raoult's law. Henry's constant gen- where ~ is the solubility parameter, V is molar volume, and
erally decreases with increase in temperature and increases x is the mole fraction. This relation may be used to calcu-
with increase in pressure. However, there are cases that that k4 late activity coefficient of component 1 in a ternary mixture.
increases with increase in temperature such as Henry's con- Herein we assume that the mixture is a binary system of com-
stant for H2S and NH3 in water [21]. Generally with good ponents 1 and M, where M represents components 2 and 3
approximation, effect of pressure on Henry's constant is ne- together (xu = 1 - x0. Activity coefficient at infinite dilution
glected and ki is considered only as a function of temperature. is calculated through Eq. (6.187) as Yl,M~176= kx,M/fL. Once Yl,g~176
Henry's law constant for a solute (component i) in a solvent is known, it can be used to calculate parameters in an activity
can be estimated from an EOS through liquid phase fugacity coefficient model as discussed earlier.
coefficient at infinite dilution (q~L,~ = lim~_~0 q~L)[21]. The main application of Henry's law is to calculate solu-
bility of gases in liquids where the solubility is limited (small
(6.186) k~ = ~L,~p
xl). For example, solubilities of hydrocarbons in water or light
P16cker et al. [33] calculated/q using Lee-Kesler EOS through hydrocarbons in heavy oils are very limited and Henry's law
calculation of ~/L'~ the above equation for solute hydro- may be used to estimate the solubility of a solute in a solvent.
gen (component 1) in various solvents versus in temperature The general relation for calculation of solubility is through
range of 295-475 K. Their calculated values of/q for HE in Eq. (6.147). For various homologous groups, Eq. (6.149) may
n-C16 are presented in Fig. 6.16 for the temperature range of be used to estimate solubility parameter at 25~ One ma-
0-200 ~ These calculated values are in good agreement with jor problem in using Eq. (6.179) occurs when it is used to
the measured values. The equation used for extrapolation of calculate solubility of light gases (C1, C2, or C3) in oils at tem-
data is also given in the same figure that reproduce original peratures greater than Tc of these components. In such cases
data with an average deviation of 1%. Another useful rela- calculation of p~at is not possible since the component is not
tion for the Henry's constant is obtained by combining Eqs. in a liquid form. For such situations Eq. (6.175) must be used
(6.177) and (6.186): and f/L represents fugacity of component i in a hypothetical
liquid state. If solute (light gas) is indicated as component 1,
(6.187) /q = ),/~ f L the following equation should be used to calculate fugacity
268 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
10.0
t~~ 1.0
o Data
O.1
0.5 1.0 1.5 2.0 2.5 3.0
Tr
FIG. 6.17--Fugacity of hypothetical liquid at 1.013 bar. Data taken from
Ref. [21] for CH4, C2H4, C2Hs, and Na.
of pure c o m p o n e n t 1 as a hypothetical liquid when Tr > 1 an initial value of xl is normally obtained from Eq. (6.191) by
[21]: assuming 7/1 = 1. As an alternative method, since values of Xl
are normally small, initial value ofxl can be assumed as zero.
(6.189) fL= fr~ ] For h y d r o c a r b o n systems Y1 m a y be calculated from regular
solution theory. The following example shows the method.
where f~L is the reduced hypothetical liquid fugacity at pres-
sure of 1 atm (fr~ = f~ and it should be calculated from Example 6./O--Estimate solubility of m e t h a n e
in n-pentane
Fig. 6.17 as explained in Ref. [21]. Data on foe Of C1, C2, N2, at 100~ when the pressure of methane is 0.01 bar.
CO, and CO2 have been used to construct this figure. For con-
venience, values obtained from Fig. 6.17 are represented by SolutionnMethane is considered as the solute (component
the following equation [23]: 1) and n-pentane is the solvent (component 2). Properties of
m e t h a n e are taken from Table 2.1 as M = 16, Tc = 190.4 K,
(6.190) f;L=exp(7.902 8.19643Tr 3.081nTr) and Pc = 46 bar. T = 373.15 K (Tr = 1.9598) and P = 0.01
bar. Since the pressure is quite low the gas phase is ideal
where Tr is the reduced temperature. Data obtained from Fig. gas, thus ~bv = 1.0. In Eq. (6.191) only Y1 and f~ must be
6.17 in the range of 0.95 < T~ < 2.6 are used to generate the calculated. For C1-C5 system, the regular solution theory
above correlation and it reproduces the graph with %AAD of can be used to calculate Yx t h r o u g h Eq. (6.145). F r o m Ta-
1.3. This equation is not valid for T~ > 3 and for c o m p o u n d s ble 6.11, at 298 K, V( = 52 and VL = 116 cm3/mol, ~i -- 11.6,
such as H2. and ~2 = 14.52 (J/cm3) 1/z. Assuming ~a ~ 0 (~2 -~ 1), from
If the vapor phase is pure c o m p o n e n t 1 and is in contact Eq. (6.145), In 7/1 = 0.143 or Yl = 1.154. Since Tr > 1, fL is
with solvent 2 at pressure P and temperature T, then its solu- calculated from Eq. (6.189). From Eq. (6.190), f7 L = 5.107
bility in terms of mole fraction, xl, is found from Eq. (6.168) and from Eq. (6.189): fL = 234.8 bar. Therefore, the solu-
as bility is xl = 0.01/(1.1519 x 234.8) = 3.7 x 10 -5. Since xl is
very small, the initial guess for @t = 0 is acceptable and there
~'P
(6.191) xl - Yl fL is no need for recalculation of y~. Therefore, the answer is
3.7 x 10 -5, which is close to value of 4 • 10 -5 as given in Ref.
where ~bv is the fugacity coefficient of pure gas (component [21]. #
1) at T and P. Yi is the activity coefficient of solute 1 in sol-
vent 2, which is a function of xl. f~ is the fugacity of pure One type of useful data is correlation of mole fraction sol-
c o m p o n e n t 1 as liquid at T and P and it m a y be calculated ubility of gases in water at 1.013 bar (1 atm). Once this infor-
from Eq. (6.189) for light gases when T > 0.95Tcl. It is clear m a t i o n is available, it can be used to determine solubility at
that to find xl from Eq. (6.191) a trial-and-error procedure is other elevated pressures through Henry's law. Mole fraction
required since ~/1 is a function of xl. To start the calculations solubility is given in the following correlations for a n u m b e r
6. T H E R M O D Y N A M I C RELATIONS FOR PROPERTY ESTIMATIONS 269
TABLE 6.12--Constants for Eq. (6.193) for estimation of Henry's constant for light gases in water [5].
Gas T range, K Pressure range, bar A1 A2 A3 A4 %AAD
Methane 274-444 1-31 569.29 0.107305 - 19537 -92.17 3.6
Ethane 279--444 1-28 109.42 -0.023090 - 8006.3 - 11.467 7.5
Propane 278-428 1-28 1114.68 0.205942 -39162.2 - 181.505 5.3
n-Butane 277-444 1-28 182.41 -0.018160 - 11418.06 -22.455 6.2
/-Butane 278-378 1-10 1731.13 0.429534 -52318.06 -293.567 5.3
270 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
I0
1
~ z f!
0,~
0,1
. . . . . . . . . . . . . . . . . . ~. o..J _
0,0~ ~ a
3~
_
.....................................
m
II ..........
1
300
0.0~
3O 3OO
P(BARs)
(a) (b)
FIG. 6.18--Comparison of predicted equilibrium ratios (K~ values)
from PR EOS without (a) and with (b) use of interaction parameters, 9
Experimental data for a crude oil. Taken with permission from Ref. [32],
(i.e., C1 and some heavy compounds) use of BIPs is required. must be evaluated with the Scatchard-Hildebrand regu-
Effect of BIPs in calculation of Ki values is demonstrated in lar solution relationship (Eq. 6.150). q~v must be evaluated
Fig. 6.18. If both the vapor and liquid phases are assumed as with the original Redlich-Kwong equation of state. Further-
ideal solutions, then by applying Eq. (6.132) the Lewis rule, more, Chao and Seader developed a generalized correlation
Eq. (6.197) becomes for calculation of ~0~ in terms of reduced temperature, pres-
sure, and acentric factor of pure component i (T,~, P~i, and
~(~, P) wi). Later Grayson and Streed [35] reformulated the corre-
(6.198) Ki - 4~V(T,p)
lation for o/L to temperatures about 430~ (~ 800~ Some
where q~v and ~0/Lare pure component fugacity coefficients process simulators (i.e., PRI/II [36]) use the Greyson-Streed
and Ki is independent of composition and depends only on expression for q~/L.This method found wide industrial appli-
T and P. The main application of this equation is for light cations in the 1960s and 1970s; however, it should not be
hydrocarbons where their mixtures may be assumed as ideal used for systems containing polar compounds or compounds
solution. For systems following Raoult's law (Eq. 6.180) the with close boiling points (i.e., i-C4/n-C4). It should not be
Ki values can be calculated from the relation: used for temperatures below -17~ (0~ nor near the criti-
cal region where it does not recognize x~ = yi at the critical
(6.199) Ki(T, P) = p/sat(T) point [37]. For systems composed of complex molecules such
P as very heavy hydrocarbons, water, alcohol, ionic (i.e., salt,
Equilibrium ratios may also be calculated from Eq. (6.181) surfactant), and polymeric systems, SAFT EOS may be used
through calculation of activity coefficients for the liquid for phase equilibrium calculations. Relations for convenient
phase. calculation of fugacity coefficients and compressibility factor
Another method for calculation of Ki values of nonpolar are given by Li and Englezos [38].
systems was developed by Chao and Seader in 1961 [34]. They Once Ki values for all components are known, various VLE
suggested a modification of Eq. (6.197) by replacing q~/Lwith calculations can be made from the following general relation-
(Yi4J~),where q~]-is the fugacity coefficient of pure liquid i and ship between xi and Yi:
Fi is the activity coefficient of component i.
(6.201) Yi = Kix~
(6.200) K~ = Y~ = ),~ck~(Ta,Pa, r Assuming ideal solution for hydrocarbons, Ki values at var-
x, ious temperature and pressure have been calculated for n-
yi must be evaluated from Eq. (6.150) paraffins from C1 to C10 and are presented graphically for
~v must be evaluated from the Redlich-Kwong EOS quick estimation. These charts as given by Gas Processor As-
~biL empirically developed correlation in terms of T~/, P~/, COl sociation (GPA) [28] are given in Figs. 6.19-6.31 for various
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 271
PRESSURE, PSIA
10 2 30 4 50 6 7 8 9100 2 300 4 500 6 7 8 9 1,000 2 3,000 4 6 7 8 910,000
1,0001 1,000
XXxXX
\ X~X\ X" X X
%% "-,
x "~\xx x
x x ,
"~% x x ~x x x
x x , " xx"x
x\ ~ x x xx~ x x~
x x x ~x
\ x xx ,
x
x
\ x x
x
\
K = Y/x K=Y/x
0/] [ 01
10 2 30 4 50 6 7 8 9100 2 300 4 500 6 7 8 9 1,000 2 3,000 4 6 7 8 910,000
K = Y/x K = Y/x
7
'6
~5
14
!3
I
0.C 0.01
10 2 30 4 50 6 7 8 @ 1 0 0 2 300 45006 7891,000 it 3,000 4 6 7 8910,000
where P is the pressure in bar and T is the temperature in where T is in kelvin and P is in bar. It should be noted that
kelvin. These equations are restricted to pressures below 69 all the original equations and constants in Table 6.11 were
bar (-~1000 psia) and temperatures between 278-366 K (40- given in the English units and have been converted to the SI
200~ Values ofb and TBfor these T and P ranges are given in units as presented here. As it can be seen in these equations
Tables 6.13 for some pure compounds and lumped C6 group. Ki is related only to T and P and they are independent of
These equations reproduce original data within 3.5% error. composition and are based on the assumption that mixtures
For C7+ fractions the following equations are provided by behave like ideal solutions. These equations are referred as
Standing [40]: Standing method and they are recommended for gas conden-
0 = 3.85 + 0.0135T + 0.02321P sate systems and are useful in calculations for surface sep-
arators. Katz and Hachmuth [41] originally recommended
(6.203) bT+ = 562.78 + 1800 - 2.36402
that K7+ = 0.15Kn-c7, which has been used by Glaso [42] with
Ta,7+ = 167.22 + 33.250 - 0.539402 satisfactory results. As will be seen in Chapter 9, in VLE
6. THERMODYNAMIC ~ L A T I O N S FOR PROPERTY ESTIMATIONS 273
PRESSURE, PSIA
10 2 30 4 50 6 7 8 9 1 0 0 2 300 45006 7 891,0C0 2 3,000 4 6 78910,000
,t,,
0.0( 001
10 2 30 4 50 6 7 8 9100 2 300 4 500 6 7 8 9 1,0(X] 2 3,000 4 8 7 8 910,000
calculations some initial Ki values are needed. Whitson [31] equation when the Edmister equation (Eq. 2.108) is used for
suggests use of Wilson correlation for calculation of initial Ki the acentric factor (see Problem 6.39).
Values:
Example 6.11--Pure propane is in contact with a nonvolatile
(6.204) Ki = exp [5.37 (1 + wi) (1 - Tffl)] oil (M = 550) at 134~ and pressure of 10 bar. Calculate Ki
P~/ value using the regular solution theory and Standing correla-
tion.
where Tu and Pu are the reduced temperature and pressure
as defined in Eq. (5.100) and wi is the acentric factor. It Solution--Consider the system as a binary system of com-
can be shown that Wilson equation reduces to Hoffman-type ponent 1 (propane) and component 2 (oil). Component 2 is
274 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
,%
O,OI 001
10 2 30 4 50 6 7 8 9 1 0 0 2 300 4 5008 7891,000 2 3,0~ 4 6 7 8910,000
0.0( 001
10 2 30 4 50 6 7 8 9100 8 300 4 500 6 7 8 9 1,C00 8 3,000 4 6 7 8 910,000
be calculated from its density. Substituting SG = 0.5063 a n d we have Vf = 550/0.9123 = 602.9 cm3/mol. To calculate Yx
T = 298 K in Eq. (5.127) gives density at 25~ as 0.493 g/cm 3 from Eq. (6.145), Xl is required. The initial value of Xl is
a n d the m o l a r volume is VL = 44.1/0.493 = 89.45 cma/mol. calculated t h r o u g h Eq. (6.191) a s s u m i n g Yl = 1. Since Tr >
Similarly at 134~ we get V1L = 128.7 cm3/mol. Value of 1, the value of fL is calculated t h r o u g h Eqs. (6.189) a n d
for Ca is given in Table 6.10 as 8 = 13.9 (J/cm3) 1/2. F r o m (6.190) as flL = 51.13 bar. Finally, the value of yl is calculated
Eq. (4.7) a n d coefficients in Table 4.5 for oil of M = 550, we as 1.285, which gives Xl = 0.94 x 10/(1.285 x 51.13) = 0.144.
get d20 = 0.9234 g/cm 3 a n d 82 = 8.342 (cal/cm3) 1/2. These val- Thus, K1 = 1/0.144 = 6.9. To calculate K1 from the Stand-
ues are very approximate as oil is a s s u m e d as a single c a r b o n ing method, Eq. (6.202) should he used. F r o m Table 6.13 for
n u m b e r . Density is corrected to 25~ t h r o u g h Eq. (2.115) as propane, b = 999.4 K a n d TB = 231.1 K, a n d from Eq. (6.202)
d25 = 0.9123 g/cm 3. Thus at 298 K for c o m p o n e n t 2 (solvent) at 407 K a n d 10 bar, a = 0.1069, c = 0.8646, a n d K1 = 5.3. r
276 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
PRESSURE, PSIA
10 ~ 30 4 50 6 7 8 9100 2 300 4 560 6 7 a 9 1,000 2 3,000 4 6 7 8 910~00,
K=Y/x K=Y/x
l
0.001
10 2 30 4 50 s 7 a 9100 2 300
PRESSURE, PSIA
4 500 s 7 a 9 1,000
~
2 3,000
i - PENTANE
4
K= Y/x K= Y/x
It should be noted that this equation can be used to calculate The above equation provides a quick way of calculating sol-
solubility of a pure solid into a solvent, yi (a function ofx~) can ubility of a solid into a solvent where chemical nature of so-
be calculated from methods given in Section 6.6.6 and x~ must lute is similar to that of solvent; therefore, only properties of
be found by trial-and-error procedure with initial value of xl solute are needed. Calculation of heat of fusion (AH/f) was
calculated at yi = 1.0. However, for ideal solutions where yi discussed in Section 6.6.6 and calculation of freezing point
is equal to unity the above equation can be used to calculate (TMi) was discussed in Section 2.6.4. Solubility of naphthalene
solubility directly. Since actual values of ACpi/R are gener- in several hydrocarbons is given in Ref. [43]. At 20~ mole
ally small (see Fig. 6.11) with a fair approximation the above fraction of solid naphthalene in solvents hexane, benzene,
relation for ideal solutions can be simplified as [17, 21, 43] and toluene is 0.09, 0.241, and 0.224, respectively. Naphtha-
lene (aromatics) has higher solubility in benzene and toluene
(6.207) x/ = exp L~L<IMi 1 - - (also aromatics) than in hexane (a paraffinic). Naphthalene
278 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
PRESSURE, PSIA
K=Yj = Y/x
000( .0001
10 2 30 4 50 6 7 8 9 1 0 0 2 300 4 5006 7891,000 2 3,000 4 6 78910,000
has A/-//f of 18.58 kJ/mol and its melting point is 80~ [43]. For solid precipitation such as wax precipitation, the solid
Solubility of naphthalene calctflated through Eq. (6.207) is phase is a mixture and the general relation for SLE is given by
xl = 0.27. Better prediction can be obtained by calculating yl Eq. (6.174). A SLE ratio, K sL, can he defined similar to VLE
for each system. For example, through regular solution theory ratio as
for naphthalene (1)-toluene (2) system at 20 ~C yl is calculated (6.208) xs ----KiSLxL
as yl = 1.17 (method of calculation was shown in Example
6.10). The corrected solubility is 0.27/1.17 = 0.23, which is in By combining Eqs. (6.114) and (6.161) with Eq. (6.174) and
good agreement with the experimental value of 0.224. As dis- use of the above definition we get [17]
cussed before, compounds with similar structures have better (6.209) K sL = f~L(T' P)yiL(T' P' xL)
solubility. f/S(T, P )yiS(T, P, x s)
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 279
K = Y/x K = Y/x
"~0
,0001
10 2 30 4 50 6 7 8 9 1 0 0 2 300 45006 7891,000 2 3,000 4 6 78910,000
where for ideal liquid and solid solutions yiL and yis are begins for a mixture (see Problem 6.29). Multicomponent
unity and the pure component fugacity ratio f/L/f/S can be SLE calculations become very easy once the stability analysis
calculated from Eq. (6.155). For nonideal solutions, yi s may is made. From stability analysis consideration a component
be calculated from Eq. (6.153). However, 8s calculated from in a liquid mixture with mole fraction zi may exist as a pure
Eq. (6.153) must be used for 8 in Eq. (6.154). Equations solid if the following inequality holds [ 17]:
(6.208) and (6.209) can be used to construct freezing/melting
points or liquid-solid phase diagram ( T x S x L) based on SLE (6.210) f i ( T , P, zi) - ~S(T, P) > 0
calculations as shown in Chapter 9. For an ideal binary sys- where ~s is fugacity of pure solid i. This equation is the ba-
tem the T x S x L diagram is shown in Fig. 6.32. Such figures sis of judgment to see if a component in a liquid mixture
are useful to determine the temperature at which freezing will precipitate as solid or not. The answer is yes if the above
280 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
PRESSURE, PSIA
10 2 30 4 506 789100 2 300 45006 7891,000 2 3,000 4 6 7 89110(~000
Plotted ftom19471al~ulalk~s ~f 9
B G.G. Brown, University of Mich-
Egan.Exlmpo/ated and drawn by
The FruorCotp, Ltd, ~n1957,
inequality holds for that component. Equality in the above of the oil. Applications of Eqs. (6.208) and (6.209) for calcu-
equation is equivalent to Eq. (6.206). The same criteria apply lation of cloud point and wax formation are demonstrated
to precipitation of a component from a gas mixture, where in in Chapter 9. Full description of a thermodynamic model for
the above inequality fi(T, P, zi) would refer to fugacity o f / i n wax precipitation is provided in Ref. [ 17]. Application of these
the gas phase with mole fraction zi. Similar principle applies relations to calculate cloud point of crude oils and reservoir
in formation of liquid i from a gas mixture when the tem- fluids are given in Chapter 9.
perature decreases. One main application of this inequality is
to determine the temperature at which solid begins to form Example 6.12--How much (in grams) n-hexacontane (n-C36)
from an oil. This temperature is equivalent to cloud point can be dissolved in 100 g of n-heptane, so that when the
6. T H E R M O D Y N A M I C R E L A T I O N S FOR P R O P E R T Y E S T I M A T I O N S 281
PRESSURE, PSIA
10
1%
2 30 4 50 6 7 8 91 O0 2 300 4 500 6 7 8 9 1,000 2 3,000 4
[ .........................
G. G Brown, universityof Mich-
igan. Extrapolatedand drawn by
The Fluor Corp. Lid. in 1957.
I0
]!
6 7 8 910,000
K=Y/x3I~. ~ ~ ~ ~ ~ \2
~ ~ ~ J/////lllllll/I 19 K=Y/x
0.0001
10 2 30 4 50 6 7 9 9100 2 300 r 500 6 7 9 9 1,000 2 3,000 4 6 7 8 910,000
t e m p e r a t u r e of the solution is r e d u c e d to 15 ~C, the solid p h a s e M1 = 506.9 a n d M2 = 100.2, f r o m Eq. (1.15) we g e t x ~ = 0.08,
begins to form. w h i c h is equivalent to 0.807 g of n-C36 in 100 g of n-C7. 0
Solution--We have a b i n a r y system of c o m p o n e n t 1 (C36) 6.8.4 Freezing Point Depression and Boiling
a n d c o m p o n e n t 2 (n-CT). C o m p o n e n t 1 is solute a n d c o m -
Point Elevation
p o n e n t 2 is c o n s i d e r e d as solvent. F r o m Table 6.10, Mz =
100.2, TM2 = 182.57 K, AH~2/RTM2 = 9 . 2 5 5 7 , M1 = 506.95, W h e n a small a m o u n t of a p u r e solid (solute) is a d d e d to
Tun = 349 K, a n d AH~/RTM1 = 30.6066. A s s u m i n g ideal so- a solvent, the freezing p o i n t of solvent decreases while its
lution we use Eq. (6.207) for calculation of xl at T = 288.2 K: boiling p o i n t increases. U p o n a d d i t i o n of a solute (compo-
xl = exp[30.6066(1 - 349/288.2)] = 0.0016. W i t h respect to nent 1) to a solvent ( c o m p o n e n t 2) m o l e fraction of solvent
282 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
PRESSURE, PSIA
10 2 30 4 50 6 7 8 9100 2 300 4 500 6 7 8 9 1,000 2 3,000 4 6 7 8 9110(~000
(x2) reduces from unity. This slight reduction in mole fraction neglecting ACpi:
of x2 causes slight reduction in chemical potential according
to Eq. (6.120). Therefore, at the freezing point when liquid
solvent is in equilibrium with its solid, the activity of pure
solid must be lower than its value that corresponds to nor-
mal freezing point. This decrease in freezing point is called where A/-/~2is the molar heat of fusion for pure solvent and
freezing point depression. At freezing point temperature, liq- TM2is the solvent melting point. T is the temperature at which
uid and solid phases are in equilibrium and Eq. (6.206) ap- solid and liquid phases are in equilibrium and is the same
plies. If the solution is assumed ideal, Eq. (6.206) can be as the freezing point of solution after addition of solute. The
written for the solvent (component 2) in the following form amount of decrease in freezing point is shown by ATe2, which
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 283
,%
K= Y/x K = Y/x
0.0
10 2 30 4 50 6 7 8 9 1 0 0 2 300 45006 7891,000 2 3,000 4 6 78910,000
is equal to (TM2 - T). The amount of solute in the solution is approximate but it is quite useful for calculation of freezing
x~ (= 1 - x2), which is very small (xl (< 1), and from a math- point depression for hydrocarbon systems. For nonideal so-
ematical approximation we have lutions, Eq. (6.211) must be used by replacing x2 with y2x2,
(6.212) In(1 - xl) ~ - x l where )'2 is a function of x2.
By similar analysis from VLE relation for ideal solutions it
Therefore, Eq. (6.211) can be solved to find ATM2 [43]: can be shown that boiling point elevation may be estimated
(6.213) ATM2 -----x1RT22 from the following simplified relation [43]:
In deriving this equation since the change in the freezing point (6.214) ATb2 ~ xlRT~2
is small, TTM2is approximated by T~2. Equation (6.213) is AH~ p
284 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
L Freezing l i n e ~
TM
~ ~T~xSMeltingline I~ Meltingline
TM2
0 1.0 0 1.0
X~1, yl S x s] , yl s
(6.226) c2 = (cnS)2 -
f (o2r (os dr
--M J \-O-~ ]s k OTJv By converting the HS EOS, Eq. (5.93), into the above virial
r form one gets [51]
cHs can be calculated from Eq. (6.216) using the CS EOS, 2 NAt73
Eq. (5.93). Derivative (SS/ST)v can be determined from
BHS= ~7r
(6.231)
Eq. (6.218) or (6.219) as a function of T and V only. (02T/O V2)s
C ns = ~5z r 2 N~o
2 6
can be determined from Eq. (6.225) as a function of T and V.
Therefore, the RHS of above equation is in terms of only T
Substituting Eqs. (5.76) and (6.230) for real and hard-sphere
and V, which can be written as
fluids virial equations into Eq. (6.39) one can calculate en-
(6.227) c= c(T, V) tropy departures for real and hard-sphere fluids as
Equation (6.226) or (6.227) is a cVT relation and is called ve- [ ' S - ~ g'] I" P B+ TB' C + TC"~
locity of sound based equation of state [44]. One direct appli- (6.232) /--1 = - [ln + ~V--- q
L R Area/fluid k 2-V--~ ',]
cation of this equation is that when a set of experimental data [" S - S ig "] [ B Ms C IJs \
on cVT or cPT for a fluid or a fluid mixture of constant compo- _ _ =_ lnP +-- + q77~
sition are available they can be used with the above relations (6.233) [ R J hardsphere ~ V 2V )
to obtain the PVT relation of the fluid. This is the essence of Since it is assumed that the left sides of the above two equa-
use of velocity of sound in obtaining PVT relations. This is tions are equal, so the RHSs must also be equal, which result
demonstrated in the next section by use of velocity of sound in the following relations:
data to obtain EOS parameters. Once the PVT relation for a
fluid is determined all other thermodynamic properties can TB' + B = B Hs
(6.234)
be calculated from various methods presented in this chapter. TC' + C = C HS
288 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 6.1$--Constants in Eq. (6.239) for calculation of second 0.6 ~/Dymond& SmirchData
virial coefficient. /
Compound a B c %AADfor Z
Methane 0.02854 19.4 1.6582 0.5 .....2 o s6o
Ethane 0.16 250 0.88 1.1 Temper~dure,K
Propane 0.22 230 1.29 1.4
Taken with permissionfromRef. [44]. FIG. 6.36--Prediction of Z factor of methane at 30 bar from
Number of data points for each compound:150;pressure range: 0.1-200bar; truncated virial EOS with second coefficient from velocity of
temperature range: 90-500K for C1, 90-600K for C2, and 90-700K for Ca. sound data. Taken with permission from Ref. [44].
6. T H E R M O D Y N A M I C R E L A T I O N S F O R P R O P E R T Y E S T I M A T I O N S 289
1.05
/••"
0.96.
o9,1 Original Parameter
:~ o,0"9
ast
=
~-~ Expedment~l
0.75,
i o.8 ~ 0.7'
0.751 0.65"
8 0.r~ ~/Experimental . . . . . . . . . . . . 1,4 0.6.
0.65 ~ ~o ~o ~ ~o ~ ~o Too
o,~ 2OO 25O 3OO 350 400 45O 5O0
Temperature, K
Temperature, K FIG. 6.39--Prediction of Z factor of propane at 30 bar from LJ
EOS with parameters from different methods. Taken with per-
FIG. 6.37--Prediction of Z factor for methane at 30 bar from
mission from Ref. [44].
vdW EOS using parameters from velocity of sound data. Taken
with permission from Ref. [44].
6. 9.2.3 R K a n d P R E O S P a r a m e t e r s - - P r o p e r t y
in Table 6.17. E r r o r s for c a l c u l a t e d Z v a l u e s w i t h u s e of L J Estimation
p a r a m e t e r s f r o m d i f f e r e n t m e t h o d s are also g i v e n in this ta-
To f u r t h e r i n v e s t i g a t e t h e p o s s i b i l i t y o f u s i n g v e l o c i t y o f s o u n d
ble. Van d e r W a a l s E O S p a r a m e t e r s d e t e r m i n e d f r o m v e l o c i t y for c a l c u l a t i o n o f PVT a n d t h e r m o d y n a m i c data, R K a n d P R
of s o u n d a r e g i v e n in Table 6.18 a n d p r e d i c t e d Z v a l u e s f o r
E O S p a r a m e t e r s w e r e d e t e r m i n e d for b o t h gases a n d l i q u i d s
m e t h a n e a n d e t h a n e a r e s h o w n in Figs. 6.37 a n d 6.38, re- t h r o u g h v e l o c i t y o f s o u n d data. U s i n g p a r a m e t e r s d e f i n e d in
spectively. P r e d i c t e d Z f a c t o r for p r o p a n e f r o m C S L J E O S
Table 6.1 for c a l c u l a t i o n of y, V, a n d (aP/aV)r a n d substi-
(Eq. 5.96) is s h o w n in Fig. 6.39. R e s u l t s p r e s e n t e d in Ta-
t u t i n g t h e m i n t o Eq. (6.216), v e l o c i t y of s o u n d , c, c a n be es-
bles 6.15-6.17 a n d Figs. 6.36-6.39 s h o w t h a t E O S p a r a m e -
t i m a t e d . F o r b o t h R K a n d S R K e q u a t i o n s t h e r e l a t i o n for c
ters d e t e r m i n e d f r o m v e l o c i t y of s o u n d p r o v i d e r e l i a b l e PVT
becomes
d a t a a n d m a y be u s e d to c a l c u l a t e o t h e r t h e r m o d y n a m i c
properties. 2
CRK,SRK = - - - ~
v2[ - R- T +
(V - b) 2
a(2V + b)
V2(V + b) 2
V .~-lq
08, ooo . -T V-b V2+bv)2x(C~-R---K-'nv-~J
al
J 9 Ta2 ,
.~ 0.9
(6.240)
"~cL 0.8
..,.,~"Otiginal Parameter a n d for P R E O S t h e r e l a t i o n for c b e c o m e s
E V2 [ RT 2a(V + b)
8 o.75i , " C2R = ---M (V
~ - b) + (V 2 + 2bV - b2) 2
r4 L
~ ~o 5~
Temperature, K
66o 6so 7o0 al
- T V b V 2 + 2bV - b 2
FIG. 6.38--Prediction of Z factor for ethane at 100 bar from
vdW EOS using parameters from velocity of sound data. Taken (6.241) •
with permission from Ref. [44].
TABLE 6.16---The Lennard-Jones parameters from the velocity of sound data and other sources.
Velocity of sound Second virial coefficienta Viscosity dataa
Compound e/ka, K o,A %AADfor Z e/k~,K a,A %AADfor Z e/kB, K cr, A %AADfor Z
Methane 178.1 3.97 0.8 148.2 3.817 4.0 144.0 3.796 4.7
Ethane 300.0 4.25 0.5 243.0 3.594 3.0 230.0 4.418 3.4
Propane 350.0 5.0 1.1 242.0 5.637 11.5 254.0 5.061 8.0
Taken with permission from Ref. [44].
aThe LJ parameters are used with Eq. (5.96) to calculate Z. The LJ parameters from the second virial coefficient and viscosity are taken from
Hirschfelder et al. [56]. kB is the Boltzman constant (1.381 x 10-23 J/K) and 1 A = 10-1~ m.
TABLE 6.17--The van der Waals constants from the velocity of sound data.
Velocity of sound Original constants a
Compounds a x 10-~ b %AADfor Z a x 10_6 b %AADfor Z
Methane 1.88583 44.78 1.0 2.27209 43.05 0.8
Ethane 3.84613 57.18 1.8 5.49447 51.98 2.4
Propane 8.34060 90.51 1.4 9.26734 90.51 1.5
Taken with permission from Ref. [44]. a is in cmT/mol2. bar and b is in cm3/mol.
=From Table 5.1.
290 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
where parameters al and a2 are first and second derivatives of T A B L E 6.18.--RK and PR EOS parameters (Eq. 6.242) from velocity
EOS parameter a with respect to temperature and for both RK of sound in gases and liquids.
and PR equations are given in Table 6.1. In terms of parame- Compound No. of RK EOS a PR EOS ~
(gas) points as fls as fls
ters a and b, velocity of sound equation for both RK and SRK
Methane 77 1.025 1.111 0.936 1.093
are the same. Their difference lies in calculation of parameter Ethane 119 1.043 1.123 0.956 0.993
a through Eq. (5.41), where for RK EOS, a = 1. Now we de- Propane 63 0.992 1.026 1.013 1.031
fine EOS parameters determined from cVT data in terms of Isobutane 80 0.993 1.019 0.983 0.983
original EOS parameters (aEOS and bEos as given in Table 5.1) n-Butane 86 0.987 1.015 0.941 0.912
n-Pentane (liquid) 1.04 0.9
in the following forms: n-Decane (liquid) 1.06 0.99
Taken with permission from Ref. [8].
(6.242) as = asaF.os ~ parameters must be used for gaseous phase with Eq. (5.40) and param-
bs = ~,b~os eters aEOS and baos from Tables 5.1.
I t
I o0
~ %F.xpetimen~l ...+'~
OriginalParameter
290 310 330 3~4) 3"10 390 410 430 450 4/0 490 510
Temperature, K
FIG. 6.40--Prediction of constant pressure heat capacity of
ethane gas at 30 bar from RK EOS using parameters from veloc-
ity o f s o u n d d a t a .
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 291
39
6 7 $ 9 IO
Pressure, MPa
320 330 340 350 360 37O
FIG. 6.41--Prediction of constant volume heat capacity Temperature, K
of 69 mol% methane and 31 mol% ethane gas at 260 K from
RK EOS using parameters from velocity of sound data. FIG. 6.43~Prediction of liquid heat capacity of n-octane at
100 bar from PR EOS using parameters from velocity of sound
data.
to calculate various physical properties errors very similar
to those obtained from original parameters are obtained as of sound for liquids can be estimated from original PR param-
shown in Table 6.19 [8]. Predicted constant pressure heat ca- eters with AAD of 9.7%; while using parameters calculated
pacity from RK EOS with parameters determined from veloc- from sonic velocity, an error of 3.9% was obtained for 569
ity of sound for ethane a gas mixture of m e t h a n e and ethane data points [8]. Graphical evaluations for prediction of liq-
is shown in Figs, 6.40 and 6.41, respectively. uid density of a mixture and constant volume heat capacity
Similarly sonic velocity data for some liquids from Cs to C t0 of n-octane are shown in Figs, 6.42 and 6.43. Results shown
were used to calculate EOS parameters. Calculated as and fls in these figures and Table 6,18 indicate that EOS parameters
parameters for use with PR EOS through Eq. (6.242) are also determined from velocity of sound are capable of predicting
given in Table 6.18. When EOS parameters from velocity of t h e r m o d y n a m i c properties. It should be noted that data on ve-
sound are used to calculate Cp of liquids ranging from C5 locity of sound were obtained either for c o m p o u n d s as gases
to C10 an average error of 6.4% is obtained in comparison or liquids but not for a single c o m p o u n d data on sonic veloc-
with 7.6% error obtained from original parameters. Velocity ity of both liquids and gases were available in this evaluation
IS
Experimen~ ....
17
15
,3
14
t3
12
tO 2O 30 40
Pressm,'e+ MP'=
FIG. 6.42--Prediction of liquid density of 10 mol% n-hexadecane
and 90 mol% carbon dioxide at 20~ from PR EOS using parameters
from velocity of sound data.
292 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
process. For this reason there is no continuity in use of pa- calculation of heat of fusion, molar volume, and solubility
rameters at and fls for use in both phases. For the same reason parameters for paraffinic, naphthenic, and aromatic groups.
when parameters obtained from gas sonic velocity were used These relations are useful in VLE and SLE calculations for
to calculate vapor pressure errors larger than original EOS petroleum fractions through the pseudocomponent method
parameters were obtained [8]. of Chapter 3. Data on the enthalpy of fusion and freezing
Research on using the velocity of sound to obtain thermo- pointd can be used to calculate freezing point of a mixture or
dynamic properties of fluids are underway, and as more data the temperature at which first solid particles begin to form.
on the speed of sound in heavy petroleum mixtures become Application of methods presented in this chapter require in-
available usefulness of this technique of calculating proper- put parameters (critical properties, molecular weight, and
ties of undefined and heavy mixtures becomes more clear. acentric factor) that for defined mixtures should be calcu-
From the analysis shown here, one may conclude that use lated from mixing rules given in Chapter 5. For undefined
of sonic velocity is a promising method for prediction and petroleum fractions these parameters should he calculated
calculation of thermodynamic properties of fluids and fluid from methods given in Chapters 2-4. Main application of
mixtures. methods presented in this chapter will be shown in the next
chapter for calculation of thermodynamic and physical prop-
erties of hydrocarbons and undefined petroleum fractions.
6.10 SUMMARY AND RECOMMENDATIONS The main characteristic of relations shown in this chapter
is that they can be used for prediction of properties of both
In this chapter fundamental thermodynamic relations that gases and liquids through an equation of state. However, as
are needed in calculation of various physical and thermody- it will be seen in the next chapter there are some empirically
namic properties are presented. Through these relations var- developed correlations that are mainly used for liquids with
ious properties can be calculated from knowledge of a PVT higher degree of accuracy. Generally properties of liquids are
relation or an equation of state. Methods of calculation of calculated with lesser accuracy than properties of gases.
vapor pressure, enthalpy, heat capacity, entropy, fugacity, ac- With the help of fundamental relations presented in this
tivity coefficient, and equilibrium ratios suitable for hydro- chapter a generalized cVT relation based on the velocity of
carbon systems and petroleum fractions are presented in this sound is developed. It has been shown that when EOS param-
chapter. These methods should be used in conjunction with eters are calculated through a measurable property such as
equations of states or generalized correlations presented in velocity of sound, thermophysical properties such as density,
Chapter 5. In use of cubic equations of state for phase equilib- enthalpy, heat capacity, and vapor pressure have been calcu-
rium calculations and calculation of Ki values, binary interac- lated with better accuracy for both liquid and vapor phases
tion parameters recommended in Chapter 5 should be used. through the use of velocity of sound data. This technique
Cubic equations are recommended for high-pressure phase is particularly useful for mixtures of unknown composition
equilibrium calculations while activity coefficient models are and reservoir fluids and it is a promising approach for esti-
recommended for low-pressure systems. Methods of calcu- mation of thermodynamic properties of complex undefined
lation of activity coefficient and Henry's law constants from mixtures.
a cubic EOS are presented. Recent studies show that cubic
equations are not the best type of PVT relation for prediction
of derivative properties such as enthalpy, Joule-Thomson co~
efficient, or heat capacity. For this reason noncubic equations 6.11 P R O B L E M S
such as statistical associating fluid theory (SAFT) are being
investigated for prediction of such properties [38, 57]. The 6.1. Develop an equation of state in terms of parameters fl
main purpose of this chapter was to demonstrate the role that and K.
theory plays in estimation of physical properties of petroleum 6.2. In storage of hydrocarbons in cylinders always a mix-
fluids. However, among the methods presented in this chapter, tures of both vapor and liquid (but not a single phase)
the LK generalized correlations are the most suitable meth- are stored. Can you justify this?
ods for calculation of enthalpy, heat capacity, and fugacity for 6.3. Derive a relation for calculation of (G - Gig)/RTin terms
both liquid and gas phases at elevated pressures. of PVT and then combine with Eq. (6.33) to derive
While the cubic equations (i.e., SRK or PR) are useful Eq. (6.50).
for phase behavior calculations, the LK corresponding state 6.4. Derive Edmister equation for acentric factor (Eq. 2.108)
correlations are recommended for calculation of density, from Eq. (6.101).
enthalpy, entropy, and heat capacity of hydrocarbons and 6.5. a. Derive a relation for molar enthalpy from PR EOS.
petroleum fractions. Partial molar properties and their meth- b. Use the result from part a to derive a relation for par-
ods of calculation have been presented for estimation of mix- tial molar enthalpy from PR EOS.
ture properties. Calculation of volume change due to mixing c. Repeat part a assuming parameter b is a temperature-
or heat of mixing is shown. Fundamental phase equilibria re- dependent parameter.
lations especially for vapor-liquid and solid-liquid systems 6.6. Derive a relation for partial molar volume from PR EOS
are developed. Through these relations calculation of vapor (Eq. 6.88).
pressure of pure substances, solubility of gases and solids 6.7. Derive fugacity coefficient relation from SRK EOS for
in liquids are demonstrated. Solubility parameters for pure a pure substance and compare it with results from
compounds are given for calculation of activity coefficients Eq. (6.126).
without use of any VLE data. Correlations are presented for 6.8. Derive Eq. (6.26) for the relation between Cv and Cv.
6. THERMODYNAMIC RELATIONS FOR PROPERTY ESTIMATIONS 2 9 3
6.9. Show that 6.18. a. For a gas mixture that follows truncated virial EOS,
show that
V
Cv-C~ = f
V=oo
l dV
L \ or2 ]vJr
VE =
1 I212y,
i j
G E = -~
PZ Z yiyj8ij
Use this relation with truncated virial equation to derive i j
Eq. (6.65). P
6.10. The Joule-Thomson coefficient is defined as j ~7
where &i is defined in Eq. (5.70) in Chapter 5.
b. Derive a relation for heat of mixing of a binary gas
that obeys truncated virial EOS.
6.19. In general for mixtures, equality of mixture fugacity be-
a. Show that it can be related to PVT in the following tween two phases is not valid in VLE calculations:
form:
fmV,x= f ~
8V) _ V However, only under a certain condition this relation is
TS-~p
true. What is that condition?
T1-- C~ 6.20. With the use of PR EOS and definition of solubility pa-
rameter (8) by Eq. (6.147) one can derive the following
b. Calculate 0 for methane at 320 K and 10 bar from the relation for calculation of 8 [17]:
SRK EOS.
6.1 I. Similar to derivation of Eq. (6.38) for enthalpy departure
8=
I ( ~ ) VL+(l+~/-i)b l
at T and V, derive the following relation for the heat
capacity departure and use it to calculate residual heat
capacity from RK EOS. How do you judge validity of where da/dT for PR EOS can be obtained from Table 6. I.
your result? With use of volume translation for VL estimate values
of VL and 8 at 25~ for hydrocarbons C5 and C10 and
OP 2 compare with values given in Table 6.10.
V
6.21. Calculate freezing point depression of toluene when it
R
(cp -Ce)r,v
ig = 7"V---~f oo \OT2]vdV is saturated with solid naphthalene at 20~
6.22. Derive Eq. (6.240) for calculation of velocity of sound
from RK/SRK EOS.
6.12. Show that Eqs. (6.50) and (6.51) for calculation of resid- 6.23. Consider the dry natural gas (fluid 1) and black oil (fluid
ual entropy are equivalent. 5) samples whose compositions are given in Table 1.2.
6.13. Prove Eq. (6.81) for the Gibbs-Duhem equation. Assume there are two reservoirs, one containing the nat-
6.14. Derive Eq. (6.126) for fugacity coefficient of i in a mix- ural gas and the other containing black oil, both at 400 K
ture using SRK EOS. and 300 bar. Calculate velocity of sound in these two flu-
6.15. Derive the following relation for calculation of fugacity ids using SRK EOS.
of pure solids at T <Ttp. 6.24. Calculate (U - uig), (H - Hig), and (S - Sig) for steam at
500~ and I00 bar from SRK and PR EOS. How do you
] 6.25.
evaluate your results?
Calculate the increase in enthalpy of n-pentane when
its pressure increases from 600 to 2000 psia at 190.6~
using the following methods:
where Pisub is the vapor pressure of pure solid i at tem- a. SRK EOS.
perature T. b. LK method.
6.16. Derive Eq. (6.216) for the velocity of sound. c. Compare the results with the measured value of 188
6.17. A mixture of C1 and C5 exists at 311 K and 69.5 bar in Btu/lbmol [17].
both gas and liquid phases in equilibrium in a closed 6.26. Estimate Cp, Cv, and the speed sound in liquid hexane
vessel. The mole fraction of C 1 in the mixture is Zl = at atmospheric pressure and 269 and 300K from the
0.541. In the gas Yl = 0.953 and in the liquid Xl = 0.33. following methods:
Calculate K~ and K5 from the following methods: a. SRK EOS
a. Regular solution theory b. PR EOS
b. Standing correlation c. Compare calculated sonic velocities with reported
c. GPA/NIST graphs values of 1200 (at 269 K) and 1071 m/s (at 300 K)
d. PR EOS (Fig. 3.33, Ref. [17]).
In using PR EOS, use shift parameters of -0.2044 and 6.27. Estimate Fi~ for the system of n-C4 and n-C32 at
-0.045 for C1 and n-C5, respectively. Also use BIP value 100~ from PR EOS and compare with the value from
of kl-5 = 0.054. Fig. 6.8.
294 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
6.28. Using results from Problem 6.27, estimate the Henry's c. Clausius-Clapeyron equation
law constant for the system of n-C4 and n-C32 at 100~ d. Compare the results with a reported value.
from PR EOS. 6.37. A natural gas is c o m p o s e d of 85% methane, 10% ethane,
6.29. For the mixture of benzene and toluene construct the and 5% propane. What are the mole fractions of each of
freezing/meking diagram similar to Fig. 6.32. the components in water (gas solubility) at 300 bar and
6.30. One of the advanced liquid solution theories is k n o w n as 298 K.
the quasichemical approximation, which is particularly 6.38. Ninety barrels of n-C36 are diluted with addition of 10
useful for mixtures containing molecules quantitatively bbl of n-C5 at 25~ Calculate volume of the solution at
different in size and shape. According to this theory the 1 bar from the following methods:
m o l a r excess Gibbs energy for a binary system is given a. Using partial molar volume from PR EOS.
by [21]: b. Using API procedure.
6.39. Show that if Edmister equation (Eq. 2.108) is used
GE (w) [ l(2w) 1 for acentric factor, Wilson correlation for Ki-values
RT- ~ XlX2 1 - ~ ~ XlX2-~-''"
(Eq. 6.204) reduces to H o f f m a n type correlation (Eq.
where the higher terms are neglected. Xl and x2 are mole 6.202).
fractions of 1 and 2 and k is the Boltzman's constant. W 6.40. Solubility of water in a gasoline sample at 1 atm can
and z are model parameters that must be determined for be determined approximately by Eq. (6.195). However,
each system. W/zkT is less than unity and z is called co- accurate solubility of water can be estimated through
ordination number and varies from 6 to 12 [21]. Typical a t h e r m o d y n a m i c model with activity coefficient calcu-
value of z is 10. Use Eq. (6.137) to derive the relations lations. A gasoline from California oil has mid boiling
for F1 and Y2. point of 404~ and API gravity of 43.5 with PNA compo-
6.31. Show that at constant T and P, the Gibbs-Duhem equa- sition of 30.9, 64.3, and 4.8% as reported by Lenoir and
tion in a m u l t i c o m p o n e n t mixture can be written in the Hipkin [ 12]. Estimate solubility of water in this gasoline
following forms: sample at 100~ and 1 atm from appropriate thermody-
namic model and compare the predicted value with the
}--~x4dln y~ = 0
value estimated from Eq. (6.195).
i
or
~x~dln]~ = 0 REFERENCES
i
6.32. Consider a binary solution of components 1 and 2. Show [1] Smith, J. M., Van Ness, H. C., and Abbott, M. M., Introduction
that in the region that Raoult's law is valid for compo- to Chemical Engineering Thermodynamics, 5th ed.,
nent I, Henry's law m u s t be valid for c o m p o n e n t 2. McGraw-Hill, New York, 1996.
6.33. In a binary liquid, mixtures of 1 and 2, fugacity of com- [2] Elliott, J. R. and Lira, C. T., Introductory Chemical Engineering
ponent 1 at 20~ can be approximately presented by the Thermodynamics, Prentice Hall, New Jersey, 1999
equation: (www.phptr.com).
[3] Daubert, T. E., Danner, R. E, Sibul, H. M., and Stebbins, C. C.,
fl~ = 30xl - 2 0 x ~ Physical and Thermodynamic Properties of Pure Compounds:
Data Compilation, DIPPR-AIChE, Taylor & Francis, Bristol, PA,
where fl% is the fugacity of 1 in the mixture in bar. At 1994 (extant) (www.aiche.org/dippr). Updated reference:
20~ and 30 bar determine: Rowley, R. L., Wilding, W. V., Oscarson, J. L., Zundel, N. A.,
a. The fugacity of pure c o m p o n e n t 1, fL. Marshall, T. L., Daubert, T. E., and Danner, R. E, DIPPR Data
b. The fugacity coefficient of pure c o m p o n e n t 1, qh. Compilation of Pure Compound Properties, Design Institute for
c. The Henry's law constant for c o m p o n e n t 1, kl. Physical Properties (DIPPR), Taylor & Francis, New York, 2002
d. Relation for the activity coefficient yl in terms of xl (http://dippr.byu.edu).
(based on the standard state of Lewis rule). [4] Poling, B. E., Prausnitz, J. M., and O'Connell, J. E, Properties of
Gases and Liquids, 5th ed., Mc-Graw Hill, New York, 2000.
e. Relation for f2
~.
[5] Daubert, T. E. and Danner, R. E, Eds., API Technical Data
6.34. Consider three h y d r o c a r b o n components benzene, cy- Book~Petroleum Refining, 6th ed., American Petroleum
clohexane, and n-hexane all having six carbon atoms. Institute (API), Washington, DC, 1997.
Both quantitatively and qualitatively state that solubility [6] Garvin, J., "Use the Correct Constant-Volume Specific Heat,"
of benzene in methylcylopentane is higher or benzene Chemical Engineering Progress, Vol. 98, No. 7, 2002, pp. 64-65.
in n-hexane. [7] Polt, A., Platzer, B., and Maurer, G., "Parameter der
6.35. A solution is made at the temperature of 298 K by adding thermischen Zustandsgl eichung von Bender fuer 14
5 g of naphthalene to a mixture of 50 g benzene and 50 g mehratomige reine Stoffe," Chemische Technik (Leipzig),
n-heptane. The temperature is gradually lowered until Vol. 44, No. 6, 1982, pp. 216-224.
the particles of solid are observed. What is the temper- [8] Shabani, M. R., Riazi, M. R., and Shaban, H. I., "Use of Velocity
of Sound in Predicting Thermodynamic Properties from Cubic
ature at this point? What is the temperature if 10 g of
Equations of State," Canadian Journal of Chemical Engineering,
naphthalene is added? Vol. 76, 1998, pp. 281-289.
6.36. Estimate vapor pressure of isobutane at 50~ from the [9] Lee, B. I. and Kesler, M. G., "A Generalized Thermodynamic
following methods: Correlation Based on Three-Parameter Corresponding States,"
a. a R K EOS. American Institute of Chemical Engineers Journal, Vol. 21, 1975,
b. PR EOS. pp. 510-527.
6. T H E R M O D Y N A M I C R E L A T I O N S F O R P R O P E R T Y E S T I M A T I O N S 295
[10] Alberty, R. A. and Silbey, R. J., Physical Chemistry, 2nd ed., Characterization Data," Journal of Petroleum Science and
Wiley, New York, 1999. Engineering, Vol. 22, 1999, pp. 67-93.
[11] Reid, R. C., Prausnitz, J. M., and Poling, B. E., Properties of [33] P16cker, U., Knapp, H., and Prausnitz, J. M., "Calculation of
Gases and Liquids, 4th ed., Mc-Graw Hill, New York, 1987. High-Pressure Vapor-Liquid Equilibria from a Corresponding
[12] Lenoir, J. M. and Hipkin, H. G., "Measured Enthalpies of Eight States Correlation with Emphasis on Asymmetric Mixtures,"
Hydrocarbon Fractions," Journal of Chemical and Engineering Industrial and Engineering Chemistry, Process Design and
Data, Vol. 18, No. 2, 1973, pp. 195-202. Development, Vol. 17, No. 13, 1978, pp. 324-332.
[13] AIChE DIPPR| Database, Design Institute for Physical Property [34] Chao, K. C. and Seader, J. D., "A General Correlation of
Data (DIPPR), EPCON International, Houston, TX, 1996 (for Vapor-Liquid Equilibria in Hydrocarbon Mixtures," American
more updated database from NIST see: http://www.nist.gov/ Institute of Chemical Engineers Journal, Vol. 7, 1961,
srd/fluids.htm and http://webbook.nist.gov/chemistry/fluid/). pp. 598-605.
[14] Wagner, W. and Pruss, A., "The IAPWS Formulation for the [35] Grayson, H. G. and Streed, C. W., "Vapor-Liquid Equlibria for
Thermodynamic Properties of Ordinary Water Substance for High Temperature, High Pressure Hydrogen-Hydrocarbon
General and Scientific Use," Journal of Physical and Chemical Systems," Proceedings of the 6th World Petroleum Congress,
Reference Data, Vol. 31, No. 2, 2002, pp. 387-535 (also see: Frankfurt/Main III, Paper 20-PD7, 1963, pp. 233-245.
http://www.iapws.org/). [36] PRO/II, Keyword Manual, Simulation Sciences Inc., Fullerton,
[15] Kesler, M. G. and Lee, B. I., "Improve Prediction of Enthalpy of CA, October 1992.
Fractions," Hydrocarbon Processing, Vol. 55, 1976, [37] Tsonopoulos, C. and Riazi, M. R., Physical Properties for
pp. 153-158. Engineers, Industrial Course Note for CHE7, OCCD, Kuwait
[16] Tsonopoulos, C., Heidman, J. L., and Hwang, S.-C., University, Kuwait, March 2002.
Thermodynamic and Transport Properties of Coal Liquids, An [38] Li, X.-S. and Englezos, P., "Vapor-Liquid Equilibrium of
Exxon Monograph, Wiley, New York, 1986. Systems Containing Alcohols Using the Statistical Associating
[17] Firoozahadi, A., Thermodynamics of Hydrocarbon Reservoirs, Fluid Theory Equation of State," Industrial and Engineering
Mc-Graw Hill, New York, 1999. Chemistry Research, Vol. 42, 2003, pp. 4953-4961.
[18] Riazi, M. R. and Daubert, T. E., "Simplify Property Predictions," [39] Hoffman, A. E., Crump, J. S., and Hocott, C. R., "Equilibrium
Hydrocarbon Processing, Vol. 59, No, 3, 1980, pp.115-116. Constants for a Gas Condensate System," Transactions of
[19] Denbigh, K., The Principles of Chemical Equilibrium, AIME, Vol. 198, 1953, pp. 1-10.
Cambridge University Press, Cambridge, UK, 1992. [40] Standing, M. B., "A Set of Equations for Computing
[20] Danesh, A., PVT and Phase Behavior of Petroleum Reservoir Equilibrium Ratios of a Crude Oil/Natural Gas System at
Fluids, Elsevier, Amsterdam, 1998. Pressures Below 1000 psia," Journal of Petroleum Technology,
[21] Prausnitz, J. M., Lichtenthaler, R. N., and de Azevedo, E. G., Vol. 31, 1979, pp. 1193-1195.
Molecular Thermodynamics of Fluid-Phase Equilibria, [41] Katz, D. L. and Hachmuth, K. H., "Vaporization Equilibrium
Prentice-Hall, New Jersey, 2nd ed., 1986, 3rd ed., 1999. Constants in a Crude-Oil-Natural-Gas System," Industrial
[22] Sandler, S. I., Chemical and Engineering Thermodynamics, 3rd and Engineering Chemistry Research, Vol. 29, 1937, pp. 1072-
ed., Wiley, New York, 1999. 1077.
[23] Riazi, M. R. and Vera, J. H., "Method to Calculate the Solubility [42] Glaso, O. S., "The Accuracy of PVT Parameters Calculated from
of Light Gases in Petroleum and Coal Liquid Fractions Based Computer Flash Separation at Pressures Less than 1000 psia,"
on their P/N/A Composition," Industrial and Engineering Journal of Petroleum Technology, Vol. 34, 1982, pp. 1811-1813.
Chemistry Research, Vol. 44, 2005. [43] Winnick, J., Chemical Engineering Thermodynamics, Wiley, New
[24] Won, K. W., "Thermodynamics for Solid Solution-Liquid- York, 1997.
Vapor Equilibria: Wax Phase Formation from Heavy [44] Riazi, M. R. and Mansoori, G. A., "Use of the Velocity of Sound
Hydrocarbon Mixtures," Fluid Phase Equilibria, Vol. 30, in Predicting the PVT Relations," Fluid Phase Equilibria,
1986, pp. 265-279. Vol. 90, 1993, pp. 251-264.
[25] Pan, H., Firoozabadi, A., and Fotland, E, "Pressure and [45] Riazi, M. R. and Roomi, Y. A., "Use of Velocity of Sound in
Composition Effect on Wax Precipitation: Experimental Data Estimating Thermodynamic Properties of Petroleum
and Model Results," Society of Petroleum Engineers (SPE) Fractions," Preprints of Division of Petroleum Chemistry
Production and Facilities, Vol. 12, No. 4, 1993, pp. 250-258. Symposia, American Chemical Society (ACS), Vol. 45, No. 4,
[26] Lira-Galeana, C. A., Firoozabadi, A., and Prausnitz, J. M., 2000, pp. 661-664.
"Thermodynamics of Wax Precipitation in Petroleum [46] Colgate, S. O., Silvarman, A., and Desjupa, C., "Acoustic
Mixtures," American Institute of Chemical Engineers Journal, Resonance Determination of Sonic Speed and the Critical
Vol. 42, 1996, pp. 239-248. Point," Paper presented at the AIChE Spring Meeting, Paper
[27] Pedersen, K. S., Skovborg, E, and Ronningsen, H. E, "Wax No. 29c, New Orleans, 1992.
Precipitation from North Sea Crude Oils, 4: Thermodynamic [47] Colgate, S. O., Silvarman, A., and Desjupa, C., "Sonic Speed
Modeling," Energy and Fuels, Vol. 5, No. 6, 1991, and Critical Point Measurements in Ethane by the Acoustic
pp. 924-932. Resonance Method," Fluid Phase Equilibria, Vol. 76, 1992,
[28] Engineering Data Book, 10th ed., Volume II, Sections 17-26, pp. 175-185.
Gas Processors (Suppliers) Association, Tulsa, OK, 1987. [48] Ball, S. J., Goodwin, A. R. H., and Trusler, J. P. M., "Phase
[29] National Institute of Standards and Technology (NIST), Behavior and Physical Properties of Petroleum Reservoir
http://www.nist.gov/srd/fluids.htm and Fluids from Acoustic Measurements," Journal of Petroleum
ht tp://webbook.nist.gov/chemistry/fluid/ Science and Engineering, Vol. 34, No. 1-4, 2002, pp. 1-11.
[30] Baker, L. E., Pierce, A. C., and Luks, K. D., "Gibbs Energy [49] Mittal Enterprises, New Delhi, India, 2004
Analysis of Phase Equilibria," Society of Petroleum Engineers (http://www.mittalenterprises.com).
Journal (Trans. AIME), Vol. 273, 1982, pp. 731-742. [50] Alem, A. H. and Mansoori, G. A., "The VIM Theory of
[31] Whitson, C. H. and Brule, M. R., Phase Behavior, Monograph Molecular Thermodynamics," American Institute of Chemical
Vol. 20, Society of Petroleum Engineers, Richardson, TX, 2000. Engineers Journal, VoL 30, 1984, pp. 468-480.
[32] Manafi, H., Mansoori, G. A., and Ghotbi, S., "Phase Behavior [51] Walas, S. M., Phase Equilibria in Chemical Engineering,
Prediction of Petroleum Fluids with Minimum Butterworth, London, 1985.
296 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
[52] Goodwin, R. D., The Thermophysical Properties of Methane from [55] Mossaad, E. and Eubank, E T., "New Method for Conversion
90 to 500 K at Pressures to 700 Bar, NBS Technical Note 653, of Gas Sonic Velocities to Density Second Virial Coefficients,"
National Bureau of Standards, April 1974. American Institute of Chemical Engineers Journal, Vol. 47,
[53] Goodwin, R. D., Roder, H. M., and Starry, G. C., The No. 1, 2001, pp. 222-229.
Thermophysical Properties of Ethane from 90 to 600 K at [56] Hirschfelder, O. J., Curtiss, C. E, and Bird, R. B., Molecular
Pressures to 700 Bar, NBS Technical Note 684, National Bureau Theory of Gases and Liquids, Wiley, New York, 1964.
of Standards, August 1976. [57] Konttorp, K., Peters, C. J., and O'Connell, J. P., "Derivative
[54] Goodwin, R. D. and Hayens, W. M., The Thermophysical Properties from Model Equations of State," Paper presented at
Properties of Propane from 85 to 700 K at Pressures to 70 MPa, the Ninth International Conference on Properties and Phase
NBS Monograph 170, National Bureau of Standards, April Equilibria for Product and Process Design (PPEPPD 2001),
1982. Kurashiki, Japan, May 20-25, 2001.
MNL50-EB/Jan. 2005
Applications: Estimation of
Thermophysical Properties
297
vap Change in value of a property due to vaporiza- stb Stock tank barrel (unit for volume of liquid oil
tion at 1 atm and 60~
S Value of a property at solid phase TVP True vapor pressure
sat Value of a property at saturation pressure VABP Volume average boiling point defined by Eq.
sub Value of a property at sublimation pressure (3.3)
[](0) A dimensionless term in a generalized correla- %AAD Average absolute deviation percentage defined
tion for a property of simple fluids by Eq. (2.135)
[](1) A dimensionless term in a generalized correla- %AD Absolute deviation percentage defined by Eq.
tion for a property of acentric fluids (2.134)
o Value of a property at low pressure (ideal gas wt% Weight percent
state) condition at a given temperature
THE LASTTHREE CHAPTERSof this book deal with application
Subscripts of methods presented in previous chapters to estimate var-
ious thermodynamic, physical, and transport properties of
A Value of a property for component A petroleum fractions. In this chapter, various methods for pre-
B Value of a property for component B diction of physical and thermodynamic properties of pure
b Value of a property at the normal boiling point hydrocarbons and their mixtures, petroleum fractions, crude
C Value of a property at the critical point oils, natural gases, and reservoir fluids are presented. As it was
i,j Value of a property for component i or j in a discussed in Chapters 5 and 6, properties of gases may be esti-
mixture mated more accurately than properties of liquids. Theoretical
L Value of a property for liquid phase methods of Chapters 5 and 6 for estimation of thermophysical
m Molar property (quantity per unit mote) properties generally can be applied to both liquids and gases;
m Mixture property however, more accurate properties can be predicted through
mix Value of a property for a mixture empirical correlations particularly developed for liquids.
nbp Value of a liquid phase property at the normal When these correlations are developed with some theoretical
boiling point of a substance basis, they are more accurate and have wider range of appli-
pc Pseudocritical property cations. In this chapter some of these semitheoretical corre-
r Reduced property lations are presented. Methods presented in Chapters 5 and 6
ref Value of a property at the reference state can be used to estimate properties such as density, enthalpy,
S Value of a property at the solid phase heat capacity, heat of vaporization, and vapor pressure.
S Value of a property for solvent (LMP) Characterization methods of Chapters 2-4 are used to de-
s Specific property (quantity per unit mass) termine the input parameters needed for various predictive
T Values of property at temperature T methods. One important part of this chapter is prediction of
tp Value of a property at the triple point vapor pressure that is needed for vapor-liquid equilibrium
W Values of a property for water calculations of Chapter 9.
20 Values of property at 20~
7+ Values of a property for C7+ fraction of an oil
can be considered as defined mixtures. Examples of unde- number of components in the mixture is large since each
fined mixtures are petroleum fractions and reservoir fluids property must be calculated for each component in the mix-
whose compositions are not known. For such mixtures, some ture. In applying a mixing rule, the role of binary interaction
bulk properties are usually known. parameters (BIPs) is important when the mixture contains
Theoretically developed methods are generally more ac- components of different size and structure. For example, in a
curate for gases than for liquids. Kinetic theory provides reservoir fluid containing C1 and a heavy component such as
sound predictive methods for physical properties of ideal C30 the role of BIP between these two components cannot be
gases [2, 3]. For this reason, empirical correlations for calcu- ignored. Similarly when nonhydrocarbon components such
lation of physical properties of liquids have been proposed. as H2S, N:, H20, and CO2 exist in the mixture, the BIPs of
Similarly, theoretical methods provide a more accurate es- these compounds with hydrocarbons must be considered. For
timation of physical properties of pure compounds than of some empirically developed correlations specific interaction
their mixtures. This is mainly due to the complexity of inter- parameters are recommended that should be used.
action of components in the mixtures especially in the liquid Theoretically developed thermodynamic relations of Chap-
phase. For undefined mixtures such as petroleum fractions, ters 5 and 6 give thermodynamic properties in molar quan-
properties can be calculated in three ways. One method is to tifies. They should be converted into specific properties by
consider them as a single pseudocomponent and to use the using Eq. (5.3) and molecular weight. In cases that no spe-
methods developed for pure components. The second method cific mixing rule is available for a specific property the simple
is to develop empirical correlations for petroleum fractions. Kay's mixing rule (Section 3.4.1) may be used to calculate
Such empirically developed methods usually have limited ap- mixture properties from pure component properties at the
plications and should be used with caution. They are accurate same conditions of T and P. If molar properties for all com-
for those data for which correlation coefficients have been ponents (0r~) are known, the mixture molar property (0m)
obtained but may not provide reliable values for properties may be calculated as
of other fractions. These two approaches cannot be applied
to mixtures with wide boiling range, such as wide fractions, (7.1) 0m = ZXmiOmi
crude oils, or reservoir fluids. The third approach is used i
However, one important limitation in use of such correlations book. Equation (7.3) is valid for both liquids and gases once
is boiling point range, carbon number, or molecular weight their Z values are calculated from an equation of state or a
of fractions and compounds used in the development of generalized correlation. If Z is known for all components in
the correlations. For example, correlations that are based a mixture, then Zm can be calculated from Eq. (7.1) and Pm
on properties of petroleum fractions and pure components from Eq. (7.3). Specific methods and recommendations for
with carbon number range of C5-C20 cannot be used for calculation of density of gases and liquids are given in the
estimation of properties of light gases (natural gases or LPG), following sections.
heavy residues, or crude oils. Another limitation of empir-
ically developed correlations is the method of calculation
7.2.1 Density of Gases
of input parameters, For example, a generalized correlation
developed for properties of heavy fractions requires critical Generally both equations of state and the Lee-Kesler gener-
properties as input parameter. For such correlations the same alized correlation (Section 5.7) provide reliable prediction of
method of estimation of input parameters as the one used gaseous densities. For high-pressure gases, cubic EOS such
in the development of correlation should be used. The most as PR or SRK EOS give acceptable values of density for both
reliable correlations are those that have some theoretical pure and mixtures and no volume translation (Section 5.5.3)
background, but the coefficients have been determined is needed. For practical calculations, properties of gases can
empirically from data on petroleum fractions as well as pure be calculated from simple equations of state. For example,
compounds. One technique that is often used in recent years Press [7] has shown that the original simple two-parameter
to develop correlations for physical properties of both pure Redlich-Kowng equation of state (RK EOS) gives reasonably
compounds and complex mixtures is the artificial neural acceptable results for predicting gas compressibility factors
network method [4-6]. These methods are called neural net- needed for calculation of valve sizes. For moderate pressures
works methods because artificial neural networks mimic the truncated virial equation (Eq. 5.76) can be used with coeffi-
behavior of biological neurons. Although neural nets can be cients (B and C) calculated from Eqs. (5.71) and (5.78). For
used to correlate data accurately and to identify correlative low-pressure gases (<5 bar), virial equation truncated after
patterns between input and target values and the impact of the second term (Eq. 5.75) with predicted second virial coef-
each input parameter on the correlation, they lack necessary ficient from Tsonopoulos correlation (Eq. 5.71) is sufficient
theoretical basis needed in physical property predictions. to predict gas densities. For light hydrocarbons and natural
The resulting correlations from neural nets are complex and gases, the Hall-Yarborough correlation (Eq. 5.102) gives a
involve a large number of coefficients. For this reason corre- good estimate of density. For defined gas mixtures the mixing
lations are inconvenient for practical applications and they rule may be applied to the input parameters (Tc, Pc, and o)) and
have very limited power of extrapolation outside the ranges the mixture Z value can be directly calculated from an EOS.
of data used in their developments. However, the neural For undefined natural gases, the input parameters may be
net model can be used to identify correlating parameters in calculated from gas-specific gravity using correlations given
order to simplify theoretically developed correlations. in Chapter 4 (see Section 4.2).
LO
~9 0,9
0,8 0~
0.7
06
0.4
T~em~ra~re."F
10 I i t
"Tp
9 1 tl 11 I1 ....... I
!1 II III1 llIll I fill IIIII I II1t 11 I I 1 Ilfl
!1 li I ~ ' I I11 II Itl f-1 I11 [1111
11 I I I-I I I l []-t [I ! 11 I I1 II '1"..1.1 II 1 H:ltl II
lill 11 Ill I !I1!f I t111 Itlll I I!It It II I II
1%11 III t l..l 1~ ~ Itl] I 1I !I1 I lill I I II i I !
9 1 r%l I Ill I II!11 t l 1.11 lllll LI It I It !1 I Iill III
llli I ! I 1 I I !]1 t 1 17~i1 I I I I Vl !1 fllll I fill III
! 1 lid I i I I I I il 1 t I I II I I I [ I l I I 1 ] l] ! ] 1 1 II l:lll
il Ililll I I11t t t I II I 111f} [ till tl II I I II I Ill
=ltI lIMI] II ll]l I]lll I I I1 i t I [ I I I I I I I ! 1 1 I l l
9 I l l d _ l J l l l I I I i 1 i t 1 l . l 11 11 I t t 1 t i I I I I I1 l ] ! I I I [ X l
lliIllill l I t 1,~1 t 1 1 1 1 1 1 1 [ 1 l I I1 l II il 1 I I11111
lr%l] llll I I11tt itl]l IIIII I 1 !11 I1 II I I I! f Ill
8 ~ ] _ l _ 1 2 i t till lllll 1 i ili 11 ill .....| . . . i l l Ill
, ~ r i 1 f I I i 11 I I 1 11 ill 1 !11 Ifi
iiI Pd ! I r % ~ ; l q I 1 t I11 t 1 !I Ill ,, i,lltl'
[]li/ll-ll llii I 1111111 lit t 1 II I tl it I ilf lli
rlilllll II1tl II!1t illll I11f111 II IIII ill
llll Ilillll I 1 M t I I t !1 f I I I I l I I l i 1 1 1 !1 I I11 II!
l ~ Ilkt 11 [ il L I l !i 1 II I i 1 l l I i
~ 1 I ill t t II Pl[ ll]"~ t [I !1 I II t ill
~ F 1 K ~ J I % I I l i t t illlI ~ f!I111 !1 l II 1 III
!I%111111t'~]!t%1 ~ II t I111t i I lit iJ it I IfI ili
5liJ__liI3t_l_]Tl'~l]i t~t II 111 itl 11 I I 1 1 1 It I NI Ill
~ ' k ~ L l l I lil I i 11 I l l 1 ~ IJ I I I 111 I i I 1111
i - I K I T N - [ I " I - I TM ~ t . _ ~ k L L J _ L ~ L [ ] I I i 1 t I 1 ! 1 1 1 1 I 1 1 1 I I I f 1I
~ K I - P t L ~ T 11 i l t I t I111 I !11 I I I I III
~ ~ t - :~' !1 i I t I I I I I II I I ! I I I II
4 ~ ~ R ~ ~ ~ I I t t !11 I t !1 t I ! I 11 l l
E
ea
~ ) ~ . 1 ill r%LI I I I%-%" 1 i I ! I i "[ I Jl l I I I t l I I
i~j__~j_~:~ii_l_~i~l_~kl 1~tl 111~l:~/el ] !I I /I !1 t I I t I lli
g4-4-3d~2~,~.ll~ i ~t li i~ ~ i lii ii =ii illz:]:l,
i'7-I-]KI-I~Lr.I%II~II I%i1%1111F~kl Ill II II I I I11 Ill
. Ikl I I~=l I~I~i~%t I fNtl I~1 il I I%Jl I I I !1 t I I ! I1 ill
~ ~ " l ~ ~I I~ f ~ t i .I t ~ ~ . k i ! !1 i I 1 II t II!
~ ~ . . ~ i i~ t~i I]~d 1 gr~LI I II ll~l~dll ] ti Ill Ill
~ ~ ] ~ I ! N:l~Ikl t 1~1~11 r " - ~ l i l t I I ~ ! f I Ill
ii t. I~ i I l-~ i ~:i ~%l~il l-li~l i T~ I i I !1 I I i ~ I t i
! I t I I i%~iA ~ I ~ k I P*t~A l-_'~l~LI 1 " ~ t 4 J 1F - - ~ ~
I I I ! 11%l 1 ! 7%~1 1--11~Cid l'~llil~Ti-k~.JL i I ]~'-]-4~.lJ 1 I i ~ I 1!
~ . 1 1 1 ! ~_1111~111~7 il~l_ F'~l~d._l I I / ~ , _ "11 II!
~ - I L ~ I I I~1 I t lr~d~l r'l~k..L.~i ]-l~'~i[- ~ ! I I I 1- I 1 1
0 Ill ! | - 9 9 II 9 ! 9
25 30 35 40 45 50 55 60 65
Density at NY"F ancf14.7 ~ . IbPercu I1
FIG. 7.2--Effect of pressure on the liquid density at 60~ (15.5~ Unit conver-
sion: p [g/cm 3] = p [Ib/ft3]/62.4; ~ = (~ x 1.8 + 32; psia = bar x 14.504. Taken
with permission from Ref. [8].
where/9mi x is the m i x t u r e liquid density (i.e., g/cm 3) a n d Xwi (7.5) dT = dro - 10 -3 x (2.34 - 1.9dT) x (T - To)
is weight fraction of c o m p o n e n t i in the liquid mixture, pi
s h o u l d be k n o w n f r o m database, experiment, or m a y be cal- where b o t h T a n d To are in kelvin or in ~ a n d dr a n d dTo
c u l a t e d from Rackett equation. E q u a t i o n (7.4) can be a p p l i e d are in g/cm 3. This m e t h o d provides reliable d e n s i t y values at
to specific gravity b u t n o t to m o l a r density. This e q u a t i o n is t e m p e r a t u r e s n e a r the reference t e m p e r a t u r e at w h i c h d e n s i t y
p a r t i c u l a r l y useful for calculation of specific gravity a n d den- is k n o w n (when T is n e a r To). W h e n the a c t u a l t e m p e r a t u r e is
sity of c r u d e oils with k n o w n c o m p o s i t i o n at a t m o s p h e r i c far from the reference t e m p e r a t u r e , this e q u a t i o n s h o u l d b e
pressure, as it is s h o w n later in this chapter. used with caution.
W h e n only a m i n i m u m of one d a t a p o i n t for liquid d e n s i t y F o r quick density calculations several g r a p h i c a l m e t h o d s
of a p e t r o l e u m fraction at a t m o s p h e r i c p r e s s u r e is k n o w n (i.e., have been developed, w h i c h are less a c c u r a t e t h a n the meth-
SG, d20, o r d25), t h e n Eq. (2.1 10) m a y be used to calculate liq- ods outlined above. G r a p h i c a l m e t h o d s for liquid density cal-
u i d density at a t m o s p h e r i c p r e s s u r e a n d o t h e r t e m p e r a t u r e s . culations r e c o m m e n d e d b y GPA [8] are s h o w n in Figs. 7.1-7.3.
7. APPLICATIONS: E S T I M A T I O N OF THERMOPHYSICAL PROPERTIES 303
30 35 40 ~5 ~0 55 60
8
E
30 35 40 45 50 55 6O
C,r at 60~Fand P~essute, Ib p~ c~ (I
Figure 7.1 gives effect of temperature on the liquid specific figure may be used for calculation of density of liquids at low
gravity (SG). Once SG at 60~ (15.5~ is known, with the use pressures where the effect of pressure on liquid density can be
of this figure specific gravity at temperature T (SGT) can be neglected (pressures less than ~50-70 bar). Figure 7.2 shows
determined. Then density at T and 1 atm can be determined effect of pressure on density at 60~ and Fig. 7.3 shows effect
through multiplying specific gravity by density of water. This of T on liquid density at any pressure P (P > 1000 psia or
304 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 7.1--Properties of saturated liquid and solid at the freezing point for some hydrocarbons [10].
No. Compound Formula M T~,t/Ttp,K Ptp, bar pL, g/cm3 pS, g/cm3 C~, J/g-K C~, J/g. K
n -Paraffins
1 n-Pentane C5H12 72.15 143.42 6.8642 x 10-7 0.7557 0.9137 1.9509 1.4035
2 n-Hexane C6H14 86.17 177.83 9.011 • 10-6 0.7538 0.8471 1.9437 1.4386
3 n-Heptane C7H16 100.20 182.57 1.8269 • 10-6 0.7715 0.8636 1.9949 1.4628
4 n-Octane C8H18 114.22 216.38 2.108 x 10-5 0.7603 0.8749 2.0077 1.5699
5 n-Nonane C9H20 128.25 219.66 4.3058 x 10-6 0.7705 0.8860 2.0543 1.6276
6 n-Decane C10H22 142.28 243.51 1.39297• 10-5 0.7656 0.8962 2.0669 1.6995
7 n-Tetradecane C14H30 198.38 279.01 2.5269 • 10-6 0.7722 0.9140 2.1589 1.8136
8 n-Pentadecane C15H32 212.41 283.07 1 . 2 8 8 7x 10 6 0.7752 0.9134 2.1713
9 n-Eicosane C20H42 282.54 309.58 9.2574 x 10-8 0.7769 0.8732 2.2049 2.2656
10 n-Hexacosane C26H54 366.69 329.25 5.1582 • 10-9 0.7803 0.9254 2.3094 2.2653
11 n-Nonacosane C29H60 408.77 336.85 6.8462 • 10-10 0.7804 0.9116 2.2553 1.8811
12 n-Triacontane C30H62 422.80 338.65 2.0985 x 10-1~ 0.7823 0.9133 2.2632
13 n-Hexacontane C36H74 506.95 349.05 2.8975 • 10-12 0.7819 0.9610 2.3960 2.4443
n-Alkylcyclohexanes (naphthenes)
14 Cyclohexane C6H12 84.16 279.69 5.3802 x 10.2 0.7894 0.8561 1.7627 1.6124
15 n-Decylcyclohexane C16H32 224.42 271.42 4.5202 x 10-8 0.8327 0.9740 1.9291 1.5398
n-Alkylbenzenes (aromatics)
16 Benzene C6H6 78.11 278.65 4.764 x 10.4 0.8922 1.0125 1.6964 1.6793
17 n-Butylbenzene C10H14 134.22 185.25 1 . 5 4 3 9x 10-9 0.9431 1.1033 1.5268 1.1309
18 n-Nonylbenzene Ct5H24 204.36 248.95 6.603 x 10-9 0.8857 1.0361 1.7270 1.6882
19 n-Tetradecylbenzene C20H34 274.49 289.15 9.8069 x 10-9 0.858 1.0046 1.8799 1.7305
1-n-Alkylnaphthalenes (aromatics)
20 Naphthalene C10H8 128.16 353.43 9.913 x 10-3 0.9783 1.157 1.687 1.6183
21 1-Methylnaphthalene CllH10 142.19 242.67 4.3382 x 10-7 1.0555 1.2343 1.4237 1.0796
22 1-n-Decylnaphthalene C20H28 268.42 288.15 8.4212 x 10-9 0.9348 1.0952 1.7289 1.5601
Other organic compounds
23 Benzoic acid C7H602 122.12 395.52 7.955x 10-3 1.0861 1.2946 2.0506 1.5684
24 Diphenylmethane C13H12 168.24 298.39 1 . 9 5 2 9x 10-5 1.0020 1.0900 1.5727 1.3816
25 Antheracene C14H10 190.32 488.93 4.951 x 10.2 0.9745 1.2167 2.0339 2.0182
Nonhydrocarbons
26 Water H20 18.02 273.15/ 6.117 x 10.3 1.0013 0.9168 4.227 2.1161
273.16
27 Carbon dioxide CO2 44.01 216.58 5.187 1.1807 1.5140 1.698 1.3844
70 bar). With use of Figs. 7.2 a n d 7.3, one m a y calculate den- equivalent to 55.15/62.4 - 0.8838 g/cm 3, w h i c h is exactly the
sity of a liquid p e t r o l e u m fraction with m i n i m u m i n f o r m a t i o n s a m e as the e x p e r i m e n t a l value. #
on specific gravity as s h o w n in the following example. These
figures are m a i n l y useful for density of undefined p e t r o l e u m Once specific gravity of a h y d r o c a r b o n at a t e m p e r a t u r e
fractions by h a n d calculation. is known, density of h y d r o c a r b o n s at the s a m e t e m p e r a t u r e
can be calculated using Eq. (2.1), w h i c h requires the density
of w a t e r at the s a m e t e m p e r a t u r e (i.e., 0.999 g/cm 3 at 60~ A
E x a m p l e 7 . I - - A p e t r o l e u m fraction has API gravity of 31.4. co r r el at i o n for calculation of density of liquid w a t e r at 1 a t m
Calculate density of this fraction at 20~ (68~ a n d 372.3 b a r for t e m p e r a t u r e s in the range of 0-60~ is given by DIPPR-
(5400 psia). C o m p a r e the e s t i m a t e d value with the e x p e r i m e n - EP CO N [10] as
tal value of 0.8838 g/cm 3 as given in C h a p te r 6 of Ref. [9].
(7.6) dr = A x B-[ 1+O-tIc)D]
point temperature and pressure (Ttp, Ptp) are also needed in 7.3.1 P u r e C o m p o n e n t s
SLE calculations. Values of density and heat capacity of liquid
Experimental data for vapor pressure of pure hydrocarbons
and solid phases for some compounds at their melting points
are given in the TRC Thermodynamic Tables [11]. Figures 7.4
are given in Table 7.1, as obtained from DIPPR [ 10]. The triple
and 7.5 show vapor pressure of some pure hydrocarbons from
point temperature (Ttp) is exactly the same as the melting or
praffinic and aromatic groups as given in the API-TDB [9].
freezing point temperature (TM). As seen from Fig. 5.2a and
Further data on vapor pressure of pure compounds at 37.8~
from calculations in Example 6.5, the effect of pressure on the
(100~ were given earlier in Table 2.2. For pure compounds
melting point of a substance is very small and for a pressure
the following dimensionless equation can be used to estimate
change of a few bars no change in TM is observed. Normal
vapor pressure [9]:
freezing point TM represents melting point at pressure of 1
atm. Ptp for a pure substance is very small and the maximum (7.8) l n P r "p = (Tr') x (ar + b r 15 + c r 26 + d r s)
difference between atmospheric pressure and Ptp is less than
1 atm. For this reason as it is seen in Table 7.1 values of TM where r = 1 - T r and p~ap is the reduced vapor pressure
and Ttp are identical (except for water). (pvap/Pc) , and Tr is the reduced temperature. Coefficients a -
Effect of temperature on solid density in a limited temper- d with corresponding temperature ranges are given in Ta-
ature range can be expressed in the following linear form: ble 7.2 for a number of pure compounds. Equation (7.8) is
a linearized form of Wagner equation. In the original Wagner
(7.7) pS
m = A - (10 -6 • B) T equation, exponents 3 and 6 are used instead of 2.6 and 5 [ 12].
The primary correlation recommended in the API-TDB [9]
where psm is the solid molar density at T in mol/cm 3. A and for vapor pressure of pure compounds is given as
B are constants specific for each compound, and T is the B E
absolute temperature in kelvin. Values of B for some com- (7.9) In pv,p = A + ~ + C In T + D T 2 + T-~
pounds as given by DIPPR [10] are n-Cs: 6.0608; n-C10: 2.46; n-
where coefficients A - E are given in the API-TDB for some 300
C20: 2.663; benzene: 0.3571; naphthalene: 2.276; benzoic acid:
2.32; and water (ice): 7.841. These values with Eq. (7.7) and compounds (hydrocarbons and nonhydrocarbons) with spec-
ified temperature range. This equation is a modified version
values of solid density at the melting point given in Table 7.1
of correlation originally developed by Abrams and Prausnitz
can be used to obtain density at any temperature as shown in
the following example. based on the kinetic theory of gases. Note that performance
of these correlations outside the temperature ranges specified
is quite weak. In DIPPR [10], vapor pressure of pure hydro-
E x a m p l e 7.2--Estimate density of ice at -50~ carbons is correlated by the following equation:
B
S o l u t i o n - - F r o m Table 7.1 the values for water are obtained (7.10) l n P vap= A + -~ + C l n T + D T E
as M = 18.02, TM ----273.15 K, p S = 0.9168 g/cm 3 (at TM).
where coefficients A - E are given for various compounds in
In Eq. (7.7) for water (ice) B ----7.841 and pS is the molar
Ref. [ 10]. In this equation, when E = 6, it reduces to the Riedel
density. At 273.15 K, pS = 0.050877 mol/cm 3. Substituting
equation [12]. Another simple and commonly used relation
in Eq. (7.7) we get A = 0.053019. With use of A and B in
to estimate vapor pressure of pure compounds is the Antoine
Eq. (7.7) at 223.15 K (-50~ we get pS = 0.051269 mol/cm 3
or ps = 0.051269 x 18.02 = 0.9238 g/cm 3. # equation given by Eq. (6.102). Antoine parameters for some
700 pure compounds are given by Yaws and Yang [13]. An-
toine equation can be written as
B
7.3 VAPOR PRESSURE (7.11) In PVap(bar) = A - - -
T+C
As shown in Chapters 2, 3, and 6, vapor pressure is required where T is in kelvin. Antoine proposed this simple modifi-
in many calculations related to safety as well as design and cation of the Clasius-Clapeyron equation in 1888. The lower
operation of various units. In Chapter 3, vapor pressure rela- temperature range gives the higher accuracy. For some com-
tions were introduced to convert distillation data at reduced pounds, coefficients of Eq. (7.11) are given in Table 7.3. Equa-
pressures to normal boiling point at atmospheric pressure. In tion (7.11) is convenient for hand calculations. Coefficients
Chapter 2, vapor pressure was used for calculation of flamma- may vary from one source to another depending on the tem-
bility potential of a fuel. Major applications of vapor pres- perature range at which data have been used in the regres-
sure were shown in Chapter 6 for VLE and calculation of sion process. Antoine equation is reliable from about 10 to
equilibrium ratios. As it was shown in Fig. 1.5, prediction 1500 m m H g (0.013-2 bars); however, the accuracy deteri-
of vapor pressure is very sensitive to the input data, partic- orates rapidly beyond this range. It usually underpredicts
ularly the critical temperature. Also it was shown in Fig. 1.7 vapor pressure at high pressures and overpredicts vapor pres-
that small errors in calculation of vapor pressure (or relative sures at low pressures. One of the convenient features of this
volatility) could lead to large errors in calculation of the height equation is that either vapor pressure or the temperature can
of absorption/distillation columns. Methods of calculation of be directly calculated without iterative calculations. No gener-
vapor pressure of pure compounds and estimation methods alized correlation has been reported on the Antoine constants
using generalized correlations and calculation of vapor pres- and they should be determined from regression of experimen-
sure of petroleum fractions are presented hereafter. tal data.
306 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
1,000
100
cO
tr
CO
G0
UJ 10
tr
Q.
n-
O
Q.
X
VAPOR PRESSURE OF
NORMALPARAFFIN
HYDROCARBONS
j•I)
TECHNICALDATABOOK
February1994
0.1 L
-50 -25 0 25 50 75 100 150 200 250 300 400 500
TEMPERATURE, F
An expanded form of Antoine equation, which covers a where Pc is in bar. Equations (7.13) and (7.14) work better
wider temperature range by including two additional terms at superatmospheric pressures (T > Tb) rather than at sub-
and a fourth parameter, is given in the following form as sug- atmospheric pressures. The main advantage of this equation
gested by Cox [12]: is that it has only two constants. This was the reason that
B it was used to develop Eq. (3.102) in Section (3.6.1.1) for
(7.12) In pvap : A + ~ + CT + DT 2 calculation of Reid vapor pressure (RVP) of petroleum fuels.
For RVP prediction, a vapor pressure correlation is applied at
Another correlation is the Miller equation, which has the fol- a single temperature (100~ or 311 K) and a two-parameter
lowing form [12]: correlation should be sufficient. Some other forms of equa-
tions used to correlate vapor pressure data are given in
(7.13) In pvap = - ~ [ 1 - T f + B ( 3 + T r ) ( 1 -T~) 3]
Ref. [12].
where A and B are two constants specific for each compound.
These coefficients have been correlated to the reduced boiling 7.3.2 P r e d i c t i v e M e t h o d s - - G e n e r a l i z e d
point Tbr (=Tb/Tc) and Pc of pure hydrocarbon vapor pressure Correlations
in the following form:
In Section 6.5, estimation of vapor pressure from an equa-
Tbrln(PJl.O1325) tion of state (EOS) through Eq. (6.105) was shown. When an
AI =
1 - Tbr appropriate EOS with accurate input parameters is used, ac-
(7.14) A -----0.4835 + 0.4605A1 curate vapor pressure can be estimated through Eq. (6.105)
or Eq. (7.65) [see Problem 7.13]. As an example, prediction of
A/A~ - (1 + Tbr) vapor pressure of p-xylene from a modified PR EOS is shown
B=
(3 + Tbr)(1 - Tbr) 2 in Fig. 7.6 [14].
7. APPLICATIONS: ESTIMATION OF THERMOPHYSICAL PROPERTIES 3 0 7
100
VAPOR PRESSURE
OF ALKYLBENZENE
HYDROCARBONS
I I TECHNICALDATABOOK
10 February1994
rt"
E~
o9
cO
LM
n"
Q..
tr
o
t)..
X
0.1
25 50 75 1O0 150 200 250 300 400 500
TEMPERATURE, F
Generally, vapor pressure is predicted through correlations The origin of most of predictive methods for vapor pres-
similar to those presented in Section 7.3.2. These correlations sure calculations is the Clapeyron equation (Eq. 6.99). The
require coefficients for individual components. A more use- simplest method of prediction of vapor pressure is through
ful correlation for vapor pressure is a generalized correlation Eq. (6.101), which is derived from the Clapeyron equation.
for all compounds that use component basic properties (i.e., Two parameters of this equation can be determined from two
Tb) as an input parameter. A perfect relation for prediction data points on the vapor pressure. This equation is very ap-
of vapor pressure of compounds should be valid from triple proximate due to the assumptions made (ideal gas law, ne-
point to the critical point of the substance. Generally no sin- glecting liquid volume, and constant heat of vaporization) in
gle correlation is valid for all compounds in this wide tem- its derivation and is usually useful when two reference points
perature range. As the number of coefficients in a correlation on the vapor pressure curve are near each other. However,
increases it is expected that it can be applied to a wider tem- the two points that are usually known are the critical point
perature range. However, a correct correlation for the vapor (Tc, Pc) and normal boiling point (Tb and 1.013 bar) as demon-
pressure in terms of reduced temperature and pressure is ex- strated by Eq. (6.103). Equations (6.101) and (6.103) may be
pected to satisfy the conditions that at T = To, pvap = Pc and at combined to yield the following re]ation in a dimensionless
T = Tb, pvap = 1.0133 bar. The temperature range Tb _< T _< Tc forln:
is usually needed in practical engineering calculations. How- Pc
ever, when a correlation is used for calculation of vapor pres- (7.15) lnP~aP=[ln(~)]x \ ~ ]Tbr
( ~x (1-~)
sure at T < Tb (PvaP _< 1.0133 bar), it is necessary to satisfy
the following conditions: at T = Ttp, pvap =Ptp and at T = Tb, where Pc is the critical pressure in bar and Tbr is the reduced
pwp = 1.0133 bar, where Ttp and Ptp are the triple point tem- normal boiling point (Tbr = Tb/Tc). The main advantage of
perature and pressure of the substance of interest. Eq. (7.15) is simplicity and availability of input parameters
308 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 7.2---Coefficients of Eq. (7.8) for vapor pressure of pure compounds [9].
(7.8) lnpVap = (T~-l ) x (ar + b~l'5 + c~2"6 + dr 5)
Compound name a b c d Ttp, K Train,K Tmax,K T~,K Max% err Ave%err
Nonhydrocarbon
1 Oxygen -6.0896 1.3376 -0.8462 - 1.2860 54 54 154 154 2.0 0.2
2 Hydrogen -4.7322 0.5547 1.5353 -1.1391 14 14 33 33 6.0 0.8
3 Water -7.8310 1.7399 -2.2505 -1.9828 273 273 647 647 0.4 0.0
4 Hydrogen chloride -6.2600 0.1021 1.0793 -4.8162 159 159 324 324 4.6 0.6
5 Hydrogen sulfide -5.7185 -0.4928 1.0044 -4.5547 188 188 373 373 7.3 1.0
6 Carbon monoxide -6.2604 1.5811 - 1.5740 -0.9427 68 68 133 133 10.1 0.9
7 Carbon dioxide -6.9903 1.3912 -2.2046 -3.3649 217 217 304 304 0.8 0.1
8 Sulfur dioxlde -6.8929 1.3119 -3.5225 0.6865 198 203 431 431 3.8 1.7
Parafms
9 Methane -5,9999 1.2027 -0,5310 -1,3447 91 91 191 191 0.1 0.0
10 Ethane -6.4812 1.4042 -1.2166 -1.7143 91 91 306 306 0.1 0.0
11 Propane -6.8092 1.6377 -1.8173 -1.8094 86 86 370 370 2.7 0.9
12 n-Butane -7.0524 1.6799 -2.0398 -2.0630 135 135 425 425 6.5 0.4
13 Isobutane -6.7710 1.0669 -0.9201 -3.8903 113 125 408 408 1.6 1.2
14 n-Pentane -7.2048 1.3503 -1.5540 -4.2828 143 157 469 469 6.5 0.8
15 Isopentane -7.1383 1.5320 -1.8896 -2.7290 113 178 461 461 0.3 0.1
16 Neopentane -6.9677 1.5464 -1.9563 -2.6057 257 257 434 434 3.2 0.2
17 n-Hexane -7.3505 0.9275 -0.7303 -6.7135 178 178 507 507 5.8 1.0
18 n-Heptane -7.4103 0.7296 -1.3081 -5.9021 183 183 540 54O 4.6 0.4
19 2-Methylhexane -7.6340 1.6113 -2.4895 -3.7538 155 222 531 531 1.9 0.2
20 n-Octane -8.0092 1.8442 -3.2907 -3.5457 216 286 569 569 3.1 0.3
21 2,2-Dimethylhexane -7.5996 1.4415 -2.3822 -4.2077 152 236 550 550 0.7 0.1
22 n-Nonane -9.5734 5.7040 -8.9745 3.3386 219 233 596 596 3.9 0.6
23 2-Methyloctane -9.4111 5.6082 -9.1179 3.9544 193 303 587 587 1.4 0.3
23a 2,2,4-Trimethylpentane -7.4717 1.5074 -2.2532 -3.5291 166 225 544 544 5.0 0.3
24 n-Decane -8.4734 2.0043 -3.9338 -4.5270 243 286 618 618 3.8 0.5
25 n-Undecane -8.6767 1.8339 -3.6173 -6.5674 248 322 639 639 4.2 0.2
26 n-Dodecane -9.1638 2.8127 -5.5268 -4.1240 263 294 658 658 5.4 0.2
27 n-Tridecane -11.5580 9,5675 -17,8080 23.9100 268 333 676 676 3.2 1.5
28 n-Tetradecane -9.5592 2,6739 -5.3261 -7.2218 279 369 692 692 0.4 0.1
29 n-Pentadecane -9.8836 2.9809 -5.8999 -7.3690 283 383 707 707 0.2 0.1
30 n-Hexadecane -10.1580 3.4349 -7.2350 -4.7220 291 294 721 721 5.9 0.3
31 n-Heptadecane -8.7518 -1.2524 0.6392 -21.3230 295 311 733 733 4.9 0.7
32 n-Octadecane -11.30200 6.3651 -12.4510 0.2790 301 322 745 745 6.6 1.9
33 n-Nonadecane -10.0790 2,7305 -7.8556 -5.3836 306 400 756 756 2.8 0.5
34 n-Eicosane -9.2912 0.7364 -8.1737 -0.4546 309 353 767 767 8.0 1.2
35 n-Tetracosane -14.4290 12.0240 -21.5550 11.2160 324 447 810 810 4.0 0.8
36 n-Octacosane -11.4490 2.0664 -7.4138 -15.4770 334 417 843 843 5.7 1.2
Naphthenes
37 Cyclopentane -7.2042 2,2227 -2.8579 -1.2980 179 203 512 512 3.9 0.6
38 Methylcyclopentane -7.1157 1.5063 -2.0252 -2.9670 131 178 533 533 2.0 0.1
39 Ethylcyclopentane -7.2608 1.3487 -1.8800 -3.7286 134 183 569 569 1.5 0.1
40 n-Propylcyclopentane -1.3961 0.2383 -5.7723 -6.0536 156 244 603 603 0.1 0.0
41 Cyclohexane -7.0118 1.5792 -2.2610 -2.4077 279 279 553 553 4.3 0.7
42 Methylcyclohexane -7.1204 t.4340 -1.9015 -3.3273 147 217 572 572 1.2 0.2
43 Ethylcyclohexane -5.9783 -1.2708 0.2099 -5.3117 162 228 609 609 0.2 0.0
44 n-Propylcyclohexane -5.6364 -2.1313 0.6054 -6.0405 178 228 639 639 0.4 0.0
45 Isopropylcyclohexane -7.8041 2.0024 -2.8297 -3.4032 184 208 627 627 3.5 0.1
46 n-Butylcyclohexane -4.9386 -3.9025 2.0300 -7.8420 198 286 667 667 0.2 0.0
47 n-Decylcyclohexane -9.5188 2.4189 -4.5835 -7.7062 272 322 751 751 9.2 0.4
48 Cycloheptane -7.3231 1.8407 -2.2637 -3.4498 265 265 604 604 0.7 0.1
Olefms
49 Ethylene -6.3778 1.3298 -1.1667 -2.0209 104 104 282 282 8.9 0.6
50 Propylene -6.7920 1.7836 -2.0451 -1.5370 88 88 366 366 5.2 0.3
51 1-Butene -6.9041 1.3587 -1.3839 -3.7388 88 125 420 420 2.4 0.5
52 1-Pentene -6.6117 0.0720 0.0003 -5.4313 108 167 465 465 6.5 1.5
53 1,3-Butadiene -5.6060 -0.9772 -0.3358 -3.1876 127 156 484 484 0.3 0.1
Diolefins and Acetylenes
54 Acetylene -7.3515 2.8334 -4.5075 6.8797 192 192 308 308 2.4 0.4
Aromatics
55 Benzene -7.0200 1.5156 -1.9176 -3.5572 279 279 562 562 1.6 0.3
56 Toluene -7.2827 1.5031 -2.0743 -3.1867 178 244 592 592 3.0 0.3
57 Ethylbenzene -7.5640 1.7919 -2.7040 -2.8573 178 236 1 617 0.8 0.1
58 m-Xylene -7.6212 1.6059 -2.4451 -3.0594 226 250 617 617 3.9 0.2
59 o-Xylene -7.5579 1.5648 -2.1826 -3.7093 248 248 631 631 2.2 0.2
60 p-Xylene -7.6935 1.8093 -2.5583 -3.0662 287 287 616 616 7.5 0.4
(Continued)
7. A P P L I C A T I O N S : E S T I M A T I O N OF THERMOPHYSICAL PROPERTIES 309
TABLE 7.2--(Continued).
Compoundname a b c d Ttp,K Tmin,K Tmax,K To K Max%err Ave%err
61 i-PropyIbenzene -8.1015 2.6607 -3.8585 -2.2594 173 236 638 638 5.4 0.4
62 n-Butylbenzene -7.8413 1.3055 -2.1437 -5.3415 186 233 661 661 5.6 0.7
63 n-Pentylbenzene -8.7573 3.1808 -4.7169 -2.7442 198 311 680 680 2.8 0.2
64 n-Hexylbenzene -8.0460 0.6792 -1.4190 -8.1068 212 333 698 698 1.8 0.2
65 n-Heptylbenzene -9.1822 3.1454 -4.8927 -4.5218 225 356 714 714 2.0 0.2
66 n-Octylbenzene -10.7760 7.0482 -10.5930 1.7304 237 311 729 729 8.0 0.8
67 Styrene -6.3281 -1.2630 0.9920 -7.1282 243 243 636 636 0.6 0.1
68 n-Nonylbenzene -10.7760 7.0038 -10.4060 1.1027 249 311 741 741 1.4 0.4
69 n-Decylbenzene -10.5490 4.7502 -7.2424 -4.8469 259 333 753 753 0.1 0.0
70 n-Undecylbenzene -11.8950 8.0001 -12.7000 4.6027 268 383 764 764 1.1 0.2
71 n-Dodecylbenzene -10.6650 3.9860 -7.6855 -1.7721 276 333 659 774 9.6 1.8
72 n-Tridecylbenzene -11.995 6.5968 -10.1880 -5.2923 283 417 783 783 2.1 0.4
73 Cumene -7.4655 1.2449 -2.0897 -4.5973 177 228 631 631 2.6 0.3
Diaromatics
74 Naphthalene -7.6159 1.8626 -2.6125 -3.1470 353 353 748 748 17.5 0.8
75 1-Methylnaphthalene -7.4654 1.3322 -3.4401 -0.8854 243 261 772 772 7.1 1.7
76 2-Methylnaphthalene -7.6745 1.0179 -1.3791 -5.6038 308 308 761 761 7.8 0.9
77 2,6-Dimethylnaphthalene -7.8198 -2.5419 9.2934 -24.3130 383 383 777 777 0.1 0.0
78 i -Ethylnaphthalene -6.7968 -0.5546 -1.2844 -5.4126 259 322 776 776 11.1 0.6
89 Anthracene -8.4533 1.3409 -1.5302 -3.9310 489 489 873 873 5.6 0.5
80 Phenanthrene -11.6620 9.2590 -10.0050 1.2110 372 372 869 869 1.0 0.2
Oxygenated compounds
81 Methanol -8.6413 1 . 0 6 7 1 -2.3184 -1.6780 176 176 513 513 5.9 0.7
82 Ethanol -8.6857 1 . 0 2 1 2 -4.9694 1.8866 159 194 514 514 4.9 0.4
83 Isopropanol -7.9087 -0.6226 -4.8301 0.3828 186 200 508 508 8.4 1.6
84 Methyl-tert-hutyl ether -7.8925 3.3001 -4.9399 0.2242 164 172 497 497 8.0 1.3
85 tert-Butyl ethyl ether -6.1886 -1.0802 -0.9282 -2.9318 179 179 514 514 8.7 4.8
86 Diisopropyl ether -7.2695 0.4489 -0.9475 -5.2803 188 188 500 500 22.7 2.7
87 Methyl tert-pentyl ether -7.8502 2.8081 -4.5318 -0.3252 158 534 534 1.3 0.4
Ttp is the triple point temperature and Tcis the critical temperature. Trainand Tmaxindicate the range at which Eq. (7.8) can be used with these coefficients.For
quick and more convenientmethod use Antoineequation with coefficientsgivenin Table7.3.
(Tb, T0, a n d Pc) for pure c o m p o u n d s . However, one should Eq. (7.16) c a n be d e t e r m i n e d from Eq. (6.103) with replacing
realize that since the base points in deriving the constants Tb a n d To by T~3 a n d T~3. The linear relationship b e t w e e n
given by Eq. (6.103) are Tb a n d To this equation should be In pvap a n d 1/T 13 for large n u m b e r of pure h y d r o c a r b o n s is
used in the t e m p e r a t u r e range of Tb _< T _< To Theoretically, s h o w n in Fig. 7.7.
a vapor pressure relation should be valid from triple p o i n t Preliminary evaluation of Eq. (7.16) shows n o m a j o r advan-
t e m p e r a t u r e to the critical temperature. But most vapor pres- tage over Eq. (7.15). A c o m p a r i s o n of Eqs. (7.15) a n d (7.16)
sure correlations are very poor at t e m p e r a t u r e s n e a r the triple for n-hexane is s h o w n in Fig. 7.8. Predicted vapor pressure
p o i n t temperature. Using Eq. (7.15) at t e m p e r a t u r e s below Tb from the m e t h o d r e c o m m e n d e d i n the API-TDB is also s h o w n
usually leads to u n a c c e p t a b l e predicted values. For better pre- i n Fig. 7.8. Clapeyron m e t h o d refers to Eq. (7.15), while the
diction of vapor pressure n e a r the triple point, the two base Korsten m e t h o d refers to Eq. (7.16), with p a r a m e t e r s A a n d
points should be n o r m a l boiling p o i n t (T = Tb, P = 1.01325 B d e t e r m i n e d from Tb, To a n d Pc. E q u a t i o n (7.15) agrees bet-
bar) a n d triple p o i n t (Ttp, Ptp). Values of Ttp a n d Ptp for some ter t h a n Eq. (7.16) with the API-TDB method. S u b s t i t u t i o n of
c o m p o u n d s are given i n Table 7.1. Similarly if vapor pressure Eq. (6.16) into Eq. (2.10) leads to Eq. (2.109) for prediction of
prediction n e a r 37.8~ (100~ is required the vapor pressure acentric factor by Korsten method. Evaluation of m e t h o d s of
data given i n Table 2 should be used as one of the reference prediction of acentric factor presented i n Section 2.9.4 also
points along with Tb, To, or Ttp to obtain the constants A a n d gives some idea o n accuracy of vapor pressure correlations
B i n Eq. (6.101). for pure hydrocarbons.
One of the latest developments for correlation of vapor pres- Korsten d e t e r m i n e d that all h y d r o c a r b o n s exhibit a vapor
sure of pure h y d r o c a r b o n s was proposed by Korsten [ 15]. He pressure of 1867.68 b a r at 1994.49 K as s h o w n i n Fig. 7.7.
investigated modification of Eq. (6.101) with vapor pressure This data p o i n t for all h y d r o c a r b o n s a n d the boiling p o i n t
data of h y d r o c a r b o n s a n d he f o u n d that l n P vap varies linearly data can be used to d e t e r m i n e p a r a m e t e r s A a n d B i n Eq.
with 1/T 1.3 for all hydrocarbons. (7.16). I n this way, the resulting e q u a t i o n requires only one
i n p u t p a r a m e t e r (Tb) similar to Eq. (3.33), which is also s h o w n
B
(7.16) In pvap = A - - - i n Section 7.3.3.1 (Eq. 7.25). Evaluation of Eqs. (7.25) a n d
TI.3
(7.16) with use of Tb as sole i n p u t p a r a m e t e r indicates that Eq.
where T is absolute t e m p e r a t u r e i n kelvin a n d pvap is the va- (7.25) is m o r e accurate t h a n Eq. (7.16) as s h o w n i n Fig. 7.8.
por pressure i n bar. I n fact the m a i n difference b e t w e e n this However, note that Eq. (7.25) was developed for p e t r o l e u m
e q u a t i o n a n d Eq. (6.101) is the exponent of T, which in this fractions a n d it m a y be used for pure h y d r o c a r b o n s with
case is 1.3 (rather t h a n 1 in the Clapeyron type equations). Pa- Nc_>5.
rameters A a n d B can be d e t e r m i n e d from boiling a n d critical Perhaps the most successful generalized correlation for
points as it was s h o w n in Example 6.6. Parameters A a n d B i n prediction of vapor pressure was based o n the theory of
TABLE 7.3--Antione coefficients for calculation of vapor pressure from Eq. (7.11).
B
In pvap = A - - - Units: bar a n d K
T+C
No. Compound ~,K A B C
n-Alkanes
1 M e t h a n e (C1) 111.66 8.677752 911.2342 -6.340
2 E t h a n e (C2) 184.55 9.104537 1528.272 -16.469
3 Propane (Ca) 231.02 9.045199 1851.272 -26.110
4 B u t a n e (n-C4) 272.66 9.055284 2154.697 -34.361
5 I s o b u t a n e (i-C4 ) 261.34 9.216603 2181.791 -24.280
6 P e n t a n e (n-Cs) 309.22 9.159361 2451.885 -41.136
7 Hexane (n-C6) 341.88 9.213541 2696.039 -48.833
8 H e p t a n e (n-C7) 371.57 9.256922 2910.258 -56.718
9 Octane (n-Ca) 398.82 9.327197 3123.134 -63.515
10 N o n a n e (n-C9) 423.97 9.379719 3311.186 -70.456
11 Decane (n-C10) 447.3 9.368137 3442.756 -79.292
12 U n d e c a n e (n-C11) 469.08 9.433921 3614.068 -85.450
13 Dodecane (n-C12) 489.48 9.493213 3774.559 -91.310
14 Tridecane (n-C13) 508.63 9.515341 3892.912 -98.930
15 Tetradecane (n-C14) 526.76 9.527867 4008.524 - 105.430
16 Pentadecane (n-C15) 543.83 9.552251 4121.512 - 111.770
17 Hexadecane (n-C16) 559.98 9.563948 4214.905 -118.700
18 H e p t a d e c a n e (n-C17) 574.56 9.53086 4294.551 - 123.950
19 Octadecane (n-C18) 588.3 9.502999 4361.787 - 129.850
20 N o n a d e c a n e (n-C19) 602.34 9.533163 4450.436 - 135.550
21 Eicosane (n-C20) 616.84 9.848387 4680.465 - 141.050
1-Alkenes
22 Ethylene (C2H4) 169.42 9.011904 1373.561 -16.780
23 Propylene (Ca H6 ) 225.46 9.109165 1818. t 76 -25.570
24 1-butane (C4H8) 266.92 9.021068 2092.589 -34.610
Naphthenes
25 Cyclopentane 322.38 9.366525 2653.900 - 38.640
26 Methylcyclopentane 344.98 9.629388 2983.098 -34.760
27 Ethylcyclopentane 376.59 9.219735 2978.882 -53.030
28 Cyclohexane 353.93 9.049205 2723.438 -52.532
29 Methylcyclohexane 374.09 9.169631 2972.564 -49.449
Aromatics
30 Benzene (C6H6) 353.24 9.176331 2726.813 -55.578
31 Toluene (C7H8) 383.79 9.32646 3056.958 -55.525
32 Ethy]benzene 409.36 9.368321 3259.931 -60.850
33 Propylbenzene 432.35 9.38681 3434.996 -65.900
34 Butylbenzene 456.42 9.448543 3627.654 -71.950
35 o-Xylene (C8H10) 417.59 9.43574 3358.795 -61.109
36 m-Xylene (C8H10) 412.34 9.533877 3381.814 -57.030
37 p-Xylene (Call10) 411.53 9.451974 3331.454 -58.523
Other hydrocarbons
38 Isooctane 372.39 9.064034 2896.307 -52.383
39 Acetylene (C2H2) 188.40 8.459099 1217.308 -44.360
40 Naphthalene 491.16 9.522456 3992.015 -71.291
Organics
41 Acetone (C3H60) 329.22 9.713225 2756.217 -45.090
42 Pyridine (C5H5N) 388.37 9.59600 3161.509 -58.460
43 Aniline (C6HTN) 457.17 10.15141 3897.747 -72.710
44 Methanol 337.69 11.97982 3638.269 -33.650
45 Ethanol 351.80 12.28832 3795.167 -42.232
46 Propanol 370.93 11.51272 3483.673 -67.343
Nonhydrocarbons
47 H y d r o g e n (H2) 20.38 6.768541 153.8021 2.500
48 Oxygen (O2) 90.17 8.787448 734.5546 -6.450
49 Nitrogen (N2) 77.35 8.334138 588.7250 -6.600
50 H e l i u m (He) 4.30 3.876632 18.77712 0.560
51 CO 81.66 8.793849 671.7631 -5.154
52 CO2 194.65 10.77163 1956.250 -2.1117
53 A m m o n i a (NH3) 239.82 10.32802 2132.498 -32.980
54 H2 S 212.84 9.737218 1858.032 -21.760
55 Sulfur (S) 717.75 9.137878 5756.739 -86.850
56 CC14 349.79 9.450845 2914.225 -41.002
57 Water (HE O) 373.15 11.77920 3885.698 -42.980
The above coefficients may be used for pressure range of 0.02-2.0 bar except for water for which the pressure range is 0.01-16
bar as reported in Ref. [ 12]. These coefficents can generate vapor pressure near atmospheric pressure with error of less than
0.1%. There are other reported coefficients that give slightly more accurate results near the boiling point. For example, some
other reported values for A, B, and C are given here. For water: 11.6568, 3799.89, and -46.8000; for acetone: 9.7864, 2795.82,
and -43.15 or 10.11193, 2975.95, and-34.5228; for ethanol: 12.0706, 3674.49, and-46.70.
310
7. APPLICATIONS: E S T I M A T I O N OF THERMOPHYSICAL PROPERTIES 311
40-
3O
<~ ~ P , c~tlc
= 20
Reduced Temperature, Tr
Tr(ln p~ap)
= --5.97616r + 1.29874r 15 -- 0.60394r 2s -- 1.06841r 5
Locz~of acentricfactors .
+ w ( - - 5 . 0 3 3 6 5 r + 1.11505r 15 -- 5.41217r 2"5 -- 7.46628r 5)
I l ir
I l i. /" I " ~ # . , ~///1///,~
/ ~ 1 / i + w2(--0.64771r + 2.41539r 1 5 - 4.26979r2.5+ 3.25259r 5)
(7.19)
> o-o5i-
/,
....,-
~
/- . . , ' ~ ~ / / ~
/ . * .'/,r
30
o API
0.001 L._ ~ _ ~ ~ , .L-~---,J . . . . , .
'%i ~,u
i n-alk341:~r~.
c > 10
& 1 Locus
5
~
o.1!
0,05
6
,/
~t ' t
0
200
. . . .- . .
250
~
300 350
'
400
~
450 500
001 7
0.005~- Nt Temperature, K
. . . . .
~49
2663.129Q - 5.994296
0.02 . . . . Antoine
logl0 pvap = 9 5 . 7 6 Q - 0.972546
0.002 ,,,i .... , .... J .... i .... i .... i .... i ....
for 0.0013 _< Q < 0.0022(2 m m H g < pvap _< 7 6 0 m m H g )
0 50 100 150 200 250 300 350 400
TABLE 7.4---Prediction of vapor pressure of benzene at 400 K (260~ from different methods in Example 7.3.
API [9] MillerEqs. (7.13) Lee-Kesler Ambrose Riedel Clapeyron Korsten Maxwell API
Fig. 7.5 and (7.14) Eq. (7.18) Eq. (7.19) Eq. (7.24) Eq. (7.25) Eq. (7.15) Eq. (7.16) Eqs.(7.20-7.22) Eq. (7.8)
pvap, b a r 3.45 3.74 3.48 3.44 3.50 3.53 3.43 3.11 3.44 3.53
%Error ... 8.4 0.9 -0.3 1.4 2.3 -0.6 -9.9 -0.3 2.3
-O.2
O3
O.4
0.5 ASTM 10% Slope
110-3
~- 0.6
a2! 0
0,.8 100 -
:-1
90-
2 "}.
.m_
N-P 80-
2
09
[3. 7o-_- LL
U)
~3 60 - oJ
CL
O-
E
B
-:-4 50"
a3
J 5
I.-
6
30"
8
L 20"
-L lO
True Vapor Pressure of
:-- 12
Gasolines and Finished
:-- 14 Petroleum Products
o-
="16
--18
II TECHNICAL
DATABOOK
~. 20 June 1993
-- 22
~- 24
Fig. 7.10 should be experimental rather than estimated from coefficients are given as A1 = 7.78511307, A 2 = - 1.08100387,
methods of Section 3.6.1.1. If no experimental data on RVP A3 = 0.05319502, A4 =0.00451316, B1 = - 5756.8562305, BE=
are available the TVP should be calculated directly from meth- 1104.41248797, and B3 =-0.00068023. There is no in-
ods discussed in Sections 7.3.2 and 7.3.3.1. formation on reliability of these methods. Figures 7.10 and
For computer applications, analytical correlations have 7.11 or Eq. (7.26) are particularly useful in obtaining values
been developed from these two figures for calculation of vapor of vapor pressure of products and crude oils needed in esti-
pressure of petroleum products and crude oils from RVP data mation of hydrocarbon losses from storage tanks [20].
[9]. For petroleum products, Fig. 7.10 has been presented by
a complex correlation with 15 constants in terms of RVP and
7.3.4 Vapor Pressure of Solids
slope of ASTM D 86 curve at 10%. Similarly for crude oils the
mathematical relation developed based on Fig. 7.11 is given Figure 5.2a shows the equilibrium curve between solid and
as [9] vapor phases, which is known as a sublimation curve. In fact,
at pressures below triple point pressure (P <Ptp), a solid di-
In pvap = A1 + A2 In(RVP) + A3(RVP) + A4T rectly vaporizes without going through the liquid phase. This
(7.26) + [B1 + BEln(RVP) + Ba(RVP)4] type of vaporization is called sublimation and the enthalpy
T change is called heat of sublimation (AHSUb). For ice, heat
of sublimation is about 50.97 kJ/tool. Through phase equilib-
where pvap and RVP are in psia, T is in ~ Ranges of appli- rium analysis similar to the analysis made for VLE of pure
cation are OF < T(~ < 140F and 2 psi < RVP < 15 psi. The substances in Section 6.5 and beginning with Eq. (6.96) for
7. A P P L I C A T I O N S : E S T I M A T I O N OF THERMOPHYSICAL PROPERTIES 315
- o
u0 -
--1
130
120
-2
-2
100-
t~ --3 ._m
(/) _ 3
o-- a
t~
9o=_:
g --4 rm LL
n
2-5
70 03
o-
mb > E
0)
~7 ~ -lo 6O
p- w
--8
----9
---lO
__ 15 4O
~2o
lI TECHNICAL DATAB O O K 0--
J u n e 1993
FIG. 7.11--True vapor pressure of crude oils from RVP. Unit conversion:
~ = (~ • 1.8 + 32; psia = bar x 14.504. Taken with permission from
Ref, [9].
v a p o r - s o l i d equilibria (VSE) one c a n derive a relation s i m i l a r as the t e m p e r a t u r e variation along the s u b l i m a t i o n curve is
to Eq. (6.101) for e s t i m a t i o n of v a p o r p r e s s u r e of solids: limited, In a d d i t i o n it is a s s u m e d that AV sub = V yap - V s ~-
B
RT/P sub. This a s s u m p t i o n is r e a s o n a b l e as V s << V v a n d psub
(7.27) In psub = A - -- is very small so t h a t the v a p o r is c o n s i d e r e d as a n ideal gas.
T
In fact a c c u r a c y of Eq. (7.27) is m o r e t h a n Eq. (6.101) be-
w h e r e p~ub is the v a p o r p r e s s u r e of a p u r e solid also k n o w n as cause the a s s u m p t i o n s m a d e in derivation of this e q u a t i o n
sublimation pressure a n d A a n d B are two c o n s t a n t s specific are m o r e realistic. The relations for s u b l i m a t i o n p r e s s u r e of
for each c o m p o u n d . Values of psub are less t h a n Ptp a n d one n a p h t h a l e n e is given as [21 ]
b a s e p o i n t to o b t a i n c o n s t a n t A is the triple p o i n t (Ttp, Ptp).
3783
Values of Ttp a n d Ptp for s o m e selected c o m p o u n d s are given in In PSUb(bar) = 8.722 - - -
Table 7.1. If a value of s a t u r a t i o n p r e s s u r e (p~ub) at a reference T
t e m p e r a t u r e of T1 is k n o w n it c a n b e u s e d along with the triple (7.29) (T in kelvin) for solid n a p h t h a l e n e
p o i n t to o b t a i n A a n d B in Eq. (7.27) as follows: Vapor pressures of solid CO2 are given at several t e m p e r a -
[P,p~ tures as: 9.81 t o r t (at -120~ 34.63 ( - 1 1 0 ) , 104.81 ( - 1 0 0 ) ,
I n k e[ub} 279.5 ( - 9 0 ) w h e r e the pressures are in m m H g (tort) (1 b a r =
B-- - -
1 1 750.06 m m Hg) a n d the n u m b e r s in the p a r e n t h e s e s are the
(7.28) r, r,p c o r r e s p o n d i n g t e m p e r a t u r e s in ~ as given b y Levine [22].
L i n e a r regression of In psub versus 1/T gives constants A a n d
A = In (p~ub) + ~1 B in Eq. (7.27) as
w h e r e Ttp a n d T1 are in kelvin. P a r a m e t e r B is equivalent to 3131.97
In Psub(bar) = 16.117 (T in kelvin) for solid CO2
AH~ub/R. I n deriving this equation, it is a s s u m e d t h a t A H ~ub is T
c o n s t a n t w i t h t e m p e r a t u r e . This a s s u m p t i o n can be justified (7.30)
316 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
7.4.2 Heat Capacity Cp for gases, but for liquids more specific correlations espe-
cially at low pressures have been proposed in the literature.
Heat capacity is one of the most important thermal properties
Estimation of Cr and Cv from equations of state was demon-
and is defined at both constant pressure (Cp) and constant vol-
strated in Example 6.2.
u m e (Cv) by Eqs. (6.17) and (6.18). It can be measured using
For solids the effect of pressure on heat capacity is ne-
a calorimeter. For constant pressure processes, Cp and in con-
glected and it varies only with temperature: C s -- C s = f(T).
stant volume processes, Cv is needed. Cp can be obtained from
At moderate and low pressures the effect of pressure on liq-
enthalpy using Eq. (6.20).
uids m a y also be neglected as CpL ~- CvL = f ( T ) . However, this
Experimental data on liquid heat capacity of some pure
assumption is not valid for liquids at high pressures. Some
hydrocarbons are given in Table 7.6 as reported by Poling
specific correlations are given in the literature for calculation
et al. [12]. For defined mixtures where specific heat capacity
of heat capacity of h y d r o c a r b o n liquids and solids at atmo-
for each c o m p o u n d in the mixture is known, the mixing rule
spheric pressures. At low pressures a generalized expression
given by Eq. (7.2) m a y be used to calculate mixture heat ca-
in a polynomial form of up to fourth orders is used to correlate
pacity of liquids cL.. Heat capacities of gases are lower than
Cp with temperature:
liquid heat capacities under the same conditions of T and P.
For example, for propane at low pressures (ideal gas state)
C L / R = C L / R = A1 + A2T + A3T 2 + A4T 3 -1- A5T 4
the value of C ;g is 1.677 J / g - K at 298 K and 3.52 J/g. K at (7.38)
800 K. Values of Cpg of n-heptane are 1.658 J/g. K at 298 K C S l R = CSvlR = B1 + BzT + B3T 2 + B4T 3 + B5T 4
and 3.403 J / g - K at 800 K. However, for liquid state and at
300 K, C L of C3 is 3.04 and that of n-C5 is 2.71 J/g. K as re- where T is in kelvin. Coefficients A1-As and B1-Bs for a num-
ported by Reid et al. [12]. While molar heat capacity increases ber of c o m p o u n d s are given in Table 7.7 as given by DIPPR
with M, specific heat capacity decreases with increase in M. [10]. Some of the coefficients are zero for some c o m p o u n d s
Heat capacity increases with temperature. and for most solids the polynomial up to T 3 is needed. In
The general approach to calculate Cp is to estimate heat fact Debye's statistical-mechanical theory of solids and exper-
capacity departure from ideal gas [Cp - C~g] and combine it imental data show that specific heats of nonmetallic solids at
with ideal gas heat capacity (cpg)9 A similar approach can be very low temperatures obey the following [22]:
used to calculate Cv. The relation for calculation of C~g of
CSp = a T 3
petroleum fractions was given by Eq. (6.72), which requires. (7.39)
9 lg
Kw and ~0 as input parameters. Cv can be calculated from Cpg
ig ig where T is the absolute temperature in kelvin. In this relation
through Eq. (6.23). Both Cp and C v are functions of only tem-
there is only one coefficient that can be determined from one
perature. For petroleum fractions, Cpg can also be calculated
data point on solid heat capacity. Values of heat capacity of
from the p s e u d o c o m p o u n d m e t h o d of Chapter 3 (Eq. 3.39) by
solids at melting point given in Table 7.1 m a y be used as the
using Eq. (6.66) for pure hydrocarbons from different fami-
reference point to find coefficient a in Eq. (7.39). Equation
lies similar to calculation of ideal gas enthalpy (Eq" 7.35). The
(7.39) can be used for a very narrow temperature range near
most accurate method for calculation of [Cp - Cpg] is t h r o u g h
the point where coefficient a is determined.
generalized correlation of Lee-Kesler (Eq. 6.57). Relations
Cubic equations of states or the generalized correlation of
for calculation of [Cp - Cpg] and [Cv - C~] from cubic equa-
Lee-Kesler for calculation of the residual heat capacity of
tions of state are given in Table 6.1. For gases at moderate
liquids [CL - C~g] do not provide very accurate values espe-
pressures the departure functions for heat capacity can be es-
cially at low pressures. For this reason, attempts have been
timated through virial equation of state (Eqs. 6.64 and 6.65).
made to develop separate correlations for liquid heat capac-
Once heat capacity departure and ideal gas properties are de-
ity. Based on principle of corresponding states and using pure
termined, Cr is calculated from the following relation:
c o m p o u n d s ' liquid heat capacity data, Bondi modified previ-
(7.37) Cp = [Cp - Cpg] + Cpg ous correlations into the following form [12]:
Relations given in Chapter 6 for the calculation of [Cp - Cpg] CL - CPg - 1.586 + 0 . 4 9
and Cpg are in molar units. If specific unit of J/g. ~ for heat R 1-Tr
capacity is needed, calculated values from Eq. (7.37) should
be divided by molecular weight of the substance. Generalized (7.40) + 0)
[4.2775 + 6.3 (1 - Tr) ~/3 0.4355 ]
correlation of Lee-Kesler normally provide reliable values of L rr
TABLE 7.7--Coefficients of Eq. (7.38) for liquid (Ai, s) and solid (Bi, s) heat capacity for some selected compounds [10].
c L / R = CL/R = A1 + A2T + A3 T2 + A4T 3 + A5T 4
(7.38) CS / R = CSv/ R = B1 + B2 T + BaT 2 + B4 T 3 + Bs T 4
Compound. M Ai A~ A3 A4 A5 Train, K Tmax,K
Liquid heat capacity, C~
n-Pentane 72.2 19.134 -3.254 x 10 -2 1.197 x 10 .4 0 0 143 390
n-Hexane 86.2 20.702 -2,210 x 10 -2 1.067 x 10 -4 0 0 178 460
n-Decane 142.3 33.512 -2,380 x 10 -2 1.291 x 10 -4 0 0 243 460
n-Pentadecane 212.4 41.726 2,641 x 10 -2 7.894 x 10 -5 0 0 283 544
n-Eicosane 282,5 42.425 9.710 x 10 -2 2.552 x 10 -5 0 0 309 617
n-Hexatriacontane (C36) 507.0 84.311 1.771 x 10 -1 0 0 0 353 770
Cyclohexane 84.2 -26.534 3.751 x 10 -1 - 1 . 1 3 x 10 -3 1.285x10 -6 0 280 400
Methylcyclohexane 98.2 15.797 -7.590 x 10 -3 9.773 x 10 -5 0 0 146 320
Benzene 78.1 19.598 -4.149 x 10 -2 1.029 x 10 -4 0 0 279 500
Toluene 92.1 16.856 -1.832 x 10 -2 8.359 x 10 .5 0 0 178 500
Naphthalene 128.2 3.584 6.345 x 10 -2 0 0 0 353 491
Anthracene 178.2 9.203 7.325 x 10 .2 - 5 . 9 3 x 10-6 0 0 489 655
Carbon dioxide 44.0 -998.833 1.255 x 10 -5.21 x 10 -2 7.223 x 10 -5 0 220 290
Water 18.0 33.242 -2.514 x 10 -1 9.77 x 10 -4 -1.698 x 10 -6 1.127 x 10 .9 273 533
B1 /32 /33 /34 B5 Train, K Tmax,K
Solid heat capacity, Cps
n-Pentane - 1.209 O. 1215 5.136 x 10 -4 - 1 . 2 2 x 10-5 5.08 x 10 -8 12 134
n-Hexane -2.330 0.1992 -1.01 x 10 -3 2.43 x 10 -6 0 20 178
n-Decane -4.198 0.3041 - 1 . 5 2 x 10 -3 3.43 • 10 -6 0 20 240
n-Pentadecane -311.823 1.3822 0 0 0 271 283
n-Eicosane -0.650 0.3877 - 1 . 5 7 x 10 -3 3.65 x 10 -6 0 93 268
n-Hexatriacontane (C36) --200.000 1.0000 0 0 0 300 325
Cyclohexane 15.763 -0.0469 1.747 • 10 -4 0 0 191 271
Methylcyclohexane - 1.471 0.1597 - 9 . 5 5 x 10 -4 3.06 x 10 -6 0 12 146
Benzene 0.890 0.0752 - 3 . 2 3 x 10 -4 8.80 x 10 -7 0 40 279
Toluene -0.433 0.1557 - 1 . 0 5 x 10 -3 2.97 x 10 -6 0 40 274
Naphthalene 0.341 0,0949 - 3 . 7 9 x 10 -4 1.34 x 10 -6 - 1 . 3 4 x 10 -9 30 353
Anthracene 2.436 0.0531 1.04 x 10 -4 - 8 . 8 2 x 10 .8 3.69 • 10 -12 40 489
Carbon dioxide -2.199 0.1636 - 1 . 4 6 x 10 .3 6.20 x 10.6 - 9 . 2 6 x 10 -9 25 216
Water - 3 . 1 5 7 • 10 -2 0.0169 0 0 0 3 273
CL CL w h e r e Kw is t h e W a t s o n c h a r a c t e r i z a t i o n f a c t o r d e f i n e d i n
(7.41) -r -sat _ e x p ( 2 . 1 T r - 17.9) + e x p ( 8 . 6 5 5 T r - 8.385)
R Eq. (2.13). P r e l i m i n a r y c a l c u l a t i o n s s h o w t h a t t h i s e q u a t i o n
overpredicts values of CL of pure hydrocarbons and accuracy
w h e r e C } s h o u l d b e c a l c u l a t e d f r o m E q . (7.40). W h e n Tr < o f t h i s e q u a t i o n is a b o u t 5%. E q u a t i o n (7.42) is r e c o m m e n d e d
0,8, i t c a n b e a s s u m e d t h a t C} ~- CsLt a n d t h e c o r r e c t i o n t e r m i n t h e A S T M D 2 8 9 0 t e s t m e t h o d f o r c a l c u l a t i o n o f h e a t ca-
m a y b e n e g l e c t e d , csLt r e p r e s e n t s t h e e n e r g y r e q u i r e d w h i l e p a c i t y o f p e t r o l e u m d i s t i l l a t e f u e l s [25]. T h e r e a r e o t h e r f o r m s
m a i n t a i n i n g t h e l i q u i d i n a s a t u r a t e d state, M o s t o f t e n csLt s i m i l a r t o E q . (7.42) c o r r e l a t i n g CpL of p e t r o l e u m f r a c t i o n s t o
is m e a s u r e d e x p e r i m e n t a l l y w h i l e m o s t p r e d i c t i v e m e t h o d s SG, Kw, a n d T u s i n g h i g h e r t e r m s a n d o r d e r s f o r t e m p e r a t u r e
e s t i m a t e C } [12]. b u t g e n e r a l l y give s i m i l a r r e s u l t s as t h a t of Eq. (7.42). S i m p l e r
For petroleum fractions the pseudocomponent method f o r m s of r e l a t i o n s f o r e s t i m a t i o n o f CeL o f l i q u i d p e t r o l e u m
s i m i l a r t o Eq, (7.35) c a n b e u s e d w i t h M o r Tu o f t h e f r a c t i o n f r a c t i o n s i n t e r m s o f S G a n d T a r e also a v a i l a b l e i n t h e lit-
as a c h a r a c t e r i s t i c p a r a m e t e r . H o w e v e r , t h e r e a r e s o m e g e n - e r a t u r e [26]. B u t t h e i r a b i l i t y t o p r e d i c t CpL is v e r y p o o r a n d
eralized correlations developed particularly for estimation of in some cases lack information on the units or involve with
h e a t c a p a c i t y o f l i q u i d p e t r o l e u m f r a c t i o n s . K e s l e r a n d Lee some errors in the coefficients reported. The corresponding
[24] d e v e l o p e d t h e f o l l o w i n g c o r r e l a t i o n f o r C } o f p e t r o l e u m s t a t e s c o r r e l a t i o n o f Eq. (7.40) m a y also b e u s e d f o r c a l c u l a -
fractions at low pressures: t i o n o f h e a t c a p a c i t y o f l i q u i d p e t r o l e u m f r a c t i o n s u s i n g To,
w, a n d C~g o f t h e f r a c t i o n . T h e API m e t h o d [9] f o r c a l c u l a t i o n
CLp = a (b + cT) o f C L of l i q u i d p e t r o l e u m f r a c t i o n s is g i v e n i n t h e f o l l o w i n g
7. APPLICATIONS: E S T I M A T I O N OF T H E R M O P H Y S I C A L P R O P E R T I E S 321
form for Tr _< 0.85: and Cpg --- 1.419 J/g. ~ (at 20~ From Tb and SG, Kw =
12.264.
C~ = A1 + AzT + A3T2
(a) To use SRK EOS use equations given in Table 6.1 and
Ax = -4.90383 + (0.099319 + 0.104281SG)Kw follow similar calculations as in Example 6.2: A = 0.039685,
B = 0.003835, Z L = 0.00492, V L = 118.3 cm3/mol, the volume
(4.81407 - 0.194833 K~ translation is c = 13.5 cm3/mol, VL(corrected) = 104.8 cm3/
+ mol, and ZL(correc.)=0.00436. From Table 6.1, P1 =
\
6.9868, P2 = -103.976, P3 = 0.5445, and [Cp - C~g] = 21.41
0.27634)
A2 = (7.53624 + 6.214610Kw) x (1.12172 - ] • 10 -4 J/mot-K =21.41/68.12 = 0.3144 J/g.K. CL = 0 . 3 1 4 4 +
1.419 = 1.733 J / g . K (error of -13%). (b) DIPPR [10] gives
( 0.70958] the value of CL = 2.138 J/g-K (error of +7%). (c) From the
A3 = -(1.35652 + 1.11863Kw)x 2.9027 ~ ] x 10 -7
Lee-Kesler correlation of Eq. (6.57), Tr = 0.612 and Pr =
(7.43) 0.0271. From Tables 6.4 and 6.5, using interpolation (for Pr)
and extrapolation (for Tr, extrapolation from the .liquid re-
where C~ is in kJ/kg 9K and T is in kelvin. This equation was gion) we get [(Cp - c~g)/R] (~ = 1.291 and [(Cv - c~g)/R] (1) =
developed by Lee and Kesler of Mobile Oil Corporation in 5.558. In obtaining these values special care should be made
1975. From this relation, the following equation for estima- not to use values in the gas regions. From Eqs. (6.57) and
tion of enthalpy of liquid petroleum fractions can be obtained. (7.38) using parameters R, M, and Cpg we get CpL = 1.633
T (error of-18%). (d) From Eq. (7.40), [(CL - Cpg)/R] = 3.9287,
H L = f CLdT + H~Lf= AI(T - Tref)+ 2 ( T - Tref) 2
CpL = 1.899 (error of -4.8%). (e) From Eq. (7.42), a = 4.2884,
b = 0.19547, c = 0.0011, C L = 2.223 J/g. K (error of +11.5%).
Tref
This is the same as ASTM D 2890 test method. (f) From
Tsonopoulos correlation, Eq. (7.45), CL = 2.127 J / g - K (error
(7.44) + ~-~3(T- T~f)3 + H;Lf
of +6.6%). The generalized Lee-Kesler correlation (Eq.
where HrL~is usually zero at the reference temperature of T~f. 6.57) gives very high error because this method is mainly
Equation (7.43) is not recommended for pure hydrocarbons. accurate for gases. For liquids, Eq. (7.40) is more accurate
The following modified form of Watson and Nelson correla- than is Eq. (6.57). Equation (7.45) although recommended
tion is recommended by Tsonopoulos et al. [18] for calcula- for coal liquids predicts liquid heat capacity of hydrocarbons
tion of liquid heat capacity of coal liquids and aromatics: relatively with relative good accuracy. r
C L = (0.28299 + 0.23605Kw) There are some other methods developed for calculation of
C~. In general heat capacity of a substance is proportional to
x I0.645 - 0.05959 SG + (2.32056 - 0.94752 SG) molar volume and can be related to the free space between
molecules. As this space increases the heat capacity decreases.
Since parameter I (defined by Eq. 2.36) also represents mo-
• (1-~00 - 0-25537)] lar volume occupied by the molecules Riazi et al. [27] showed
that CL varies linearly with 1/(1 - I). They obtained the fol-
(7.45)
lowing relation for heat capacity of homologous hydrocarbon
where C~ is in kJ/kg- K and T is in kelvin. This equation pre- groups:
dicts heat capacity of coal liquids with an average error of
about 3.7% for about 400 data points [18]. The following ex- (7.46) C+A =c( a l M
l +M
b l ) • + Id~ - Il) R
ample shows various methods of calculation of heat capacity
of liquids. In the above relation M is molecular weight, R is the gas con-
stant, and coefficients al-dl are specific for each hydrocarbon
Example 7.6--Calculate CL of 1,4-pentadiene at 20~ using family. Parameters I is calculated throughout Eqs. (2.36) and
the following methods and compare with the value of 1.994 (2.118) at the same temperature at which CL is being cal-
J/g. ~ reported by Reid et al. [12]. culated. Parameters al-dl for different hydrocarbon families
a. SRK EOS and solid phase are given in Table 7.8.
b. DIPPR correlation [10]
c. Lee-Kesler generalized corresponding states correlation
7.4.3 Heats of Phase Changes--Heat
(Eq. 6.57)
of Vaporization
d. Bondi's correlation (Eq. 7.40)
e. Kesler-Lee correlation (Eq. 7.42)--ASTM D 2890 method Generally there are three types of phase changes: solid to
f. Tsonopoulos et al. correlation (Eq. 7.45) liquid known as fusion (or melting), liquid to vapor (vapor-
ization), and solid to vapor (sublimation), which occurs at
Solution--Basic properties of 1,4-pentadiene are not given in pressures below triple point pressure as shown in Fig. 5.2a.
Table 2.1. Its properties obtained from other sources such as During phase change for a pure substance or mixtures of con-
DIPPR [10] are as follows: M = 68.1185, Tb = 25.96~ SG = stant composition, the temperature and pressure remain con-
0.6633, Tc = 205.85~ Pc = 37.4 bar, Zc = 0.285, w = 0.08365, stant. According to the first law of thermodynamics, the heat
322 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 7.8--Constants for estimation of heat capacity from refractive index (Eq. 7.46).
Cp/R = (aiM + bl)[I/(1 - 1)] + cl M + dl
No. of
Group State Carbon range Temp. r a n g e , ~ al b~ cl d~ d a t a points AAD% MAD%
n-Alkanes Liquid C5-C20 - 15-344 -0.9861 -43.692 0.6509 5.457 225 0.89 1.36
1-Alkenes Liquid C5-C20 -60-330 -1.533 40.357 0.836 -21.683 210 1.5 5.93
n-Alkyl-cyclopentane Liquid Cs-C20 -75-340 -1.815 56.671 0.941 -28.884 225 1.05 2.7
n-Alkyl-cyclohexane Liquid C6-C20 -100-290 -2.725 165.644 1.270 -68.186 225 1.93 2.3
n-Alkyl-benzene Liquid C6-C20 -250-354 -1.149 4.357 0.692 -3.065 225 1.06 4.71
n-Alkanes Solid C5-C20 -180-3 -1.288 -66.33 0.704 14.678 195 2.3 5.84
AAD%: Average absolute deviation percent. MAD%: M a x i m u m a b s o l u t e deviation percent. Coefficients are t a k e n f r o m Ref. [27]. D a t a source: D I P P R [10].
TABLE 7.9~Coefficients of Eq. 7.50 for calculation of enthalpy of vaporization of pure compounds versus temperature f9].
AHTaPEkJ/kg] = A(l - Tr)B+CTr
Compound A B C Compound A B C
Water 2612.982 -0.0577 0.3870 Methylcyclopentane 527.6931 0.3967 0.0000
Ammonia 1644.157 - 0.017 0.3739 Ethylcyclopentane 502.1246 0.3912 0.0000
HaS 754.073 0.3736 0.0000 Pentylcyclopentane 442.0789 0.3800 0.0000
CO2 346.1986 -0.6692 0.9386 Decylcyclopentane 397.8670 0.3800 0.0000
N2 228.9177 - 0.1137 0.4281 Pentadecylcyclopentane 372.6050 0.3800 0.0000
CH4 570.8220 -0. l 119 0.4127 Cyclohexane 534.5225 0.3974 0.0000
C2H6 588.1554 0.0045 0.3236 Methylcyclohexane 503.9656 0.4152 0.0000
C3H8 610.2175 0.3649 0.0000 Ethylene 679.2083 0.3746 0.0000
n-C4Hl0 568.6540 0.3769 0.0000 Propylene 539.9479 0.0169 0.0000
n-CsH12 540.6440 0.3838 0.0000 Benzene 651.8210 0.6775 -0.2695
n-C6H14 515.2685 0.3861 0.0000 Toluene 544.7929 0.3859 0.0000
n-C7H16 497.0039 0.3834 0.0000 Ethylbenzene 515.2839 0.3922 0.0000
n-Call18 489.0450 0.4004 0.0000 o-Xylene 521.7788 0.3771 0.0000
n-C10H22 461.4396 0.3909 0.0000 Propylbenzene 500.4582 0.3967 0.0000
n-C15H32 431.6786 0.4185 0.0000 n-Butylbenzene 470.0009 0.3808 0.0000
n-C20H42 407.3617 0.4089 0.0000 n-Octylbenzene 456.0581 0.4281 0.0000
Cyclopentane 517.7318 0.1808 0.1706 Naphthalene 371.4852 -0.3910 0.0000
f(~ a n d f(1)(T~) are c o r r e l a t e d to (I - Tr), w h e r e as T~ ~ 1, meters. Once the value of A "~bp ~rvap is calculated, it s h o u l d be
AHvap/RTc ~ O. However, m o r e a c c u r a t e predictive m e t h o d s divided b y M to convert its unit from kJ/krnol to kJ/kg.
are developed in two steps. In the first step heat of vaporiza- E q u a t i o n (7.56) with coefficients given in Table 7.10 can be
tion at n o r m a l boiling point, A/-/n~bp, is c a l c u l a t e d a n d then used for fractions with m o l e c u l a r weight range of 70-300 (~Tb
c o r r e c t e d to the desired t e m p e r a t u r e b y a second correla- of 300-600 K) with a c c u r a c y of a b o u t 2% w h e n tested against
tion. One of the m o s t successful correlations for p r e d i c t i o n 138 p u r e h y d r o c a r b o n s . Application of the e q u a t i o n can be
of AHnV~p was p r o p o s e d by Riedel [12]: extended up to 700 K with r e a s o n a b l e accuracy. Once AH~b~p v
is d e t e r m i n e d , the W a t s o n relation can be used to calculate
(7.54) AHn~pp -- 1.093RTcTbrl n-0-~-_-~
e o - 1.013 A/-Fap at the desired t e m p e r a t u r e (T).
TABLE 7.11--Experimental data on heat of vaporization of some coal liquid fractions with calculated
basic parameters [28].
Fraction (a) Tb, K SG AH~ p, kJ/kg M Tr K Pc, bar co
5HC 433.2 0.8827 309.4 121.8 649.1 33.1 0.302
8HC 519,8 0.9718 281,4 162.7 748.1 27.1 0,394
11HC 612,6 1.0359 269,6 223,1 843.1 21.5 0,512
16HC 658.7 1,0910 245,4 247.9 896.4 20.5 0.552
17HC 692.6 1.1204 239,3 272.0 932.2 19.4 0,590
M from Eq. (2.51), Tr and Pc from Eqs. (2.63) and (2.64), ~ofrom Eq. (2.108). Experimental value on Tb, SG, and A Hnbt~
are taken from J. A. Gray, Report DOEfET/10104-7, April 1981; Department of Energy, Washington, DC and are also
given in Ref. [28].
AHT~p in the temperature range of 350-550 K are given in Solution--The enthalpy departure from SRK is given in Table
Ref. [28]. Predicted values from Eq. (7.57) with use of differ- 6.1. If it is applied to both saturated vapor and saturated liquid
ent methods for calculation of AHV~p p as given in Table 7.12 at the same temperature and pressure and subtracted from
are c o m p a r e d graphically in Fig. 7.15. The average deviations each other based on Eq, (7.49) we get:
for the Riedel, Vetre, Riazi-Daubert, and Lee-Kesler are 1.5,
H v - H L = A H yap = RT (Z v - Z L)
1.8, 1,9 and 1.7%, respectively. The data show that the Riedel
m e t h o d gives the best result for both A Hvap
~bp and A HTvap when a _ Zv ZL
the latter is calculated from the Watson method. (7.59) +(~ Tb) [ln ( ~ ) - In ( Z ~ - - ~ B
+) ]
As a final method, A/-F~p can be calculated from Eq. (7.49)
by calculating residual enthalpy for both saturated vapor and where al is da/dT as given in Table 6.1 for the SRK EOS.
liquid from an equation of state. This is demonstrated in the Replacing for Z = PV/RT and B = bP/RT and considering that
following example for calculation of A n yap from SRK EOS. the ratio of VV/(v v + b) is nearly unity (since b << VV):
one c o m p o u n d (i.e., H20 or CO2). However, when a hydro- HHV = 121 + 0 . 2 2 x 1 0 0 - - 1 4 3 lo/g or 61000 Btu/lb. The
carbon (CxHy) is burned the only products are H20 and CO2. heating values of other fuels are calculated as follows:
Combustion is a reaction in which the enthalpy of products is
less than enthalpy of reactants and as a result the heat of reac- Hydrogen, Methane, Propane, Carbon, Sulfur,
tion (enthalpy of products - enthalpy of reactants) is always Fuel H2 CH4 C3I-I8 C S
negative. This heat of reaction is called heat of combustion LI-I~, kJ/g 121 50 46.4 32.8 9.3
HHV, kJ/g 143 55.5 50.4 32.8 9.3
and is shown by zXHc. Heat of combustion depends on the
temperature at which the combustion takes place. The stan-
dard temperature at which usually values of A H c are reported As it can be seen from these calculations, hydrogen has the
is 25~ (298 K). highest heating value and carbon has the lowest heating value.
Amount of heat released by burning one unit mass (i.e., kg, Thus hydrogen is the best, while carbon is considered as the
g, or lb) of a fuel is called heating value or calorific value and worst fuel. Sulfur heating value is even less than carbon but
has the unit of lo/kg or Btu/lb (1 lo/kg = 0.42993 Btu/lb). In sulfur is not really considered as a fuel. Some values of HHV
some cases for liquid fuels the heating values are given per for several other fuels as reported by Felder and Rousseau
unit volume (i.e., lOlL of fuel), which differs from specific [32] are given in Table 7.13. The calculated value of HHV of
(mass unit) heating values by liquid density. If in the combus- C is near the HHV of hard coal (i.e., solid form) as given in
tion process produced H20 is considered as liquid, then the Table 7.13. In natural gases since there are some hydrocar-
heat produced is called gross heat of combustion or higher bons heavier than methane, its heating value is somewhat
heating value (HHV). When produced H20 is considered as lower than that of pure methane, t
vapor (as in the actual cases), then the heat produced is called
lower heating value (LHV). The LHV is also k n o w n as the net Example 7.8 shows that the heating value generally in-
heating value (NHV). The difference between HHV and LHV is creases as the hydrogen content of fuel increases and car-
due to the heat required to vaporize produced water from liq- bon content decreases. In other words, as CH weight ratio
uid to vapor form at the standard temperature (43.97 kJ/mol increases the heating value decreases. Furthermore, presence
or 2.443 lo/g of H20). The a m o u n t of H20 formed depends on of sulfur further reduces the heating value. For this reason,
the hydrogen content of fuel. If the hydrogen wt% of fuel is some researchers have correlated HHV to wt% of C, H, S, N,
H% then the relation between HHV and LHV is given as [30]: and O content of fuel. For example, Tsonopoulos et al. [18]
proposed the following relation for estimation of HHV of coal
(7.61) LHV = HHV - 0.22H% liquids:
HHV [lo/g] = 0.3506 (C%) + 1.1453 (H%) + 0.2054 (S%)
where both LHV and HHV are in kJ/g. The heating values can
(7.62) + 0.0617 (N%) - 0.0873 (0%)
also be determined from standard heats of formation (A H2f98).
Values of AHaf98 for any element (i.e., H2, 02, C, S, etc.) is S, N, and O are usually found in heavy fuels and aromatic rich
zero and for formed molecules such as H20 are given in fuels such as coal liquids. This equation predicts HHV of coal
most thermodynamics references [12, 21, 31]. For example, liquids with %AAD of 0.55 for some 130 fuels. This equation
for H20(g), CO2(g), CO(g), SO2, CH4(g), C2H6(g), C3H8(g), predicts HHV of pure C as 35 lo/g. However, this equation is
and n-CmH22 the respective values of AHf98 are -241.81, not r e c o m m e n d e d for light fuels, petroleum fraction, or pure
-393.51, -110.53, -296.81, -74.52, -83.82, -104.68, and compounds. There is a simpler relation for calculation of LHV
- 2 4 9 . 4 6 lo/mol. The following example shows calculation of of heavy fuels and petroleum fractions [30]:
heating values from heats of formations.
(7.63) LHV[lo/g] = 55.5 - 14.4 • SG - 0.32S%
where S% is the sulfur wt% in the fuel. A very simple but
Example 7.8--Calculate HHV and LHV of hydrogen, me-
approximate formula for calculation of HHV of crude oils is
thane, propane, carbon, and sulfur from heats of formation.
[26]:
Solution--Here the calculation of heating value of CH4 (7.64) HHV = 51.9 - 8.8 • SG 2
is demonstrated and a similar approach can be used for where HHV is in lo/g (or MJ/kg) and SG is the specific gravity
other fuels. The chemical reaction of combustion of CH4 is of crude and S% is the sulfur wt% of the crude. Accuracy of
CH4(g) + 202(g) -+ 2H20(g) + CO2(g) + A H c, where A H c = these equations is usually about 1%. A typical crude oil has
2AH2f98(H2 O) + AH2f98(CO2) -- z~sf98(CH4) - AS2f98(O2) ~---2x heating value of about 10 500 cal/g (~44 lo/g). Increase in hy-
(-241.835) + (-393.51) - (-74.8936) - ( 0 ) = - 8 0 2 . 2 8 6 k J / drogen content of a fuel not only increases the heating value
mol. Since the produced water is assumed to be in gas phase
so the LHV is calculated as 802.286/16.04 = 50.01 lo/g.
This is equivalent to 11953 cal/g or 21500 Btu/lb. The TABLE 7.13--Heating values of some fitels.
H H V can be calculated by adding heat of vaporization of Taken with permission from Ref. [32].
Higher heating value
water (2 x 43.97 = 87.94 kJ/mol) to the molar LHV. HHV =
Fuel kJ/g Btu/lb
802.286 + 87.94 = 890.2 kJ/mol or 55.5 kJ/g of CH4. Equa-
Wood 18 7700
tion (7.61) to convert LHV to HI-IV or vice versa using Hard coal 35 15 000
H% of fuel m a y also be used. In this case, H% of CH4 = Crude oil 44 19 000
(4/16) x 100 = 25 wt%. Thus HHV = 50 + 0.22 x 25 = 55.5 Natural gas 54 23 000
kJ/g. Similarly for H2, LHV = 241.81/2.0 = 121 kJ/g. The Hydrogen 143 61 000
326 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
of the fuel but also decreases a m o u n t of u n b u r n e d hydrocar- can be estimated with greater accuracy than for properties of
bon and CO/CO2 production. For these reasons, natural gas liquids. This is mainly due to better understanding of inter-
is considered to be a clean fuel but the cleanest and m o s t molecular forces in gaseous systems. Similarly properties of
valuable fuel is hydrogen. This is the reason for global ac- gases at low pressures can be estimated with better accuracy
celeration of development of hydrogen fuel cells as a clean in comparison with gases at high pressures. Effect of pres-
energy, although still production of energy from hydrogen is sure on properties of gases is m u c h greater than the effect
very costly [33]. of pressure on properties of liquids. At high pressures as we
HHV can be measured in the laboratory t h r o u g h combus- approach the critical region properties of gases and liquids
tion of the fuel in a b o m b calorimeter surrounded by water. approach each other and under such conditions a unique gen-
Heat produced can be calculated from the rise in the tempera- eralized correlation for both gases and liquids termed dense
ture of water. The experimental procedure to measure heating fluids m a y be used for prediction of properties of both gases
value is explained in ASTM D 240 test method. The heating and liquids. For wide boiling range fractions or crude oils the
value of a fuel is one of the characteristics that determines mixture should be split into a n u m b e r of p s e u d o c o m p o n e n t s
price of a fuel. and treat the fluid as a defined mixture. A more accurate ap-
proach would be to consider the fluid as a continuous mixture
W h e n using a t h e r m o d y n a m i c model, cubic equations of
7.5 S U M M A R Y A N D R E C O M M E N D A T I O N S state (i.e., PR or SRK) should be used for calculation of
PVT and equilibrium properties at pressures greater than
In this chapter, application of methods and procedures pre- about 13 bars (~200 psia). At low pressures and especially
sented in the b o o k for calculation and estimation of various for liquids, properties calculated from a cubic EOS are not
thermophysical properties are shown for pure hydrocarbons reliable. For liquid systems specific generalized correlations
and their defined mixtures, natural gases and nonhydrocar- developed based on liquid properties are m o r e accurate than
bon gases associated with t h e m (i.e. H2S, CO2, N2, H20), other methods. It is on this basis that Rackett equation pro-
defined and undefined petroleum fractions, crude oils, coal rides more accurate data on liquid density than any other
liquids, and reservoir fluids. Characterization methods of correlation. In application of EOS to petroleum mixtures the
Chapters 2-4 and t h e r m o d y n a m i c relations of Chapters 5 and BIPs especially for the key components have significant im-
6 are essential for such property calculations. Basically, ther- pact on accuracy of predicted results. Wherever possible BIP
mophysical properties can be estimated through equations of of key components (i.e., C1-C7+ in a reservoir fluid) can be
state or generalized correlations. However, for some special tuned with available experimental data (i.e., density or satu-
cases empirical methods in the forms of graphical or analyt- ration pressure) to improve prediction by an EOS model [34].
ical correlations have been presented for quick estimation of
certain properties. TABLE 7.14---Summary of recommended methods ]or various
Methods of prediction of properties introduced in the pre- properties.
vious chapters such as density, enthalpy, heats of vaporization Property Methods of estimation for various fluids
and melting, heat capacity at constant pressure and volume, Density 9 Eq. (7.3) for gases with Lee-Kesler
generalized correlation for calculation
vapor pressure, and fuels' heating values are presented.
of Z (Ch 5), also see Section 7.2.1.
For calculation of properties of pure components w h e n a 9 For pure liquid hydrocarbons, Table
correlation for a specific c o m p o u n d is available it must be 2.1 and Eq. (7.5) or Rackett equation.
used wherever applicable. Generalized correlations should 9 Eq. (7.4) for defined liquid mixture.
be used for calculation of properties of pure hydrocarbons 9 For petroleum fractions use Rackett
equation, Eq. (7.5), or Figs. 7.1-7.3.
w h e n specific correlation (analytical or graphical) for the 9 See Section 7.2.2 for other cases.
given c o m p o u n d is not available. For defined mixtures the Vapor pressure 9 Eq. (7.8) for pure compounds and
best way of calculation of mixture properties when experi- if coefficients are not known use
mental data on properties of individual components of the Eqs. (7.18) or (7.19).
mixture are available is t h r o u g h appropriate mixing rules for 9 Use Eqs. (7.20) and (7.22) for
petroleum fractions and Eqs. (7.21)
a given property using pure components properties and mix- and (7.23) for coal liquids.
ture composition. For defined mixtures wherein properties 9 For crude oils use Eq. (7.26) or
of pure components are not available, the basic input pa- Fig. 7.11.
rameters for equations of states or generalized correlations Enthalpy 9 Use Eq (7.32) and Fig. 7.13 with
Lee-Kesler correlations of Ch 6 for
should be calculated from appropriate mixing rules given in petroleum fractions.
Chapter 5. These basic properties are generally Tr Pc, Vc, w, 9 Use Eq. (7.34) for A H ig.
M, and C~g, which are k n o w n for pure components. For 9 For special cases see Section 7.4.1.
petroleum fractions these parameters should be estimated Liquid heat capacity, CL 9 Eq. (7.40) for pure compounds.
and the m e t h o d of their estimations has a great impact on 9 Eq. (7.43) for petroleum fractions.
9 Eq. (7.45) for coal liquids.
accuracy of predicted physical properties. In fact the impact AHvap 9 Eq. (7.50) or Eqs. (7.54) and (7.57) for
of estimation of basic input properties is greater than the im- pure compounds.
pact of selected t h e r m o d y n a m i c m e t h o d on the accuracy of 9 Eqs. (7.54) and (7.57) for petroleum
property predictions. fractions.
For prediction of properties of petroleum fractions, spe- 9 Eqs. (7.58) and (7.57) for coal liquids.
Heating value See Section 7.4.4.
cial methods are provided for undefined mixtures. For both These recommendations are not general and for special cases one should see
pure c o m p o u n d s and petroleum mixtures, properties of gases specific recommendationsin each section.
7. APPLICATIONS: E S T I M A T I O N OF THERMOPHYSICAL P R O P E R T IE S 3 2 7
7.6 PROBLEMS
VV
(7.65) psat(vV -- vL) ~ fvL pEOS(dV) T
Applications: Estimation
of Transport Properties
329
recovery and for process engineers it can be used to determine for fluid-flow calculations, it is needed in calculation of other
foaming characteristics of hydrocarbons in separation units. properties such as diffusion coefficient. Experimental values
As was discussed in Chapter 7, properties of gases may of gas viscosity at 1 atm versus temperature for several hydro-
be estimated more accurately than can the properties of liq- carbon gases and liquids are shown in Fig. 8.1. As it is seen
uids. Kinetic theory provides a good approach for develop- from this figure, viscosity of liquids increases with molecular
ment of predictive methods for transport properties of gases. weight of hydrocarbon while viscosity of gases decreases.
However, for liquids more empirically developed methods are
used for accurate prediction of transport properties. Perhaps
8.1.1 Viscosity of Gases
combination of both approaches provides most reliable and
general methods for estimation of transport properties of flu- Viscosity of gases can be predicted more accurately than can
ids. For petroleum fractions and crude oils, characterization the viscosity of liquids. At low pressures (ideal gas condition)
methods should be used to estimate the input parameters. It viscosity can be well predicted from the kinetic theory of gases
is shown that choice of characterization method may have [1, 3, 4].
a significant impact on the accuracy of predicted transport
property. Use of methods given in Chapters 5 and 6 on the 2
(8.2) / ~ - 3jr2/3 d2
development of a new experimental technique for measure-
ment of diffusion coefficients in high-pressure fluids is also where m is the mass of one molecule in kg (m = O.O01M/NA),
demonstrated. kB is the Boltzmann constant (= R/NA), and d is the molecular
diameter. In this relation if m is in kg, kB in J/K, T in K, and d
in m, then/x would be in kg/m. s (1000 cp). This relation has
8.1 E S T I M A T I O N OF VISCOSITY been obtained for hard-sphere molecules. Similar relations
can be derived for viscosity based on other relations for the
Viscosity is defined according to the Newton's law of viscosity: interrnolecular forces [ 1]. The well-known Chapman-Enskog
avx a @vx) equations for transport properties of gases at low densities
(8.1) ~yx=-~-y = - ~ Oy ~=-p (low pressure) are developed on this basis by using Lennard-
Jones potential function (Eq. 5.11). The relation is very sim-
where Zyxis the x component of flux of momentum in the y di- ilar to Eq. (8.2), where/x is proportional to (MT)I/2/(a2S2) in
rection in which y is perpendicular to the direction of flow x. which a is the molecular collision diameter and f2 is a func-
Velocity component in the x direction is V~. p is the density tion of kBT/~. Parameters a and E are the size and energy
and pV~ is the specific momentum (momentum per unit vol- parameters in the Lennard-Jones potential (Eq. 5.1 i). From
ume). OV~/Oyis the velocity gradient or shear rate (with di- such relations, one may obtain molecular collision diameters
mension of reciprocal of time, i.e., s -1) and 3(pVx)/Oy is gra- or potential energy parameters from viscosity data. At low
dient of specific momentum, r represents tangent force to pressures, viscosity of gases changes with temperature. As
the fluid layers and is called shear stress with the dimension shown by the above equation as T increases, gas viscosity also
of force per unit area (same as pressure). Velocity is a vector increases. This is mainly due to increase in the intermolecular
quantity, while shear stress is a tensor quantity. While pres- collision that is caused by an increase in molecular friction.
sure represents normal force per unit area, r represents tan- At high pressures, the behavior of the viscosity of gases and
gent (stress) force per unit area, which is in fact the same as liquids approach each other.
momentum (mass x velocity) per unit area per unit time or For pure vapor compounds, the following correlation was
momentum flux. Thus/x has the dimension of mass per unit developed by the API-TDB group at Penn State and is recom-
length per unit time. For example, in the cgs unit system it has mended in the API-TDB for temperature ranges specified for
unit of g/cm. s, which is called poise. The most widely used each compound [5]:
unit for viscosity is centipoise (1 cp = 0.01 p = 10-4 micro-
poise). The ratio of Ix/p is called kinematic viscosity and is 1000AT a
(8.3) " = (1 + ~C + ~)
D
usually shown by v with the unit of stoke (cm2/s) in the cgs
unit system. The common unit for v is cSt (0.01 St), which
where correlation coefficients A-D are given for some selected
is equivalent to mm2/s. Liquid water at 20~ exhibits a vis-
compounds in Table 8.1. The average error over the entire
cosity of about 1 cp, while its vapor at atmospheric pressure
temperature range is about 5% but usually errors are less
has viscosity of about 0.01 cp. More viscous fluids (i.e., oils)
than 2%. This equation should not be applied at pressures in
have viscosities higher than the viscosity of water at the same
which Pr > 0.6.
temperature. Fluids that follow linear relation between shear
The following relation developed originally by Yoon and
stress and shear rate (i.e., Eq. 8.1) are called Newtonian. Poly-
Thodos [6] is recommended in the previous editions of API-
mer solutions and many heavy oils with large amount of wax
TDB and DIPPR manuals for estimation of viscosity of hy-
or asphahene contents are considered non-Newtonian and
drocarbons as well as nonhydrocarbons and nonpolar gases
follow other relations between shear stress and shear rate.
at atmospheric pressures:
Viscosity is in fact a measure of resistance to motion and the
reciprocal of viscosity is called fluidity. Fluids with higher vis- /z~ x 10s = 1 + 46.1T~ - 20.4 exp(-0.449Tr)
cosity require more power for their transportation. Viscosity
(8.4) + 19.4 exp(-4.058Tr)
is undoubtedly the most important transport property and it
has been studied both experimentally and theoretically more 1 1 2
than other transport properties. In addition to its direct use (8.5) = Tc6M-2 (0.987Pc)-~
332 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
3o0
250 9 J
u)
f j
0
r
u2 200'
_._ .,,,"1~ f J
E
"6
~0 lo0
YS
>
50'
~oise = 0.0672E- 6 Ib/fts ]
o
[
200 0 200 400 600 800 1000
Temperature, C
(a) Gases
\ ,' I
i[1 centipoise= 0.000672Ib/ftsj
I
O
rl
\
C X'x,,~
8
10
~2r
."1 " ~ i'~. N'%"~.,~ IPr~ IButanq
"6
0.1 L
i \
13-,
8
--~Metl~ane; ~.~. AEthanq
0,01
-200 -150 -1 oo -50 0 50 1oo 150 200
Temperature, C
(b) Liquids
FIG. 8.1--Viscosity of several light hydrocarbons versus temperature at atmo-
spheric pressure. Taken with permission from Ref. [2].
w h e r e / z is in cp, T~ is the reduced temperature, and ~ is a where units of ~ and ~ are the same as in Eq. (8.4). For de-
parameter that has a dimension of inverse viscosity and is ob- fined gas mixtures at low pressures, Eq. (8.6) m a y be used
tained from kinetic theory of gases. The factor 0.987 comes with To, Pc, and M calculated from Kay's mixing rule (Eq. 7.1).
from the original definition that unit of atm was used for Pc. In However, w h e n viscosity of individual gases in a mixture are
the above equation Pc, Tc, and M are in bar, kelvin, and g/reel, known, a more accurate m e t h o d of estimation of mixture vis-
respectively. In cases where data on gas viscosity is available, cosity is provided by Wilke, which can be applied for pressures
it would be more appropriate to determine ~ from viscosity with Pr < 0.6 [1, 5]:
data rather than to calculate it from the above equation. Reli-
N X//~ i
ability of this equation is about 3-5%. For some specific com- #m = Y~
pounds such as hydrogen, the numerical coefficients in Eq. i=l '~N=I Xj~ij
(8.4) are slightly different and in the same order as given in the (8.7)
DIPPR manual are 47.65, -20.0, -0.858, +19.0, and -3.995.
In the 1997 edition of the API-TDB [5], the more c o m m o n l y
#~ij=~ l+Mi } I+ # \Mi.]J
used correlation developed earlier by Stiel and Thodos [7] is
This semiempirical method is r e c o m m e n d e d by both API-
recommended:
TDB and DIPPR for calculation of viscosity of gas mixtures
/*~ = 3.4 x 10-4TrTM for Tr _< 1.5 of k n o w n composition at low pressures. Accuracy of this
(8.6)
/,~ = 1.778 x 10-4(4.58Tr - 1.67) 0.625 f o r Tr > 1.5 equation is about 3% [5, 8]. In the above relation dPii:# dPii.
8. APPLICATIONS: ESTIMATION OF TRANSPORT PROPERTIES 333
TABLE 8.1---Coefficients of Eq. (8.3) for viscosity of pure vapor compounds. (Taken with permission from Ref. [5].)
1000ArB Units: cp a n d kelvin (8.3)
/* = (l+Y.+TT_)
API No. Compound A B C D Train , K Tmax,K
794 Oxygen 1.1010E-06 5.6340E-01 9.6278E+0t 0.0000E+00 54 1500
781 Hydrogen 1.7964E-07 6.8500E-01 -5.8889E-01 1.4000E+02 14 3000
845 Water 6.1842E-07 6.7780E-01 8.4722E+02 -7.4074E+04 273 1073
771 Ammonia 4.1856E-08 9.8060E-01 3.0800E+01 0.0000E+00 196 1000
786 Hydrogen sulfide 5.8597E-08 1.0170E+00 3.7239E+02 -6.4198E+04 250 480
789 Nitrogen 6.5593E-07 6.0810E-01 5.4711E+01 0.0000E+00 63 1970
774 Carbon monoxide 1.1131E-06 5.3380E-01 9.4722E+01 0.0000E+00 68 1250
775 Carbon dioxide 2.1479E-06 4.6000E-01 2.9000E+02 0.0000E+00 194 1500
797 Sulfur trioxide 3.9062E-06 3.8450E-01 4.7011E+02 0.0000E+00 298 694
770 Air 1.4241E-06 5.0390E-01 1.0828E+02 0.0000E+00 80 2000
Paraffins
1 Methane 5.2553E-07 5.9010E-01 1.0572E+02 0,0000E+00 91 1000
2 Ethane 2.5904E-07 6.7990E-01 9.8889E+01 0.0000E+00 91 1000
3 Propane 2.4995E-07 6.8610E-01 1.7928E+02 -8,2407E+03 86 1000
4 n-Butane 2.2982E-07 6.9440E-01 2.2772E+02 -1,4599E+04 135 1000
5 Isobutane 6.9154E-07 5.2140E-01 2.2900E+02 0.0000E+00 150 1000
6 n-Pentane 6.3411E-08 8.4760E-01 4.1722E+01 0.0000E+00 143 1000
7 Isopentane 1.1490E-06 4.5720E-01 3.6261E+02 -4.9691E+03 113 1000
8 Neopentane 4.8643E-07 5.6780E-01 2.1289E+02 0,0000E+00 257 1000
9 n-Hexane 1.7505E-07 7.0740E-01 1.5711E+02 0,0000E+00 178 1000
10 2-Methylpentane 1.1160E-06 4.5370E-01 3.7472E+02 0,0000E+00 119 1000
14 n-Heptane 6.6719E-08 8.2840E-01 8.5778E+01 0,0000E+00 183 1000
15 2-Methylhexane 1.0130E-06 4.5610E-01 3.5978E+02 0.0000E+00 155 1000
23 n-Octane 3.1183E-08 9.2920E-01 5.5089E+01 0.0000E+00 216 1000
24 2-Methylheptane 4.4595E-07 5.5350E-01 2.2222E+02 0.0000E+00 164 1000
37 2,2,4-Trimethylpentane 1.1070E-07 7.4600E-01 7.2389E+01 0.0000E+00 166 1000
41 n-Nonane 1.0339E-07 7.7300E-01 2.2050E+02 0.0000E+00 219 i000
62 n-Decane 2.6408E-08 9.4870E-01 7.1000E+01 0.0000E+00 243 1000
73 n-Undecane 3.5939E-08 9.0520E-01 1.2500E+02 0.0000E+00 248 1000
74 n-Dodecane 6.3443E-08 8.2870E-01 2.1950E+02 0.0000E+00 263 1000
75 n-Tridecane 3.5581E-08 8.9870E-01 1.6528E+02 0.0000E+00 268 1000
76 n-Tetradecane 4.4566E-08 8.6840E-01 2.2822E+02 4.3519E+03 279 1000
77 n-Pentadecane 4.0830E-08 8.7660E-01 2.1272E+02 0.0000E+00 283 1000
78 n-Hexadecane 1.2460E-07 7.3220E-01 3.9500E+02 6.0000E+03 291 1000
79 n-Heptadecane 3.1340E-07 6.2380E-01 6.9222E+02 0.0000E+00 295 1000
80 n-Octadecane 3.2089E-07 6.1840E-01 7.0889E+02 0.0000E+00 301 1000
81 n-Nonadecane 3.0460E-07 6.2220E-01 7.0556E+02 0.0000E+00 305 1000
82 n-Eicosane 2.9247E-07 6.2460E-01 7.0278E+02 0.0000E+00 309 1000
86 n-Tetracosane 2.6674E-07 6.2530E-01 7.0000E+02 0.0000E+00 324 1000
9O n-Octacosane 2.5864E-07 6.1860E-01 6.9833E+02 0.0000E+00 334 1000
Naphthenes
101 Cyclopentane 2.3623E-07 6.7460E-01 1.3900E+02 0.0000E+00 179 1000
102 Methylcyclopentane 9.0803E-07 4.9500E-01 3.5589E+02 0.0000E+00 131 1000
103 Ethylcyclopentane 2.1695E-06 3.8120E-01 5.7778E+02 0.0000E+00 134 1000
109 n-Propylcyclopentane 2.6053E-06 3.4590E-01 5.8556E+02 0.0000E+00 156 1000
146 Cyclohexane 6.7700E-08 8.3670E-01 3.6700E+01 0.0000E+00 279 900
147 Methylcyclohexane 6.5276E-07 5.2940E-01 3.1061E+02 0.0000E+00 147 1000
148 Ethylcyclohexane 4.1065E-07 5.7140E-01 2.3011E+02 0.0000E+00 162 1000
156 n-Propylcyclohexane 9.7976E-07 4.5420E-01 3.8589E+02 0.0000E+00 178 1000
157 Isopropylcyclohexane 5.7125E-07 5.2610E-01 2.7989E+02 0.0000E+00 184 1000
158 n-Butylcyclohexane 5.3514E-07 5.2090E-01 2.7711E+02 0.0000E+00 198 1000
168 n-Oecylcyclohexane 3.3761E-07 5.4480E-01 2.0728E+02 0.0000E+00 272 1000
Olefins
192 Ethylene 2.0793E-06 4.1630E-01 3.5272E+02 0.0000E+00 169 1000
193 Propylene 8.3395E-07 5.2700E-01 2.8339E+02 0.0000E+00 88 1000
194 1-Butene 1.0320E-06 4.8960E-01 3.4739E+02 0.0000E+00 175 1000
198 l~ 1.6706E-06 4.1110E-01 4.3028E+02 0.0000E+00 108 1000
204 1-Hexene 1.3137E-06 4.3220E-01 4.0211E+02 0.0000E+00 133 1000
Diolefins and acetylene
292 1,3-Butadiene 2.6963E-07 6.7150E-01 1.3472E+02 0.0000E+O0 164 1000
322 Acetylene 1.2019E-06 4.9520E-01 2.9139E+02 0.0000E+00 192 600
Aromatics
335 Benzene 3.1347E-08 9.6760E-01 7.9000E+00 0.O000E+00 279 1000
336 Toluene 8.7274E-07 4.9400E-01 3.2378E+02 0.0000E+00 178 1000
337 Ethylbenzene 3.8777E-07 5.9270E-01 2.2772E+02 0.0000E+00 178 1000
(Continued)
334 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 8.1--(Continued)
API No. Compound A B C D Tmi~, K Tr~ax, K
338 o-Xylene 3.8080E-06 3.1520E-01 7.7444E+02 0.0000E+00 248 1000
339 m-Xylene 4.3098E-07 5.7490E-01 2.3861E+02 0.0000E+00 226 1000
340 p-Xylene 5.7656E-07 5.3820E-01 2.8700E+02 0.0000E+00 287 1000
341 n-Propylbenzene 1.6304E-06 4.1170E-01 5.4722E+02 0.0000E+00 173 1000
349 n-Butylbenzene 9.9652E-07 4.6320E-01 4.3278E-02 0.0000E+00 186 1000
371 n-Pentylbenzene 4.2643E-07 5.5740E-01 2.5900E+02 0.0000E+00 198 1000
372 n-Hexylbenzene 5.5928E-07 5.1090E-01 2.8722E+02 0.0000E+00 212 1000
373 n-Heptylbenzene 4.3188E-07 5.3580E-01 2.4561E-02 0.0000E+00 225 1000
374 n-Octylbenzene 5.4301 E - 07 4.9890E- 01 2.7711E - 0 2 0.0000E+00 237 1000
375 n-Nonylbenzene 4.8731E-07 5.0900E-01 2.6178E-02 0.0000E+00 249 1000
376 n-Decylbenzene 4.6333E-07 5.1060E-01 2.5611E-02 0.0000E+00 259 1000
377 n-Undecylbenzene 4.3614E-07 5.1410E-01 2.4761E-02 0.0000E+00 268 1000
378 n-Dodecylbenzene 3.7485E-07 5.2390E-01 2.1878E-02 0.0000E+00 276 1000
379 n-Tridecylbenzene 3.5290E-07 5.2760E-01 2.1039E-02 0.0000E+00 283 1000
384 Styrene 6.3856E-07 5.2540E-01 2.9511E+02 0,0000E+00 243 1000
342 Cumene 4.1805E-06 3.0520E-01 8.8000E+02 0.0000E+00 177 1000
Diaromatics
427 Naphthalene 6.4323E-07 5.3890E-01 4.0022E+02 0.0000E+00 353 1000
428 l-Methylnaphthalene 2.6217E-07 6.4260E-01 2.3522E+02 0.0000E+00 243 1000
474 Anthracene 7.3176E-08 7.5320E-01 1.0000E+00 0.0000E+00 489 1000
475 Phenanthrene 4.3474E-07 5.2720E-01 2.3828E+02 0.0000E+00 372 1000
Aromatics amines
746 Pyridine 5.2402E-08 9.0080E-01 6.2722E+01 0.0000E+00 232 1000
749 Quinoline 1.3725E-06 4.8350E-01 9.2389E+02 -6.7901E+04 511 1000
Sulfur
776 Carbonyl sulfide 2.2405E-05 2.0430E-01 1.3728E+03 0.0000E+00 134 1000
828 Methyl mercaptan 1.6372E-07 7.6710E-01 1.0800E+02 0.0000E+00 150 1000
891 Thiophene 1.0300E-06 5.4970E-01 5.6944E+02 0.0000E+00 235 1000
892 Tetrahydrothiophene 1.6446E-07 7.4400E-01 1.4472E+02 0.0000E+00 394 1000
Alcohols
709 Methanol 3.07E-007 6.9650E-001 2.0500E+02 0.0000E+00 240 1000
710 Ethanol 1.06E-006 8.0660E-001 5.2700E+02 0.0000E§ 200 1000
712 Isopropanol 1.99E-007 7.2330E-001 1.7800E+02 0.0000E+00 186 1000
766 Methyl-ten-butyl ether 1.54E-007 7.3600E-001 1.0822E+02 0.0000E+00 164 1000
A s i m p l e r version of Eq. (8.7) for a gas m i x t u r e is given as [9]: viscosity is (/z -/Za)~, w h i c h is also called as residual viscosity
(similar to residual h e a t capacity) a n d is usually c o r r e l a t e d to
(8.8) /J~om -- N the r e d u c e d density (Pr = P/Pc = Vc/V). F o r p u r e h y d r o c a r b o n
~i=1 xi~ gases at high pressures the following m e t h o d is r e c o m m e n d e d
w h e r e N is the total n u m b e r of c o m p o u n d s in the mix- in the API-TDB [5]:
ture, r = MiX/z, a n d s u b s c r i p t o indicates low p r e s s u r e (at-
(/s -- # a ) ~ ---- 1 . 0 8 X 10 -4 [exp (1.439p~) - exp ( - 1 . 1 lp1S58)]
m o s p h e r i c a n d below) while s u b s c r i p t m indicates m i x t u r e
property. By a s s u m i n g q~i = 1 this equation reduces to Kay's (8.11)
m i x i n g rule (#m = ~ xitzi), w h i c h usually gives a r e a s o n a b l y
a c c e p t a b l e result at very low pressure. The s a m e e q u a t i o n c a n he a p p l i e d to mixtures if To, Pc, M, a n d
Pressure has a good effect on the viscosity of real gases a n d Vc of the m i x t u r e are calculated from Eq. (7.1). V o r p can be
at a c o n s t a n t t e m p e r a t u r e w i t h increase in p r e s s u r e viscosity e s t i m a t e d f r o m m e t h o d s of C h a p t e r 5. F o r mixtures, in cases
also increases. F o r simple gases at high pressures, r e d u c e d t h a t there is at least one d a t a p o i n t on tz, it can be u s e d to ob-
viscosity (/~r) is usually c o r r e l a t e d to Tr a n d Pr b a s e d on the tain IZa r a t h e r t h a n to use its e s t i m a t e d value. E q u a t i o n (8.11)
t h e o r y of c o r r e s p o n d i n g states [1]. /Zr is defined as the ratio m a y also be u s e d for n o n p o l a r n o n h y d r o c a r b o n s as r e c o m -
of tz/tzc, w h e r e / z c is called critical viscosity a n d represents m e n d e d in the DIPPR m a n u a l [ 10]. However, in the API-TDB
viscosity of a gas at its critical p o i n t (Tc a n d Pc). a n o t h e r generalized c o r r e l a t i o n for n o n h y d r o c a r b o n s is given
3 1 2 in the form of tz/tZa versus Tr a n d Pr with s o m e 22 n u m e r i c a l
(8.9) /Zc = 6.16 • 10- (MTc)~(Vc)-~ constants. The advantage of this m e t h o d is m a i n l y simplicity
(8.10) /z~ = 7.7 • 10-4~ -1 in calculations since there is no n e e d to calculate Pr a n d IZ c a n
be directly c a l c u l a t e d t h r o u g h ~a a n d Tr a n d Pr.
I n the above relations,/zr is in cp, Tc in kelvin, Vc is in cm3/mol, I n the p e t r o l e u m i n d u s t r y one of the m o s t widely u s e d cor-
a n d ~ is defined by Eq. (8.5). E q u a t i o n (8.10) can be o b t a i n e d relations for e s t i m a t i o n of viscosity of dense h y d r o c a r b o n s is
b y c o m b i n i n g Eqs. (8.9) a n d (8.5) with Eq. (2.8) a s s u m i n g p r o p o s e d b y Jossi et al. [1 l]:
Z~ = 0.27. In s o m e predictive m e t h o d s , r e d u c e d viscosity is
defined with respect to viscosity at a t m o s p h e r i c p r e s s u r e (i.e.,
[(/z - - / Z o ) ~ -at- 1 0 - 4 ] ~ = 0.1023 + 0.023364pr + 0.058533Pr2
/Zr = IZ//Za),where/~a is the viscosity at I a t m a n d t e m p e r a t u r e
T at w h i c h / z m u s t be calculated. A n o t h e r r e d u c e d form of (8.12) - 0.040758p~ + 0.0093324p 4
8. APPLICATIONS: ESTIMATION OF TRANSPORT PROPERTIES 335
This equation is, in fact, a modification of Eq. (8.11) and nature. An approximate theory for liquid transport properties
was originally developed for nonpolar gases in the range of is the Eyring rate theory [1, 4]. Effect of pressure on the liq-
0.1 < Pr < 3. /Zo is the viscosity at low pressure and at the uid viscosity is less than its effect on viscosity of gases. At low
same temperature at w h i c h / z is to be calculated./Zo m a y be and moderate pressure, liquid viscosity m a y be considered as
calculated from Eqs. (8.6)-(8.8). However, this equation is a function of temperature only. Viscosity of liquids decreases
also used by reservoir engineers for the calculation of the vis- with increase in temperature. According to the Eyring rate
cosity of reservoir fluids u n d e r reservoir conditions [9, 12]. model the following relation can be derived on a semitheo-
Later Stiel and Thodos [13] proposed similar correlations for retical basis:
the residual viscosity of polar gases:
Ngh f 3.8Tb'~
(/z -/Zo)~ = 1.656 x 10-4p T M forpr _< 0.1 (8.16) tt = - ~ - exp ~ - - )
N
where/~g is the viscosity of natural gas in cp, M is the gas In/z m = ~ x~ In/zi for liquid n o n h y d r o c a r b o n s
molecular weight, T is absolute temperature in kelvin, and p i=1
is the gas density in g/cm 3 at the same T and P that ~g should
where /~m is the mixture viscosity in cp and x~ is the mole
be calculated. This equation m a y be used up to 550 bar and
fraction of c o m p o n e n t i with viscosity /zi. There are some
in the temperature range of 300-450 K. For cases where M is
other mixing rules that are available in the literature for liquid
not known, it m a y be calculated from specific gravity of the
viscosity of mixtures [ 18].
gas as discussed in Chapter 3 (M = 29 SGg). For sour natural
For liquid petroleum fractions (undefined mixtures), usu-
gases, correlations in terms of H2S content of natural gas are
ally kinematic viscosity v is either available from experimen-
available in handbooks of reservoir engineering [ 15, 16].
tal measurements or can be estimated from Eqs. (2.128)-
(2.130), at low pressures and temperatures. The following
8.1.2 Viscosity of Liquids equation developed by Singh m a y also be used to estimate
v at any T as r e c o m m e n d e d in the API-TDB [5]:
Methods for the prediction of the viscosity of liquids are less
accurate than the methods for gases, especially for the estima-
tion of viscosity of undefined petroleum fractions and crude log10 (vr) = A ( 3- 1- 1 ) B - 0.8696
oils. Errors of 20-50% or even 100% in prediction of liquid
viscosity are not unusual. Crude oil viscosity at r o o m temper- (8.19) A = log10 (v3s000)) + 0.8696
ature varies from less than 10 cp (light oils) to m a n y thou- B = 0.28008 x log~0 (v38(100)) + 1.8616
sands of cp (very heavy oils). Usually conventional oils with
API gravities from 35 to 20 have viscosities from 10 to 100 cp where T is in kelvin and v38000) is the kinematic viscosity at
and heavy crude oils with API gravities from 20 to 10 have 100~ (37.8~ or 311 K) in cSt, which is usually k n o w n from
viscosities from 100 to 10000 cp [17]. Most of the methods experiment. The average error for this m e t h o d is about 6%.
developed for estimation of liquid viscosity are empirical in For blending of petroleum fractions the simplest m e t h o d is
336 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
TABLE 8.2---Coefficients of Eq. (8.17) for viscosity of pure liquid compounds. (Taken with permission from Ref. [5].)
lz= IO00exp(A + B/T +C lnT + DT E) (8.17)
API No. Compound A B C D E Train, K Tmax,K
794 Oxygen -4.1480E+00 9.4039E+01 - 1.2070E+00 0.0000E+00 0.0000E+00 54 150
781 Hydrogen -1.1660E+01 2.4700E+01 -2.6100E-01 -4.1000E-16 1.0000E+01 14 33
845 Water -5,2840E+01 3.7040E+03 5.8660E+00 -5.8791E-29 1.0000E+01 273 646
771 Ammonia -6.7430E+00 5.9828E+02 -7.3410E-01 -3.6901E-27 1.0000E+01 196 393
786 Hydrogen sulfide - 1.0900E+01 7.6211E+02 - 1.1860E-01 0.0000E+00 0.0000E+00 188 350
798 Nitrogen 1.6000E+01 -1.8160E+02 -5.1550E+00 0,0000E+00 0.0000E+00 63 124
775 Carbon dioxide 1.8770E+01 -4.0290E+02 -4.6850E+00 -6.9999E-26 1.0000E+01 219 304
Paraffins
1 Methane -6.1570E+00 1.7810E+02 -9.5240E-01 -9.0611E-24 1.0000E+01 91 188
2 Ethane -3.4130E+00 1.9700E+02 -1.2190E+00 -9,2022E-26 1.0000E+01 91 300
3 Propane -6.9280E+00 4.2080E+02 -6.3280E-01 -1,7130E-26 1,0000E+01 86 360
4 n-Butane -7.2470E+00 5.3480E+02 -5.7470E-01 -4,6620E-27 1.0000E+01 135 420
5 Isobutane -1.8340E+01 1.0200E+03 1.0980E+00 -6.1001E-27 1.0000E+01 190 400
6 n-Pentane -2.0380E+01 1.0500E+03 1.4870E+00 -2.0170E-27 1.0000E+01 143 465
7 Isopentane - 1.2600E+01 8.8911E+02 2.0470E-01 0.0000E+00 0.0000E+00 150 310
8 Neopentane -5.6060E+01 3.0290E+03 6.5860E+00 0.0000E+00 0.0000E+00 257 304
9 n-Hexame -2.0710E+01 1.2080E+03 1.4990E+00 0.0000E+00 0.0000E+00 178 343
10 2-Methylpentane -1.2860E+01 9.4689E-04 2.6190E-01 0.0000E+00 0.0000E+00 119 333
14 n-Heptane -2.4450E+01 1.5330E+03 2.0090E+00 0.0000E+00 0.0000E+00 183 373
15 2-Methylhexane -1.2220E+01 1.0210E+03 1.5190E-01 0.0000E+00 0.0000E+00 155 363
23 n-Octane -2.0460E+01 1,4970E+03 1.3790E+00 0.0000E+00 0.0000E+00 216 399
24 2-Methylheptane -1.1340E+01 1.0740E+03 1.3050E-02 0.0000E+00 0.0000E+00 164 391
37 2,2,4-Trimethylpentane - 1.2770E+01 1.1300E+03 2.3460E-01 -3.7069E-28 1.0000E+01 166 541
41 n-Nonane -2.1150E+01 1.6580E+03 1.4540E+00 0.0000E+00 0.0000E+00 219 424
62 n-Decane - 1.6470E+01 1.5340E+03 7.5110E-01 0.0000E+00 0.0000E+00 243 448
73 n-Undecane - 1.9320E+01 1.7930E+03 1. t 430E+00 0.0000E+00 0.0000E+00 248 469
74 n-Dodecane -2.1386E+05 1.9430E+03 1.3200E+00 0.0000E+00 0.0000E+00 263 489
75 n-Tridecane -2.1010E+01 2.0430E+03 1.3690E+00 0.0000E+00 0.0000E+00 268 509
76 n-Tetradecane -2.0490E+01 2.0880E+03 1.2850E+00 0.0000E+00 0.0000E+00 279 528
77 n-Pentadecane - 1.9300E+01 2.0890E+03 1.1090E+00 0.0000E+00 0.0000E+00 283 544
78 n-Hexadecane -2.0180E+01 2.2040E+03 1.2290E+00 0.0000E+00 0.0000E+00 291 564
79 n-Heptadecane - 1.9990E+01 2.2450E+03 1.1980E+00 0.0000E+00 0.0000E+00 295 576
80 n-Octadecane -2.2690E+01 2.4660E+03 1.5700E+00 0.0000E+00 0.0000E+00 301 590
81 n-Nonadecane - 1.63995E+01 2.1200E+03 6.8810E-01 0.0000E+00 0.0000E+00 305 603
82 n-Eicosane -1.8310E+01 2.2840E+03 9.5480E-01 0.0000E+00 0.0000E+00 309 617
86 n-Tetracosane -2.0610E+01 2.5360E+03 1.2940E+00 -7.0442E-30 1,0000E+01 324 793
Naphthenes
101 Cyclopentane -3.2610E+00 6.1422E+02 -1.1560E+00 0.0000E+00 0.0000E+00 225 325
102 Methylcyclopentane - 1.8550E+00 6.1261E+02 - 1.3770E+00 0.0000E+00 0.0000E+00 248 353
103 Ethylcyclopentane -6.8940E+00 8.1861E+02 -5.9410E-01 0.0000E+00 0.0000E+00 253 378
109 n-Propylcyclopentane -2.3300E+01 1.6180E+03 1.8470E+00 0.0000E+00 0.0000E§ 200 404
110 Isopropylcyclopentane - 1.0500E+01 1.0840E+03 -8.2650E-02 0.0000E+00 0.0000E+00 162 399
146 Cyclohexane -6.9310E+01 4.0860E+03 8.5250E+00 0.0000E+00 0.0000E+00 285 354
147 Methylcyclohexane -1.5920E+01 1.4440E+03 6.6120E-01 2.1830E-27 1.0000E+01 200 393
148 Ethylcyclohexane -2.2110E+01 1.6730E+03 1.6410E+00 0.0000E+00 0.0000E+00 200 405
156 n-Propylcyclohexane -3.1230E+01 2.1790E+03 2.9730E+00 0.0000E+00 0.0000E+00 248 430
158 n-Butylcydohexane -3.9820E+01 2.6870E+03 4.2270E+00 0.0000E+00 0.0000E+00 253 454
168 n-Decylcyclohexane -2.7670E+01 2.9210E+03 2.1910E+00 0.0000E+00 0.0000E+00 272 420
Oleflns
192 Ethylene 1.8880E+00 7.8861E+01 -2.1550E+00 0.0000E+00 0.0000E+00 104 250
193 Propylene -9.1480E+00 5.0090E+02 -3.1740E-01 0,0000E+00 0.0000E+00 88 320
Diolefms a n d a c e t y l e n e s
322 Acetylene 6.2240E+00 -1.5180E§ -2.6550E§ 0.0000E§ 0.0000E+00 204 384
Aromatics
335 Benzene -7.3700E+00 1.0380E+03 -6.1810E-01 - 1.1020E-28 1.0000E+01 279 545
336 Toluene -6.0670E+01 3.1490E+03 7.4820E+00 -5.7092E-27 1.0000E+01 178 384
337 Ethylbenzene - 1.0450E+01 1.0480E+03 -7.1500E-02 0.0000E+00 0.0000E+00 248 413
338 o-Xylene -1.5680E+01 1.4040E+03 6.6410E-01 0.0000E+00 0.0000E+00 248 418
341 n-Propylbenzene -1.8280E+01 1.5500E+03 1.0450E+00 0.0000E+00 0.0000E+00 200 432
349 n-Butylbenzene -2.3800E+01 1.8870E+03 1.8480E+00 0.0000E+00 0.0000E+00 200 457
371 n-Pentylbenzene -7.8290E+01 4.4840E+03 9.9270E+00 -2.3490E-27 1.0000E+01 220 478
372 n-Hexylbenzene -8.8060E+01 5.0320E+03 1.1360E+01 -2.6390E-27 1.0000E+01 220 499
373 n-Heptylbenzene -9.5724E+01 5.4770E+03 1.2480E+01 -2.8510E-27 1.0000E+01 336 519
374 n-Octylbenzene -9.4614E+01 5.5678E+03 1.2260E+01 -1.8370E-27 1.0000E+01 237 538
(Con~nued)
8. A P P L I C A T I O N S : ESTIMATION OF TRANSPORT PROPERTIES 337
TABLE 8.2--(Continued)
API No. Compound A B C D E Train, K Tmax, K
375 n-Nonylbenzene 1.0510E+02 6.1272E+03 1.3820E+01 -2.8910E-27 1.0000E+01 360 555
376 n-Decylbenzene 1.0710E+02 6.3311E+03 1.4080E§ -2.7260E-27 1.0000E+01 253 571
377 n-Undecylbenzene - 1.0260E+02 6.2200E+03 1.3380E+01 -2.4450E-27 1.0000E+01 258 587
378 n-Dodecylbenzene 8.8250E+01 5.6472E+03 1.1230E+01 -1.8200E-27 1.0000E+01 268 601
379 n-Tridecylbenzene 4.5740E§ 3.6870E+03 4.9450E+00 -5.8391E-28 1.0000E§ 328 614
383 Cyclohexylbenzene -4.3530E+00 1.4700E+03 - 1.1600E+00 0.0000E+00 0.0000E+00 280 513
386 Styrene -2.2670E+01 1.7580E+03 1.6700E+00 0.0000E+00 0.0000E+00 243 418
342 Cumene -2.4962E+01 1.8079E+03 2.0556E+00 0.0000E+00 0.0000E+00 200 400
Diaromatics and condensed rings
427 Naphthalene 1.9310E+01 1.8230E+03 1.2180E+00 0.0000E+00 0.0000E+00 353 633
472 Acenaphthene 2.0430E+01 1.0380E+02 -4.6070E+00 0.0000E+00 0.0000E+00 367 551
473 Fluorene 4.1850E+00 7.2328E+02 -2.1490E+00 0.0000E+00 0.0000E+00 388 571
474 Anthracene 2.7430E+02 2.1060E+04 3.6180E+01 0.0000E+00 0.0000E+00 489 595
709 Methanol 1.2135E+04 1.7890E+03 2.0690E+04 0.0000E+00 0.0000E+00 176 338
710 Ethanol 7.8750E+00 7.8200E+02 -3.0420E+00 0.0000E+00 0.0000E+00 200 440
to use Eq. (3.105) by calculating blending index of the mix- More recently a corresponding state correlation similar to
ture. The viscosity-blending index can be calculated from the this equation was proposed for estimation of viscosity of hy-
following relation proposed by Chevron Research C o m p a n y drocarbon fluids at elevated pressures in which the reduced
[193: molar refraction (parameter r defined by Eq. 5.129) was used
instead of co [20], Parameters [/zr] (~ and [/zr] (1) have been
BI,~- logl0v correlated to Tr and Pr. Results show that for h y d r o c a r b o n
(8.20) 3 + log10 v systems, parameter co can be replaced by r in the correspond-
BImix = Y~.xviBIi ing states correlations. Such correlations have higher power
of extrapolation to heavier hydrocarbons. Moreover, param-
in which v is the kinematic viscosity in cSt. Once v is de- eter r can be accurately calculated for heavy petroleum frac-
termined absolute viscosity of a petroleum fraction can be tions and undefined h y d r o c a r b o n mixtures as discussed in
estimated from density (# -- p x v). It should be noted that Section 5.9.
Eqs. (2.128)-(2.130) or Eqs. (8.19) and (8.20) are not suitable Equation (8.21) is r e c o m m e n d e d for low-molecular-weight
for pure hydrocarbons. hydrocarbons [5]. For such systems, Jossi's correlation
To consider the effect of pressure on liquid viscosity of (Eq. 8.12) can also be used for calculation of viscosity of high-
hydrocarbons, the three-parameter corresponding states pressure liquids. However, this approach is not appropriate
correlations m a y be used for prediction of viscosity of high- for heavy or high-molecular-weight liquid hydrocarbons and
pressure liquids [5]: their mixtures. For such liquids the Kouzel correlation is rec-
(8.21) //Jr = /Z ~ [/~r](0) "}- co [/Zr](1)
o m m e n d e d in the API-TDB [5]:
(/Zp) P - 1 . 0 1 3 3 (_1.48+5.86,0.181)
where [#r] (~ and [/x~](1) are functions of Tr and Pr. These func- (8.22) log10 ~a -- 10000
tions are given in the API-TDB [5] in the form of polynomials
in terms of Tr and Pr with more than 70 numerical constants. where P is pressure in bar and #a is low-pressure (1 atm) vis-
cosity at a given temperature in cp./zr is the viscosity at pres-
sure P and given temperature in cp. The m a x i m u m pressure
10
for use in the above equation is about 1380 bar (~20000 psi)
n-Pentane
and average error is about 10% [5].
", - - ......... n-Decane
When a gas is dissolved in a pure or mixed liquid hydrocar-
. . . . . n-Eicosane
bons viscosity of solution can be calculated from viscosity of
\~ ""- - - - - Cyclohexane
1 gas-free h y d r o c a r b o n (/za) and gas-to-liquid ratio (GLR) using
\..~. 4-. - - Benzene the following relation [5]:
"~ "'\~ ........ . m Water
P ~ _-- / _I.651(GLR) + 137#~/3 ~_ 538.4 / 3
;_q 0A ~a | ,~/31137 + 4.891(GLR)] + 538.4 ]
(8.23)
log0 ( )
0.01 ~ ~ where both/Z m and/z a are at 37.8~ (100~ in cp and GLR is in
200 300 400 500 600 700 m 3/m3. # r is the viscosity of solution at temperature T, where
Temperature, K
T is in kelvin. This equation should not be used for pressures
above 350 bar. If/Za at 37.8~ (100~ is not available, it m a y
FIG. 8.2--Liquid viscosity of several compounds be estimated; however, if/xa at the same temperature at which
versus temperature at atmospheric pressure. /x is to be calculated is available then/x m a y be estimated from
338 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
/g = A(Ix,) B, where A and B are functions of GLR (see Prob- The term (V-Vo) represents the free space between molecules.
lem 8.4). GLR were calculated from the following relation: As temperature increases V also increases a n d / g decreases.
This theory is applicable to liquids at low pressures. In Chap-
379xA
(8.24) GLR ----- ter 2 it was shown that parameter I (defined by Eq. 2.36) is
MB proportional with fraction of liquid occupied by molecules.
Therefore parameter I is proportional to Vo/V and thus
where XA is the mole fraction of dissolved gas in liquid, MB
is molecular weight of liquid, and SGB is the specific grav- (8.27) /g-1 = C (1-1 - 1)
ity of liquid. In this relation GLR is calculated as stm 3 of
gas/stm 3 of liquid (1 m3/m 3 = 1 scf/st, ft 3 = 5.615 scf/bbl). where/g and I are evaluated at given temperature. Methods of
Units of GLR are discussed in Section 1.7.23. Prediction of calculation of I were discussed in Chapter 2 (see Eqs. (2.36)
viscosity of crude oils (gas free dead oils at 1 atm) is quite dif- and (2.118)). On this basis, one can see that 1/# varies lin-
ficult due to complexity of mixtures. However, there are m a n y early with 1/I for any substance. This relation has been also
empirical correlations developed for calculation of crude oils confirmed with experimental data [23]. Similar correlations
[15, 16]. For example the Glaso's correlation for viscosity of for thermal conductivity and diffusivity were developed and
crude oils is given as the coefficients were related to h y d r o c a r b o n properties such
/god = (3.141 x 1010) • [ ( 1 . 8 T - 4 6 0 ) -3"444] • [logl0(API)] ~ as molecular weight [23, 24]. Equation (8.27) is applicable
only to nonpolar and h y d r o c a r b o n liquid systems in which the
n = 10.313 [logl0(1.8T - 460)] - 36.447
intermolecular forces can be determined by L o n d o n forces.
(8.25) Other developments in the calculation of liquid viscosity are
reported by Chung et al. (generalized correlations for polar
where/zoo is the viscosity of dead oil (gas free at 1 atm.), T is
and nonpolar compounds) [25] and Quinones-Cisneros et al.
temperature in kelvin, and API is the oil gravity. This equation
(pure hydrocarbons and their mixtures) [26].
should be used for crude oils with API gravity in the range of
20-48 and in the temperature range of 283-422 K (50-300~
More advanced and accurate methods of calculation of viscos- Example & / - - C o n s i d e r a liquid mixture of 74.2 mol%
ity of crude oils is based on splitting the oil into several pseu- acetone and 25.8 mol% carbon tetrachloride (CCI4) at 298.2 K
docomponents and to use methods discussed in Chapter 4 and 1 atm. Estimate its viscosity assuming the only informa-
for calculation of the mixture properties. Accurate prediction tion k n o w n for this system are To Pc, Vo ~o, M, and ZRn of each
of viscosities of heavy crude oils is a difficult task and most compound. Compare estimated value with the experimental
correlations result in large errors and errors of 50-100% are value of 0.395 m P a . s (cp) [10].
quite c o m m o n in such predictions.
As seen from Eqs. (8.11) and (8.25), viscosity of liquids
and oils is mainly related to density. In general, heavier oils Solution--CC14 and acetone are n o n h y d r o c a r b o n s whose
(lower API gravity) exhibit higher viscosity. Pure hydrocar- critical properties are not given in Table 2.1 and for this rea-
b o n paraffins have viscosity of about 0.35 cp (0.5 cSt.), naph- son they are obtained from other sources such as DIPPR [ 10]
thenes about 0.6 cp, n-alkylbenzenes (aromatics) about 0.8 cp or any chemical engineering thermodynamics text as [ 18,27]:
(I.1 cSt.), gasoline about 0.6 cp, kerosene about 2 cp, and for acetone, Tc = 508.2 K, Pc = 47.01 bar, Vc = 209 cm3/mol,
residual oils' viscosity is in the range of 10-100 000 cp [17]. w = 0.3065, M = 58.08 g/tool, and ZRA = 0.2477; for CC14,
The methods of m e a s u r e m e n t of viscosity of oils are given in Tc = 556.4 K, Pc -= 45.6 bar, Vc = 276 cm3/mol, o) = 0.1926,
ASTM D 445 and D 446. A graphical m e t h o d for calculation M = 153.82 g/mol, and ZRA = 0.2722 [18]. Using the Kay's
of viscosity of the blend is given by ASTM D 341. For light mixing rule (Eq. 7.1) with xl = 0.742 and x2 = 0.258: Tc =
oils capillary viscometers are suitable for measuring liquid 520.6 K, Pc = 46.6 bar, Vc = 226.3 cm3/mol, 09 = 0.2274, M =
viscosity in which viscosity is proportional to the pressure 82.8, and ZRa = 0.254. Mixture liquid density at 298 K is cal-
difference in two tubes. culated from Racket equation (Eq. 5.121): V s = 80.5 cm3/mol
Most recently Riazi et al. [21] developed a relation for es- (P25 = 1.0286 g/cm3). This gives Pr = VJV = 226.3/80.5 =
timation of viscosity of liquid petroleum fractions by using 2.8112. For calculation of residual viscosity a generalized cor-
refractive index at 20 ~C as one of the input parameters in addi- relation in terms of Pr m a y be used. Although Eq. (8.12) is pro-
tion to molecular weight and boiling point (see Problem 8.3). posed for hydrocarbons and nonpolar fluids, for liquids/o r is
Another development on the prediction of viscosity and other quite high and the equation can be used up to Pr of 3.0. F r o m
transport properties for liquid h y d r o c a r b o n systems was to Eq. (8.5), ~ = 0.02428 and Tr = T/Tc = 0.5724 < 1.5. F r o m
use refractive index to estimate a transport property at the Eq. (8.6), /go = 0.00829 cp. F r o m Eq. (8.12), /g = 0.374 cp,
same temperature in which relative index is available. The- which in comparison with experimental value of 0.395 cp
ory of Hildebrand [22] suggests that fluidity (1//g) of a liquid gives an error of only - 5.3%. This is a good prediction consid-
is proportional to the free space between the molecules. ering the fact that the mixture contains a highly polar com-
p o u n d (acetone) and predicted density was used instead of
(8.26) --/gl= E ( ~ 0 V 0 ) a measured value. If actual values of p2s [18] for pure com-
pounds were used (p25 = 0.784 for acetone and p2s = 1.584
where E is a constant, V is the liquid volume (i.e., molar), and g/cm 3 for CC14) and density is calculated from Eq. (7.4) we get
V0 is the value of V at zero fluidity (/g -+ 0). Parameters E and p25 = 1.03446 g/cm 3 (Pr = 2.828), which predicts/gmix = 0.392
V0 m a y be determined from regression of experimental data. cp (error of only -0.8%). r
8. A P P L I C A T I O N S : E S T I M A T I O N O F T R A N S P O R T P R O P E R T I E S 339
0.25
.. 0.2 J
J
o
15 0.15 f
f
1D 0.1
t.-
8
0.05"
w/m K = 0.5788~TU,'hrft RI
l
o 200 400 600 800 1000
Temperature, C
(a)Gases
,,,,,
0.22
0.2'
\ 1,1 W/rn K = 0.5788 BTU/Ilr ft RI
0.18
0.16
'5
, IP,opa~ iB ~
0.14
0.12
o
o
0.1
E
0.08 \
0.06
-200 -150 -1 O0 -50 0 50 1O0 150 200
Temperature, C
(b) Liquids
FIG. 8.3mThermal conductivity of several light hydrocarbons versus temperature
at atmospheric pressure. Taken with permission from Ref. [2].
w h e r e Aij m a y be set equal to ~ii given in Eq. (8.7). A n o t h e r e q u a t i o n is oversimplified a n d should be used w h e n o t h e r
m i x i n g rule that does not r e q u i r e viscosity of p u r e c o m p o n e n t m e t h o d s are n o t applicable. Riazi a n d F a g h r i [30] u s e d the
is given by Poling et al. [18]. A m o r e a d v a n c e d mixing rule general r e l a t i o n s h i p b e t w e e n k, T, a n d P at the critical p o i n t
for calculation of mixture t h e r m a l conductivity of gases a n d (To, Pc) to develop a n e q u a t i o n s i m i l a r to Eq. (2.38) for es-
liquids is p r o v i d e d by M a t h i a s et al. [29]. F o r vapors from un- t i m a t i o n of t h e r m a l conductivity of p e t r o l e u m fractions a n d
defined p e t r o l e u m faction, the following equation has b e e n pure hydrocarbons.
derived from regression of a n old figure developed in the
k = 1.7307A(1.8Tb)BSG c
1940s [5]:
A = exp (21.78 - 8.07986t + 1.12981t 2 - 0.05309t 3)
k = A + B ( T - 255.4)
0.42624 1.9891 (8.37) B = - 4 . 1 3 9 4 8 + 1.29924t - 0.17813t 2 + 0.00833t 3
(8.36) A = 0.00231 + ~ 4 M2 C = 0.19876 - 0.0312t - 0.00567t 2
1.3047 x 10 -4 0.00574 1.8T - 460
B---- 1.0208 x 10 -4 + + - - t=
M M2 100
w h e r e k is in W / m . K a n d T is in kelvin. The e q u a t i o n should w h e r e k is in W / m . K, Tb a n d T are in kelvin. F a c t o r s 1.7307
be u s e d for p r e s s u r e below 3.45 bar, for p e t r o l e u m fractions a n d 1.8 c o m e f r o m the fact that the original units were in
with M b e t w e e n 50-150 a n d T in the r a n g e of 260-811 K. This English. This equation can be a p p l i e d to p u r e h y d r o c a r b o n s
8. A P P L I C A T I O N S : E S T I M A T I O N O F T R A N S P O R T P R O P E R T I E S 341
TABLE 8.3--Coefficients of Eq. (8.33) for thermal conductivity of pure gases [5].
k = A + B T + C T 2 (8.33)
No. Compound name A x 10-1 B x 10 4 C x 10 -7 Range, K
1 Methane -0.0076 0.9753 0.7486 97-800
2 Ethane -0.1444 0.9623 0.7649 273-728
3 Propane -0.0649 0.4829 1.1050 233-811
4 n-Butane 0.0000 0.0614 1.5930 273-444
5 n-Pentane 0.0327 -0.0676 1.5580 273-444
6 n-Hexane 0.0147 0.0654 1.2220 273-683
7 n-Heptane -0.0471 0.2788 0.9449 378-694
8 n-Octane -0.1105 0.5077 0.6589 416-672
9 n-Nonane -0.0876 0.4099 0.6937 450-678
10 n-Decane -0.2249 0.8623 0.2636 450-678
11 n-Undecane -0.1245 0.4485 0.6230 472-672
12 n-Dodecane -0.2535 0.8778 0.2271 516-666
13 n-Pentadecane -0.3972 1.3280 -0.2523 566-644
14 Ethene -0.0174 0.3939 1.1990 178-589
15 Propene -0.0844 0.6138 0.8086 294-644
16 Cyclohexane -0.0201 0.0154 1.4420 372-633
17 Benzene -0.2069 0.9620 0.0897 372-666
18 Toluene -0.3124 1.3260 -0.1542 422-661
19 Ethylbenzene -0.3383 1.3240 -0.1295 455-678
20 1,2-Dimethylbenzene(o-Xylene) -0.1430 0.8962 0.0533 461-694
21 n-Propylbenzene -0.3012 0.9695 0.7099 455-616
TABLE 8.S--Vatues ofk<ru and k(r2) for Eq. (8.39). (Taken with permission from Ref. [32].)
kr = ~ ----(0.5 -- (.o)k~ + o ) k ~ (8.101)
Pr
Tr 0.2 0.5 1.0 1.5 2.0 3.0 4.0 6.0 8.0 10.0
Values of k!1) versus Tr and Pr
1.00 1.1880 1.3307 2.0000 4.15t7 4.4282 4.7900 5.2140 5.7989 6.2080 6.5132
1.05 1.3002 1.3640 1.8922 3.2806 3.7990 4.4915 4.7590 5.2817 5.7710 6.2040
1.10 1.4300 1.4810 1.8660 2.5989 3.3334 4.1068 4.4746 4.9502 5.3740 5.8812
1.15 1.5182 1.5365 1.8356 2.2978 2.9769 3.8583 4.4676 4.9404 5.3734 5.8760
1.20 1.8311 1.8956 2.1200 2.3983 2.8809 3.5626 4.2067 4.9285 5.3731 5.8699
1.40 2.1838 2.2520 2.3589 2.5291 2.7120 3.3000 4.0020 4.6327 5.2404 5.7656
1.60 2.5971 2.6589 2.7305 2.8572 3.0035 3.3760 3.8239 4.4385 4.8967 5.3031
2.00 3.6763 3.6984 3.7418 3.9161 3.9594 4.1370 4.3768 4.7138 5.0462 5.3614
3.00 6.9896 7.0010 7.0310 7.0617 7.1079 7.1452 7.2197 7.4077 7.5915 7.7685
Values of kr(2) versus Tr and Pr
1.00 1.6900 1.6990 2.0000 2.0619 2.3112 2.3140 2.3160 2.3180 2.3210 2.3212
1.05 1.7200 1.7290 1.8100 1.8170 2.1318 2.1912 2.3010 2,8380 2.3398 2.3400
1.10 1.8001 1.8211 1.8300 1.8310 1.9672 2.1384 2.1369 2.3614 2.3988 2.4105
1.15 2.0599 2.0601 2.0661 2.0700 2.0801 2.1269 2.2246 2.3780 2.4618 2.4622
1.20 2.1441 2.1539 2.1629 2.1681 2.1689 2.1901 2.2319 2.3981 2.4640 2.4701
1.40 2.6496 2.6772 2.6865 2,6889 2.6900 2.6911 2.7001 2.7119 2.8079 2.8810
1.60 3.2184 3.2448 3.2559 3.2886 3.3142 8.8292 3.3343 3.3352 3.8869 3.4525
2.00 4.5222 4.5330 4.5465 4.6871 4.6378 4.7108 4.8148 4.8119 4.8850 4.9885
3.00 8.4002 8.4158 8.4234 8.4503 8.4504 8.5038 8.6083 8.6204 8.6732 8.7454
T = 573.2 K (300~ k ~ 0.048 W / m . K . F r o m Lee-Kesler hydrocarbons, t h e r m a l conductivity varies linearly with tem-
correlation (Eq. 5.107), the m o l a r volume at 573.2 K a n d perature:
100 b a r is calculated as Z = 0.59 or V = 281 cm3/mol. Thus
Pr = Vc/V = 313.05/281 = 1.114. Since 0.5 < pr < 2, from (8.41) k = A + BT
Eq. (8.38) F = 151.82, A = 2.702, B = 0.67, C = - 1 . 0 6 9 , a n d Coefficients A a n d B can be d e t e r m i n e d if at least two data
k = 0.048 + 0.017 = 0.065 W/m- K. points o n t h e r m a l conductivity are available. Values of ther-
To calculate k from Eq. (8.39), kr is required. Since in mal conductivity of some c o m p o u n d s at m e l t i n g a n d boiling
Table 8.4 value of k~ for n-C5 is not given, one c a n ob- points are given in Table 8.6, as given in the API-TDB [5]. Liq-
tain it from i n t e r p o l a t i o n of values given for C4 a n d C7 uid t h e r m a l conductivity of several n-paraffins as calculated
by a s s u m i n g a linear relation b e t w e e n / ~ a n d To. For C4, from Eq. (8.41) (or Eq. 8.42) is s h o w n in Fig. 8.4.
= 0.0478 a n d Tc = 425.2 K a n d for C7, k~ = 0.0535 a n d If values of t h e r m a l conductivity at melting a n d boiling
ire = 540.2 K. For C5 with T~ = 469.7 by linear interpolation, points are taken as reference points, then Eq. (8.41) can be
/~ = [(0.0535 - 0.0478)/(540.2 - 425.2)] • (469.7 - 425.2) + used to o b t a i n value of t h e r m a l conductivity at any other tem-
0.0478 = 0.05 W / m . K. Extrapolation between values ofkc for perature:
C3 a n d Ca to k~ of C5 gives a slightly different value. At T a n d P
of interest, Tr = 1.22 a n d Pr = 2.97. F r o m Table 8.5, k~1) = 3.5 T - TM
(8.42) krr = k~ + (k~ - k~) Tb Z
a n d k~2) = 2.2. F r o m Eq. (8.39), /q = 1.42 a n d k = 0.05 •
1.42 = 0.071 W / m . K . Stiel-Thodos m e t h o d varies by 8.5%
where TM a n d Tb are n o r m a l melting (or triple) a n d boiling
from Riazi-Faghri method, which represents a reasonable
points, respectively, k~ a n d k~ are values of liquid t h e r m a l
deviation. I n this case the Stiel-Thodos m e t h o d is more ac-
conductivity at TM a n d Tb, respectively, krL is value of liquid
curate since the value of k ~ is calculated m o r e accurately. #
t h e r m a l conductivity at t e m p e r a t u r e T. According to API-TDB
8.2.2 Thermal Conductivity of Liquids [5] this e q u a t i o n can predict values of liquid t h e r m a l conduc-
tivity of pure c o m p o u n d s u p to pressure of 35 b a r with a n
Theory of t h e r m a l conductivity of liquids was proposed by accuracy of a b o u t 5% [5]. There are a n u m b e r of generalized
B r i d g m a n [1]. I n this theory, it is a s s u m e d that molecules correlations developed for prediction of t h e r m a l conductivity
are arranged as cubic lattice with center-to-center spacing of pure h y d r o c a r b o n liquids. The Riedel m e t h o d is included
of (V/NA) 1/3, in which V is the m o l a r v o l u m e a n d NA is the in the API-TDB [5]:
Avogadro number. Furthermore, it is a s s u m e d that energy
is transferred from one lattice to a n o t h e r at the speed of C M"
sound, cs. This theory provides the basis of prediction of ther- (8.43) kL -
w h e r e VL is the liquid m o l a r v o l u m e at 25~ (298 K) in API-TDB [5]. This e q u a t i o n can be used for t e m p e r a t u r e s at
cm3/mol. F o r s o m e c o m p o u n d s these values of V~ are given Tr < 0.8 an d pressure b el o w 35 bar. F o r e s t i m a t i o n of k L at
in Table 6.10. M is the m o l e c u l a r weight in g/mol and k L t e m p e r a t u r e s above n o r m a l boiling p o i n t ( c o m p r e s s e d or sat-
is desired liquid t h e r m a l conductivity at T in W / m . K. Av- u r a t e d liquids), there are a n u m b e r of m e t h o d s that use re-
erage err o r for this e q u a t i o n is a b o u t 5% as r e p o r t e d in the d u c e d density Pr as a correlating p a r a m e t e r [5, 8]. Riazi a nd
Faghri [30] also developed a m e t h o d similar to Eq. (8.37) for
prediction of t h e r m a l conductivity of liquid h y d r o c a r b o n s for
0.8
-- - -- n-Pentane
p e n t a n e s an d heavier.
n-Decane
f
k = 1.7307A(1.8Tb)BSG c
. . . . . n-Eicosane j
~0.6 A = exp ( - 4 . 5 0 9 3 - 0.6844t - 0.1305t a)
- - Benzene 7
TABLE 8.7--Comparison of various methods of calculation of liquid thermal conductivity at 20~ (Example 8.3).
Linear,a Eq. (8.42) API, Eq. (8.43) Eq. (8.52) RF,b Eq. (8.44) RF, Eq. (8.46)
Nc k~xp K %Dev k %Dev k %Dev k %Dev k %Dev
5 0.114 0.114 0.4 0.113 --0.8 0.107 --6.6 0.107 --6.5 0.113 --0.6
6 0.121 0.121 0.2 0.119 --1.4 0.111 --8.4 0.112 --7.8 0.118 --2.4
7 0.1262 0.126 --0.1 0.124 --1.7 0.114 --9.7 0.116 --8.3 0.122 --3.3
8 0.1292 0.129 0.0 0.127 --1.5 0.116 --10.0 0.119 --7.6 0.126 --2.8
9 0.1316 0.132 0.0 0.130 --1.2 0.118 -10.2 0.123 -6.9 0.129 --2.2
10 0.133 0.133 0.0 0.132 --0.6 0.120 --9.9 0.125 --5.8 0.132 --1.1
Overall 0.1 1.2 9.1 7.1 2.1
aLinear refers to linear relation betweern k and T.
bRF referes to Riazi-Faghri methods.
w h e r e k256 refers to the value of k at 256 K (0~ a n d k422 is liquid t h e r m a l conductivity of h y d r o c a r b o n systems:
the value of k at 422 K (300~ Using Eq. (8.41) a n d on the
= EZ~i~ik L
basis of linear i n t e r p o l a t i o n of t h e r m a l conductivity f r o m the i J
above equations, the following relation was also derived for
the t e m p e r a t u r e range a n d m o l e c u l a r weight ranges specified + w h e r e k/i = kii a n d k/i = ki
for Eq. (8,44):
conductivity and when no data are available it can be con- recommended for practical applications in the petroleum
sidered as unity [18]. industry.
For calculation of thermal conductivity of liquid petroleum
fractions, if the PNA composition is available the pseudocom-
ponent method using Eq. (8.42) and Table 8.4 may be applied. 8.3 DIFFUSION COEFFICIENTS
The simplest method of calculation of kL for petroleum frac-
tions when there is no information on a fraction is provided Diffusion coefficient or diffusivity is the third transport prop-
in the API-TDB [5]: erty that is required in calculations related to molecular dif-
fusion and mass transfer in processes such as mixing and
(8.50) k~ = 0.164 - 1.277 x 10-4T
dissolution. In the petroleum and chemical processing, dif-
where T is in kelvin and kL in W/re. K. In other references this fusion coefficients of gases in liquids are needed in design
equation is reported with slight difference in the coefficients. and operation of gas absorption columns and gas-gas diffu-
For example, Wauquier [8] gives kz = 0.17 - 1.418 x 10-4T. sion coefficients are required to determine rate of reactions
At 298 K (25~ this relation gives a value of 0.128 W/re. K in catalytic-gas-phase reactions, where mass transfer is a con-
(near k of n-C8), while Eq. (8.50) gives a value of 0.126, which trolling step. In the petroleum production, knowledge of dif-
is the same as the value of k for n-heptane. The error for this fusion coefficient of a gas in oil is needed in the study of gas
equation is high, especially for light and branched hydrocar- injection projects for improved oil recovery. If a binary sys-
bons. Average error of 10% is reported for this equation [5] tem of components A and B is considered, where there is a
and it may be used in absence of any information on a frac- gradient of concentration of A in the fluid, then it diffuses
tion. A more accurate relation uses average boiling point of in the direction of decreasing concentration (or density)--a
the fraction as an input parameter and was developed by the process similar to heat conduction due to temperature gradi-
API group at Penn State [5]: ent. In this case the diffusion coefficient of component A in
the system of A and B is called binary (or mutual) diffusion
(8.51) kL = Tb0"2904 X (2.551 • 10-2 - 1.982 • 10-ST) coefficient and is usually shown by DAB,which is defined by
the Fick's law [1]:
where both Tb and T are in kelvin. This equation gives an av-
erage error of about 6%. For n-C5 of Example 8.3, this equa-
(8.54) JAy = -- DAB~ = - p DAB~dd~WAy
tion gives a value of 0.104 W/m. K with - 9% error. However,
this equation is not recommended for pure hydrocarbons.
where JAy is the mass flux of component A in the y direction
For petroleum fractions Eq. (8.46) can be used with better
(i.e., g/cm 2. s) and dpA/dy is the gradient of mass density in
accuracy with the specified ranges of boiling point and tem-
the y direction, p is the mass density (g/cm3) of mixture and
perature when both Tb and SG are available. For coal liquids
XwAis the weight fraction of component A in the mixture. In
and heavy fractions, Tsonopoulos et al. [33] developed the
the above relation, the second equality holds when p is con-
following relation based on the corresponding states method
stant with respect to y. JA represents the rate of transport of
of Sato and Riedel:
mass in the direction of reducing density of A. It can be shown
(8.52) kL = 0.05351 + 0.10177 (I - Tr)2/3 that in binary systems DaB is the same as DBA [1]. From the
above equation, it can be seen that the unit of diffusivity in
where kL is in W/re. K. This equation is not recommended for the cgs unit system is cm2/s (or 1 cm2/s --- 10-4 m2/s). Diffu-
pure hydrocarbons. For some eight coal liquid samples and sion of a component within its own molecules (DAA)is called
74 data points this equation gives an average error of about self-diffusion coefficient. From thermodynamic equilibrium
3% [33]. point of view, the driving force behind molecular diffusion is
For liquid hydrocarbons and petroleum fractions when gradient of chemical potential O#A/Oy. Since chemical poten-
pressure exceeds 30-35 atm, effect of pressure on liquid ther- tial is a function of T, P, and concentration, for systems with
mal conductivity should be considered. However, this effect uniform temperature and pressure,/za is only a function of
is not significant for pressures up to 70-100 atm. For the re- concentration (see Eq. 6.12 i) and Eq. (8.54) is justified. Var-
duced temperature range of 0.4-0.8 and pressures above 35 ious forms of Fick's law can be established in the forms of
atm, the following correction factor for the effect of pressure gradients of molar concentration, mole, weight, or mass frac-
on liquid thermal conductivity is recommended in the API- tions [1]. A comparison between Eqs. (8.1), (8.28), and (8.54)
TDB [5]: shows the similarity in momentum, heat, and mass trans-
fer processes. The corresponding molecular properties (i.e.,
L L C2
kinematic viscosity (v), thermal diffusivity (a), and diffusion
(8.53) ~ = k~ coefficient (D)) that characterize the rate of these processes
C = 17.77 + 0.065Pr - 7.764Tr 2054T2 have the same unit (i.e., cm2/s or ft2/h). This is the reason that
exp (0.2Pr)
these physical properties are called transport properties. The
To calculate value of k~ at I"2 and P2, value of kL at/'1 and P1 diffusion process may also be termed mass transfer by con-
must be known. In case of lack of an experimental value, the duction. The ratio of v/D or (Iz/pD) is a dimensionless num-
value of k~ at T1 and P~ can be calculated from Eqs. (8.41)- ber called Schmidt number (Ns~) and is similar to the Prandfl
(8.43). There are some other generalized correlations based number (Npr) in heat transfer (see Eq. 8.29). Schmidt number
on the theory of corresponding states for prediction of both represents the ratio of mass transfer by convection to mass
viscosity and thermal conductivity of dense fluids [25, 34]. transfer by diffusion. Values of Nsc of methane, propane, and
However, these methods, although complex, are not widely n-octane in the air at 0~ and 1 atm are 0.69, 1.42, and 2.62,
346 CHARACTERIZATION AND P R O P E R T I E S OF P E T R O L E U M FRACTIONS
TABLE 8.8--Order of magnitude of binary diffusion coefficient and its concentration dependency
for various systems [35].
Order of magnitude Activation energy (E), Concentration
Type of system of D, cm2/s kcal/mol dependence
Gas-gas (vapor-gas) 0.1-1.0 E <5 Very weak
Gas-liquid ~10 -5 E _<5 Weak
Normal liquids 10-5-10 -6 5-10 +100%
Polymer solutions 10-5-10 -8 10-20 + 1000%
Gas or liquid in polymer or solids ~10-12-10 -15 E _>40 Factor of 1000%
respectively. The order of magnitude of Nsc in liquids such as relation is developed for gas-gas diffusivities [ 1, 3]:
water is 103 .
Diffusion coefficient like any other t h e r m o d y n a m i c prop- 3xn • (k~ ~ 1/2 ( 1 1 ~,/2 T3/2
DAB - 8 • p( )z
erty is a function of the state of a system and depends on
T, P, and concentration (i.e., xi). One theory that describes (8.56)
molecular diffusion is based on the assumption that molecu-
lar diffusion requires a j u m p in their energy level. This energy where DAB is in cm2/s, kB is the Boltzman's constant (1.381 •
10 -23 J/K), T is temperature in kelvin, P is the pressure in
is called activation energy and is shown by EA. This activation
energy, although not the same, is very similar to the activa- bar, m is the molecular mass in kg [M/NA, i.e., mA = MA •
tion energy required for a chemical reaction to occur. Heavier
10-3/(6.022 x 1023)], and d is the hard sphere molecular diam-
molecules have higher activation energy and as a result lower eter in m (1 n m = 1 x 10 -9 m). Values o f d m a y be determined
from measured viscosity or thermal conductivity data by Eqs.
diffusion coefficients. Based on this theory, dependency of D
with T can be expressed by Arrhenius-type equation in the (8.2) and (8.30), respectively. For example, for CH4 value of d
following form: from viscosity is 0.414 n m while from thermal conductivity is
0.405 nm. For 02, Ha and COa, values o f d are 0.36, 0.272, and
0.464 nm, respectively [3]. As an example, the self-diffusion
coefficient of CH4 at 1 bar and 298 K from the kinetic the-
(8.55) D = D oexp ~
f - ~EA~
-~) ory is calculated as mA = mB = m = 2.66 x 10 -26 kg, dn = d B =
d = 0.414 x 10 -9 m and from Eq. (8.56) DAB = 0.194 cm2/s.
Thus one can calculate diffusion coefficient from viscosity
where Do is a constant (with respect to T), EA is the activation data through calculation of molecular diameter. For gases
energy, R is the gas constant, and T is the absolute tempera- at low pressures D varies inversely with pressure, while it is
ture. The order of magnitude of D and EA in various systems proportional to T 3/2. Furthermore, DA varies with M~ 1/2, that
and concentration dependency of D are shown in Table 8.8. is, heavier molecules have lower diffusivity u n d e r the same
Diffusion coefficients depend on the ability of molecules to conditions of T and P. In practical cases molecular diame-
move. Therefore, larger molecules have more difficulty to ters can be estimated from liquid molar volumes in which
move and consequently their diffusivity is lower. Similarly in actual data are available, as will be seen in Eq. (8.59). A more
liquids where the space between molecules is small, diffusion accurate equation for estimation of diffusivity of ideal gases
coefficients are lower than in the gases. Increase in T would was derived independently by C h a p m a n and Enskog from the
increase diffusion coefficients, while increase in P decreases kinetic theory and is known as C h a p m a n - E n s k o g equation,
diffusivity. The effect of P on diffusivity of liquids is less than which m a y be written as [1, 9]
its effect on the diffusivity of gases. At very high pressures,
values of diffusion coefficients of liquids approach their val- 2.2648• 5T~ ( ~ + u~) ~
ues for the gas phase. At the critical point, both liquid and (pDAB)~ = a~2~
gas phases have the same diffusion coefficient called critical
diffusion coefficient and it is represented by Dc. aA +orb
O'AB--
In this section, methods of estimation of diffusion coef- 2
ficients in gases and liquids as well as in m u l t i c o m p o n e n t ~/ = 0.1866V1/3 Z~ 6/5
systems and the effect of porous media on diffusivity are pre- 1.06036
sented. In the last part a new m e t h o d different from conven- g2AB -- (T~)0.1561 + 0.193 exp ( - 0 . 4 7 6 3 5 T ~ ) + 1.76474
tional methods for experimental m e a s u r e m e n t of diffusion
coefficients in dense hydrocarbon fluids (both gases and liq- x exp ( - 3 . 8 9 4 1 1 T ~ ) + 1.03587 exp ( - 1 . 5 2 9 9 6 T ~ )
uids) is presented.
T~m = T / e ~
e A B = (eAeB) 1/2
ei = 65.3T~iZ~i8/5
8.3.1 Diffusivity o f Gases at Low Pressures
(8.57)
Similar to viscosity and thermal conductivity, kinetic theory
provides a relatively accurate relation for diffusivity of rigid where (pDAB)~ represents the product of density-diffusivity
(hard) molecules with different size. Based on this theory, for of ideal gas at low-pressure conditions according to the
gases at low pressures (ideal gas conditions) the following C h a p m a n - E n s k o g theory and is in mol/cm 9s. e and a are the
8. APPLICATIONS: ESTIMATION OF TRANSPORT PROPERTIES 347
energy and size parameters in the potential energy relation hydrodynamic theories that in liquid systems diffusion co-
(i.e., Eq. 5.11). a is in ~,, T and Tc are in kelvin, and Vc is in efficient is inversely proportional to viscosity of solvent. For
cma/mol. The correlations for calculation of Lennard-Jones example, based on the hydrodynamic theory and the Stokes-
(LJ) parameters (e and a) from critical constants as given in Einstein equation, Wilke and Chang developed the following
Eq. (8.57) were developed by Stiel and Thodos [36]. There are relation for estimation of diffusion coefficient at infinite dilu-
some other correlations given in the literature for calculation tion [18, 28]:
of LJ parameters [18]. Typical values of e and a determined
from various properties are given in Table 6.16. In the above (8.60) D~BL = 7.4 x 10-8 (tI/BMB)I/2 T
relation low-pressure diffusivity can be calculated through di- /ZBV2"6
viding (pDAB)~ by p~ (= 83.14T/P) in which T is in kelvin and where D~BL is the diffusion coefficient (in cm2/s) of solute A in
P is in bar. Calculated DAB would be in cm2/s. solvent B, when concentration of A is small (dilute solution).
For practical calculations, a more accurate estimation The superscript oc indicates the system is dilute in solute and
method is required. Most of these correlations are based on for this reason concentration of solute is not included in this
the modified version of Chapman-Enskog theory [18]. The equation. MB is the molecular weight of solvent (g/mol), T
empirical correlation of Chen-Othmer for estimation of DAB is absolute temperature in kelvin, and #B is the viscosity of
of gases at low pressures is in the following form [28]: solvent B (in cp). Because the solution is dilute,/Zs is almost
(1
1.518 x 10 2T181 ~ -F
1) 1/2 the same as viscosity of solution. VA is the molar volume of
solute A at its normal boiling point in cm3/mol and it may
(8.58) D~a3 = be calculated from Vr according to the relation given in Eq.
P (TcATCB)~176 (VcO~4 -~-VcO~4)2
(8.59). qrB is called association parameter for solvent where for
where D~d3is the diffusivity of A in B at low pressures in cm2/s, water the value of 2.6 is recommended [28]. For methanol and
T is in kelvin, P is in bar, M in g/mol, and TeAand VCAare the ethanol, qra is 1.9 and 1.5, respectively. For benzene, heptane,
critical temperature and volume of A in kelvin and cm3/mol, and unassociated solvents (most hydrocarbons) its value is
respectively. This method can be used safely up to pressure of 1.0 [18, 28]. The average error for this equation for some 250
about 5 bar. This equation predicts self-diffusion coefficient systems is about 10% [18].
of methane at 298 K and 1 bar as 0.248 cm2/s versus the value Another simple method derived from Tyn and Calus equa-
of 0.194 from the kinetic theory (Eq. 8.56). For hydrocarbon- tion and is given as follows [18]:
hydrocarbon systems the API-TDB recommends the Gilliland (vo. 67 r
method in the following form [5]: (8.61) D~a3L = 8.93 • 10-s \ ~ ] --/ZB
(1
4.36 x 10-3T 15 MAA+
1) 1/2
where the parameters and units are the same as those given in
(8.59) DAB Eq. (8.60). VBis the molar volume of solvent at its boiling point
p (v1/3 q- V1/3) 2
and can be calculated from VCBsimilar to VA. Equation (8.61)
V/ = 0.285V1"048 is suitable for organic and hydrocarbon systems. Because of
higher accuracy, the Wilke-Chang method (Eq. 8.60) is widely
where V/is the liquid molar volume of component i at its nor- used for calculation of diffusion coefficient of liquids and it
mal boiling point and Va is the molar critical volume and both is also recommended in the API-TDB [5].
are in cm3/mol. Other units are the same as in Eq. (8.58). This As shown in Table 8.8 diffusion coefficient of a binary liquid
equation can be used up to pressure of 35 bar with an accu- system depends on the concentration of solute. This is the rea-
racy of about 4% as reported in the API-TDB [5]. Several other son that most experimental data on liquid diffusivity are re-
methods for prediction of gas diffusivity at low pressures are ported for dilute solutions without concentration dependency
given by Poling et al. [18]. and for the same reason predictive methods (Eqs. (8.60) and
(8.61)) are developed for diffusion coefficients of dilute solu-
8.3.2 Diffusivity of Liquids at Low Pressures tions. There are a number of relations that are proposed to
calculate DL at different concentrations. The Vignes method
Calculation of diffusion coefficients for liquids is less accurate suggests calculation of DABfrom D~BL and DffL as follows [35]:
than gases as for any other physical property. This is mainly
(8.62) DL = (D~,BL)xB (D~L) xAOtAB
due to the lack of a perfect theory for liquids. Generally there
are three theories for diffusivity in liquids: (i) hydrodynamic where Xg is the mole fraction of solute andxB is equal to 1 -- XA.
theory, which usually applies to systems of solids dissolved in Parameter aAB is a dimensionless thermodynamic factor in-
liquids, (2) Eyring rate theory, and (3) the free-volume theory. dicating nonideality of a solution defined as
In the hydrodynamic theory, it is assumed that fluid slides
(01nyA~ = l+(~ln~
over a particle according to the Stoke's law of motion. The
(8.63) ~AB = 1 + \Olnx~]r,p \81nxB]r,e
Eyring theory was presented earlier by Eq. (8.55) in which
molecules require an energy jump before being able to dif- where YAis the activity coefficient of solute A and can be esti-
fuse. The free-volume theory says that for a molecule to jump mated from methods of Chapter 6. For ideal systems or dilute
to a higher energy level (activation energy), it needs a critical solutions (XA ------0), ags = 1.0. For simplicity in calculations for
free-volume (V2) and Eq. (8.55) can be modified by multiply- hydrocarbon-hydrocarbon systems this parameter is taken as
ing the right-hand side by factor exp (V~/V), where V is the unity.
apparent molar volume of liquid. None of these theories is Another simple relation is suggested by Caldwell and Babb
perfect; however, it can be shown by both the Eyring rate and and is also recommended in the API-TDB for hydrocarbon
348 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
a=1.07 b=-0.27-0.38~o
o.8 Nk~ ;
(8.67) c = -0.05 + 0.1~o Pr = P/Pc
Tc = xATcA + XBTcB Pc = XAPcA + XBPcB
0.6
60 = XA0)A + XBO)B
a b
2L _ 1 1 1 =~-=1.4
2LCos0 Cos0 Cos45~ ~r2J
2
FIG. 8.9--Approximate calculation of tortuosity (r).
size a n d shape of particles are all different, value of r varies [50] suggest that for calculation of diffusion coefficients of
f r o m 3 to 5. gases in p o r o u s solids (i.e., catalytic reactors) effective diffu-
In a p o r o u s m e d i a r is r e l a t e d to the formation resistivity sion coefficients can be calculated f r o m the following equa-
factor a n d porosity as tion:
(8.72) r = (Fr ~ (8.75) Deft = r
w h e r e F is the resistivity a n d r is the porosity, b o t h are di- This e q u a t i o n can be o b t a i n e d from Eq. (8.70) b y a s s u m i n g
m e n s i o n l e s s p a r a m e t e r s . r is the fraction of c o n n e c t e d e m p t y ~n ~___(bl.5.
space in a p o r o u s m e d i a a n d F is an i n d i c a t i o n of electrical
resistance of m a t e r i a l s that f o r m the p o r o u s m e d i a a n d is
always g r e a t e r t h a n unity, nl is a d i m e n s i o n l e s s e m p i r i c a l pa- 8.4 I N T E R R E L A T I O N S H I P A M O N G
r a m e t e r that d e p e n d s on the type of p o r o u s media. Theoreti- TRANSPORT PROPERTIES
cally, value of nl in Eq. (8.72) is one; however, in practice nl is
t a k e n as 1.2. Various relations b e t w e e n z a n d r are given b y In previous sections three t r a n s p o r t p r o p e r t i e s of/z, k, a n d D
Amyx et al. [48] a n d Langness et al. [49]. One general r e l a t i o n were introduced. In the predictive m e t h o d s for these molecu-
is given as follows [48]: lar properties, there exist s o m e similarities a m o n g these prop-
(8.73) r = ar -m erties. Most of the predictive m e t h o d s for t r a n s p o r t p r o p e r t i e s
of dense fluids are developed t h r o u g h r e d u c e d density, Pr. I n
w h e r e p a r a m e t e r s a a n d m are specific of a p o r o u s media. addition, diffusion coefficients of dense fluids a n d liquids are
P a r a m e t e r m is called cementation factor a n d it is specifi- related to viscosity. Riazi a n d D a u b e r t developed several re-
cally a characteristic of a p o r o u s m e d i a a n d it usually varies lationships b e t w e e n / z , k, a n d D b a s e d on the principle of
f r o m 1.3 to 2.5. S o m e r e s e a r c h e r s have a t t e m p t e d to correlate d i m e n s i o n a l analysis [37]. F o r example, they f o u n d that for
p a r a m e t e r m with p o r o s i t y a n d resistivity. F o r s o m e reser- liquids In (#2/3D/T) versus In (T/Tb) is linear a n d o b t a i n e d the
voirs a = 0.62 a n d m = 2.15, while for s o m e o t h e r reservoirs, following relations:
w h e n ~b > 0.15, a = 0.75 a n d m = 2 a n d for r < 0.15, a = 1
/Z2/3 ( T ~ 0"7805
a n d m = 2. By c o m b i n i n g Eqs. (8.72) a n d (8.73) with nl = 1.2
--D = 6.3 x 10 -8 for liquids except w a t e r
a n d a = 1:
(8.74) r = ~i.2-1.2m ;/3 (r
--D = 10.03 x 10 -s for liquid w a t e r
E q u a t i o n (8.74) can be c o m b i n e d with Eq. (8.70) to e s t i m a t e
effective diffusion coefficients in a p o r o u s media. P a r a m e t e r (8.76)
r a i n Eq. (8.74) can be taken as an a d j u s t a b l e parameter, while w h e r e / ~ is liquid viscosity in cp ( m P a . s), T is t e m p e r a t u r e
for simplicity, p a r a m e t e r n in Eq. (8.70) can be t a k e n as unity. in kelvin, D is liquid self-diffusivity in cruZ/s, a n d Tb is nor-
In practical applications, engineers use s i m p l e r relations m a l boiling p o i n t in kelvin. F o r example, for n-C5 in w h i c h Tb
b e t w e e n tortuosity a n d porosity. F o r example, Fontes et al. is 309 K the viscosity a n d self-diffusion coefficient at 25~
352 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
(298.2 K) are 0.215 cp and 5.5 x 10 -4 cm2/s, respectively. in which A, B, and p are constants specific for each prop-
Equation (8.76) gives value of D = 5.1 x 10 -5 cm2/s. This erty and each compound. These constants for a large number
equation is developed based on very few compounds includ- of compounds are given in Ref. [23]. Equation (8.78) is de-
ing polar and nonpolar substances and is not recommended veloped for liquid hydrocarbons. Parameter I is defined in
for accurate estimation of diffusivity. However, it gives a gen- terms of refractive index (n) by Eq. (2.36) and n must be eval-
eral trend between viscosity and diffusivity. Similarly the uated from n20 using Eq. (2.114) at the same temperature at
following relation was derived between/~, k, and D [37]: which a transport property is desired. Methods of estimation
of refractive index were discussed in Section 2.6.2. The linear
relationships between 1//z or D and (1/I- 1) are shown in
(8.77) -- D = 4.2868 x 10 -9 - ~ - Figs. 8.10 and 8.11, respectively. Similar relations are shown
\Cs /
for k of several hydrocarbons in Ref. [23]. Equation (8.78)
where k is the liquid thermal conductivity in W/m. s, cs is the can reproduce original data with an average deviation of less
velocity of sound in m/s, and/z is liquid viscosity in cp. Values than 1% for hydrocarbons from Cs to C20.
of cs can be calculated from methods given in Section 6.9. Equation (8.78) is applicable for calculation of transport
Again it should be emphasized that this equation is based properties of liquid hydrocarbons at atmospheric pressures.
on very few compounds and data and it is not appropriate Coefficients A, B, and p for a number of compounds are given
for accurate prediction of D from k. However, it shows the in Table 8.11. As shown in this table, parameter p for ther-
interrelationship among the transport properties. mal conductivity is the same for all compounds as 0.1. For
Riazi et al. [23] developed a generalized relation for predic- n-alkanes coefficients of p, A, and -B/A have been correlated
tion of/z, k, and D in terms of refractive index parameter, I. to M as given in Table 8.12 [23]. Equation (8.78) with coeffi-
In fact, the Hildebrand relation for fluidity (Eq. 8.26) can cients given in Table 8.12 give average deviations of 0.7, 2.1,
be extended to thermal conductivity and diffusivity and the and 5.2% for prediction of/z, k, and D, respectively. Example
following relation can be derived based on Eq. (8.27): 8.5 shows application of this method of prediction of trans-
port properties.
/ /
j'~
; N"
z~ 2
v
O/i- I) p
FIG. 8.10---Relationship between fluidity (1//~) (/~ is in mPa. s (cp)) and refractive index
parameter I from Eq. (8.78). Adopted from Ref. [23].
9
0 n-nonane, data (p=0,546) /
- - n-flOr~13e, predicted / /
op
1'
tan -t) P
FIG. 8.11--Relationship between diffusivity (D) (D is in 10s cm2/s) and refractive
index parameter I from Eq. (8.78). With permission from Ref. [23].
TABLE 8.11--~oefficients o f Eq. (8. 78) for some liquid hydrocarbons with permission
from Ref. [231.
Compound M n2o p A -B/A T range, K
Coefficients for viscosity (mPa. s or cp)
n-Pentane 72.2 1.3575 0.747 7.8802 2.2040 144-297
n-Decane 142.3 1.4119 0.709 6.3394 2,0226 256-436
n-Eicosane 282.6 1.4424 0.649 4.5250 1.8791 311-603
Cyclopentane 70.1 1.4065 0.525 8.3935 1.6169 250-322
Methylcyclopentane 84.2 1,4097 0.584 7.9856 1.7272 255-345
n-Decylcyclopentane 210.4 1.4487 0.349 8.5664 1.3444 255-378
Cyclohexane 84.2 1.4262 0.567 8.8898 1.7153 288-345
n-Pentylcyclohexane 154.3 1.4434 0.650 6.5114 1.8300 255-378
n-Decylcyclohexane 224.4 1.4534 0.443 7.7700 t.4899 255-378
Benzene 78.1 1.5011 0.863 11.2888 1.9936 278-344
Toluene 92.1 1.4969 0.777 9.9699 1.8321 233-389
n-Pentylbenzene 148.2 1.4882 0.740 7.6244 1.8472 255-411
n-Decylbenzene 218.4 1.4832 0.565 7.0362 1.6117 255-411
Water 18 1.3330 0.750 6.3827 2.5979 273-373
Methanol 32.0 1.3288 0.919 4.8375 3.1701 268-328
Ethanol 46.1 1.3610 0.440 8.2649 1.6273 280-338
Coefficients for thermal conductivity (W/rnK)
n-Pentane 72.2 1.3575 0.1 2.6357 0.6638 335-513
n-Decane 142.3 1.4119 0.1 2.0358 0.5152 256-436
n-Eicosane 282.6 1.4424 0.1 1.6308 0.3661 427-672
Cyclopentane 70.1 1.4065 0.1 2.5246 0.6335 328-551
Methylcyclopentane 84.2 1.4097 0.1 2.3954 0.6010 328-551
Cyclohexane 84.2 1.4262 0.1 1.9327 0.4755 411-544
Benzene 78.1 1.5011 0.I 2.6750 0.6384 410-566
Toluene 92.1 1.4969 0.I 2.1977 0.5358 354-577
Ethylbenzene 106.2 1.4959 0.1 2.0965 0.5072 354-577
Coefficients for self-diffusion coefficients (105 x crn2/s)
n-Pentane 72.2 1.3575 0.270 10.4596 1.2595 195-309
n-Decane 142.3 1.4119 0.555 10.0126 1.7130 227-417
Benzene 78.1 1.5011 0.481 16.8022 1.4379 288-313
Water 18 1.3330 0.633 14.6030 2.2396 273-373
Methano] 32.0 1.3288 0.241 11.6705 1.2875 268-328
Ethanol 46.1 1.3610 0.220 15.t893 1.2548 280-338
354 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 8.12--Coefficients of Eq. (8. 78) for estimation of transport properties of liquid n-alkanes
with permission from Ref. [23].
0 Coefficientsof Eq. (8.78) for n-alkanes
1//z (cp) -1 p = 0.8036 - 5.8492 x 10-4M
A ~ 2.638 + 5.2141nM + 0.0458M - 2.408M ~
- B / A = 2.216 - 1.235 x 10-3M - 94(lnM) -5 +2.1809 • 103M-2"2, if M < 185
- B / A = 5.9644-3.625x lO-3M+788(lnM)-3-71.441M-0.4, if M > 185
1/k (W/mK) -a p = 0.1
A = 3.27857- 0.01174M+ 1.6 x 10 5M2
B = -2.50942 + 0.0139M - 2.0 x 10-SM2
105D, cm2/s p = -0.99259 + 0.02706M - 1.4936 x 10-4M 2 + 2.5383 x 10-7M 3
A = 10.06464 + 0.02191M - 2.6223 x 10-4M 2 + 6.17943 x 10-7M 3
-n = -9.80924 + 0.518156M - 3.31368 x 10-3M 2 + 5.70209 x 10-6M 3
S o l u t i o n - - F o r n-Clo, from Table 2.1, M = 142.3. F r o m the system's deviation f r o m equilibrium. Therefore, once a
Eq. (2.42) with coefficients given in Table 2.6 for 120 of n o n e q u i l i b r i u m gas is b r o u g h t into contact with a liquid, the
n-alkanes we get 120 = 0 . 2 4 8 7 5 . F r o m Eq. (2.114), n20 = system tends to a p p r o a c h e q u i l i b r i u m so t h a t the Gibbs en-
1.41185. F r o m Eq. (2.118) at T = 349 K, n r = 1.38945 a n d ergy, a n d therefore pressure, decreases with time. Once the
from Eq. (2.14) we calculate IT = 0.2368. F r o m Table 8.12 system has r e a c h e d an e q u i l i b r i u m state the p r e s s u r e as well
for 1/k we get p = 0.1, A = 1.93196, a n d B = -0.9364. Sub- as c o m p o s i t i o n of b o t h gas a n d liquid phases r e m a i n s un-
stituting these values in Eq. (8.78) we get (l/k) ~ = 1.93196 changed. S c h e m a t i c of the process is s h o w n in Fig. 8.12. If
(1/0.2368 - 1) ~ - 0.9364 or k = 0.1206 W / m . K , w h i c h dif- the gas p h a s e is h y d r o c a r b o n , dissolution of a h y d r o c a r b o n
fers from the e x p e r i m e n t a l value b y 1.3%. DIPPR gives gas in an oil causes increase in oil volume a n d height of liquid
value of 0.1215 W / m . K [45]. r (Lo) increases. F o r the case of nitrogen, the result is opposite
a n d dissolution of N2 causes decrease in the oil volume. In
f o r m u l a t i o n of diffusion process in each phase, the Fick's law
8.5 M E A S U R E M E N T OF D I F F U S I O N a n d m a t e r i a l b a l a n c e equations are a p p l i e d for each c o m p o -
COEFFICIENTS IN RESERVOIR FLUIDS nent in the system. At the interphase, e q u i l i b r i u m criterion
is i m p o s e d on each c o m p o n e n t ( [ /od =/~igas ). In addition, at
M o l e c u l a r diffusion is an i m p o r t a n t p r o p e r t y n e e d e d in simu- the i n t e r p h a s e the rates of diffusion in each p h a s e are equal
lation a n d evaluation of several oil recovery processes. E x a m - for each c o m p o n e n t . A semianalytical m o d e l for calculation
ples are vertical miscible gas flooding, n o n t h e r m a l recovery of of rates of diffusion process in b o t h phases of gas a n d liquid
heavy oil b y solvent injection, a n d solution-gas-derived reser- is given b y Riazi [56]. The m o d e l is a c o m b i n a t i o n of mate-
voirs. In these cases w h e n p r e s s u r e is r e d u c e d b e l o w b u b b l e rial b a l a n c e a n d v a p o r - l i q u i d e q u i l i b r i u m calculations. W h e n
p o i n t of oil, gas b u b b l e s are f o r m e d a n d the rate of t h e i r diffu- the diffusion processes c o m e to a n end the system will be at
sion is the controlling step. Attempts in m e a s u r e m e n t of gas equilibrium. Diffusion coefficients n e e d e d in the m o d e l are
diffusivity in h y d r o c a r b o n s u n d e r high-pressure conditions c a l c u l a t e d t h r o u g h a m e t h o d such as Eq. (8.67). The m o d e l
goes b a c k to the early 1930s a n d has c o n t i n u e d to the recent predicts c o m p o s i t i o n of each phase, location of the liquid in-
years [37, 51-57]. I n general, m e t h o d s of m e a s u r i n g diffusion terface, a n d p r e s s u r e of the system versus time.
coefficients in h y d r o c a r b o n systems can be divided into two To evaluate the p r o p o s e d m e t h o d , p u r e m e t h a n e was p l a c e d
categories. In the first category, d u r i n g the e x p e r i m e n t sam- on p u r e n-pentane at 311 K (100~ a n d 102 b a r in a PVT cell
ples of the fluid are t a k e n at various times a n d are analyzed b y of 21.943 c m height a n d 2.56 c m diameter. The initial vol-
gas c h r o m a t o g r a p h y o r o t h e r analytical tools [37, 55]. In the u m e of liquid was 35% of the cell volume. Pressures were
s e c o n d category, s a m p l e s are not a n a l y z e d b u t self-diffusion m e a s u r e d a n d r e c o r d e d m a n u a l l y at selected times a n d con-
coeff• are m e a s u r e d by e q u i p m e n t such as N M R a n d t i n u o u s l y on a strip chart. The liquid level was m e a s u r e d
t h e n b i n a r y diffusion coefficients are c a l c u l a t e d [41]. Other m a n u a l l y with a p r e c i s i o n of 4-0.02 mm. M e a s u r e m e n t s were
m e t h o d s involve m e a s u r i n g v o l u m e of gas dissolved in oil c o n t i n u e d until there is no c h a n g e in b o t h p r e s s u r e a n d liq-
versus t i m e at c o n s t a n t p r e s s u r e in o r d e r to d e t e r m i n e gas u i d length at w h i c h the system reaches equilibrium. Diffu-
diffusivity in reservoir fluids [43]. sion coefficients were c o r r e c t e d so that p r e d i c t e d p r e s s u r e
In the early 1990s a s i m p l e m e t h o d to d e t e r m i n e diffu- curve versus t i m e m a t c h e s the e x p e r i m e n t a l d a t a as s h o w n
sion coefficients in b o t h g a s - g a s a n d g a s - l i q u i d for b i n a r y in Fig. 8.13. W h e n diffusivities calculated by Eq. (8.67) are
a n d m u l t i c o m p o n e n t systems at high pressures w i t h o u t com- m u l t i p l i e d by 1.1 the m o d e l p r e d i c t i o n perfectly m a t c h e s
p o s i t i o n a l m e a s u r e m e n t was p r o p o s e d b y Riazi [56]. I n this e x p e r i m e n t a l data. This t e c h n i q u e m e a s u r e s diffusion coef-
m e t h o d , gas a n d oil are initially p l a c e d in a PVT u n d e r con- ficient of Ca-C5 in liquid p h a s e at 311 K a n d 71 b a r s as
stant t e m p e r a t u r e condition. As the system a p p r o a c h e s its 1.51 x 10 -4 cm2/s, while the e x p e r i m e n t a l d a t a r e p o r t e d by
e q u i l i b r i u m the p r e s s u r e as well as g a s - l i q u i d i n t e r p h a s e po- R e a m e r et al. [52] is 1.43 • 10 4 cm2/s. Diffusion coefficients
sition in the cell vary a n d are m e a s u r e d versus time. B a s e d on of C1-C5, in b o t h gas a n d liquid phases, versus pressure, a n d
the rate of change of p r e s s u r e o r the liquid level, rate of diffu- c o m p o s i t i o n are s h o w n in Figs. 8.14 a n d 8.15, respectively.
sion in each p h a s e can be d e t e r m i n e d [56]. The m e c h a n i s m of Diffusivity of m e t h a n e in heavy oils (bitumens) at 50 b a r a n d
diffusion process is b a s e d on the principle of t h e r m o d y n a m i c 50~ is within the o r d e r of m a g n i t u d e of 5 • 10 -4 cm2/s, while
e q u i l i b r i u m a n d the deriving force in m o l e c u l a r diffusion is e t h a n e diffusivity in such oils is a b o u t 2 • 10 -4 cm2/s [55].
8. APPLICATIONS: ESTIMATION OF TRANSPORT PROPERTIES 355
O_<z_<L
GAS 1 zg z = Zo if z ~ Lo
Z= Zg + Lo if z >_Lo
Lo
!!!ill!, i!i!i!!iiiiii!i
FIG. 8.12--Schematic and dimensions of a constant volume cell. Taken with per-
mission from Ref. [23].
This m e t h o d can be extended to m u k i c o m p o n e n t systems model and the assumptions made in its formulation. Methods
and it has been successfully used to measure gas diffusiv- that use unrealistic assumptions, i.e., neglecting natural con-
ity in heavy oils [57, 58]. In this method, pressure mea- vection terms when it exists, or oversimplified b o u n d a r y con-
surement is more accurate than measuring the interphase ditions (i.e., semiinfinite assumption) lead to predictions that
location. Furthermore, the initial measurements are m o r e do not m a t c h the entire curve. In these cases, reported diffu-
critical than measurements near the final equilibrium condi- sion coefficients are based on a portion of experimental data
tion. The a m o u n t of initial liquid or gas determines diffusivity and this is the reason that in such cases differences as large
of which phase can be measured more accurately [56]. As it as ~ 100% are reported for diffusion coefficients in liquids at
can be seen from Fig. 8.13 once a correct value of diffusion co- high pressures for the same systems u n d e r the same condi-
efficient is used, the model prediction matches experimental tions. The technique can also be used to measure diffusivity
data throughout the curve. This confirms the validity of the in porous media by placing a reservoir core in the b o t t o m of a
PVT cell saturated initially with liquid oil. With such experi-
ments and availability of more data, Eq. (8.67) can be further
studied, modified, and improved.
o ex~nentaf
100" : R~WP~s~n I~;~ CoeH~ts
. . . . . . . . . . . . ::
i- ~ ----,-4=~ GAS I
j,''""" |I
gO"
t.5- /
/
k'- 85 ~
I
/
/t .,
O
Q
75 ~
-8
0.3~
-0,3
E
o ""X "--"
x%4-
-0.2 ~
\
i LI.
o
\'\. 0.t
CI
"x
-0.05
Hg
H20
-_-_-_ pV, -_-_
//////////////////////I
L-___- ______-
10 11 12
31 .... I
1[ II
(1) anti-vibration table; (2) light source; (3) optical rail; (4) light diffuser; (5)
iris; (6) green filter; (7) thermostated interfacial tension cell with optical flats;
(8) syringe to form pendant drops; (9) thermostat; (10) photomacrographic
Tessovar zoom lens; (11) CCD camera; (12) computer with digitizing board,
358 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 8.15--Effect of characterization method on prediction of interracial tension of some petroleum fractions through Eq. (8.91).
% E r r o r on prediction of IFT ~
Fraction Tb, K SG P25, g/cm 3 M cr at 25~ mN/m Method 1 Method 2 Method 3
1 429 0.769 0.761 130.9 22.3 26.5 2.7 14.8
2 499 0.870 0.863 167.7 30.7 -29.3 -7.5 -2.0
3 433 0.865 0.858 120.2 29.2 -15.4 3.4 22.9
4 505 0.764 0.756 184.4 25.6 -4.7 7.8 -10.9
Overall 19.0 5.4 12.7
aExperimental data are taken from Miqueu et al. [59]. Method 1: Tc and Pc from Kesler-Lee (Eqs. (2.69) and (2.70)) and w from Lee-Kesler (Eq. 2.105). Method 2:
Tc and Pc from API-TDB (Eqs. (2.65) and (2.66)) [5] and o) from Lee-Kesler (Eq. 2.105). Method 3: Tc and Pc from Twu (Eqs. (2.80) and (2.86)) and ~0from Lee-Kesler
(Eq. 2.105).
for calculation of surface tension of undefined petroleum frac- An evaluation of various methods for prediction of surface
tions: tension of n-alkanes is shown in Fig. 8.2 I. Data are taken from
DIPPR [45]. The most accurate method for calculation of sur-
face tension of pure hydrocarbons is through Eq. (8.85) with
(8.92) values of parachor from Table 8.14 or Eq. (8.86). Method of
(0.85 - 0.19o>) T12/11 Block and Bird (Eq. 8.88) or Eq. (8.90) for petroleum fractions
Pa= also provide reliable values for surface tension of pure hydro-
(Pc/lO) 9/11
carbons with average errors of about 3%. Equation (8.90) is
In this method, n in Eq. (8.85) is equal to 11/3 or 3.6667. In the perhaps the most accurate method as it gives the lowest er-
above equation, Tc and Pc are in kelvin and bar, respectively, ror for surface tension of n-alkanes (error of ~2%), while it
is in mN/m (dyn/cm), and p is in g/cm 3. Predicted values of is proposed for petroleum fractions. Equations (8.92)-(8.94)
surface tension by this method strongly depend on the char- give generally very large errors, especially for hydrocarbons
acterization method used to calculate To, Pc, and M. For four heavier than C10. Equation (8.93) is developed for petroleum
petroleum fractions predicted values of surface tension by fractions ranging from Cs to C10 and Eq. (8.94) is not suitable
three different characterization methods described in Chap- for heavy hydrocarbons as shown in Fig. 8.21.
ter 2 are given in Table 8.15. As it is seen from this table, the Interfacial tension (IFT) between hydrocarbon and water
API method of calculating Tc, Pc, w, and M (Section 2.5) yields is important in understanding the calculations related to oil
the lowest error for estimation of surface tension. Miqueu recovery processes. The following simple relation is suggested
et al. [59] used the pseudocomponent method (Section 3.3.4, in the API-TDB [5] to calculate anw from surface tension of
Eq. 3.39) to develop the following equation for estimation of hydrocarbon ~H and that of water ~w:
parachor and surface tension of defined petroleum fractions
(8.95) aHw = ffH -~ GW -- 1.10 (GHGW)1/2
with known PNA composition.
Use of this method is also demonstrated in Example 8.7. An-
other relation for IFT of hydrocarbon-water systems under
reservoir conditions was proposed by Firoozabadi and Ramey
Pa = x p P a , p + XNPa,N + XAPa,A [16, 65] in the following form:
(8.93) Pa,e = 27.503 + 2.9963M (8.96) amv = 111 (~, - p H ) 1"024 (T/TcH) -1"25
Pa,s = 18.384 + 2.7367M where cq-iwis the hydrocarbon-water IFT in dyn/cm (mN/m),
Pa,A = 25.511 + 2.8332M Pw and ~ are water and hydrocarbon densities in g/cm 3, T is
temperature in kelvin, and TcHis the pure hydrocarbon critical
where xe, XN, and XA are mole fractions of paraffins, naph- temperature in kelvon. Errors as high as 30% are reported for
thenes, and aromatics in the fraction. Units are the same as this correlation [16]. IFT similar to surface tension decreases
in Eq. (8.92). Experimental data of Darwish et al. [63] on with increase in temperature. For liquid-liquids, such as oil-
surface tension consist PNA distribution of some petroleum water systems, IFT usually increases slightly with pressure;
fractions. For undefined fractions, the PNA composition may however, for gas-liquid systems, such as methane-water, the
be estimated from methods of Chapter 3. For cases where ac- IFT slightly decreases with increase in pressure.
curate PNA composition data are not available the parachor
number of an undefined petroleum fraction may be directly Example 8.7--A kerosene sample has boiling point and spe-
calculated from molecular weight of the fraction (M), using cific gravity of 499 K and 0.87, respectively. Calculate the IFT
the following correlation originally provided by Fawcett and of this oil with water at 25~ Liquid density of the fraction
recommended by Miqueu et al. [59]: at this temperature is 0.863 g/cm 3.
(8.94) Pa = 81.2+2.448M value of n in Eq. (8.85) = 11/3
Solution--Tb = 499 K and SG = 0.87. From Eq. (2.51), M =
In this method, only M and liquid density are needed to 167.7. Parachor can be calculated from the Fawcett method as
calculate surface tension at atmospheric pressure. Firooz- given in Eq. (8.94): Pa = 491.73. From data p25 = 0.863 g/cm 3.
abadi [64] also provided a similar correlation (Pa = 11.4 + Substituting values of M, Pa, and p25 (for pL) in Eq. (8.85)
3.23M - 0.0022M2), which is reliable up to C10, but for heav- with n = 11/3 gives a2s = 30.1 mN/m, where in comparison
ier hydrocarbons it seriously underpredicts values of surface with the experimental value of 30.74 mN/m [59] the error is
tension. -2.1%. When using Eq. (8.85), the value of pV is neglected
8. APPLICATIONS: E S T I M A T I O N OF T R A N S P O R T P R O P E RT IE S 361
35
30 0 ,..~176
"";'; ~ ~ i
O ~ _ t Y """ 0 Data (DIPPR)
#2/ "-., Method,
~ "\ ~ Method 2
~ 20 ~ \ - ~ -Method3
\-. Method 4
q ~ ~.~ ~ "\ 9 Method 5
.4[
15 l~ ~ 4h. o.- ~ ~- . . . . . Method 6
m ~ Method7
m ~ . . Method 8
10
0 5 10 15 20 25 30 35 40
Carbon Number
with respect to pL at atmospheric pressure. To calculate IFT To estimate viscosity of pure liquids, Eq. (8.17) is recom-
of water-oil, Eq. (8.95) can be used. From Table 8.13 for mended and for a defined hydrocarbon mixture Eq. (8.18)
water at 25~ aw = 72.8 mN/m. From Eq. (8.95), aW-oil = can be used. For petroleum fractions when kinematic vis-
72.8 + 30.1 - 1.1(72.8 • 30.1) 1/2 = 51.4 mN/m. To calculate cosity at 100~ (37.8~ is available, Eq. (8.19) can be used.
aw-oi] from Eq. (8.96), TcH is calculated from the API method When two petroleum fractions are mixed, Eq. (8.20) is use-
(Eq. 2.65) as 705 K and at 25~ Pw--0.995 g/cm 3. From ful. Viscosity of liquid hydrocarbons at high pressure can
Eq. (8.96), aW-oil = 40.9 mN/m. This is about 20% less than be calculated from Eq. (8.22). For crude oil at atmospheric
the value calculated from Eq. (8.95). As mentioned before pressure Eq. (8.25) is useful; however, for reservoir fluids
large error may be observed from Eq. (8.96) for calculation of Eq. (8.12) can be used for both gases and liquids or their
IFT. mixtures.
Thermal conductivity of pure hydrocarbon gases at low
pressures should be calculated from Eq. (8.33) and for those
8.7 SUMMARY AND RECOMMENDATIONS for which the coefficients are not available, Eq. (8.34) should
be used. For defined hydrocarbon gas mixtures Eq. (8.35)
In this chapter, methods and procedures presented in the pre- and for undefined petroleum vapor fractions Eq. (8.37) is
vious chapters are used for estimation of four transport prop- recommended. For vapor fractions at temperatures in which
erties: viscosity, thermal conductivity, diffusion coefficient, Eq. (8.37) is not applicable, Eq. (8.36) is recommended. For
and surface tension. In general semitheoretical methods for hydrocarbon gases at high pressures Eq. (8.39) may be used
estimation of transport properties have wider range of ap- and if not possible Eq. (8.38) can be used for both pure gases
plications than do pure empirical correlations and their de- and undefined gas mixtures.
velopment and applications are discussed in this chapter. A For pure hydrocarbon liquids at low pressures, Eq. (8.42) is
s u m m a r y of recommended methods is given below. recommended and for those compounds whose thermal con-
For calculation of viscosity of pure gases at atmospheric ductivity at two reference points are not known, Eq. (8.43)
pressure, Eq. (8.3) should be used and for compounds for is recommended. For undefined liquid petroleum fractions,
which the coefficients are not available, Eq. (8.6) may be used. Eq. (8.46) and for defined liquid mixtures Eq. (8.48) can
For defined gas mixture when viscosity of components are be used. For fractions without any characterization data,
known Eq. (8.7) or (8.8) can be used. For hydrocarbon gases at Eq. (8.50) can be used for determination of approximate value
high pressure, viscosity can be calculated from Eq. (8.12) and of thermal conductivity. For fractions with only boiling point
for nonhydrocarbons Eq. (8.13) can be used. For estimation available, Eq. (8.51) should be used and for coal liquid frac-
of viscosity of natural gas at atmospheric pressure, Eq. (8.14) tions Eq. (8.52) is recommended. For liquid fractions at high
and at higher pressure Eq. (8.15) are recommended. pressures, Eq. (8.53) is recommended.
362 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
[6] Yoon, E and Thodos, G., "Viscosity of Nonpolar Gaseous [27] Smith, J. M., Van Ness, H. C., and Abbott, M. M., Introduction
Mixtures at Normal Pressures," American Institute of Chemical to Chemical Engineering Thermodynamics, 5th ed,
Engineers Journal, Vol. 16, No. 2, 1970, pp. 300-304. McGraw-Hill, New York, 1996.
[7] Stiel, L. I. and Thodos, G., "The Viscosity of Nonpolar Gases at [28] McCabe, W. L, Smith, J. C., and Harriot, P., Unit Operations
Normal Pressures," American Institute of Chemical Engineers of Chemical Engineering, 5th ed., McGraw-Hill, New York, 2001.
Journal, Vol. 7, No. 4, 1961, pp. 611-615. [29] Mathias, E M., Parekh, V. S., and Miller, E. J., "Prediction and
[8] Wauquier, J.-P., Petroleum Refining, VoI. 1: Crude Oil, Petroleum Correlation of the Thermal Conductivity of Pure Fluids and
Products, Process Flowsheets, Editions Technip, Pads, 1995. Mixtures, Including the Critical Region," Industrial and
[9] Riazi, M. R. and Whitson, C. H., "Estimating Diffusion Engineering Chemistry Research, Vol. 41, 2002, pp. 989-999.
Coefficients of Dense Fluids," Industrial and Engineering [30] Riazi, M. R. and Faghri, A., "Thermal Conductivity of Liquid
Chemistry Research, Vol. 32, No. 12, 1993, pp. 3081-3088. and Vapor Hydrocarbon Systems: Pentanes and Heavier at Low
[10] Danner, R. P. and Daubert, T. E., Manual for Predicting Chemical Pressures," Industrial and Engineering Chemistry, Process
Process Design Data, Design Institute for Physical Property Design and Development, Vol. 24, No. 2, 1985, pp. 398-401.
Data (DIPPR), AIChE, New York, 1986. Rowley, R. L., Wilding, [31] Stiel, L. I. and Thodos, G., "The Thermal Conductivity of
W. V., Oscarson, J. L., Zundel, N. A., Marshall, T. L., Dauhert, Nonpolar Substances in the Dense Gaseous and Liquid
T. E., and Danner, R. P., DIPPR Data Compilation of Pure Regions," American Institute of Chemical Engineers Journal,
Compound Properties, Design Institute for Physical Properties, Vol. 10, No. 1, 1964, pp. 26-30.
Taylor & Francis, New York, 2002 (http://dippr.byu.edu/). [32] Riazi, M. R. and Faghri, A., "Prediction of Thermal
[11] Jossi, J. A., Stiel, L. I., and Thodos, G., "The Viscosity of Pure Conductivity of Gases at High Pressures," American Institute
Substances in the Dense Gaseous and Liquid Phases," of Chemical Engineers Journal, Vol. 31, No. 1, 1985,
American Institute of Chemical Engineers Journal, Vol. 8, 1962, pp. 164-167.
pp. 59-63. [33] Tsonopoulos, C., Heidman, J. L., and Hwang, S.-C.,
[12] Lohrenz, J., Bray, B. G., and Clark, C. R., "Calculating Viscosity Thermodynamic and Transport Properties of Coal Liquids, An
of Reservoir Fluids from Their Compositions," Journal of Exxon Monograph, Wiley, New York, 1986.
Petroleum Technology, October 1964, pp. 1171-1176. [34] Pedersen, K. S. and Fredenslund, Aa., "An Improved
[13] Stiel, L. I. and Thodos, G., "The Viscosity of Polar Substances Corresponding States Model for the Prediction of Oil and Gas
in the Dense Gaseous and Liquid Regions," American Institute Viscosities and Thermal Conductivities," Chemical Engineering
of Chemical Engineers Journal, Vol. 10, No. 2, 1964, Science, Vol. 42, 1987, pp. 182-186.
pp. 275-277. [35] Duda, J. L., "Personal Notes in Transport Phenomena,"
[14] Lee, A., Gonzalez, M. H., and Eakin, B. E., "The Viscosity of Instructor's note for ChE 545 Course, Chemical Engineering
Natural Gases," Journal of Petroleum Technology, August 1966, Department, Pennsylvania State University, University Park,
pp. 997-1000. PA, 1977.
[15] Ahmed, T., Reservoir Engineering Handbook, Gulf Publishing, [36] Stiel, L. I. and Thodos, G., "Lennard-Jones Constants Predicted
Houston, TX, 2000. from Critical Properties," Journal of Chemical and Engineering
[16] Danesh, A., PVT and Phase Behavior of Petroleum Reservoir Data, Vol. 7, No. 2, 1962, pp. 234-236.
Fluids, Elsevier, Amsterdam, 1998. [37] Riazi, M. R. and Daubert, T. E., "Application of Corresponding
[17] Speight, J. G., The Chemistry and Technology of Petroleum, States Principles for Prediction of Self-Diffusion Coefficients in
3rd ed., Marcel Dekker, New York, 1999. Liquids," American Institute of Chemical Engineers Journal,
[18] Poling, B. E., Prausnitz, J. M., and O'Connell, J. P., Properties of Vol. 26, No. 3, 1980, pp. 386-391.
Gases and Liquids, 5th ed., Mc-Graw Hill, New York, 2000. [38] Slattery, J. C. and Bird, R. B., "Calculation of the Diffusion
Reid, R. C., Prausnitz, J. M., and Poling, B. E., Properties of Coefficients of Dilute Gases and of the Self-Diffusion
Gases and Liquids, 4th ed., Mc-Graw Hill, New York, 1987. Coefficients of Dense Gases," American Institute of Chemical
[19] Crude Oil I~elds and Product Properties, Cud Thomas Baird IV, Engineers Journal, Vol. 4, 1958, pp. 137 142.
1222 Vezenaz, Geneva, Switzerland, June 1981. [39] Takahashi, S., "Preparation of a Generalized Chart for the
[20] Riazi, M. R., Mahdi, K., and Alqallaf, M., "A Generalized Diffusion Coefficients of Gases at High Pressures," Journal of
Correlation for Viscosity of Hydrocarbons Based on Chemical Engineering of Japan, Vol. 7, 1974, pp. 417-421.
Corresponding States Principles and Molar Refraction," [40] Sigmund, E M., "Prediction of Molecular Diffusion at Reservoir
Journal of Chemical and Engineering Data, American Chemical Conditions. Part I---Measurement and Prediction of Binary
Society, Vol. 50, 2005. Dense Gas Diffusion Coefficients," Canadian Journal of
[21] Riazi, M. R. and A1-Otaibi, G. N., "Estimation of Viscosity of Petroleum Technology, Vol. 15, No. 2, 1976, pp. 48-57.
Petroleum Fractions," Fuel, Vol. 80, 2001, pp. 27-32. [41] Dawson R., Khoury, E, and Kobayashi, R., "Self-Diffusion
[22] Hildebrand, J. H., "Motions of Molecules in Liquids: Viscosity Measurements in Methane by Pulsed Nuclear Magnetic
and Diffusivity," Science, Vol. 174, 1971, pp. 490-493. Resonance," American Institute of Chemical Engineers Journal,
[23] Riazi, M. R., A1-Enezi, G., and Soleimani, S., "Estimation of Vol. 16, No. 5, 1970, pp. 725-729.
Transport Properties of Liquids," Chemical Engineering [42] Whitson, C. H. and Brule, M. R., Phase Behavior, Monograph
Communications, Vol. 176, 1999, pp. 175-193. Vol. 20, Society of Petroleum Engineers, Richardson, TX, 2000.
[24] Riazi, M. R. and Roomi, Y., "Use of the Refractive Index in the [43] Renner, T. A., "Measurement and Correlation of Diffusion
Estimation of Thermophysical Properties of Hydrocarbons and Coefficients for CO2 and Rich Gas Applications," SPE Reservoir
Their Mixtures," Industrial and Engineering Chemistry Engineering, No. 3 (May), 1988, pp. 517-523.
Research, Vol. 40, No. 8, 2001, pp. 1975-1984. [44] Eyring, H., Significant Liquid Structure, Wiley, New York, 1969.
[25] Chung, T. H., Ajlan, M., Lee, L. L., and Starling, K. E., [45] AIChE DIPPR| Database, Design Institute for Physical
"Generalized Multiparameter Correlation for Nonpolar and Property Data (DIPPR), EPCON International, Houston, TX,
Polar Fluid Transport Properties," Industrial and Engineering 1996.
Chemistry Research, Vol. 27, 1988, pp. 671-679. [46] Wilke, C. R., "Diffusional Properties of Multicomponent Gases,"
[26] Quinones-Cisneros, S. E., Zeberg-Mikkelsen, C. K., and Stenby, Chemical Engineering Progress, Vol. 46, No. 2, 1950, pp. 95-104.
E. H., "The Friction Theory (f-Theory) for Viscosity Modeling," [47] Saidi, A. M., Reservoir Engineering of Fracture Reservoirs,
Fluid Phase Equilibria, Vol. 169, 2000, pp. 249-276. TOTAL Edition Presse, Paris, 1987, Ch. 8.
364 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
[48] Amyx, J. W., Bass, D. M., and Whiting, R. L., Petroleum [58] Upreti, S. R. and Mehrotra, A. K., "Diffusivity of CO2, CH4,
Reservoir Engineering, Physical Properties, McGraw-Hill, C2H6 and N2 in Athabasca Bitumen," The Canadian Journal of
New York, pp. 129 (total of 600 pp.), 1960. Chemical Engineering, Vol. 80, 2002, pp. 117-125.
[49] Langness, G. L., Robertson, J. O., Jr., and Chilingar, G. V., [59] Miqueu, C., Satherley, J., Mendiboure, B., Lachiase, J., and
Secondary Recovery and Carbonate Reservoirs, Elsevier, Graciaa, A., "The Effect of P/N/A Distribution on the Parachors
New York, 304 pp. 1972. of Petroleum Fractions," Fluid Phase Equilibria, Vol. 180, 2001,
[50] Fontes, E. D., Byrne, E, and Hernell, O., "Put More Punch Into pp. 327-344.
Catalytic Reactors," Chemical Engineering Progress, Vol. 99, [60] Millette, J. P., Scott, D. S., Reilly, I. G., Majerski, P., Piskorz, J.,
No. 3, 2003, pp. 48-53. Radlein, D., and de Bruijin, T. J. W., "An Apparatus for the
[51] Hill, E. S. and Lacy, W. N., "Rate of Solution of Methane in Measurement of Surface Tensions at High Pressures and
Quiescent Liquid Hydrocarbons," Industrial and Engineering Temperatures," The Canadian Journal of Chemical Engineering,
Chemistry, Vol. 26, 1934, pp. 1324-1327. Vol. 80, 2002, pp. 126-134.
[52] Reamer, H. H., Duffy, C. H., and Sage, B. H., "Diffusion [61] Rao, D. N., Girard, M., and Sayegh, S. G., "The Influence of
Coefficients in Hydrocarbon Systems: Methane-Pentane in Reservoir Wettability on Waterflood and Miscible Flood
Liquid Phase," Industrial and Engineering Chemistry, Vol. 48, Performance," Journal of Canadian Petroleum Technology,
1956, pp. 275-282. Vol. 31, No. 6, 1992, pp. 47-55.
[53] Lo, H. Y., "Diffusion Coefficients in Binary Liquid n-Alkanes [62] Miqueu, C., Broseta, D., Satherley, J., Mendiboure, B.,
Systems," Journal of Chemical and Engineering Data, Vol. 19, Lachiase, J., and Graciaa, A., "An Extended Scaled
No. 3, 1974, pp. 239-241. Equation for the Temperature Dependence of the Surface
[54] McKay, W. N., "Experiments Concerning Diffusion of Tension of Pure Compounds Inferred From an Analysis of
Mulficomponent Systems at Reservoir Conditions," Journal of Experimental Data," Fluid Phase Equilibria, Vol. 172, 2000,
Canadian Petroleum Technology, Vol. 10 (April-June), 197 t, pp. 169-182.
pp. 25-32. [63] Darwish, E., A1-Sahhaf, T. A., and Fahim, M. A., "Prediction
[55] Nguyen, T. A. and Farouq All, S. M., "Effect of Nitrogen on the and Correlation of Surface Tension of Naphtha Reformate
Solubility and Diffusivity of Carbon Dioxide into Oil and Oil and Crude Oil," Fuel, Vol. 74, No. 4, 1995,
Recovery by the Immiscible WAG Process," Journal of Canadian pp. 575-581.
Petroleum Technology, Vol. 37, No. 2, 1998, pp. 24-31. [64] Firoozabadi, A., Katz, D. L., and Soroosh, H., SPE
[56] Riazi, M. R., "A New Method for Experimental Measurement of Reservoir Engineering, Paper No. 13826, February 1988,
Diffusion Coefficient in Reservoir Fluids," Journal of Petroleum pp. 265-272.
Science and Engineering, Vol. 14, 1996, pp. 235-250. [65] Firoozabadi, A. and Ramey, Jr., H. J., "Surface Tension of
[57] Zhang, Y. P., Hyndman, C. L., and Maini, B. B., "Measurement Water-Hydrocarbon Systems at Reservoir Conditions," Journal
of Gas Diffusivity in Heavy Oils," Journal of Petroleum Science of Canadian Petroleum Technology, Vol. 27, No. 3, May-June
and Engineering, Vol. 25, 2000, pp. 37-47. 1988, pp. 41-48.
MNL50-EB/Jan. 2005
Applications: Phase
Equilibrium Calculations
365
P Density at a given temperature and pressure, GC Gas condensate (a type of reservoir fluid defined
g/cm 3 (molar density unit: cma/mol) in Chapter 1)
PM Molar density at a given temperature and pres- GOR Gas-to-oil ratio, scf/labl
sure, m o l / c m 3 HFT Hydrate formation temperature
CO Acentric factor defined by Eq. (2.10), dimen- IFT Interracial tension
sionless LLE Liquid-liquid equilibria
Chemical potential of component i in a mixture LMP Low molecular weight n-paraffins (i.e., C3, n-Cs,
defined by Eq. (6.115) n-C7)
8i Solubility parameter for i defined in Eq. (6.147), LVS liquid-vapor-solid
(J/cm3) 1/2 o r (cal/cm3) 1/2 LS Liquid-solid
Yi Activity coefficient of component i in liquid so- MeOH Methanol
lution defined by Eq. (6.112), dimensionless PR Peng-Robinson EOS (see Eq. 5.39)
ACpi Difference between heat capacity of liquid and SRK Soave-Redlich-Kwong EOS given by Eq. (5.38)
solid for pure component i at its melting (freez- and parameters in Table 5.1
ing) point (= cLi -- CSi), J/tool - K SAFT Statistical associating fluid theory (see
t,~f Heat of fusion (or latent heat of melting) for Eq. 5.98)
pure component i at the freezing point and SLE Solid-liquid equilibrium
1.013 bar, J/mol scf Standard cubic foot (unit for volume of gas at
1 atm and 60~
stb Stock tank barrel (unit for volume of liquid oil
Superscript at 1 atm and 60~
L Value of a property at liquid phase VABP Volume average boiling point defined by
V Value of a property at vapor phase Eq. (3.3).
S Value of a property at solid phase VLE Vapor-liquid equilibrium
VLSE Vapor-liquid-solid equilibrium
VS Vapor-solid
Subscripts VSE Vapor-solid equilibrium
A Value of a property for component A WAT Wax appearance temperature
A Value of a property for asphaltenes WPT Wax precipitation temperature
C Value of a property at the critical point %AAD Average absolute deviation percentage defined
i,i Value of a property for component i or j in a by Eq. (2.135)
mixture %AD Absolute deviation percentage defined by
L Value of a property for liquid phase Eq. (2.134)
M Value of a property at the melting point of a wt% Weight percent
substance
pc Pseudocritical property ONE OF THE MAIN APPLICATIONS o f science of thermodynamics
S Value of a property at the solid phase in the petroleum industry is for the prediction of phase behav-
S Value of a property for solvent (LMP) ior of petroleum fluids. In this chapter calculations related to
s Specific property (quantity per unit mass) vapor liquid and solid-liquid equilibrium in petroleum flu-
T Values of property at temperature T ids are presented. Their application to calculate gas-oil ratio,
tc True critical property crude oil composition, and the amount of wax or asphaltene
tr Value of a property at the triple point precipitation in oils under certain conditions of temperature,
20 Values of property at 20~ pressure, and composition is presented. Methods of calcula-
7+ Values of a property for C7+ fraction of an oil tion of wax formation temperature, cloud point temperature
of crude oils, determination of onset of asphaltene, hydrate
formation temperature, and methods of prevention of solid
Acronyms formation are also discussed. Finally application of character-
ABSA Alkyl benzene sulfonic acid ization techniques, methods of prediction of transport prop-
API-TDB American Petroleum Institute--Technical Data erties, equations of state, and phase equilibrium calculations
Book (see Ref. [12]) are demonstrated in modeling and evaluation of gas injection
BIP Binary interaction parameter projects.
bbl Barrel, unit of volume of liquid as given in Sec-
tion 1.7.11
CPT Cloud-point temperature 9.1 TYPES OF P H A S E EQUILIBRIUM
cp Centipoise, unit of viscosity, (1 cp = 0.01 p = CALCULATIONS
0.01 g- cm. s = 1 mPa. s = 10 -3 kg/m. s)
cSt Centistoke, unit of kinematic viscosity, (1 cSt = Three types of phase equilibrium, namely, vapor-liquid
0.01 St = 0.01 cm2/s) (VLE), solid-liquid (SLE), and liquid-liquid (LLE), are of
EOR Enhanced oil recovery particular interest in the petroleum industry. Furthermore,
LOS Equation of state vapor-solid (VSE), vapor-liquid-solid (VLSE), and vapor-
FH Flory-Huggins liquid-liquid (VLLE) equilibrium are also of importance in
9. A P P L I C A T I O N S : P H A S E E Q U I L I B R I U M CALCULATIONS 367
. . . . . . !
iiiiiiiiii!iiiii
::::::::::::::::::::::::::::: "9 ..' .'". ' . ' Vmoles ..'.'."
'.'."
iiii Pe(rdleumFiuid iiii ".'.'.' Yi .'.'."
:i:i F moles
t
i ii i Compositionzi i i LiquidSolution i i all phases
i:[: L moles i:i
i:i: TF,PF iii at T and P
!i!i x,L !il
....
9 .........................
iiiiiiiiiiiiiiiiiiiiiiiiiiiii
i:i:i:i:i:i:i:i:i:i:i:i:i:i:i
ili::i::iii9"';oTi~~o'l~tion""iiiiii!i 1
[i!i!i!!!! S moles i!!!!i!i]
:::::::::::::::::::::::::::::
iiiiiiiiiiiiiiiiiiiii!iiiiiii li i i i ............
i .......x.,.................iiiiii
Non-equilibrium state Equilibrium state
FIG. 9.1--Typical vapor-liquid-solid equilibrium for solid precipitation.
calculations related to petroleum and natural gas production. calculations are the most widely used VLE calculations by
VLE calculations are needed in design and operation of sepa- both chemical and reservoir engineers in the petroleum pro-
ration units such as multistage surface separators at the sur- cessing and production.
face facilities of production fields, distillation, and gas absorp-
tion columns in petroleum and natural gas processing as well
as phase determination of reservoir fluids. LLE calculations 9.2 VAPOR-LIQUID EQUILIBRIUM
are useful in determination of a m o u n t of water dissolved in CALCULATIONS
oil or a m o u n t of oil dissolved in water u n d e r reservoir con-
ditions. SLE calculations can be used to determine a m o u n t VLE calculations are perhaps the most important types
and the conditions at which a solid (wax or asphaltene) m a y of phase behavior calculations in the petroleum industry.
be formed from a petroleum fluid. Cloud-point temperature They involve calculations related to equilibrium between two
(CPT) can be accurately calculated t h r o u g h SLE calculations. phases of liquid and vapor in a multicomponent system. Con-
VSE calculation is used to calculate hydrate formation and sider a fluid mixture with mole fraction of each c o m p o n e n t
the conditions at which it can be prevented. shown by zi is available in a sealed vessel at T and P. Under
Schematic of a system at vapor-liquid-solid equilibrium these conditions assume the fluid can exist as both vapor and
(VLSE) is shown in Fig. 9.1. The system at its initial con- liquid in equilibrium. Furthermore, assume there are total of
ditions of TF and PF is in a nonequilibrium state. W h e n it F mol of fluid in the vessel at initial temperature and pres-
reaches to equilibrium state, the conditions change to T and sure of Tp and PF as shown in Fig. 9.1. The conditions of the
P and new phases m a y be formed. The initial composition of vessel change to temperature T and pressure P at which both
the fluid mixture is zi; however, at the final equilibrium con- vapor and liquid can coexist in equilibrium. Assume V mol of
ditions, compositions of vapor, liquid, and solid in terms of vapor with composition Yi and L(= F - V) mol of liquid with
mole fractions are specified as yi, x L, and x s, respectively. The composition x / a r e produced as a result of phase separation
a m o u n t of feed, vapor, liquid, and solid in terms of n u m b e r of due to equilibrium conditions. No solid exists at the equilib-
moles is specified by F, V, L, and S, respectively. Under VLE rium state and S -- 0 and for this reason composition of liquid
conditions, no solid is formed (S -- 0) and at VSE state no phase is simply shown by xi. The a m o u n t of vapor m a y be ex-
liquid exists at the final equilibrium state (L = 0). The system pressed by the ratio of V/F or VF for each mole of the mixture.
variables are F, zi , T, P, V, Yi, L, x L, S, and x s, where in a typi- The parameters involved in this equilibrium problem are T,
cal equilibrium calculation, F, zi, T, and P are known, and V, P, zi, x4, Yi, and VF (for the case of F = 1). The VLE calcu-
L, S, Yi, x L, and x s are to be calculated. In some calculations lations involve calculation of three of these parameters from
such as bubble point calculations, T or P m a y be u n k n o w n three other known parameters.
and must be calculated from given information on P or T and Generally there are five types of VLE calculations: (i) Flash,
the a m o u n t of V, L, or S. Calculations are formulated through (ii) bubble-P, (iii) bubble-T, (iv) dew-P, and (v) dew-T. (i) In
both equilibrium relations and material balance for all com- flash calculations, usually zi, T, and P are k n o w n while xi, Yi,
ponents in the system. Two-phase equilibrium such as VLE and V are the u n k n o w n parameters. Obviously calculations
or SLE calculations are somewhat simpler than three-phase can be performed so that P or T can be found for a k n o w n
equilibrium such as VLSE calculations. value of V. Flash separation is also referred as flash distilla-
In this chapter various types VLE and SLE calculations tion. (ii) In the bubble-P calculations, pressure of a liquid of
are formulated and applied to various petroleum fluids. Prin- k n o w n composition is reduced at constant T until the first
ciples of phase equilibria were discussed in Section 6.8 vapor molecules are formed. The corresponding pressure is
t h r o u g h Eqs. (6.171)-(6.174). VLE calculations are formu- called bubble point pressure (Pb) at temperature T and estima-
lated through equilibrium ratios (Ki) and Eq. (6.201), while tion of this pressure is known as bubble-P calculations. For
SLE calculations can be formulated through Eq. (6.208). In analysis of VLE properties, consider the system in Fig. 9.1
addition there are five types of VLE calculations that are dis- without solid phase (S -- 0). Also assume the feed is a liquid
cussed in the next section. Flash and bubble point pressure with composition (xi = zi) at T = TF and PF. N o w at constant
368 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
T, pressure is reduced to P at which infinitesimal a m o u n t Since vapor and liquid leaving a flash unit are in equilib-
of vapor is produced (~V = 0 or beginning of vaporization). rium from Eq. (6.201) we have
Through bubble-P calculations this pressure is calculated.
(9.1) Yi = gix4
Bubble point pressure for a mixture at temperature T is sim-
ilar to the vapor pressure of a pure substance at given T. Off) in which Ki is the equilibrium ratio of c o m p o n e n t i at T
In bubble-T calculations, liquid of k n o w n composition (x4) and P and compositions xi and Yi. Calculations of Ki values
at pressure P is heated until temperature T at which first have been discussed in Section 6.8.2.3. Mole balance equation
molecules of vapor are formed. The corresponding tempera- around a separator unit (Fig. 9.2) for c o m p o n e n t i is given by
ture is known as bubble point temperature at pressure P and the following equation:
estimation of this temperature is k n o w n as bubble-T calcu-
lations. In this type of calculations, P = PF and temperature (9.2) 1 x zi = LF x xi + VF x yi
T at which small a m o u n t of vapor is formed can be calcu- Substituting for LF = 1 -- VF, replacing for Yi from Eq. (9.1),
lated. Bubble point temperature or saturation temperature and solving for xi gives the following:
for a mixture is equivalent to the boiling point of a pure sub-
zi
stance at pressure P. (iv) In dew-P calculations a vapor of (9.3) x~ --
1 + VF(Ki -- 1)
known composition (Yi = zi) at temperature T = T~ is com-
pressed to pressure P at which infinitesimal a m o u n t of liquid Substituting Eq. (9.3) into Eq. (9.1) gives a relation for cal-
is produced (~L = 0 or beginning of condensation). T h r o u g h culation of Yi. Since for both vapor and liquid products we
dew-P calculations this pressure k n o w n as dew point pressure must have ~ x4 = ~ Yi = 1 or ~ (Yi - x4) = 0. Substituting x4
(Pd) is calculated. For a pure substance the dew point pressure and Yi from the above equations gives the following objective
at temperature T is equivalent to its vapor pressure at T. (v) In function for calculation of VF:
dew-T calculations, a vapor of k n o w n composition is cooled
at constant P until temperature T at which first molecules of (9.4) F(VF) ----
~_, 1 ~ z~(K~ - 1)
--1) -- 0
liquid are formed. The corresponding temperature is k n o w n i=1
as dew point temperature at pressure P and estimation of this Reservoir engineers usually refer to this equation as
temperature is k n o w n as dew-T calculations. In these calcu- Rachford-Rice method [ 1]. W h e n VF = 0, the fluid is a liq-
lations, P = PF and temperature T at which condensation be-
uid at its bubble point (saturated liquid) and if VF = 1, the
gins is calculated. Flash, bubble, and dew points calculations system is a vapor at its dew point (saturated vapor). Correct
are widely used in the petroleum industry and are discussed solution of Eq. (9.4) should give positive values for all x~ and
in the following sections.
Yi, which match the conditions ~ xi = ~_, yi = 1. The follow-
ing step-by-step procedure can be used to calculate VF:
9.2.1 Flash Calculations--Gas-to-Oil Ratio 1. Consider the case that values of zi (feed composition), T,
In typical flash calculations a feed fluid mixture of compo- and P (flash condition) are known.
sition zi enters a separator at T and P. Products of a flash 2. Calculate all Ki values assuming ideal solution (i.e., using
separator for F mol of feed are V mol of vapor with composi- Eqs. 6.198, 6.202, or 6.204). In this way knowledge of x4
tion Yi and L mol of liquid with composition x4. Calculations and Yi are not required.
can be performed for each mole of the feed (F = 1). By calcu- 3. Guess an estimate of VF value. A good initial guess m a y
lating vapor-to-feed mole ratio (VF ----V/F), one can calculate be calculated from the following relationship [2]: VF = A/
the gas-to-oil ratio (GOR) or gas-to-liquid ratio (GLR). This (A - B), where A = ~ [ z i (Ki - 1)] and B = ~ [ z i ( K i - 1)/
parameter is particularly important in operation of surface Ki].
separators at the oil production fields in which production of 4. Calculate F ( V ) from Eq. (9.4) using assumed value of VF in
m a x i m u m liquid (oil) is desired by having low value of GOR. Step 3.
Schematic of a continuous flash separator unit is shown in 5. If calculated F(VF) is smaller than a preset tolerance, e
Fig. 9.2. (e.g., 10-15), then assumed value of VF is the desired an-
swer. If F(VF) > e, then a new value of VF must be calcu-
lated from the following relation:
Vapor
9 V moles F(VF)
Yi (9.5) V/~ ew = VF dF(VF)
dVr
:'i'i':':':'i':'i':'i'."
Feed
:i:i:i:iT:ZZ:i:i:i
::::::::::::::::::::::::
In which dF(VF)/dVF is the first-order derivative of F(VF)
with respect to VF.
1 mole 9..................:.... 9 ,..,.......
..............................................
zi :::::::::::::::::::::::::::::::::::::::::::::::::
::::::::::::::::::::::::::::::::::::::::::::::::: dF(VF)-L{ zi(Ki-1)2]
TF, PF :::::::::::::::::::::::::::::::::::::::::::::::::
:::::::::::::::::::::::::::::::::::::::::::::::::
:::::::::::::::::::::::::::::::::::::::::::::::::
.:::,.,.,,,.:::,...........,.,............
(9.6) d~ -- i=1 [VF~i---ii-+ 1]2
i!ili!iiiii~ililili~ii!i!~i:i~i:i:i2i~i~i:i:i:i:i
:::::::::::::::::::::::::::::::::::::::::::::::::
The procedure is repeated until the correct value of VF is
Liquid obtained. Generally, if F (VF) > O, VF must be reduced and if
9 L moles
xi
F(VF) < O, V~ must be increased to approach the solution.
6. Calculate liquid composition, xi, from Eq. (9.3) and the
FIG. 9.2--A continuous flash separator. vapor phase composition, Yi, from Eq. (9.1).
9. A P P L I C A T I O N S : P H A S E E Q U I L I B R I U M C A L C U L A T I O N S 369
TABLE 9.2--Characterization parameters of the C7+ fraction of sample of Table 9.1 [7].
Pseudocomponent mol% wt% M SG Tb, K n20 Nc P% N% A%
C7+ (1) 10.0 12.5 110 0.750 391.8 1.419 8 58 22 20
C7+ (2) 9.0 17.1 168 0.810 487.9 1.450 12.3 32 35 33
C7+ (3) 7.7 23.1 263 0.862 602.1 1.478 19.3 17 37 46
C7+ (4) 2.5 11.6 402 0.903 709.0 1.501 28.9 6 34 60
C7+ (5) 1.2 8.2 608 0.949 777.6 1.538 44 0 45 55
Total C7+ 30.4 72.5 209.8 0.843 576.7 1.469 15.3 25 34 41
E x p e r i m e n t a l values o n M7+ a n d SG7+. Distribution p a r a m e t e r s (for Eq. 4.56) a n d c a l c u l a t e d values: M7+ = 209.8; Mo =
86.8; So = 0.65; $7+ = 0.844; BM = 1; As = 0.119; n7+ = 1.4698; AM = 1.417; Bs = 3; May = 209.8; Say = 0.847.
characterization parameters (i.e., M, Tb, SG, n2o, Nc, and PNA ponent of heptane-plus, C7(1). The value of BIP of this pair
composition). This is done using the distribution model de- exhibits a major impact in the calculation results. Ki values
scribed in Section 4.5.4 with M7+ and SG7+ as the input pa- are calculated from SRK EOS and flash calculations are per-
rameters. The basic parameters (Tb, n20) are calculated from formed for three stages shown in Fig. 9.3. The liquid product
the methods described in Chapter 2, while the PNA composi- from the first stage is used as the feed for the second stage
tion for each p s e u d o c o m p o n e n t is calculated from methods separator and flash calculation for this stage is performed
given in Section 3.5.1.2 (Eqs. 3.74-3.81). The calculation re- to calculate composition of feed for the last stage. Similarly,
sults with distribution parameters for Eq. (4.56) are given in the final crude oil is produced from the third stage at atmo-
Table 9.2. Molar and specific gravity distributions of the C7+ spheric pressure. Composition of C7+ in each stream can be
fraction are shown in Fig. 9.4. The PNA composition is needed calculated from s u m of mole fractions of the five pseudocom-
for calculation of properties through p s e u d o c o m p o n e n t ap- ponents of C7i. GOR for each stage is calculated from Eq. (9.8).
p r o a c h (Section 3.3.4). Such information is also needed when S u m m a r y of results are given in Table 9.3. Overall GOR is cal-
a simulator (i.e., EOR software) is used for phase behavior culated as 853 c o m p a r e d with actual value of 850 scf/stb. This
calculations [9]. is a very good prediction mainly due to adjusting BIPs with
To generate the composition of gases and liquids in sepa- liquid density of produced crude oil. The calculated composi-
rators, see Fig. 9.3, the feed to the first stage is considered tions in Table 9.3 are also in good agreement with actual data
as a mixture of 17 components (12 components listed in of Table 9.1.
Table 9.1 and 5 components listed in Table 9.2). For pure com- The method of characterization selected for treatment of
ponents (first 11 components of Table 9.1), Tc, Pc, Vc, and to are C7+ has a major impact on the results of calculations as shown
taken from Table 2.1. For C 6 fraction (SCN) and C7+ fractions by Riazi et al. [7]. Table 9.4 shows results of GOR calcu-
(Table 9.3) critical properties can be obtained from methods lations for the three stages from different characterization
of Chapter 2 (Section 2.5) or from Table 4.6. For this example, methods. In the Standing method, Eqs. (6.204) and (6.205)
Lee-Kesler correlations for calculation of To, Pc, and to and have been used to estimate K/values, assuming ideal solution
Riazi-Daubert correlations (the API methods) for calculation mixture. As shown in this table, as the n u m b e r of pseudocom-
of Vc and M (or Tb) have been used. The binary interaction pa- ponents for the C7+ fraction increases better results can be
rameters (BIPs) for n o n h y d r o c a r b o n - h y d r o c a r b o n are taken obtained. #
from Table 5.3 and for h y d r o c a r b o n - h y d r o c a r b o n pairs are
calculated from Eq. (5.63). Parameter A in this equation has
been used as an adjustable parameter so that at least one pre- 9.2.2 B u b b l e a n d D e w P o i n t s Calculations
dicted property matches the experimental data. This property
can be saturation pressure or a liquid density data. For this Bubble point pressure calculation is performed t h r o u g h the
calculation, parameter A was determined so that predicted following steps:
liquid specific gravity from last stage matches experimental
value of 0.815. Liquid SG is calculated from Eq. (7.4) using 1. Assume a liquid mixture of k n o w n xi and T is available.
SG of all components in the mixture. It was found that when 2. Calculate plat (vapor pressure) of all components at T from
A -- 0.18, a good m a t c h is obtained. Another adjustable pa- methods described in Section 7.3.
rameter can be the BIP of methane and the first pseudocom- 3. Calculate initial values of Yi and Pbub from Raoult's law as
P = ~-~x/Pisat and Yi = xipsat/p.
4. Calculate Ki from Eq. (6.197) using T, P, xi, and Yi.
0.01 4 5. Check if 1~ xiK,- - 1 [ < e, where e is a convergence toler-
0.008 ance, (e.g., 1 x 10 -12) and then go to Step 6. If not, repeat
3
0.006 calculations from Step 4 by guessing a new value for pres-
~2
~" 0 . 0 0 4 LL sure P and yi = Kixi. If ~ x i K i - 1 < 0, reduce P and if
0.002 1 xiKi - 1 > 0, increase value of P.
0
6. Write P as the bubble point pressure and yi as the com-
0 200 400 600 800 0.6 0.8 1 1.2 1.4 position of vapor phase. Bubble P can also be calculated
through flash calculations by finding a pressure at which
M SG
Vr ~ 0. In bubble T calculation x4 and P are known. The
FIG. 9.4~Probability density functions for molecular calculation procedure is similar to bubble P calculation
weight and specific gravity of the C7+ fraction given in m e t h o d except that T m u s t be guessed instead of guess-
Table 9.2 [8]. ing P.
9. A P P L I C A T I O N S : PHASE EQUILIBRIUM CALCULATIONS 371
TABLE 9.3---Calculated values for the data given in Table 9.1 using proposed characterization
method. Taken with permission from Ref [7].
No. Component Feed 1st-Stage gas 2nd-Stage gas 3rd-Stage gas 3rd-Stage liquid
1 N2 0.09 0.54 0.12 0.05 0.00
2 CO2 2.09 3.91 4.09 1.44 0.02
3 H2S 1.89 1,47 4.38 5.06 0.14
4 H20 0.00 0.00 0.00 0.00 0.00
5 C1 29.18 64.10 32.12 5.68 0.03
6 C2 13.60 19.62 32.65 25.41 0.38
7 Ca 9.20 7.41 18.24 35.47 3.05
8 n-C4 4.30 1.48 4.56 13.92 4.38
9 i-C4 0.95 0.41 1.23 3.47 0.78
10 n-C5 2.60 0.36 1.01 3.98 4.81
11 i-C6 1.38 0.24 0.68 2.61 2.37
12 C6 4.32 0.27 0.61 2.22 9.01
13 C7+ 30.40 0.19 0.31 0.69 75.03
SG at 60~ 0.8105
Temp,~ 245 105 100 90 90
Pressure, psia 2197 315 75 15 15
GOR, scf/stb 853 580 156 117
F o r vapors of k n o w n c o m p o s i t i o n d e w P o r d e w T can be gravity). These correlations are widely used b y reservoir en-
calculated as o u t l i n e d below: gineers for quick a n d convenient calculation of b u b b l e p o i n t
p r e s s u r e s [1, 3, 10]. The S t a n d i n g c o r r e l a t i o n for p r e d i c t i o n
1. Assume a v a p o r m i x t u r e of k n o w n Yi a n d T is available. of b u b b l e p o i n t p r e s s u r e is [1, 3]
2. Calculate P y (vapor pressure) of all c o m p o n e n t s at T f r o m
m e t h o d s of Section 7.3. Pb(psia) = 18.2(a x l 0 b - 1.4)
3. Calculate initial values of xi a n d Pdew f r o m Raoult's law as
a = (GOR/SGgas) 0"83
1/P = ~_, yi / P~sat a n d x / = yi P / P~sat.
(9.9)
4. Calculate Ki f r o m Eq. (6.197), using T, P, xi, a n d Yi. b = 0.00091T - 0.0125 (APIofl)
5. Check if ]~]yi/Ki - 11 < e, w h e r e e is a convergence tol-
erance, (e.g., 1 x 10 -lz) go to Step 6. If not, r e p e a t cal- T = Temperature, ~
culations from Step 4 b y guessing a n e w value for pres- w h e r e / ~ is the b u b b l e p o i n t pressure, SGgas is the gas specific
sure P a n d x~ = yi/Ki. If ~ , y i / K i - 1 < O, increase P a n d if gravity ( = Mg/29), APIoil is the API gravity of p r o d u c e d liquid
~ y i / K i - 1 > O, decrease value of P. c r u d e oil at stock t a n k condition, a n d GOR is the solution gas-
6. Write P as the dew p o i n t p r e s s u r e a n d xi as the c o m p o s i t i o n to-oil ratio in scf/stb. Use of this correlation is shown in the
of f o r m e d liquid phase. following example. A deviation of a b o u t 15% is expected f r o m
Dew P can also be calculated t h r o u g h flash calculations the above correlation [3]. M a r h o u n developed the following
b y finding a p r e s s u r e at w h i c h VF = 1. I n d e w T calculation relation for calculation of Pb b a s e d on PVT d a t a of 69 oil
Yi a n d P are known. The calculation p r o c e d u r e is s i m i l a r to s a m p l e s from the Middle East [10]:
dew P calculation m e t h o d except that T m u s t be guessed in- Pb(psia) = a (GOR) b (SGgas) c (SGoiI)d ( T ) e
stead of guessing P. In this case if ~ 3#/Ki - 1 < O, decrease
T a n d if ~ y i / K i - 1 > O, increase T. B u b b l e a n d d e w p o i n t (9.10) a = 5.38088 • 10 -3 b = 0.715082 c = -1.87784
calculations are u s e d to calculate PT d i a g r a m s as s h o w n in d = 3.1437 e = 1.32657 T = temperature, ~
the next section.
Reservoir engineers usually use e m p i r i c a l l y developed cor- w h e r e SGoi I is the specific gravity of stock t a n k oil a n d GOR is
relations to e s t i m a t e b u b b l e a n d dew points for reservoir fluid in scf/stb. The average e r r o r for this equation is a b o u t 4-4%.
mixtures. F o r example, Standing, Glaso, a n d Vazquez a n d
Beggs correlations for p r e d i c t i o n of b u b b l e p o i n t p r e s s u r e E x a m p l e 9 . 2 - - C a l c u l a t e b u b b l e p o i n t p r e s s u r e of reservoir
of reservoir fluids are given in t e r m s of t e m p e r a t u r e , GOR, fluid o f Table 9.1 at 245~ f r o m the following m e t h o d s a n d
gas specific gravity, a n d stock t a n k oil specific gravity (or API c o m p a r e the results with an e x p e r i m e n t a l value of 2387 psia.
TABLE 9.4--Calculated GOR from different C7+ characterization methods. Taken with permission
from Ref [7].
No. of C7+ OverallGOR,
Method Input for C7+ ~a~ions scffs~ Stage 1 Stage 2 Stage 3
Lab data 850 601 142 107
Proposed M7+ and SG7+ 5 853 580 156 117
Standing (Eqs. 6.202 MT+ and SG7+ 1 799 534 134 131
and 6.203)
Simulation 1a Nc & Tb 1 699 472 141 86
Simulation 2 Nc & Tb 5 750 516 142 92
Simulation 3 M & PNA 1 779 542 142 95
Simulation 4 M & PNA 5 797 559 143 95
aCalculations have been performed through PR EOS using a PVT simulator [9].
372 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
a. Thermodynamic model with use of SRK EOS similar to the TABLE 9.5--Effect of C7+ characterization methods on calculated
one used in Example 9.1. mixture critical properties [7].
b. Standing correlation, Eq. (9.9). Charac. Input for C7+ No. of C7+
scheme of Table 9.3 Fractions Tc, K Pc,bars Zc
c. Mahroun's correlation, Eq. (9.10). Simulation 1 Nc & Tb 1 634 98 0.738
Simulation 2 Nc & Tb 5 651 141 0.831
Solutions(a) The saturation pressure of the reservoir fluid Calculationshavebeen performedthroughPR EOSusinga PVTsimulator[9].
(Feed in Table 9.1) at 245~ can be calculated along flash cal-
culations, using the method outlined above. Through flash
in Chapter 4, for lighter reservoir fluids such as gas conden-
calculations (see Example 9.1) one can find a pressure at
sate samples detailed treatment of C7+ has less effect on the
245~ and that the amount of vapor produced is nearly zero
phase equilibrium calculations of the fluid.
(V~ -~ 0). The pressure is equivalent to bubble (or saturation)
The true critical temperature (Ttc) of a defined mixture may
pressure. This is a single-stage flash calculation that gives
also be calculated from the following simple mixing rule pro-
psat = 2197 psia, which differs by - 8% from the experimental
posed by Li [11]:
value of 2387 psia. (b) A simpler method is given by Eq. (9.9).
This equation requires GOR, APIoi], and SGga~. GOR is given T t c = ~_~ ~i Tci
in Table 9.1 as 850 scf/sth. APIoflis calculated from the specific i
gravity of liquid from the third stage (SG = 0.815), which gives (9,11) xiVci
APIo~ = 42.12. SGg~sis calculated from gas molecular weight, S i Xi Vci
Mg~s, and definition of gas specific gravity by Eq. (2.6). Since
where a~, Tci, and Vci are mole fraction, critical temperature,
gases are produced in three stages, Mgas for these stages are
and volume of component i in the mixture, respectively. The
calculated from the gas composition and molecular weights of
average error for this method is about 0.6% (~3 K) with max-
components as 23.92, 31.74, and 44.00, respectively. Mga~for
imum deviation of about 1.6% (--8 K) [12]. The Kreglewski-
the whole gas produced from the feed may be calculated from Kay correlation for calculation of true critical pressure, Ptc,
GOR of each stage as Mg~, = (601 x 23.92 + 142 x 31.74 +
is given as [13] follows:
107 • 44.00)/(601 + 142 + 107) = 27.76. SGg~ = 27.76/29 =
0.957. From Eq. (9.9), A = 139.18 and Pb = 2507.6 psia, which Ptc=Pp~[l + ( 5 . 8 0 8 + 4 , 9 3 w ) ( ~ - l ) ]
differs by +5.1% from the experimental value. (c) Using
(9.12)
Marhoun's correlation (Eq. 9.10) with T = 705~ SGoij =
Tpc= ExiTci Pp~ = ZxiPci and w = E x , ooi
0.815, SGgas = 0.957, and GOR = 850 we get Pb = 2292 psia i i i
(error of -4%). In this example, Marhoun's correlation gives
the best result since it was mainly developed from PVT data where Tpc and Ppc are pseudocritical temperature and pres-
of oils from the Middle East, r sure calculated through Kay's mixing rule (Eq. 7.1). The aver-
age deviation for this method is reported as 3.8% (~2 bar) for
nonmethane systems and average deviation of 50% (~48 bar)
9.2.3 Generation o f P-T Diagrams---True may be observed for methane-hydrocarbon systems [12].
Critical Properties These methods are recommended in the API-TDB [12] as well
A typical temperature-pressure (TP) diagram of a reservoir as other sources [3].
fluid was shown in Fig. 5.3. The critical temperature and pres-
sure (critical point) in a PT diagram are true critical proper- 240
ties and not the pseudocritical. For pure substances, both the BubblePoint L : Liquid
true and pseudocritical properties are identical. The main ap- ....... Dew Point V: Vapor
plication of a PT diagram is to determine the phase (liquid, 200 9 CriticalPoint
160f7+ \
L
vapor or solid) of a fluid mixture. For a mixture of known com-
position, pseudocritical temperature and pressure (Tpc, P~c)
may be calculated from the Kay's mixing rule (Eq. 7.1) or
other mixing rules presented in Chapter 5 (i.e., Table 5.17).
Methods of calculation of critical properties of undefined 12o \
petroleum fractions presented in Section 2.5 all give pseudo-
critical properties. While pseudocritical properties are useful
for generalized correlations and EOS calculations, they do not 80 ":, \ V
/ c7§(1 component) '
represent the true critical point of a mixture, which indicates
phase behavior of fluids. Calculated true critical temperature 40 : I
and pressure for the reservoir fluid of Table 9.1 by simula- -" j
For undefined petroleum fractions the following correla- discusses the nature of heavy compounds that are present
tion may also be used to estimate true critical temperature in petroleum residua and heavy oils. Precipitation of these
and pressure from specific gravity and volume average boil- heavy compounds under certain conditions of temperature
ing point (VABP) of the fraction [12]: and pressure or composition follow general principles of SLE,
which were discussed in Section 6.8.3. In this section, the
Ttc= 358.79 + 1.6667A - 1.2827(10-3)A 2
problems associated with such heavy compounds as well as
(9.13) A = SG (VABP - 199.82) methods that can be used to predict the certain conditions at
which they precipitate will be discussed. Based on the princi-
loglo(Ptc/Ppc) = 0.05 + 5.656 x logl0(Ttc/Tpc) ple of phase equilibrium discussed in Section 6.8.1, a thermo-
where Tt~, Tpr and VABP are in kelvin and Ppc and Ptc are in dynamic model is presented for accurate calculation of cloud
bars. It is important to note that both Tpc and Ppc must be point of crude oils under various conditions. Methods for cal-
calculated from the methods given in Section 2.5 for criti- culating the amount of solid precipitation from sophisticated
cal properties of undefined petroleum fractions. The average thermodynamic models as well as readily available parame-
error for calculation of Tt~ from the above method is about ters for a petroleum fluid are also discussed in this section.
0.7% (~3.3 K) with maximum error of 2.6% (~12 K). Relia-
bility of the above method for prediction of true critical pres- 9.3.1 Nature of Heavy Compounds, Mechanism
sure of undefined petroleum fractions is about 5% as reported of their Precipitation, and Prevention Methods
in the API-TDB [12]. The above equation for calculation of
Ptc is slightly modified from the correlation suggested in the Petroleum fluids, especially heavy oils and residues, contain
API-TDB. This correlation is developed based on an empirical heavy hydrocarbons from paraffinic, naphthenic, and aro-
graph of Smith and Watson proposed in the 1930s. For this matic groups. Generally, there are three types of heavy hydro-
reason it should be used with special caution. The following carbons that may exist in a heavy petroleum fluid: (1) waxes,
method is recommended for calculation of true critical vol- (2) resins, and (3) asphaltenes. As discussed in Section 1.1.3,
ume in some petroleum-related references [3]: the main type of waxes in petroleum fluids are paraffinic
ZtoRTtc waxes. They are mainly n-paraffins with carbon number range
Vtc ~ ' - - of C16-C36 and average molecular weight of about 350. Waxes
P,c that exist in petroleum distillates usually have freezing points
(9.14)
Ztc : ~ X4Z c i between 30 and 70~ Another group of waxes called crys-
i talline waxes are primarily isoparaffins and cycloparaffins
Method of calculation of true critical points (Tt~, Ptc, and Vtc) (with long-chain alkyl groups) with carbon number range of
of defined mixtures through an equation of state (i.e., SRK) 30-60 and molecular weight range of 500-800. The melting
requires rigorous vapor-liquid thermodynamic relationships points of commercial grade waxes are in the 70-90~ range.
as presented in Procedure 4.B4.1 in Chapter 4 of the API- Solvent de-oiling of petroleum or heavy residue results in
TDB [12]. At the true critical point, a correct VLE calculation dark-colored waxes or a sticky, plastic to hard nature material
should show that x~ = Yi. Most cubic EOSs fail to perform [14]. Waxes present in a petroleum fluid may precipitate when
properly at the critical point and for this reason attempts have the conditions of temperature and pressure change. When the
been made and are still continuing to improve EOS phase temperature falls, heavy hydrocarbons in a crude or even a
behavior predictions at this point. gas condensate may precipitate as wax crystals. The temper-
ature at which a wax begins to precipitate is directly related
to the cloud point of the oil [15, 16]. Effects of pressure and
9.3 VAPOR-LIQUID-SOLID EQUILIBRIUMm composition on wax precipitation are discussed by Pan et al.
SOLID PRECIPITATION [17].
Wax formation is undesirable and for this reason, different
In this section, practical application of three-phase equilib- additives usually polymer-based materials are used to lower
rium in the petroleum industry is demonstrated. Upon re- pour points of crude oils. Wax inhibitor materials include
ducing the temperature, heavy hydrocarbons present in a polyalkyl acrylates and methacrylates, low-molecular-weight
petroleum fluid may precipitate as a solid phase and the polyethylene waxes, and ethyl-vinyl acetate (EVA) copoly-
liquid becomes in equilibrium with both the solid and the reefs. The EVA copolymers are probably the most commonly
vapor phase. In such cases, the solid is at the bottom, liquid used wax inhibitors [14]. These inhibitors usually contain
is in the middle, and the vapor phase is on top of the liquid 20-40 wt% EVA. Molecular weight of such materials is usu-
phase. A general schematic of typical vapor-liquid-solid equi- ally greater than 10 000. The amount of EVA added to an oil is
librium (VLSE) during solid precipitation in a petroleum fluid important in its effect on lowering pour point. For example,
is shown in Fig. 9.1. Solid precipitation is a serious problem when 100 ppm of EVA is added to an oil it reduces pour point
in the petroleum industry and the basic question is: what is from 30 to 9~ while if 200 ppm of same inhibitor is added
the temperature at which precipitation starts and under cer- to another oil, it causes an increase in the pour point from
tain temperature, pressure, and composition how much solid 21~ to 25~ [14].
can be precipitated from a petroleum fluid? These two ques- Asphaltenes are multiring aromatics (see Fig. 1.2) that are
tions are answered in this section. Since solids are formed insoluble in low-molecular-weight n-paraffins (LMP) such as
at low temperatures, under these conditions the amount of C3, n-C4, n-C5, or even n-C7 but soluble in benzene, carbon
vapor produced is low and the problem reduces to SLE such disulfide (CS2), chloroform, or other chlorinated hydrocar-
as the case for asphaltene precipitation. Initially, this section bon solvents [15]. They exist in reservoir fluids and heavy
374 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
reversal and its understanding will help plan for more effi-
cient oil recovery processes [24]. Similarly, asphaltene depo-
90 sition negatively affects the EOR gas flooding projects [25].
Asphaltene precipitation may also occur during oil processing
Volatile in refineries or transportation in pipelines and causes major
80 84 Saturates/ problems by plugging pipes and catalysts pores [26, 27]. The
Aromatics
problem is more severe for heavy oils. Further information
7O on problems associated with asphaltene precipitation during
production, especially in the Middle East fields, is given by
Riazi et al. [28].
.~ 60-
There are a number of models and theories that are pro-
"0
posed to describe mechanism of asphaltene formation [29].
,~ 5o Understanding of kinetics of asphaltene formation is much
more difficult than wax formation. There is no universally
accepted model for asphaltene formation; however, most re-
~ 4o
searchers agree on two models: (1) colloidal and (2) micellar.
Schematic of colloidal model is shown in Fig. 9.7. The nature
30 and shape of the resulting aggregates will determine their
effect on the behavior of the petroleum fluid [30, 31]. In
this model, asphaltene particles come together to form larger
20
molecules (irreversible aggregation), which grow in size. Ac-
cording to this model the surface of asphaltene molecules
10 must be fully covered by resin molecules. For this reason
when concentration of resin exceeds from a certain level, rate
of asphaltene deposition decreases even if its concentration
0 20 40 60 80 100 is high. Because of this it is often possible that an oil with
Weight Percent higher asphaltene content results in less precipitation due to
high resin content in comparison with an oil with lower as-
FIG. 9.6--Representation of proportions of resins and as-
phaltene and resin contents. Knowledge of the concentration
phaltenes in a petroleum residua. Taken with permission from
of resin in oil is crucial in determination of the amount of
Ref. [15].
asphaltene precipitation.
In the micellar model, it is assumed that asphaltene mol-
ecules exist as micelles in crude and micellar formation is a
TABLE 9.7--Dipole moments of some compounds and oil mixtures. reversible process. Furthermore, it is assumed that the micel-
No. Compound Dipole, debye
lar shape is spherical, the micellar sizes are monodispersed
1 Methane (Ca) 0.0
2 Eicosane (Cz0) 0.0 (i.e., all having the same size), and the asphaltene micellar
3 Tetracosane (C24) 0.0 core is surrounded by a solvated shell as shown in Fig. 9.8.
4 2-methylpentane 0.1 In this model too, resins may cover asphaltene cores and pre-
5 2,3-dimethylbutane 0.2 vent precipitation. Thermodynamic models to describe phase
6 Propene 0.4 behavior of asphaltic oils depend on such models to describe
7 1-butene 0.3
8 Cyclopentane 0.0 nature of asphaltene molecules.
9 Methylcyclopentane 0.3 Asphaltenes precipitate when conditions of temperature,
10 Cyclopentene 0.9 pressure, or composition change. The condition under which
11 Benzene 0.0 precipitation begins is called the onset of asphaltene precipi-
12 Toluene 0.4
13 Ethylbenzene 0.2 tation. In general to select a right method for determination
14 o-Xylene 0.5 of asphaltene onset, asphaltene content or asphaltene preven-
15 Acetone (C3H60) 2.9 tion one must know the mechanism of asphaltene precipita-
16 Pyridine 2.3 tion, which as mentioned earlier very much depends on the oil
17 Aniline 1.6 composition. Asphaltenes flocculate due to excess amounts of
18 NH3 1.5
19 H2S 0.9 paraffins in the solution and micellization (self-association) of
20 CO2 0.0 asphaltene is mainly due to increase in aromaticity (polarity)
21 CC14 0.0 of its medium [32].
22 Methanol 1.7 During the past decade, various techniques have been de-
23 Ethanol 1.7
24 Water 1.8 veloped to determine asphaltene onset from easily measur-
able properties. These methods include measuring refractive
O i l mixtures
25 Crude Oils <0.7 index to obtain the onset [33]. Fotland et al. [34] proposed
26 Resins 2-3 measuring electric conductivity to determine the asphaltene
27 Asphaltenesa 4-8 onset. Escobedo and Mansoori [35] proposed a method to
Data source for pure compounds: Poling, B. E., Prausnitz, J. M., O'Connell, determine the onset of asphaltene by measuring viscosity of
J. E, Properties of Gases and Liquids, 5th ed., McGraw-Hill, New York, 2000.
aAsphaltenes separated by n-heptane. Data source for oil mixtures: Goul and crude oil diluted with a solvent (n-Cs, n-C7, n-C9). They showed
Firoozabadi [23]. that with a decrease in deposition rates with increasing crude
(a) Colloidal Phenomenon Due to Increase in Concentration of Polar Miscible Solvent (such as polar aromatic hydrocarbons
shown by solid ellipses) in crude oil
FIG. 9.7--Schematic of colloidal model for asphaltene formation. Taken with permission
from Ref. [29],
m Monomeric
asphaltene
Monomedc
resin
N Asphal=free
oil
4,(I r
0% Toluene
- - 9 1% Toluene
u 3% Toluene
3.0 m . 5% Toluene
-- 6% Toluene ~ , . ,."
2,0
"5,
.m
.., j'
P,
f ,,'~ /
1.0
,. // .
f / / ,~-"
O.O 9 . _ J ~ -1 / . ,-.,J
05 0.$2 0.54 0,56 I~ 4[I.$ @,G2 @.K4 @.KK 0~68
Mole fraction of C02
TABLE 9.8reEvaluation of three types of additives for asphaltene removal on a rock surface [37].
Removal efficiency (wt% of adsorbed asphaltene)
for different additives
Additivewt% in solvent Time, h AdditiveA AdditiveB AdditiveC
0.1 1 40.5 14.0 2.7
0.1 18 48.9 21.4 6.0
2.0 1 49.9 32.9 8.9
2.0 18 51.5 46.0 10.1
are in the bulk state; on the contrary, the asphaltene up-take example, Kawanaka et al. [30] used a thermodynamic ap-
by toluene is very low (10-20 wt%) when the same material proach to study the phase behavior and deposition region in
is adsorbed on a rock surface (clays, dolomia, quartz, etc.), CO2-crude mixtures at different pressures, temperatures, and
as experienced by Piro et al. [37]. For this reason, they used compositions. In the next few sections, thermodynamic mod-
additives dissolved in toluene for asphaltenes' removal when els for solid formation are presented to calculate the onset
they are adsorbed on rock surface. Three types of additives and amount of solid precipitation.
were evaluated: additive A was based on alkyl benzene sul- For the same tank oil shown in Table 9.9, Pan and
fonic acid (ABSA); additive B was based on complex poly- Firoozabai [20] used their thermodynamic model based on
mer; additive C was based on another complex polymer. As- micellar theory of asphaltene formation to calculate asphal-
phaltic materials were obtained from a crude oil of 42 API tene precipitation for various solvents. Their data are 'shown
gravity by precipitation with n-heptane. The rock on which in Fig. 9.12, where amount of precipitation is shown versus
asphaltenes were adsorbed was powdered dolomite (average dilution ratio. The dilution ratio (shown by Rs) represents
particle size of 60/zm and surface area of 10 m2/g) and toluene volume (in cm 3) of solvent added to each gram of crude oil.
was employed as solvent for the additives. Experiments were The amount of precipitated resin under the same conditions
conducted to study the effect of different types of additives, is also shown in this figure. The onset of asphaltene forma-
concentration level, and time on the amount of asphaltenes tion is clearly shown at the point where amount of precipita-
up-take. A summary of results of experiments is given in tion does not change with a further increase in solvent-to-oil
Table 9.8. The results show that addition of additive A (0.1 ratio. Lighter solvents cause higher precipitation. Generally
wt% in toluene solution) can remove up to 41% after 1 h and value of Rs at the onset for a given oil is a function of sol-
up to 49% of asphaltene after 18 h [37]. Therefore, at higher vent molecular weight (Ms) and it increases with increase in
additive concentrations the contact time can be reduced. Ms [38]. Effect of temperature on asphaltene precipitation
Deasphalted oils may also be used as asphaltene inhibitor depends on the type of solvent as shown in Fig. 9.13 [39].
since they contain resins that are effective in keeping asphal- The amount of solid deposition increases with temperature
tene molecules soluble in the oil in addition to their potential for propane, while for n-heptane the effect of temperature
for greater solvency. There are some synthetic resins such 2- is opposite. Effect of pressure on asphaltene precipitation is
hexadecyl naphthalene that can also be used as asphaltene shown in Fig. 9.14. Above the bubble point of oil, increase in
inhibitor. Most of these inhibitors are expensive and research pressure decreases the amount of precipitation, while below
on manufacturing of commercially feasible asphaltene in- bubble point precipitation increases with pressure.
hibitors is continuing. Asphaltenes or other heavy organics
are precipitated under certain conditions that can be deter-
mined through phase diagram (i.e., PT or Px diagrams). An
9.3.2 Wax Precipitation---Solid Solution Model
example of such diagrams is the Px diagram at constant tem- There are generally two models for wax formation calcula-
perature of 24~ for an oil-CO2 system as shown in Fig. 9.11. tions. The first and more commonly used model is the solid-
Some specifications for this oil are given in Table 9.9. In this solution model. In this model, the solid phase is treated as a
figure, the solid phase is indicated by S and regions of LVS homogenous solution similar to liquid solutions. Formulation
and LS are the regions that asphaltenes may precipitate and of SLE calculations according to this model is very similar
should be avoided. The best way to prevent asphaltene pre- to VLE calculations with use of Eq. (6.205) and equilibrium
cipitation is to avoid the region in the phase diagram where ratio, KsL, from Eq. (6.209) instead of Ki for the VLE. This
asphaltene precipitation can occur. It is for this reason that model was first introduced by Won [41] and later was used
phase behavior of petroleum fluids containing heavy organics to predict wax precipitation from North Sea oils by Pedersen
is important in determining the conditions in which precipi- et al. [ 14, 42]. The second model called multisolid-phase model
tation can be avoided. Construction of such phase diagrams was proposed by Lira-Galeana et al. in 1996 [43], which has
is extremely useful to determine the conditions where precip- also found some industrial applications [16]. In this model,
itation occurs. Unfortunately such diagrams for various oils the solid mixture is not considered as a solution hut it is de-
and solvents are not cited in the open literature. Figure 9.11 scribed as a mixture of pure components; each solid phase
shows that the solid phase is formed at very high concentra- does not mix with other solid phases. The multisolid-phase
tion of CO2, that is, the region that is not of practical applica- model is particularly useful for calculation of CPT of oils.
tion and should be considered with cauttion. Thermodynamic The temperature at which wax appears is known as wax ap-
models, along with appropriate characterization schemes can pearance (or precipitation) temperature (WAT or WPT), which
be applied to waxy or asphaltic oils to determine possibility theoretically is the same as the CPT. Both models are based on
and amount of precipitation under certain conditions. For the following relation expressing equilibrium between vapor,
9. APPLICATIONS: PHASE EQUILIBRIUM CALCULATIONS 379
,• 0.0 25 50 75
LV
100.0
b d - -
Mole Percent C 02
(a)
1380.0
908.5
891.5
Mole Percent CO 2
(b)
FIG. 9.11--Phase diagram for oil-CO2 mixtures at 24~ Asphaltene
precipitation occurs in the LS and LVS regions. (a) Entire composition
range. (b) Enlarged LS section. Oil properties are given in Table 9.9.
Taken with permission from Ref. [30].
380 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
N ZiK/VL
liquid, and solid phases for a multicomponent system shown
in Fig. 9.15. (9.18) Yi = Ei=t (1 -- SF) + Ve (g/vL - 1)
99.4 70.0
(a) (b)
99.3 65.0
99.2 60.0
K
99.1 55.0
99.0 50.0
.o
98.9 45.0
98.8 40.0
270 280 290 300 310 320 330 340 350 360 240 250 260 270 280 290 300 310 320
T,K T,K
(a) (b)
FIG. 9.13---Effect of temperature on asphaltene precipitation. (a) Propane diluent; (b)
n-heptane diluent. Taken with permission from Ref. [39].
c. In calculation of f/L/f/S parameters TMi, AH/f, and ACei Pedersen et al. [42], based on their data for North Sea oils,
must be calculated for each component i. showed that both Won and Hansen procedures significantly
d. Calculate TMi from Eq. (6.156), AH~/from Eq. (6.157), and overestimate both the amount of wax precipitation and CPT.
ACpi f r o m Eq. (6.161). For this reason, they suggested a number of adjustable pa-
e. Calculate both ~/L and yis from Eq. (6.154). In calculation of rameters to be used for calculation of various parameters.
yis, calculate ~s from Eq. (6.155). V/s and V/L can be obtained Chung [44] has used the following empirical set of correla-
from Table 7.1. tions for calculation of properties of C7+ fractions for the wax
f. Once f/L/f/S, yiL, and yis have been determined, calculate formation prediction:
K/SLfrom Eq. (6.209).
A/-//f = 0.9TMiM~~
formation without the use of any adjustable parameter. All pa- (9.22)
~ = 6.743 + 0.938 (In M~) - 0.0395 (In M~)2
rameters can be calculated from the molecular weight of com-
ponents or pseudocomponents as described in Sections 6.6.6. -13.039 (In Mi) -1
Using PNA composition for calculation of properties of C7+
pseudocomponents through Eqs. (6.149), (6.156), and (6.157) where TMi is the melting point in kelvin, A/~ f is the molar
improves model predictions. heat of fusion in cal/mol, V/r is the molar liquid volume in
5~0
~P~edic~ed
0
A Mea~'~s~d
J
~.2.0
0o0 i I 9 , .~
0 2OO 4O0 600 800 1003 i 200 1400
Pressure, bar
FIG. 9.14--Effect of pressure on asphaltene precipitation. Taken with
permission from Ref. [40],
382 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
330
::::::::: " 'Vapor l V moles" " : : : : : : :
o Experimental i.i-~.i~ y~ ... 9
-- Predicted
320
:::::::::::::::::::::::::::::::::::::::::
iiii~iiii ' Liquid,L'moles i:i:i:i
J
Pressure, bar i:!:i:i:!:!:!:i: NS_ 1 moles !:i:i:!:!:i:i:
FIG. 9.15--Prediction of WAT for a Chinese reser- !i! !iiii!!i!il........ i2dy ;ii; ........ !i! iiii!i!ii
voir fiuid using method of Mei et al. [46]. Absolute er- ::::::::::::::: s i:i:i:i:i:i:~:
::::::::::::::: n ~, moles +:-:-:-:+:
ror between calculated and experimental data is 1.6 :i:i:?:::::::: .... ~'~. . . . . . . ::::i:i:i:::::
K. Composition of reservoir fluid is given in Problem
9.10 (Table 9.18). FIG. 9.16~Schematic of multiphase-solid mod-
el for wax precipitation. Courtesy of Lira-Galeana
et al. [43].
cm3/mol,and 8/L is the solubility parameter in (cal/cm3) ~
Various researchers have used similar correlations but with
different numerical coefficients. Most recently, Mei et al. [46] the case for heavy liquid oils such as crude oils or dead oils.
have applied the Pedersen et al. model to calculate wax pre- Pressure causes slight increase in the a m o u n t of wax precipi-
cipitation in a live oil (oil under reservoir conditions) from tation as will be discussed in the next section where the exact
Pubei Oil field located in the western part of China where gas m e t h o d of calculation of CPT of crude oils is presented.
injection is used in EOR processes. Composition of this oil is
given in Problem 9.10 (Table 9.18). Basically, they used Won's 9.3.3 Wax Precipitation: Multisolid-Phase
correlations [41] for ACpi, T ~ , and A H / f while the Thomas Model--Calculation o f Cloud Point
et al.'s correlation [47] was used for calculation of 8s:
One of the problems with the solid-solution model in predic-
7.62 + 2.8a {1 - exp [-9.51 tion of wax formation is that without the use of adjustable
~s= x 10-4 (Mi - 48.2)] } forM/ < 4 5 0 parameters it usually overestimates a m o u n t of wax precipi-
10.30 + 1.78 x 10-3a (Mi - 394.8) for Mi > 450 tation and cloud point of crude oils. In this section, another
model that is particularly accurate for calculation of CPT of
(9.23) crude oils will be presented. In this model, the solid is consid-
where a is an adjustable parameter. They used six adjustable ered as multilayer, each layer represents a pure c o m p o n e n t
parameters for calculation of A C P i , TMi , A/-//f, and 8s in terms (or pseudocomponent) as a solid that is insoluble in other
of T and M, which were determined by matching calculated solid layers. This model was developed by Lira-Galeana et al.
and experimental data on measured WAT values for the oil [43] and is used for calculation of both the a m o u n t of wax
[46]. For the beginning of the flash calculations, the initial precipitated in terms of wt% of initial oil as well as CPT. A
values of K~zL m a y he estimated from the Wilson's correlation schematic of the model is demonstrated in Fig. 9.16. In this
(Eq. 6.204) assuming ideal solution theory. Mei et al. [46] sug- model, it is assumed that as temperature is reduced only a
gested that initial/~SL ValUeS can be set equal to the reciprocal selected n u m b e r of precipitating components will coexist in
of K/vL values also calculated from the Wilson's formula. Pre- SLE. The basis of calculations for this model is the stability
dicted WAT versus pressure is compared with measured val- criteria expressed by Eq. (6.210), which should be applied to
ues and is given in Fig. 9.15. One major problem associated all N components (pure as well as pseudocomponents) in the
with this model is that it requires experimental data on wax following form:
precipitation temperature or the a m o u n t of wax formation to (9.24) fii(T, P, zi) - ~S(T, P) > 0 i = 1, 2 . . . . . N
find the adjustable parameters. This graph is developed based
on data reported in Ref. [46]. where fii(T, P, zi) is the fugacity of c o m p o n e n t i in the orig-
Composition of this reservoir fluid (Table 9.18 in Problem inal fluid mixture at T and P, and f/s is the fugacity of pure
9.10) indicates that it is a gas condensate sample and for gases solid i at T and P. A c o m p o n e n t m a y exist as a pure solid
usually WAT declines with increase in pressure. Lower WAT phase if inequality by Eq. (9.24) is valid. This inequality can
values for an oil are always desirable. This indicates that pres- be applied only to single-solid phase and is not applicable
sure behaves as an inhibitor for wax precipitation for live oils, to solid-phase solutions. Assume c o m p o n e n t 1 is the lightest
gas condensate, or natural gas samples. However, this is not (i.e., Ct in a reservoir fluid) and N is the heaviest c o m p o n e n t
9. A P P L I C A T I O N S : PHASE EQUILIBRIUM CALCULATIONS 383
(i.e., the last pseudocomponent of a C7+ fraction). If Eq. (9.24) in the initial fluid. The ratio of S / F is the same as SF used in
is applied to all N components in the mixture the number Eq. (9.17). The ratio of V / F in Eqs. (9.26) and (9.27) is the
of components that satisfy this equation is designated as Ns same as VF in Eqs. (9.17) and (9.18).
(<N). If Ns = N it means that the mixture at T and P is initially The above set of equations can be solved by converting them
in a solid phase (100% solid). All precipitating components into equations similar to Eqs. (9.17)-(9.19). For precipitating
must satisfy the following isofugacity equations: components, x/t can be calculated directly from Eq. (9.26),
while for nonprecipitating components they must be calcu-
(9.25) f,L(T, P, x3) = f/S(r, P) i = (N - Ns + 1) . . . . . N lated from Eq. (9.27) after finding V / F and S/F. Moles of
The material balance equation for the nonprecipitating com- solid formed for each component, n s, must be calculated from
ponents is Eq. (9.27). Values of V / F and S/F must be found by trial-and-
error procedure so that Eq. (9.29) is satisfied.
(9.26)
I
zi-x~ 1-
j=(N_Ns+I) - f
ni -Ki x i ~ = 0
i = i . . . . . ( N - Ns)
The CPT of a crude can be calculated directly from
Eq. (9.24) using trial-and-error procedure as follows:
30
[N-Ns+ 1] components
O
9 CPT
N ~ components
I --~ 9
200 400
Temperature, K
TABLE 9. l O---Compositions and cloud point temperature of some oils. Taken with permission from Ref. [43].
Oil No.
1 2 3 4 5 6
Comp mol% M Mol% M tool% M mol% M mol% M tool% M
C1 0.056 0.000 0.016 0.000 0.000 0.021
C2 0.368 0.113 0.144 0.100 0.173 0.254
C3 1.171 1.224 1.385 0.118 1.605 1.236
i-C4 0.466 0.645 1.174 0.106 1.150 0.588
n-C4 1.486 2.832 3.073 0.099 3.596 2.512
i-C5 0.961 1.959 2.965 0.162 3.086 1.955
n-C5 1.396 3.335 3.783 0.038 4.171 3.486
C6 2.251 5.633 7.171 0.458 7.841 6.842
C7 6.536 88.8 9.933 92.8 11.27 94.1 2.194 90.8 11.11 94.1 12.86 92.2
C8 8.607 101.0 10.75 106.3 12.41 107.0 2.847 106.5 13.43 105.4 13.99 105.4
C9 4.882 116.0 7.179 120.0 7.745 122.0 1.932 122.3 9.420 119.0 9.195 119.0
C10 2.830 133.0 6.561 134.0 5.288 136.0 5.750 135.0 5.583 135.0 6.438 134.0
Cll 3.019 143.0 5.494 148.0 5.008 147.0 4.874 149.0 4.890 148.0 5.119 148.0
CI2 3.1t9 154.0 4.547 161.0 3.969 161.0 5.660 162.0 3.864 162.0 4.111 161.0
C13 3.687 167.0 4.837 175.0 3.850 175.0 6.607 176.0 4.300 t75.0 4.231 175.0
C14 3.687 181.0 3.700 189.0 3.609 189.0 6.149 189.0 3.272 188.0 3.682 188.0
C15 3.687 195.0 3.520 203.0 3.149 203.0 5.551 202.0 2.274 203.0 3.044 202.0
C16 3.079 207.0 2.922 216.0 2.300 214.0 5.321 213.0 2.791 216.0 2.255 214.0
C17 3.657 225.0 3.072 233.0 2.460 230.0 5.022 230.0 2.311 232.0 2.405 230.0
C18 3.289 242.0 2.214 248.0 2.801 244.0 4.016 244.0 1.960 246.0 2.006 245.0
C19 3.109 253.0 2.493 260.0 2.100 258.0 4.176 256.0 1.821 256.0 1.770 257.0
C20+ 38.4 423.0 17.0 544.0 14.33 418.0 38.80 473.0 11.33 388.0 12.00 399.0
SG20+ 0.893 0.934 0.880 0.963 0.872 0.887
CPT, K 313.15 311.15 314.15 295.15 305.15 308.15
CPT, Calc. 312.4 308.2 316.0 299.3 301.2 309.5
Error, K 0.75 2.95 -1.85 -4.15 3.95 -1.35
9. APPLICATIONS: PHASE EQUILIBRIUM CALCULATIONS 385
16 9.0
8 ,1= 41P
"~ 5,0
6
:ff
4 ~oOtq O0
i~0000 | 4.o
"u
2 _= 3,0
0
2.0
220 240 260 280 300 320
Temperature, K ~ 1,o
(a) Oil 1
0,0 ~ ~ '= ,
270,0 280,0 290.0 300.0 310,0 320.0
10
Temperature, K
0 Data
-- Model FIG. 9.20--Effect of pressure on cloud-point temperature and
wax formation in a synthetic crude oil at low pressures. Taken
with permission from Ref. [17].
4 322
320
0 ~176176
220 240 260 280 300 320 ~ 318
Temperature, K
o . . . . " ' "~" " ~ i ~n ~ ~a' : ~ ' -l ~ - ~~- ' : " Oil
(b) Oil 6
~ 316
FIG. 9.18--Prediction of wax precipitation U - - 30%C2
and cloud-point temperature for oils 1 and 6
. . . . . 30%C5
given in Table 9.10. Taken with permission from i i i i i
Ref. [43]. 314
0 100 200 300 400 500 600
Pressure, bar
0.01 FIG. 9.21mEffect of pressure on cloud-point
t~ temperature at high pressures for a crude oil di-
luted by various light hydrocarbons. Taken with
permission from Ref. [17].
,..3
of vapor produced in equilibrium with liquid is quite small so activity coefficient form (y/L) as
that the problem of asphaltene precipitation reduces to LSE.
The effect of temperature on asphaltene precipitation is the (9.34) y L = e x p In +I-~mL+RT
same as for wax precipitation, that is, as the temperature de-
creases the amount of precipitation increases. The effect of Once ~iL iS knownfi L can be calculated from Eq. (6.114) and
pressure on asphaltene precipitation depends on the type of after substituting into Eq. (9.32) we get the following relation
oil. For crude oils (flee of light gases) and live oils above their for the volume fraction of asphaltene in the liquid solution
bubble point pressure, as pressure increases the amount of as- [38]:
phaltene precipitation decreases, but for live oils at pressures
below the bubble point pressure, as the pressure increases (9.35) ap~ = exp[ EE- -- 1 -- vk (~A--~m) 2]
asphaltene precipitation also increases so that at the bubble LV~
point the amount of precipitation is m a x i m u m [48]. where apE is the volume fraction of asphaltenes in the oil
There are specific thermodynamic models developed for (liquid) phase at the time solid has been precipitated. Once
prediction of asphaltene precipitation; these are based on apE is known, amount of asphaltene precipitated can be cal-
the principles of SLE and the model adopted for the mecha- culated from the difference between the initial amount of as-
nism of precipitation. These mechanisms were discussed in phaltene in liquid and its amount after precipitation as
Section 9.3.1. Most thermodynamic models are based on two
L L
models assumed for asphaltene precipitation: colloidal and mAD = mAT - - PAf~AV~
micellization models. A molecular thermodynamic flame- mAT = 0.01 x (initial asphaltenes in liquid, wt%) x pLkVL
work based on colloid theory and the SAFT model has been
established to describe precipitation of asphaltene from crude
asphaltene precipitated wt% = 100 x PmixV~
"
oil by Wu-Prausnitz-Firoozabadi [39, 49]. Mansoori [29] also
discusses various colloidal models and proposed some ther- (9.36)
modynamic models. Pan and Firoozabadi have also devel-
oped a successful thermodynamic micellization models for where mADis the mass of asphaltenes deposited (precipitated)
asphaltene precipitation [20, 36]. and mAT is the mass of total asphaltene initially dissolved in
Most thermodynamic models consider asphaltene as poly- the liquid (before precipitation) both in g. PA and pLm~are
m e r molecules. Furthermore, it is assumed that the solid mass densities of asphaltenes and initial liquid oil (before
phase is pure asphaltene. The solution is a mixture of oil precipitation) in g/cm 3. VL is the total volume of liquid oil
(asphaltene-free) specified by component B and asphaltene before precipitation in cm 3. pLix can be calculated from an
component specified by A. Applying the principle of SLE EOS or from Eq. (7.4). In determination of mAT,the initial wt%
to asphaltene component of crude oil in terms of equality of of asphaltenes in oil (before precipitation) is needed. This
fugacity parameter may be known from experimental data or it can be
considered as one of the adjustable parameters to match other
(9.32) fL = fAs experimental data. mATcan also be determined from Eq. (9.35)
from the knowledge of asphahene composition in liquid at
(,~,L tTL'tonset
where fL is fugacity of asphahene (A) in the liquid solution tbeonsetwhenmAoiszeroandmgT=pA~,~AvrJ . A more
(A and B) and fs is fugacity of pure solid asphaltene. One accurate model for calculation of asphaltene precipitation is
good theory describing polymer-solution equilibrium is the based on Chung's model for SLE [44]. This model gives the
Flory-Huggins (FH) theory, which can be used to calculate following relation for asphaltene content of oil at temperature
solubility of a polymer in a solvent. Many investigators who T and pressure P [51]:
studied thermodynamic models for asphaltene precipitation
have used the FH theory of polymer solutions for calculation XA = exp / RTMA 1 - - - - ~-~ (~A - - t~rn) 2
of chemical potential of asphaltenes dissolved in oil [38, 50].
Nor-Azian and Adewnmi [48] also used FH theory for the as- In A Vk (vk
(9.37) - ~ -1+~+ )/-T-
phaltic oils. Moreover, they also considered the vapor phase
in their model with VLE calculations between liquid oil and
where subscript A refers to asphaltene component and PMAis
its vapor. According to the FH theory the chemical potential
the pressure at melting point TMa. All other terms are defined
of component i (polymer) in the solution is given as
previously. The last term can be neglected when assumed
(9.33) g/L _/Z~L v/L V/L V~ - Vs. This model has been implemented into some reser-
R~ = l n ( % ) + 1 - ~ + ~-~ ($i - ~m) 2 voir simulators for use in practical engineering calculations
related to petroleum production [51 ].
where api = ~ ,r n . . 8m = E .~=, . Oigi,/2 n is the chemical potential As mentioned earlier (Table 9.6), in absence of actual data,
of component ~ m the hqmd phase, and #~L lS the chemical PA and MA may be assumed as 1.1 cm3/g and I000 g/tool, re-
potential at reference state, which is normally taken as pure spectively. Other values for asphaltene density are also used
liquid i. api is the volume fraction of i, V/L is the liquid molar by some researchers. Speight [15] has given a simplified
volume of pure i at Y and P of the solution, R is the gas version of Eq. (9.37) in terms of asphaltene mole fraction
constant, VL is the liquid molar volume of mixture, and 3i is (XA) as
the solubility parameter for component i at Y of the solution. MA ( ~ L -- 8A)2
The above equation can be conveniently converted into an (9.38) lnxg =
RT pA
9. A P P L I C A T I O N S : P H A S E E Q U I L I B R I U M CALCULATIONS 387
E
In fact one may obtain Eq. (9.38) from Eq. (9.35) by assuming 0.8 Q.-
molar volumes of both oil and asphaltene in liquid phase are
equal: VA L = VL. As this assumption can hardly be justified, -~ 0.7 o
where ~ is the asphaltene solubility parameter in (cal/cm3) ~ where /~ is the refractivity intercept defined in Eq. (2.14)
and T is temperature in kelvin. MB is the molecular weight of as Ri = n2o -dzo/2. Amounts of asphaltenes and resin in a
asphaltene-free crude oil. To calculate/2s., values of 6s and crude mainly depend on the composition of the crude. In Sec-
pAs are needed. 8s can be calculated from Eq. (6.154) and ps is tion 3.5.1.2, parameters/~ and SG were used to predict the
assumed the same as p~. It should be noted that in these rela- composition of petroleum fractions. Calculation of n20 and d20
tions asphaltene-free crude refers to the liquid phase in equi- for a crude is not as accurate as for a fraction since the crude
librium with precipitated solid phase, which include added has a very wide boiling point range. For this reason, the above
solvent (i.e., C3, n-Cs, or n-C7) and the original crude. The equation gives only an approximate value of asphaltene and
phase diagram shown in Fig. 9.11 was developed based on resin contents. Coefficients a, b, and c in Eq. (9.42) are given in
this compositional model [30]. Equation (9.40) gives value of Table 9.11. These coefficients have been determined based on
9.5 (cal/cm3) ~ at 25~ which is consistent with the value re- the calculation of nzo and d20 from Eq. (4.7) and Table 4.5. Only
ported by other investigators. Equation (9.40) is named after M is required for calculation of these two properties. If M is
Hirschberg who originally proposed the relation [52]. not available it may be estimated from other properties such
Most of the thermodynamic models discussed in this sec- as viscosity and SG (i.e., Eq. (2.52) or reversed form of Eq.
tion predict data with good accuracy when the adjustable (4.7) and Table 4.5). The above correlation generally predicts
parameters in the model are determined from experimen- amount of asphahene and resin contents with absolute devi-
tal data on asphaltene precipitation. Results of a thermody- ation of 1.5-2 wt%. Experimental data points for resin con-
namic model based on the colloidal model and SAFT theory tents were very limited and for this reason predicted values
for Suffield crude oil are shown in Fig. 9.22. The crude has must be taken with caution. Data to develop these correlations
388 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
TABLE 9.11---Constants in Eq. 9.42 for estimation o f asphaltene and resin contents of crude oils.
Constants in Eq. (9.42) Absolute Dev.%
Compositiontype (wt%) Rangeof AP][gravityof oil Rangeof wt% a b c No. of oils Avg Max
Asphaltene in oil 5.9--43.4 0.1-20 -731 674 31 122 1.4 -5.2
Resin in oil 5.5--43.4 5.6--40 -2511.5 2467 -76 41 1.9 -4.9
were mostly obtained from Speight [15] and the Oil a n d Gas precipitation appropriate thermodynamic models as intro-
J o u r n a l D a t a B o o k [54]. For prediction of amount of asphal- duced in this section should be used.
tene precipitation when it is diluted by an n-alkane solvent,
the following correlation was developed based on very limited
data [53]: 9.5 V A P O R - S O L I D E Q U I L I B R I U M ~
HYDRATE FORMATION
Asphaltene predicted, wt% = a + b (/~) + c (SG)
In this section, another application of phase equilibrium in
(9.43) + d (Rs) + e (Ms)
the petroleum industry is demonstrated for prediction of hy-
where coefficients a-e are determined from experimental drate formation from vapor-solid equilibrium (VSE) calcula-
data. Parameters /~ and SG are the same as in Eq. (9.42) tions. Hydrates are molecules of gas (C~, C2, C3, iC4, nC4,
and should be calculated in the same way. Ms is the solvent N2, CO2, or H2S) dissolved in solid crystals of water. Gas
molecules, in fact, occupy the void spaces in water crystal
molecular weight (n-alkane) and Rs is the solvent-to-oil ratio
in cm3/g. This correlation was developed based on the data lattice and the form resembles wet snow. In the oil fields
available for three different Kuwaiti oils and 45 data points, hydrates look like grayish snow cone [i]. Gas hydrates are
and for this limited database the coefficients were deter- solid, semistable compounds that can cause plugging in natu-
mined as a = -2332, b = 2325, c = -112.6, d = 0.0737, and ral gas transmission pipelines, gas handling equipments, noz-
e = -0.0265. With these coefficients the above equation pre- zles, and gas separation units. Gas hydrates may be formed at
dicts asphaltene precipitation of Kuwait oils with AD of 0.5%. temperatures below 35~ when a gas is in contact with water.
The correlation is not appropriate for other crude oils and However, at high pressures (> 1000 bar), hydrate formation
to have a generalized correlation for various oils, the coeffi- has been observed at temperatures above 35~ Figure 9.23
cients in Eq. (9.43) must be reevaluated with more data points shows temperature and pressure conditions that hydrates are
for crude oils from around the world. The following example formed for natural gases. As pressure increases hydrate can
shows application of these equations. be formed at higher temperatures. Severe conditions in arctic
and deep drilling have encouraged the development of pre-
dictive and preventive methods. It is generally believed that
Example 9.3--For Suffield crude oil the asphaltene precipi- large amounts of energy is buried in hydrates, which upon
tation by various solvents is shown in Fig. 9.22. Calculate their dissociation can be released.
Hydrates are the best example of the application of
a. asphaltene content.
VSE calculations. Whitson [1] discusses various methods of
b. resin content.
calculation of the temperature at which a hydrate may form
c. amount of asphaltene (wt%) precipitated by adding
10 cm 3/g n-decane.
1000
i~~::~/~:~::i~:~84
::~ :::~ :i84
~ :~~ :~~:~! :~i84184
:: ~i::~~:~:i:~!):!:~:ii~
~!~:i:!:~i~!~:~;i84
~:i:~
Solution--For this oil, M = 360 and SG = 0.952. n20 and d20
should be calculated through M using Eq. (4.7) with coeffi-
cients in Table 4.5. The results are n2o = 1.4954, d20 = 0.888,
and/~ = 1.05115. (a) From Eq. (9.42), asphaltene wt% = 7%.
(b) From Eq. (9.42), resin wt% = 9.3%. (c) For calculation of loo i
asphaltene precipitation from Eq. (9.43) we have Ms = 142
and Rs = 10 cm3/g, thus wt% of asphaltene precipitated is cal-
culated as 1.3%. The experimental value as shown in Fig. 9.22
is 0.5%. The experimental values for asphaltene and resin con-
tents are 13 and 8%, respectively [39]. For resin content the 1~ i.i iiii
calculated value is in error by 1.3% from the experimental
data. This is considered as a good prediction. For the amount
of asphaltene precipitated, Eq. (9.43) gives %AD of 0.8. The
biggest error is for asphaltene content with %AD of 6. As men-
tioned these correlations are very approximate and based on
limited data mainly from Middle East. However, the coeffi- -15 -10 -5 0 5 10 15 20 25 30
cients may be reevaluated for other oils when experimen-
Temperature, ~
tal data are available. In this example predicted values are
relatively in good agreement with experimental data; how- FIG. 9.23---Hyflrate formation for methane and natural
ever, this is very rare. For accurate calculations of asphaltene gases. Drawn based on data provided in Ref. [1].
9. APPLICATIONS: P H A S E E Q U I L I B R I U M CALCULATIONS 389
TABLE 9.12--Coefficients (Ci) for Eq. (9.46) for estimation of liFT at very high pressures. Taken
with permission from Ref. [1].
Pressure, bar (psia) Methane Ethane Propane /-Butane n-Butane
414(6000) 18933 20806 28382 30696 17340
483(7000) 19096 20848 28709 30913 17358
552(8000) 19246 20932 28764 30935 17491
620(9000) 19367 21094 29182 31109 17868
690(10000) 19489 21105 29200 30935 17868
N 45.9039
(9.44) ~ ~Yi >_ 1 lnK; vs
H2s =
-6.051 + 0 . 1 1 1 4 6 T +
T
1.9293 x 10-4pT
i=1
+ 1.94087 ( P ) -0.64405 In ( P ) - 56.87 ( ~ 2 )
where equality holds at temperature where hydrate forma-
tion begins. In the vapor phase the amount of water is very - 7.5816 x 10-6T 3
small (<0.001 mol%) thus its presence in the vapor phase can where T = given temperature in kelvin - 255.4 and
be neglected in the calculations (yW _ 0). To find the temper-
ature at which a hydrate dissociates and hydrocarbons are P = given pressure in bar
released, a calculation similar to bubble point calculations (9.45)
can be performed so that ~ x S K vs > 1. Katz provided charts
for calculation of Kivs, which later Sloan converted into em- For pressures between 400 and 700 bars (~6000-10000 psia),
pirical correlations in terms of T and P and they are used in a simple empirical method is proposed by McLeod and
the petroleum industry [1]. It should be noted that these Ki Campbell in the following form as given in Ref. [1]:
values are not true VSE ratios as the above calculations are / ~' \~/2
based on water-free phases. This method can be applied to
pressures below 70 bar (~1000 psia). For methane, ethane, (9.46) T = 2.16 ti~=l YiCi)
propane, n-butane, and HaS the correlations for calculation
of K vs are given as follows [1]: where values of Ci for C I - C 4 a r e given in Table 9.12 at several
pressures encountered in deep-gaswell drilling.
17.59 3.403 This method can be used for quick estimation of HFT or
lnKcvs = 0.00173 + ~ p + 1.3863 x 10-4pT to check the validity of estimated temperatures from other
methods. More sophisticated methods using chemical poten-
1.0356P 0.78338 in ( P ) - 23.9804 ( ~ ) tial and equations of state are discussed in other references
+ T [1].
(p3) Because of the problems associated with hydrate for-
- 1.34136 x 10-6T 3 - 1.8834 x 10 .5 mation, hydrate inhibitors are used to reduce HFT. Com-
monly used hydrate inhibitors are methanol, ethanol, glycols,
l n K vs = 3.92157 - 161.268 + ~181.267 + 1.8933 x 10-5p a sodium chloride, and calcium chloride. These are nearly the
same materials that are used as water antifreeze inhibitors.
1.04557P 1.19703 in ( P ) 402.16 Effect of methanol (CH3OH) on the depression of HFT of
+ T p2 methane reservoir fluid is shown in Fig. 9.24 [55]. The com-
position of this condensate sample in terms of mol% is as
- 8.8157 (~) + 0.133231 (7) - 21.2354 (~-7) follows: 0.64 N2, 3.11 COa, 73.03 C1, 8.04 Ca, 4.28 C3, 0.73
i-C4, 1.5 n-C4, 0.54 i-C5, 0.6 n-Cs, and 7.53 C6+ with mixture
+46.13339(T) molecular weight of 32.4. The most commonly used equation
to calculate the degree of decrease in HFT (AT) is given by
26.1422 Hammerschmidt, which is in the following form [1, 14]:
l n K vs = - 7 . 5 9 2 2 4 + T 3.0545 • 10-5pT +2.315
Awt%
(P) 79.3379 (9.47) AT =
x10-3Te+0"123481n ~- + p---T-- M(100 - wt%)
where AT is the decrease in HFT in ~ (or in kelvin), wt% is
+ 0.05209 ( - ~ ) - 2 6 . 4 2 9 4 ( ~ 3 ) + 3.2076 x 10-ST 3 the weight percent of inhibitor in the aqueous phase, and M
is the molecular weight of the inhibitor. Values of M and A
390 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
1.0
0.9 M
0.8 N~Y
0.6 i F~
" 0.5
rx~
0.4
0.3
MATRIX FRACTURES
and amount of solid formation in petroleum fluids. Some of pure compounds have been used in their development in
guidelines for quick convergence in the calculation and de- addition to data on petroleum mixtures from oils around the
termination of key interaction parameters are given. Various world. The weakest predictive methods are perhaps those em-
models available in most recent publications for calculation pirically developed correlations that are based on a set of data
of amounts of wax and asphaltene precipitations and their on- for oils from a certain part of the world.
sets are presented. Mechanism of solid formation, their neg- As it is shown in this book the main difficulty in prediction
ative effects in the petroleum industry and methods of their of properties of petroleum fractions relies on properties of
prevention are also discussed. Results of calculations from heavy fractions containing polar multiring compounds with
various models when applied to different petroleum mixtures few experimental data available on their properties. As heavy
are given. Effects of temperature and pressure on the wax and compounds are generally polar with high boiling points, data
asphaltene precipitation for different oils are demonstrated. on specific gravity and molecular weight alone are not suffi-
Methods of calculation of the conditions at which hydrates cient for their property predictions. For such compounds it
may be formed are shown. The impact of characterization is not possible to measure critical properties or even boiling
methods of Chapters 2-4 on property and phase behavior point. Boiling points of such compounds or their mixtures are
predictions as well as methods of calculation of true critical not measurable and estimated boiling points based on distil-
properties are also presented. The chapter ends with another lation data at low pressures have little practical applications
application of methods presented in the book in evaluation as they do not represent true boiling points. For such com-
of gas injection projects for EOR. pounds one has to look at other properties that are directly
In VLE calculations, accuracy of the results basically de- measurable and represent their characteristics.
pends on the method chosen for calculation of equilibrium Reported values of critical properties of heavy compounds
ratios. In this regards suggestions given in Table 6.14 should are usually predicted from methods developed for lighter hy-
be used as a guide. For calculation of CPT and WFT the drocarbons. For example, in the API-TDB [ 12] reported values
multisolid-phase model provides a reliable method without of critical constants for heavy compounds are calculated from
the need for adjustable parameters. A good prediction of onset group contribution methods. Kesler-Lee method for calcula-
of asphaltene precipitation is possible through measurement tion of critical properties of heavy hydrocarbons are based
of kinematic viscosity. on calculated values from vapor pressure data [58]. Predicted
values of critical constants and boiling point from different
methods for heavy compounds differ significantly from each
9.8 FINAL W O R D S other, especially as carbon number increases. This leads to an
even greater difference in predicted thermodynamic proper-
Variety of methods for prediction and calculation of various ties. Presence of very heavy compounds in a mixture requires
thermophysical properties for petroleum and related fluids is a rigorous mixing rule for calculation of mixture properties.
much wider than the methods presented in this book. How- Attempts in this area should be focused on standardization
ever, attempts were made to include the most accurate and of values of critical constants for heavy hydrocarbons and
widely used methods by the people from industry and re- characterization of heavy oils.
searchers. Limitations of application of methods, points of Use of directly measurable properties in calculation of ther-
strength and weaknesses, and their degrees of accuracy have modynamic properties of heavy petroleum mixtures is an ap-
been discussed for different systems. Furthermore, the basis propriate approach as it was discussed in Chapters 5 and 6.
of development of nearly all methods discussed in this book Use of velocity of sound to determine EOS parameters was
have been discussed so the students and new researchers in demonstrated in Section 6.9 and new developments in this
this area can understand the basic concepts and fundamen- area are highly desirable [59]. Measurement and reporting
tals of property calculations. In addition, the approaches pre- of this thermodynamic property on heavy petroleum frac-
sented in the book should help researchers in expansion of the tions and crude oils would help researchers to find methods
existing methods a n d b e used as a guide in the development of of calculation of EOS parameters from measurable proper-
new predictive methods. The methods presented in the book ties. Other useful and measurable properties for heavy oils
should also help users of various simulators (process, PVT, include molecular weight, density, and refractive index. Use of
phase behavior, etc.) to be able to select the most appropriate refractive index in determination of EOS parameters has been
method for their property prediction purposes. shown in Section 5.9. It seems that more advanced equations
Empirical correlations should be used with caution and of state such as SAFT equations would be more appropriate
as a last option in absence of experimental data or accurate for prediction of thermodynamic properties of heavy oils such
fundamentally based thermodynamic models. In use of these as those containing heavy residues, asphaltenes, and complex
correlations their limitations and sensitivity to the input pa- polar compounds. Investigation of this approach should he
rameters must be considered. Some of these methods are reli- continued for more accurate estimation of thermophysical
able when the input parameters are determined through rec- properties. Newly developed methods for phase equilibrium
ommended methods. Perhaps the most accurate methods are calculations and phase determination of many-component
those based on fundamental theoretical approach combined systems are useful tools in formulation and efficient predic-
with empirically determined coefficients and parameters. In tion of hydrocarbon phase behavior and should be pursued
development of such relations availability of input parame- [60].
ters and accuracy of their measurements should be consid- Another appropriate approach in characterization of heavy
ered. Furthermore, predictive methods can have general ap- oils was taken by Goual and Firoozabadi [23] to mea-
plication for a wide range of petroleum fluids if properties sure dipole moments of such complex systems. Attempts in
9. APPLICATIONS: PHASE EQUILIBRIUM CALCULATIONS 393
measurement and reporting of such data should be contin- TABLE 9.15--Properties of gas and liquid phases in a constant
ued to enable us in our understanding of properties of heavy volume cell.
petroleum fluids. Upon availability of such data it would be Specification Initial state Final state
possible to develop more accurate and physically sound meth- Temperature, K
Pressure, bar
ods for characterization of heavy petroleum fractions and Volume of the cell, cm3
crude oils based on their degrees of polarity. Use of dipole Volume of the liquid phase,
moment in correlation of transport properties of polar fluids cm3
was shown by Chung et al. [61]. Measurement and effects of Volume of the gas phase, cm3
heteroatoms in such complex compounds on physical proper- Moles of liquid, tool
Moles of gas, mol
ties should also be considered with great emphasis. Presence Molecular weight of liquid
of heteroatoms such as S, N, or O in a hydrocarbon com- phase
pound can have appreciable impact on the properties of the Molecular weight of gas phase
compound. Mass of liquid, g
Mass of gas phase, g
The market for heavy oils and residues are limited; how-
Density of liquid phase, g/cm3
ever, production of light oil in the world is in decline. There- Density of gas phase, g/cm3
fore, heavy oil conversion becomes increasingly important. Molar density of liquid,
Theoretically, the resources for heavy oils are infinite, as it mol/cm3
is near to impossible to produce the last barrels of oils from Molar density of gas, mol/cm3
heavy oil reservoirs. Considering limited information avail- Length of the cell, cm
Length of the liquid phase, cm
able on properties of heavy compounds, the focus of future Length of the gas phase, cm
studies must be on characterization of heavy hydrocarbons Volume fraction of the liquid
and petroleum fractions. In the area of solid formation and Mole fraction of the gas
prevention methods generation and development of phase phase in the cell
Equilibrium ratio of methane
envelope diagrams for different reservoir fluids would be of Mole fraction of methane in
importance for designers and operating engineers. In this the liquid
book attempts were made to address some of the difficul- Mole fraction of methane in
ties associated with property prediction of heavy and complex the gas
petroleum mixtures and with limited data available appropri-
ate approaches are recommended; however, the challenge in
this area of petroleum research continues.
volume. The system reaches to final equilibrium state at
pressure of 270 bar when temperature is kept constant
9.9 P R O B L E M S at 403 K. Determine the bubble point pressure of oil at
403 K. Also determine the final equilibrium composi-
9.1. Three-Phase Flash--Consider three phases of water, tion of gas in terms of mole fractions of N2, CO2, C~, Ca,
hydrocarbon, and vapor in equilibrium under reservoir (Ca q- C4), and C5+ in the free volume.
conditions. Water (L1) and hydrocarbons (L2) in the 9.5. Consider a constant volume-temperature cylinder as
liquid phase form two immiscible phases. Develop ap- shown in Fig. 8.13. The volume of cylinder is 96.64 cm 3
propriate equations for three-phase flash calculations and its length is 20.5 cm. Initially the cell is filled with
and derive relations for calculation of xL1, x/L2, and 30 vol% liquid n-pentane at 311.1 K and 100 bar. The
Yi. Measurement and prediction of VLLE in water- rest of the cylinder is filled with pure methane at the
hydrocarbon systems by PR EOS has been presented same initial temperature and pressure. Since the sys-
by Eubank et al. [62]. tem is not in equilibrium it approaches to a final equi-
9.2. Derive Eq. (9.8) for calculation of GOR. librium state at a lower pressure keeping temperature of
9.3. Calculate composition of liquid and gas streams from the cell constant. Through constant volume isothermal
the third stage in Table 9.1 (also see Fig. 9.3) using Stand- flash calculations using PR EOS and information given
ing correlations for calculation of K/. in the problem complete Table 9.15.
9.4. Consider the PVT cell and the core sample shown in 9.6. Composition of a reservoir fluid (gas condensate) sep-
Fig. 9.26. The free volume is 268 cm 3 and is filled ini- arated in a separator at 300 psig and 62~ is given
tially with pure N2. The core (porous media) has porosity in Table 9.16. The C7+ properties are given as fol-
of 0.31 and is filled with saturated oil with the follow- lows: SG7+ -- 0.795 and M7+ = 143. Laboratory mea-
ing composition in terms of mole fraction (Table 9.14). sured value of produced stock tank liquid-to-well stream
The C7+ has molecular weight (MT§ and specific gravity ratio is 133.9 bbl/MMscf and the gas-to-feed ratio is
(SG7+) of 228 and 0.853, respectively. Nitrogen diffuses 801.66 Mscf/MMscf. Associated gas (separator product)
into the core and light gases from matrix into the free specific gravity is SGgas = 0.735 and the primary stage
TABLE 9.16--Composition of reservoir fluid of Problem 9.6. 9.9. The following d a t a (Table 9.17) on two types of Chinese
WelIstream, Separator liquid, Separator gas, r e c o m b i n e d c r u d e oils are given b y H u et al. [63]:
Component tool% mol% mol%
The reservoir t e m p e r a t u r e is at 339 K a n d m e a s u r e d
CO2 0.18 Trace 0.22
N2 0.13 Trace 0.16 b u b b l e p o i n t p r e s s u r e s for oils I a n d 2 are 102.8 a n d
C1 61.92 7.78 75.31 74.2 bar, respectively. Densities of Cl1+ fraction at 20~
C2 14.08 10.02 15.08 for oils 1 a n d 2 are 0.91 a n d 0.921 g/cm 3, respectively.
C3 8.35 15.08 6.68 Ml1+ for oils 1 a n d 2 are 428 a n d 443, respectively. At
i-C4 0.97 2.77 0.52 the reservoir p r e s s u r e of 150 bar, viscosities of oils 1
n-C4 3,41 11.39 1.44
i-C5 0,84 3.52 0.18 a n d 2 are 5.8 a n d 6. I cP, respectively. E s t i m a t e the bub-
n-C5 1.48 6.50 0.24 ble p o i n t p r e s s u r e s from an EOS for these two oils a n d
C6 1.79 8.61 0.11 c o m p a r e with available data.
C7+ 6.85 34.33 0.06 9.10. Mei et al. [46] have r e p o r t e d e x p e r i m e n t a l d a t a on com-
Total 100 100 100 p o s i t i o n of a well s t r e a m fluid from West China oil field
w i t h c o m p o s i t i o n s of s e p a r a t o r gas a n d p r o d u c e d oil as
given in Table 9.18. Reservoir conditions (T a n d P), sat-
GOR is 4428 scf/bbl at 60 ~ The API gravity of p r o d u c e d u r a t i o n p r e s s u r e of fluid at reservoir T, a n d the GOR of
crude oil is 58.5. Calculate the following: reservoir fluid are also given in this table. Density of the
a. C o m p o s i t i o n of s e p a r a t o r gas a n d liquid using Stand- reservoir fluid (well s t r e a m u n d e r reservoir T a n d P) h a s
ing c o r r e l a t i o n for Ki. b e e n m e a s u r e d a n d r e p o r t e d . F r o m analysis of d a t a it is
b. SGgas for s e p a r a t o r gas. observed t h a t t h e r e is an e r r o r in the c o m p o s i t i o n of well
c. API gravity of s e p a r a t o r liquid. s t r e a m as the s u m of all n u m b e r s is 90.96 r a t h e r t h a n
d. GOR in scf/bbl. 100. I n a d d i t i o n reservoir t e m p e r a t u r e of 94 K is n o t
e. Stock t a n k liquid to well s t r e a m ratio in barrels/ correct (too low). P e r f o r m the following calculations to
MMscf. get correct values for the well s t r e a m c o m p o s i t i o n a n d
f. Gas-to-feed ratio in Mscf/MMscf. reservoir t e m p e r a t u r e .
g. C o m p a r e p r e d i c t e d values with available l a b o r a t o r y a. R e c o m b i n e s e p a r a t o r gas a n d oil t a n k to get the orig-
values. inal well stream. Make a p p r o p r i a t e m a t e r i a l b a l a n c e
9.7. F o r the gas c o n d e n s a t e s a m p l e of P r o b l e m 9.6 calculate calculations, using m o l e c u l a r weight, to generate well
Z factor at the reservoir conditions of 186~ a n d 5713 s t r e a m c o m p o s i t i o n . Also d e t e r m i n e if given GOR is
psia a n d c o m p a r e it with the r e p o r t e d value of 1.107. in stm3/m 3 or m3/m 3 at s e p a r a t o r conditions.
W h a t is the value of gas c o n d e n s a t e e x p a n s i o n factor in b. Use b u b b l e - T calculations to calculate reservoir tem-
Mscf for each b b l at reservoir conditions? The m e a s u r e d p e r a t u r e at w h i c h c o r r e s p o n d i n g s a t u r a t i o n p r e s s u r e
value is 1.591 Mscf/bbl. is 311.5 bar.
9.8. F o r the gas c o n d e n s a t e s a m p l e of P r o b l e m 9.6 calculate c. Use t r i a l - a n d - e r r o r p r o c e d u r e to find a t e m p e r a t u r e
dew p o i n t pressure (Pa) at 186~ a n d c o m p a r e it w i t h at w h i c h calculated density of reservoir fluid m a t c h e s
the m e a s u r e d value of 4000 psia. m e a s u r e d r e p o r t e d value at reservoir pressure. This
[33] Buckley, J. S., "Microscopic Investigation of the Onset of [48] Nor-Azian, N. and Adewumi, M. A., "Development of
Asphaltene Precipitation," Fuel Science and Technology Asphaltene Phase Equilibria Predictive Model," SPE Paper
International, Vol. 14, No. 1/2, 1996, pp. 55-74. No. 26905, Paper presented at the Eastern Regional Conference
[34] Fotland, E, Anflndsen, H., and Fadnes, E H., "Detection of and Exhibition, Pittsburgh, PA, November 2--4, 1993.
Asphaltene Precipitation and Amounts Precipitated by [49] Wu, J., Prausnitz, J. M., and Firoozabadi, A., "Molecular
Measurement of Electrical Conductivity," Fluid Phase Thermodynamics of Asphaltene Precipitation in Reservoir
Equilibria, Vol. 82, 1993, pp. 157-164. Fluids," American Institute of Chemical Engineers Journal,
[35] Escobedo, J. and Mansoori, G. A., "Viscometric Determination Vol. 46, No. 1, 2000, pp, 197-209.
of the Onset of Asphaltene Flocculation: A Novel Method," [50] Burke, N. E., Hobbs, R. E., and Kashou, S. E, "Measurement
Paper No. 28018, Society of Petroleum Engngineers, Production and Modeling of Asphaltene Precipitation," Journal of
and Facilities, Vol. 10, May 1995, pp. 115-118. Petroleum Technology, Vol. 42, November 1990,
[36] Pan, H. and Firoozabadi, A., "Thermodynamic Micellization pp. 1440-1447.
Model for Asphaltene Precipitation Inhibition," American [51] Qin, X., "Asphaltene Precipitation and Implementation of
Institute of Chemical Engineers Journal, Vol. 46, No. 2, 2000, Group Contribution Equation of State Into UTCOME" M.Sc.
pp. 416-426. Thesis, Department of Petroleum Engineering, University of
[37] Piro, G., Rabaioli, M. R., Canonico, L. B., and Mazzolini, E. I., Texas, Austin, TX, August 1998.
"Evaluation of Asphaltene Removal Chemicals: A New Testing [52] Hirschberg, A., Delong, L. N. J., Schipper, B. A., and Meijers,
Method," Paper No. 27386, Paper presented at the SPE J. G., "Influence of Temperature and Pressure on Asphaltene
International Symposium on Formation Damage Control, Flocculation," Society of Petroleum Engineers Journal, Vol. 24,
Lafayette, LA, February 7-10, 1994. No. 3, 1984, pp. 283-293.
[38] Rassamdana, H. B., Dabir, B., Nematy, M., Farhani, M., and [53] Ghuraiba, M. M. A., "Characterization of Asphaltenes in Crude
Sahimi, M., "Asphaltene Flocculation and Deposition: I. The Oil," M.Sc. Thesis, Chemical Engineering Department, Kuwait
Onset of Precipitation," American Institute of Chemical University, Kuwait, 2000.
Engineers Journal, Vol. 42, No. 1, 1996, pp. 10-22. [54] Oil and Gas Journal Data Book, 2000 Edition, PennWell, Tulsa,
[39] Wu, J., Prausnitz, J. M., and Firoozabadi, A., OK, 2000.
"Molecular-ThermodynamicFramework for Asphaltene-Oil [55] Munck, J., Skjold-Jorgensen, S., and Rsmussen, E,
Equilibria," American Institute of Chemical Engineers Journal, "Compositions of the Formation of Gas Hydrates," Chemical
Vol. 44, No., 5, 1998, pp. 1188-1199. Engineering Science, Vol. 43, No. 10, 1988, pp. 2661-2672.
[40] Pan, H. and Firoozabadi, A., "Thermodynamic Micellization [56] Riazi, M. R., Whitson, C. H., and da Silva, E, "Modeling of
Model for Asphaltene Precipitation from Reservoir Crudes at Diffusional Mass Transfer in Naturally Fractured Reservoirs,"
High Pressures and Temperatures," Society of Petroleum Journal of Petroleum Science and Engineering, Vol, 10, 1994,
Engineers (SPE) Production & Facilities, Vol. 15, No. I, pp. 239-253.
February 2000, pp. 58-65. [57] Warren, J. E. and Root, E J., "The Behavior of Naturally
[41] Won, K. W., Thermodynamics for Solid Solution-Liquid-Vapor Fractured Reservoirs," Society of Petroleum Engineers Journal,
Equilibria: Wax Formation from Heavy Hydrocarbon Vol. 3, 1963, pp. 245-251.
Mixtures," Fluid Phase Equilibria, Vol. 30, 1986, pp. 265-279. [58] Kesler, M. G. and Lee, B. I., "Improve Prediction of Enthalpy of
[42] Pedersen, K., Skovborg, E, and Ronningsen, H. E, "Wax Fractions," Hydrocarbon Processing, March 1976, pp. 153-158.
Presipitation from North Sea Crude Otis, 4: Thermodynamic [59] Mossaad, E. and Eubank, E T., "New Method for Conversion of
Modeling," Energy and Fuels, Vol. 5, No. 6, 1991, pp. 924-932. Gas Sonic Velocities to Density Second Virial Coefficients,"
[43] Lira-Galeana, C. A., Firoozabadi, A., and Prausnitz, J. M., American Institute of Chemical Engineers'Journal, Vol. 47,
"Thermodynamics of Wax Precipitation in Petroleum No. 1, 2001, pp. 222-229.
Mixtures," American Institute of Chemical Engineers Journal, [60] Iglesias-Silva, G. A., Bonilla-Petriciolet, A., Eubank, E T.,
Vol. 42, 1996, pp. 239-248. Holste, J. C., and Hall, K. R., "An Algebraic Method that
[44] Chung, T.-H., "Thermodynamic Modeling for Organic Solid Includes Gibbs Minimization for Performing Phase
Precipitation," Paper no. 24851, Paper presented at the 67th Equilibrium Calculations for Any Number of Components
Annual Technical Conferenceand Exhibition, Washington, DC, and Phases," Fluid Phase Equilibria, Vol. 210, No. 2, 2003,
October 4-7, 1992. pp. 229-245.
[45] Hansen, J. H., Fredenslund, Aa., Pedersen, K. S., and [61] Chung, T. H., Ajlan, M., Lee, L. L., and Starling, K. E.,
Ronningsen, H. E, "A Thermodynamic Model for Predicting "Generalized Multiparameter Correlation for Nonpolar and
Wax Formation in Crude Oils," American Institute of Chemical Polar Fluid Transport Properties," Industrial Engineering and
Engineers Journal, Vol. 34, No. 12, 1988, pp. 1937-1942. Chemical Research, Vol. 27, 1988, pp. 671-679.
[46] Mei, H., Zhang, M., Li, L., Sun, L., and Li, S., "Research of [62] Eubank, E T., Wu, C. H., Alvarado, E J., Forero, A., and Beladi,
Experiment and Modeling on Wax Precipitation for Pubei M. K., "Measurements and Prediction of Three-Phase Water/
Oilfield," Paper (No. 2002-006) presented at the International Hydrocarbon Equilibria," Fluid Phase Equilibria, Vol. 102,
Petroleum Conference 2002, Canadian Petroleum Society, No. 2, 1994, pp. 181-203.
Calgary, Alberta, Canada, June 11-13, 2002. [63] Hu, Y.-E, Li, S., Liu, N., Chu, Y.-E, Park, S. J., Mansoori, G. A.,
[47] Thomas, E B., Bennion, D. B., and Bennion, D. W., and Guo, T.-M., "Measurement and Corresponding States
"Experimental and Theoretical Studies of Solids Precipitation Modeling of Asphaltene Precipitation in Jilin Reservoir Oils,"
from Reservoir Fluid," Journal of Canadian Petroleum Journal of Petroleum Science and Engineering, Vol. 41, 2004,
Technology, Vol. 31, No. 1, 1992, pp. 22-31. pp. 169-182.
MNL50-EB/Jan. 2005
Appendix
ASTM DEFINITIONS OF TERMS BTU--One British thermal unit is the amount of heat re-
quired to raise 1 lb of water I~ E 1705, E48
ASTM DICTIONARY OF SCIENCE AND TECHNOLOGY1 defines vari- Carbon blackmA material consisting essentially of elemental
o u s engineering terms in standard terminology. ASTM pro- carbon in the form of near-spherical colloidal particles and
vides several definitions for most properties by its different coalesced particle aggregates of colloidal size, obtained by
committees. The closest definitions to the properties used in partial combustion or thermal decomposition of hydrocar-
the book are given below. The identifier provided includes the bons. D 1566, D l l
standard designation in which the term appears followed by Carbon residue---The residue formed by evaporation and
the committee having jurisdiction of that standard. For ex- thermal degradation of a carbon-containing material.
ample, D02 represents the ASTM Committee on Petroleum D 4175, D02
Products and Lubricants. Catalyst--A substance whose presence initiates or changes
the rate of a chemical reaction, but does not itself enter
Additive---Any substance added in small quantities to an-
into the reaction. C 904, C03
other substance, usually to improve properties; sometimes
Cetane number (cn)----A measure of the ignition perfor-
called a modifier. D 16, D01
mance of a diesel fuel obtained by comparing it to reference
Aniline p o i n t m T h e minimum equilibrium solution temper-
fuels in a standardized engine test. D 4175, D02
ature for equal volumes of aniline (aminobenzene) and
Chemical p o t e n t i a l (/~i or r partial molar free energy
sample. D 4175, D02
of component i, that is, the change in the free energy of a
API g r a v i t y ~ A n arbitrary scale developed by the American
solution upon adding 1 tool of component i to an infinite
Petroleum Institute and frequently used in reference
amount of solution of given composition, (SG/~n4)r.v,.~ =
to petroleum insulating oil. The relationship between API
Gi =/zi, where G -- Gibbs free energy and r~ = number of
gravity and specific gravity 60/60~ is defined by the fol-
moles of the ith component. E 7, E04
lowing: Degree API gravity at 60~ = 141.5/(SG 60/60~ -
131.5. [Note: For definition see Eq. (2.4) in this book.] Cloud p o i n t - - T h e temperature at which a defined liquid mix-
D 2864, D27 ture, under controlled cooling, produces perceptible haze
Ash--Residue after the combustion of a substance under or cloudiness due to the formation of fine particles of an
specified conditions. D 2652, D28 incompatible material. D 6440, D01
Assay~Analysis of a mixture to determine the presence or Coal--A brown to black combustible sedimentary rock (in
concentration of a particular component. F 1494, F23 the geological sense) composed principally of consolidated
Autoignition--The ignition of material caused by the appli- and chemically altered plant remains. D 121, D05
cation of pressure, heat, or radiation, rather than by an CokemA carbonaceous solid produced from coal, petroleum,
external ignition source, such as a spark, flame, or incan- or other materials by thermal decomposition with passage
descent surface. D 4175, D02 through a plastic state. C 709, D02
Autoignition t e m p e r a t u r e - - T h e minimum temperature at Combustion~A chemical process of oxidation that occurs at
which autoignition occurs. D 4175, D02 a rate fast enough to produce heat and usually light either
Average (for a series of observations)~The total divided by as glow or flames. D 123, D I 3
the number of observations. D123, D13 Compressed natural gas (CNG)~Natural gas that is typi-
Bar--Unit of pressure; 14.5 lb/in 2, 1.020 kg/cm 2, 0.987 atm, cally pressurized to 3600 psi. CNG is primarily used as a
0.1 MPa. D 6161, D I 9 vehicular fuel. D 4150, D03
Bitumen--A class of black or dark-colored (solid, semisolid, Concentration--Quantity of substance in a unit quantity of
or viscous) cementitious substances, natural or manufac- sample. E 1605, E06
tured, composed principally of high-molecular-weight hy- Critical p o i n t - - I n a phase diagram, that specific value
drocarbons, of which asphalts, tars, pitches, and asphaltites of composition, temperature, pressure, or combinations
are typical. D 8, D04 thereof at which the phases of a heterogeneous equilibrium
Boiling p o i n t - - T h e temperature at which the vapor pressure become identical. E 7, E04
of an engine coolant reaches atmospheric pressure under Critical pressure---Pressure at the critical point.
equilibrium boiling conditions. [Note: This definition is ap- E 1142, E37
plicable to all types of liquids.] D 4725, D I S Critical temperature----(1) Temperature above which the
Boiling p r e s s u r e - - A t a specified temperature, the pressure vapor phase cannot be condensed to liquid by an increase
at which a liquid and its vapor are in equilibrium. in pressure. E 7, E04
E 7, E04 (2)Temperature at the critical point. E 1142, E37
ASTM Dictionary of Engineering Science and Technology, 9th ed., D e g r a d a t i o n ~ D a m a g e by weakening or loss of some prop-
ASTM International, West Conshohocken, PA, 2000. erty, quality, or capability. E 1749, E 06
397
Degree Celsius (~ unit of temperature in the In- Gibbs free energy--The maximum useful work that can be
ternational System of Units (SI). E 344, E20 obtained from a chemical system without net change in
Density--The mass per unit volume of a substrate at a spec- temperature or pressure, AF = AH - TAS. [Note: For def-
ified temperature and pressure; usually expressed in g/mE inition see Eq. (6.6) in this book; the author has used G for
kg/L, g/cm 3, g/L, kg/m 3, or lb/gal. D 16, D01 Gibbs free energy.] E 7, E04
Deposition--The chemical, mechanical, or biological pro- GrainmUnit of weight; 0.648 g, 0.000143 lb. D 6161, D19
cesses through which sediments accumulate in a resting Gross calorific value (synonym: higher heating value,
place. D 4410, D19 HI-IV)raThe energy released by combustion of a unit quan-
Dew point--The temperature at any given pressure at which tity of refuse-derived fuel at constant volume or constant
liquid initially condenses from a gas or vapor. It is specifi- pressure in a suitable calorimeter under specified condi-
cally applied to the temperature at which water vapor starts tions such that all water in the products is in liquid form.
to condense from a gas mixture (water dew point) or at This the measure of calorific value is predominately used
which hydrocarbons start to condense (hydrocarbon dew in the United States. E 856, D34
point). D 4150, D03 Heat capacity--The quantity of heat required to raise a sys-
Diffusion--(1) Spreading of a constituent in a gas, liquid, or tem 1~in temperature either at constant volume or constant
solid tending to make the composition of all parts uniform. pressure. D 5681, D34
(2) The spontaneous movement of atoms or molecules to Heat flux (q)~The heat flow rate through a surface of unit
new sites within a material. B 374, B08 area perpendicular to the direction of heat flow (q in SI
Distillation--The act of vaporizing and condensing a liq- units: W/m 2; q in inch-pound units: Btu/h/ft 2 = Btu/h 9ft2)
uid in sequential steps to effect separation from a liquid C 168, C16
mixture. E 1705, E 48 Henry's law--The principle that the mass of a gas dissolved
Distillation t e m p e r a t u r e (in a c o l u n m distillation)---The in a liquid is proportional to the pressure of the gas above
temperature of the saturated vapor measured just above the liquid. D 4175, D02
the top of the fractionating column. D 4175, D02 Higher heating value (HHV)--A synonym for gross calorific
E n d o t h e r m i e r e a c t i o n - - A chemical reaction in which heat value. D 5681, D34
is absorbed. C 1145, C 28 Inert components---Those elements or components of nat-
E n t h a l p y m A thermodynamic function defined by the equa- ural gas (fuel gas) that do not contribute to the heating
tion H = U + PV, where H is the enthalpy, U is the internal value. D 4150, D03
energy, P is the pressure, and V the volume of the system.
Inhibitor--A substance added to a material to retard or pre-
[Note: Also see Eq. (6.1) of this book.] E 1142, E37
vent deterioration. D 4790, D16
E q u i l i b r i u m ~ A state of dynamic balance between the op-
Initial boiling p o i n t - - T h e temperature observed immedi-
posing actions, reactions, or velocities of a reversible
ately after the first drop of distillate falls into the receiving
process. E 7, E04
cylinder during a distillation test. D 4790, D 16
Evaporation--Process where a liquid (water) passes from a
Interface--A boundary between two phases with different
liquid to a gaseous state. D 6161, D19
chemical or physical properties. E 673, E 42
Fire p o i n t - - T h e lowest temperature at which a liquid or solid
specimen will sustain burning for 5 s. D 4175, D02 Interracial tension (IF]F)---The force existing in a liquid-
Flammable liquid--A liquid having a flash point below liquid phase interface that tends to diminish the area of
37.8~ (100~ and having a vapor pressure not exceed- the interface. This force, which is analogous to the surface
ing 40 psi (absolute) at 37.8~ and known as a Class I liquid. tension of liquid-vapor interfaces, acts at each point on the
E 772, E44 interface in the plane tangent at that point. D 459, D I 2
Flash p o i n t - - T h e lowest temperature of a specimen cor- International System of Units, SI--A complete coherent
rected to a pressure of 760 m m H g (101.3 kPa), at which system of units whose base units are the meter, kilogram,
application of an ignition source causes any vapor from second, ampere, kelvin, mole, and candela. Other units are
the specimen to ignite under specified conditions of test. derived as combinations of the base units or are supple-
D 1711, D09 mentary units. A 340, A06
FluiditymThe reciprocal of viscosity. D 1695, D01 Interphase---The region between two distinct phases over
Freezing point--The temperature at which the liquid and which there is a variation of a property. E 673, E42
solid states of a substance are in equilibrium at a given ISO--Abbreviation for International Organization for Stan-
pressure (usually atmospheric). For pure substances it is dards: An organization that develops and publishes inter-
identical with the melting point of the solid form. national standards for a variety of technical applications,
D 4790, D16 including data processing and communications.
Gas--One of the states of matter, having neither independent E 1457, F05
shape nor volume and tending to expand indefinitely. Jet fuel--Any liquid suitable for the generation of power by
D 1356, D22 combustion in aircraft gas turbine engines. D 4175, D02
Gasification--Any chemical or heat process used to convert Joule (J)--The unit of energy in the SI system of units. One
a feedstock to a gaseous fuel. E 1126, E 48 joule is 1 W..-. A 340, A06
Gasoline---A volatile mixture of liquid hydrocarbons, nor- Kelvin (K)--The unit of thermodynamic temperature; the SI
mally containing small amounts of additives, suitable for unit of temperature for which an interval of 1 kelvin (K)
use as a fuel in spark-ignition internal combustion engines. equals exactly an interval of 1~ and for which a level of
D 4175, D02 273.15 K equals exactly 0~ D 123, D13
APPENDIX 399
Liquefied petroleum gas (LPG)--A mixture of normally that all the water in the products remains in the form of
gaseous hydrocarbons, predominantly propane or butane vapor. D 121, I)05
or both, that has been liquefied by compression or cooling, Net heat of combustion--The oxygen bomb (see Test
or both, to facilitate storage, transport, and handling. Method D 3286) value for the heat of combustion, corrected
D 4175, D02 for gaseous state of product water. E 176, E05
Liquid--A substance that has a definite volume but no defi- Octane number (for spark ignition engine fuel)mAny one
nite form, except such given by its container. It has a viscos- of several numerical indicators of resistance to knock ob-
ity of 1 x 10-3 to 1 x 103 St (1 x 10-7 to 1 x 10-1 m 2 "S-1) tained by comparison with reference fuels in standardized
at 104~ (40~ or an equivalent viscosity at agreed upon engine or vehicle tests. D 4175, D02
temperature. (This does not include powders and granular Oxygenate---An oxygen-containing ashless organic com-
materials.) Liquids are divided into two classes: pound, such as an alcohol or ether, which may be used as
a fuel or fuel supplement. D 4175, D02
(1) Class A, low viscosity--A liquid having a viscosity of
Paraffinic oilmA petroleum oil (derived from paraffin crude
1 x 10-3 to 25.00 St (1 x 10 -7 to 25.00 x 10-4 m 2 -s -I) at
oil) whose paraffinic carbon type content is typically greater
104~ (40~ or an equivalent viscosity at agreed upon
than 60%. E 1519, E35
temperature.
Partial pressure---The contribution of one component of a
(2) Class B, high viscosity--A liquid having a viscosity of
system to the total pressure of its vapor at a specified tem-
25.01 to 1 x 103 St (25.01 x 10-4 to 1 x 10-1 m 2.s 1)
perature and gross composition. E 7, E04
at 104~ (40~ or an equivalent viscosity at agreed upon
Porosity--The percentage of the total volume of a material
temperature.
occupied by both open and closed pores. [Note: In this book
D 16, D01
porosity represented by ~ (see Eq. 8.72) is the fraction of
Lower heating value (LHV)mA synonym for net calorific total volume of a material occupied by open pores and is
value. D 5681, D34 not identical to this definition.] C 709, D02
Lubricant--Any material interposed between two surfaces Pour point--The lowest temperature at which a liquid can
that reduces the friction or wear between them. be observed to flow under specified conditions.
D 4175, D02 1) 2864, 1)27
MassmThe quantity of matter in a body (also see weight). Precipitation--Separation of new phase from solid, liquid,
D 123, D13 or gaseous solutions, usually with changing conditions or
Melting p o i n t - - I n a phase diagram, the temperature at temperature or pressure, or both. E 7, E04
which the liquids and solids coincide at an invariant point. Pressure--The internal force per unit area exerted by any
E 7, E04 material. Since the pressure is directly dependent on the
Micron (/~m, micrometer)--A metric unit of measurement temperature, the latter must be specified. 1) 3064, 1)10
equivalent to 10-6 m, 10 4 cm. Pressure, saturation--The pressure, for a pure substance
1) 6161, D19 at any given temperature, at which vapor and liquid, or
Molality~Moles (gram molecular weight) of solute per vapor and solid, coexist in stable equilibrium. [Note: This
1000 g of solvent. 1) 6161, 1)19 is the definition of vapor pressure used in this book.]
Molarity~Moles (gram molecular weight) of solute per liter E 41, G03
of total solution 1) 6161, 1)19 QualitymCollection of features and characteristics of a prod-
Molecular d i f f u s i o n ~ A process of spontaneous intermixing uct, process, or service that confers its ability to satisfy
of different substances, attributable to molecular motion, stated or implied needs. E 253, E18
and tending to produce uniformity of concentration. Range---The region between the limits within which a quan-
D1356, D22 tity is measured and is expressed by stating the lower and
Mole fraction--The ratio of the number of molecules (or upper range values. E 344, E20
moles) of a compound or element to the total number of Refractive index--The ratio of the velocity of light (of speci-
molecules (or moles) present. 1) 4023, 1)22 fied wavelength) in air to its velocity in the substance under
Naphtha, aromatic solvent--A concentrate of aromatic hy- examination. This is relative refractive index of refraction.
drocarbons including C8, C9, and C10 homologs. If absolute refractive index (that is, referred to vacuum)
D 4790, 1) 16 is desired, this value should be multiplied by the factor
Napthenic oilmAn hydrocarbon process oil containing more 1.00027, the absolute refractive index of air. [Note: In this
than 30%, by mass, of naphthenic hydrocarbons. book absolute refractive index is used.] 1) 4175, 1)02
1) 1566, D l l Saturation--The condition of coexistence in stable equilib-
Natural gas---A naturally occurring mixture of hydrocarbon rium of a vapor and a liquid or a vapor and solid phase of
and nonhydrocarbon gases found in porous geological for- the same substance at the same temperature. E 41, G03
mations (reservoirs) beneath the earth's surface, often in S m o k e pointmThe maximum height of a smokeless flame of
association with petroleum. The principal constituent of fuel burned in a wick-fed lamp. 1) 4175, 1)02
natural gas is methane. 1) 4150, D03 SolidmA state of matter in which the relative motion of
Net calorific value (Net heat of combustion at constant molecules is restricted and in which molecules tend to
pressure)--The heat produced by combustion of unit retain a definite fixed position relative to each other. A
quantity of a solid or liquid fuel when burned, at constant solid may be said to have a definite shape and volume.
pressure of 1 atm (0.1 MPa), under the conditions such E 1547, E 15
400 CHARACTERIZATION AND PROPERTIES OF PETROLEUM FRACTIONS
Solubility--The extent that one material will dissolve in an- Wax a p p e a r a n c e p o i n t - - T h e temperature at which wax or
other, generally expressed as mass percent, or as volume other solid substances first begin to separate from the liquid
percent, or parts per 100 parts of solvent by mass or vol- oil when it is cooled under prescribed conditions (refer to
ume. The temperature should be specified. D 3064, D10 D 3117, Test Method for Wax Appearance Point of Distillate
Solubility p a r a m e t e r (of liquids)--The square root of the Fuels). D 2864, D27
heat of vaporization minus work of vaporization (cohesive Weight ( s y n o n y m o u s with m a s s ) - - T h e mass of a body
energy density) per unit volume of liquid at 298 K. is a measure of its inertia, or resistance to change in motion.
D 4175, D02 E 867, E17
Solutes--Matter dissolved in a solvent. D 6161, D19
Specific gravity (deprecated term of liquids)--The ratio of
density of a substance to that of a reference substance such
as water (for solids and liquids) or hydrogen (for gases)
Greek Alphabet
under specified conditions. Also called relative density. Alpha
[Note: In this book the reference substance for definition fl Beta
of gas specific gravity is air]. D 4175, D02 F G a m m a (Uppercase)
Surface t e n s i o n - - P r o p e r t y that exists due to molecular F Gamma
forces in the surface film of all liquids and tends to pre- A Delta (Uppercase)
vent the liquid from spreading. B 374, B08 8 Delta
T e m p e r a t u r e - - T h e thermal state of matter as measured on e Epsilon
a definite scale. B 713, B01 ( Zeta
T h e r m a l conductivity (X)--Time rate of heat flow, under Eta
steady conditions, through unit area, per unit temperature | Theta (Uppercase)
gradient in the direction perpendicular to the area. 0 Theta
E 1142, E37 K Kappa (Uppercase)
T h e r m a l diffusivity--Ratio of thermal conductivity of a K Kappa
substance to the product of its density and specific heat A Lambda (Uppercase)
capacity. E l 142, E37 )~ Lamhda
Vapor--The gaseous phase of matter that normally exists in # Mu
a liquid or solid state. D 1356, D22 v Nu
Vapor pressure---The pressure exerted by the vapor of a liq- Xi
uid when in equilibrium with the liquid. D 4175, D02 H Pi (Uppercase)
Viscosity, absolute (~/)--The ratio of shear stress to shear zr Pi
rate. It is the property of internal resistance of a fluid that p Rho
opposes the relative motion of adjacent layers [Note: See N Sigma (Uppercase)
Eq. (8.1) in this book.] The unit most commonly used for Sigma
insulating fluids is centipoise. D 2864, D27 r Tau
Viscosity, k i n e m a t i c - - T h e quotient of the absolute (dy- v Upsilon
namic) viscosity divided by the density, O/p both at the q~ Phi (Uppercase)
same temperature. For insulating liquids, the unit most q~ Phi
commonly unit is the centistokes (100 cSt = 1 St). [Note: Phi
See Eq. (8.1) in this book.] D 2864, D27 x Chi
Viscosity, Saybolt Universal--The efflux time in seconds of Psi (Uppercase)
60 mL of sample flowing through a calibrated Saybolt Uni- Omega (Upper case)
versal orifice under specified conditions. D 2864, D27 co Omega
MNL50-EB/Jan. 2005
Index
A ASTM D 92, 34, 133 ASTM method, 128
ASTM D 93, 34, 133-134, 144 molecular weight prediction, 56
Absolute density, 120 ASTM D 97, 135, 144 Atmospheric critical pressure, heavy
Academia, 17 ASTM D 129, 99 hydrocarbons, 51
Acentric factor, 11 ASTM D 189, 141,144 Autoignition temperature, definition, 34
aromatics, 52 ASTM D 240, 144 Avogadro number, 24
definition, 33 ASTM D 287, 93
estimation, 80-82, 115-116 ASTM D 323, 33, 144
pure hydrocarbons, prediction, 64-66 ASTM D 341, 70, 338 B
Activation energy, 346 ASTM D 357, 34, 139
Activity coefficients ASTM D 445, 100, 144, 338 Benedict-Webb-Rubin equation of state,
mixtures, 254-255 ASTM D 446, 338 modified, 214, 217-220
calculation, 257-261 ASTM D 524, 144, 141 Benzene, 4-5
Albahri et al. method, 137 ASTM D 611, 35, 137, 144 vapor pressure, 3 l 3
Alcohols, octane number, 139 ASTM D 613, 138 Binary interaction parameter, 209-210,
Alkanes ASTM D 908, 34, 139 269-270
boiling point, 58-59 ASTM D 976, 138 Binary systems, freezing-melting diagram,
critical compressibility factor, 64 ASTM D 1018, 99 285
critical temperature, 50 ASTM D 1160, 92, 100-101,106, 108, 110, Block and Bird correlation, 359
entropy of fusion, 262 114, 144 Boiling point, 11
liquid thermal conductivity, 343-344 ASTM D 1218 94, 144 n-alkanes, 58-59
surface tension, 361 ASTM D 1262 99 definition, 31
vapor pressure, 306 ASTM D 1266 99, 144 elevation, 282-284
n-Alkyl, critical pressure, 52 ASTM D 1298 93 heavy hydrocarbons, 50, 52
Alkylbenzene ASTM D 1319 144 hydrocarbon-plus fractions, 173
entropy of fusion, 262 ASTM D 1322 142 petroleum fractions, 88-93
vapor pressure, 307 ASTM D 1368 99 prediction, pure hydrocarbons, 58-59
Analytical instruments, 96-98 ASTM D 1500 144 reduced, 251
ANFOR M 15-023, 10 ASTM D 1548 99 sub- or superatmospheric pressures,
Aniline point, 11 ASTM D 1552 99 106-107
definition, 35 ASTM D 1747 95 true, 89
petroleum fractions, 137 ASTM D 2007 96 Boiling points
Antoine coefficients, 310 ASTM D 2267 10 average, 100-101
Antoine equation, 305-306 ASTM D 2270, 122-124 range, 88
API degree, 21 ASTM D 2386, 136, 144 Boiling range fractions, narrow versus
API gravity, 11 ASTM D 2500, 135, 144 wide, 112-119
crude oils, 156 ASTM D 2501, 36 Boltzman constant, 24
definition, 32 ASTM D 2502, 56 Boossens correlation, 57-58
petroleum fractions, 93 ASTM D 2503, 94 Bubble point, calculations, 370-371
prediction, pure hydrocarbons, ASTM D 2533, 133 Bubble point curve, 201
58-60 ASTM D 2549, 97 Bubble point pressure, 223, 367
API methods, 124-126 ASTM D 2700, 144 Bubble point temperature, 368
critical temperature and pressure, ASTM D 2717-95, 144 Bulk parameters, petroleum fractions, 114
prediction, 60 ASTM D 2759, 127 Butane, equilibrium ratios, 274-275
critical volume, prediciton, 63 ASTM D 2887, 12, 89-90, I00, 104-105,
molecular weight prediction, 56 110, 144
API RP 42, 37, 56 ASTM D 2890, 320-321 C
API Technical Data Book-Petroleum ASTM D 2892, 144, 154
Refining, 15 ASTM D 2983, 144 Capillary pressure, 357
Aromatics, 4-5 ASTM D 3178, 99 Carbon
Arrhenius-type equation, 346 ASTM D 3179, 99 prediction in petroleum fractions, 127
Asphalt, 10 ASTM D 3228, 99 see also SCN groups
Asphaltene, 373-378 ASTM D 3238, 121, 126 Carbon number range approach, petroleum
inhibitor, 377-378 ASTM D 3343, 99, 128, 130 fractions, 186
precipitation, 375, 377, 379 ASTM D 3431, 99 Carbon residue, petroleum fractions,
solid-liquid equilibrium, 385-388 ASTM D 3710, 90 141-142
temperature and pressure effects, 381 ASTM D 4045, 99 Carbon-to-hydrogen weight ratio, 11
Association parameter, 347 ASTM D 4052, 93, 144 definition, 36
ASTM, definitions of terms, 397-400 ASTM D 4124, 96 Carnahan-Starling equation of state,
ASTM D 56, 133 ASTM D 4530, 141 214-215
ASTM D 86-90, 92, 100-106, 108, 110, ASTM D 4737, 144 Cavett method, 61
113-115, 118, 131, 134, 140, 144, ASTM D 4953, 131 Cementation factor, 351
313-314 ASTM D 5296, 94 Cetane number, petroleum fractions,
ASTM D 88, 23 ASTM D 5985, 135 137-138
401
402 C H A R A C T E R I Z A T I O N A N D P R O P E R T I E S OF P E T R O L E U M F R A C T I O N S
Virial equation of state, 210-214 Volatility index, and vapor liquid ratio, 133 Wax precipitation
truncated, 240 Volume, units, 20 rnultisolid-phase model, 382-385
Viscosity, 12, 331-338 Volume translation, cubic equations solid solution model, 378,
gases, 331-335 of state, 207-208 380-382
heavy hydrocarbons, 44 Volumetric flow rates, units, 20 Waxes, 373
liquids, 335-338 Wet gas, 6
petroleum fractions, 99-100 Wetting liquid, 357
pressure effect, 334 W Wilke-Chang method, 347
versus temperature, 332 Wilson correlation, 273
units, 23 Walsh-Mortimer method, 137 Winn method, 137
Viscosity-blendingindex, 337 Water Winn-Mobil method, 62
Viscosity coefficients, pure liquid ideal gas heat capacity, 242-243 Winn nomogram, 73-75
compounds, 336-337 vapor pressure, 312 Won model, 380
Viscosity gravity constant, 11 Watson characterization factor, 320
definition, 35-36 Watson K, 11, 13, 323
Viscosity index, 122-124 definition, 35 X
Viscosity-temperature relation, 72 Wax appearance temperature, 378,
Volatility, properties related to, 131-135 382 Xylene, vapor pressure, 311