You are on page 1of 22

THE ROLES OF ACTIVATED CARBON IN GAS CONDITIONING

by
MARTA J. BOURKE
AND
ANTHONY F. MAZZONI

CALGON CARBON CORPORATION


PITTSBURGH, PA 15230

For presentation at the Gas Conditioning Conference


March 6-8, 1989

INTRODUCTION

Activated carbon has been successfully used in


numerous gas processing applications for many
years. Carbon's highly porous nature and large
surface area make it an ideal adsorbent for
removing trace contaminants from liquid and
gaseous streams. For some inorganic contaminants
encountered in gas processing operations, removal
can be enhanced by impregnation of the carbon to
promote both physical adsorption and
chemisorption.
This paper will review uses for activated
carbon in gas conditioning in the
following areas:
Purification of recirculating amines
and glycols
Desulfurization of natural gas
Mercury removal from natural gas
Equipment and catalyst protection.
Calgon Carbon has extensive experience in these
applications. Specific examples will be reviewed
to provide basic guidelines and understanding of
the proper utilization, selection, and design of
activated carbon and related systems. Operating
parameters, performance results, and benefits
for a number of actual cases will be discussed.
ACTIVATED CARBON AND CARBON SELECTION

Activated carbon can be manufactured from any


carbonaceous material -coal, wood, peat, coconut
shell, etc. Granular activated carbon is
generally produced by grinding the raw material,
adding a suitable binder for hardness,
recompacting, and crushing to the correct mesh
size. The carbon-based

material is converted to activated carbon by


thermal decomposition of the organic materials
present in a controlled non-oxygen atmosphere.
The resultant product has an incredibly large
surface area per unit volume and a network of
submicroscopic pores where adsorption occurs. One
pound of carbon provides a surface area
equivalent to 125 acres.
Because activated carbon can be made from
various raw materials, differences will exist in
the finished product. Domestically, most
granular carbons are manufactured from various
grades of coal. Some grades of coal need a
chemical pretreatment prior to thermal
activation to develop proper pore structure.
Table I shows some of the properties of
activated carbon and how these vary with the raw
material. These differences in carbon properties
make some products better suited than others for
specific applications.
Iodine Number is the most common standard for
indicating total surface area available for
adsorption. It is defined as the milligrams of
iodine adsorbed by one gram of carbon and it
approximates the internal surface area (micro
pores) in square meters per gram. Table I shows
how Iodine Number can vary for activated
carbons produced from different materials.
Densities of the carbon can vary drastically;
fewer pounds of a carbon with a low density will
fill the same volume as a higher density
product. This is significant because the
contaminant removal capacity of this volume of
low density carbon is severely reduced. The
concept of volume activity then becomes
important in selecting carbons. A simple
calculation for determining the volume activity
of carbon is to multiply the bulk density by the
Iodine Number. Thus, two containers having the
same volume of carbon will have different total
surface areas if the densities are different.

Abrasion resistance refers to a carbon's


ability to withstand degradation during
handling and is expressed in terms of an
Abrasion Number; the higher the Number, the
more resistant a carbon is to abrasion. This
characteristic is especially important for high
pressure applications, where carbon fines in
the effluent can cause process problems and
where carbon is on line for extended periods.
Ash content of activated carbon can be
significant as some carbons have a high total
ash content which contribute nothing for
adsorption. These inorganic materials can leach
from the carbon and cause operational problems
such as foaming.
PURIFICATION OF AMINES AND GLYCOLS
In the process of absorbing acid gas
constituents, amine streams becomes contaminated
with thermal and chemical degradation products,
organic acids and/or iron sulfides.
Contamination of the amine stream can cause
foaming and carryover in the absorber or
stripper, high steam and circulation rates,
increased system corrosion and heat stable salt
formation. A properly designed activated carbon
system can reduce, and in many cases eliminate,
the need for anti foam agents, reduce amine
make-up rates, reduce total system energy
consumption, reduce corrosion, and improve
scrubbing efficiencies and product quality.
Historically, many carbon systems have been
under-designed and/or operated incorrectly. The
carbon systems had insufficient contact time,
and in many cases, the wrong carbon was used and
the carbon was rarely changed. A properly
designed carbon system should treat a 10-20%
slipstream of the amine solution, have a minimum
of 15 minutes empty bed contact time, and a
superficial velocity of 2-4 gpm/ft2. When the
amine solution changes color and clarity, it's
usually time to change the carbon, since this is
usually indicative of the carbon being
exhausted.

Figure 1 shows an effective carbon system.


Besides having a properly designed and sized
adsorber, the system also includes mechanical
filtration both upstream and downstream of the
carbon adsorber. The upstream filter removes
suspended solids to ensure the carbon is used to
adsorb contaminants, and the downstream filter
will remove carbon fines which could cause
foaming. A feature of this system is the carbon
handling equipment which allows quick and easy
change out of the carbon without exposing
personnel to spent carbon.
The carbon recommended is a hard, small mesh,
steam activated carbon, having a broad range of
pore diameters. This prevents short bed life,
excessive fines, foaming, and the ability to
remove a wide range of organic contaminants. The
carbon should also be free of phosphorus, which
could leach from the carbon bed in the form of
phosphates. These phosphates could cause
additional foaming in the amine stream.
Cost savings which can be realized with a
properly designed carbon system were presented in
an article titled "Gas Treaters Need Clean
Amines" in the December 1987 issue of Hydrocarbon
Processing. The article by R. L. Bright and D. A.
Leister of Calgon Carbon Corporation, illustrated
actual cost savings of a refinery in the western
United States. The carbon system was installed to
reduce a severe foaming problem in the amine
contactor. The amine solution was black in
appearance and operating costs were very high.
Table II shows the operating cost elements before
and after the activated carbon system was put
into service. Because the carbon treated amine
did not carry over from foaming, the amine makeup and anti foam consumption was reduced. The
refinery was able to increase the amine strength
and reduce circulation which resulted in steam
savings. Finally, the filter cartridge usage was
reduced because the activated carbon removed
organic acids from the amine stream resulting in
lower tendency to create iron sulfide. All of
these reductions resulted in a first year savings
of $147,800 (Table III). Second year savings
increased to $205,800 as the refiner was able to
further reduce anti foam addition, filter
cartridge replacement and amine make-up.

As a side benefit, the product gas quality


improved to where the refiner was able to reduce
H2S concentration from 10-40 ppm to at or near
non-detectable levels. Heat stable salts were
reduced to a low level of 5 ppm. Although
activated carbon does not adsorb heat stable
salts, it will remove the precursors to their
formation.
A similar application for activated carbon is
the removal of dissolved impurities from glycol
streams which could cause foaming. The reduced
foaming tendency of the glycol provides better
contact with the gee which improves drying
efficiency. In this application, the activated
carbon system can be installed on the rich side.
Due to the higher viscosity of the glycol, the
maximum superficial velocity recommended is
2gpm/ft2.
DESULFURIZATION
Natural gas feed stocks are used for the
production of many chemicals including ammonia,
methanol and hydrogen. A feed stock virtually
free of sulfur is required to protect the
catalysts from deactivation. Precious metal
catalysts are highly susceptible to sulfur
poisoning and subsequent loss of production and
shortened life.
Depending on the sulfur concentration, various
sulfur control technologies can be applied.
When H2S or other low molecular weight sulfur
compounds are present at 10 ppm or less (as
total sulfur), plants have successfully used
granular activated carbon to remove the sulfur
compounds. Metal oxide impregnated carbons such
as Calgon Carbon Type FCA are specifically
designed to chemisorb hydrogen sulfide and
mercaptans.

As shown in Figure II, the metal oxides within


the carbon pores react with the sulfur compounds
to form sulfides and sulfates. The carbon can
then be regenerated in situ with hot gee
containing traces of oxygen to restore the
impregnants to the metal oxide form. If large
amounts of undesirable high molecular weight
organics are present in the feed gas, a guard
bed of unimpregnated carbon is used.
This method of sulfur control is being utilized
at more than 60 plants in the United States. At
a Louisiana chemical plant, several beds of FCA
are being used for desulfurization of the
methanol and ammonia plant feed stocks. Table
III details the design conditions. The influent
H2S concentration in the natural gas varies up
to 10 ppm. Each unit has 2 carbon vessels in
series operation. When the concentration of H2S
in the natural gee leaving the lead vessel
reaches 0.22 ppm, the vessel is regenerated with
steam and traces of oxygen. The regenerated
carbon bed is then put into the polish position.
Currently, the time between regenerations is 6
months, but this can vary depending on H2S
concentration in the natural gas.
MERCURY REMOVAL
In years past, the presence or absence of
mercury received little attention in the oil and
gas industry. In the early seventies, natural
gee processors began to use cyogenic techniques
utilizing aluminum core heat exchangers (cold
boxes). After the failure of several cold boxes,
metallurgists determined that mercury corrosion
was the source of the problem. Initially, it was
believed that the mercury was present due to
leaking instrumentation; however, further
testing revealed mercury was present in the
reservoir.

Extremely low levels of naturally occurring


elemental mercury in the inlet gas can change to
a solid phase or liquid phase under normal plant
operating conditions. When the mercury
accumulates as a liquid in the heat exchangers,
it can amalgamate with the aluminum and destroy
the heat exchangers. The exchanger is not only
extremely expensive to replace or repair, but a
failure is an obvious safety hazard as well.
Traditionally, it was believed that only
Indonesian and North African gas contained
sufficient amounts of mercury to cause equipment
problems. However, several Gulf Coast gas and
olefin plants have experienced cold box failures
which have been traced to mercury corrosion
within the past two years. Unacceptable levels of
mercury have also been detected in tertiary oil
recovery plants in the western United States.
The suspected mechanism of mercury corrosion is
multi-step. Being an extremely heavy metal with
a very high vapor pressure, mercury tends to
accumulate in pipe bends and along the welds of
elbows -areas where the aluminum may already be
stressed. For liquid metal embrittlement (LME)
or intergranular mercury corrosion to occur, the
mercury must be present in a liquid state, i.e.
above -40F, and must wet the surface of the
aluminum. Aluminum is slightly soluble in
mercury and amalgamatea with the mercury. The
aluminum then forms an aluminum oxide and frees
the mercury to attack the next layer of metal.
Mercury is not permanently bound with the
aluminum, but actually acts as a catalyst for
the corrosion or embrittlement to occur.
Even minute amounts of mercury in the gas can
threaten the integrity of the plate-fin
aluminum exchangers. Conditions which favor
amalgamation are: temperatures in excess of 40F, an oxide free or stressed aluminum
surface, presence of moisture or air at the
mercury/aluminum interface, small amounts of
mercury in the liquid state.

Activated carbon is effective at physically


adsorbing trace components from gas streams due
to its high population of micro pores; however,
the capacity is low. By impregnating the carbon
with a compound which will chemically react with
the mercury, capacity can be enhanced. It is
imperative that not only must the mercury be
removed to non-detectable levels but that it is
adsorbed in a form which will not elute from the
carbon bed under fluctuating conditions. In
order to obtain long bed life and high removal
efficiencies, a carbon is required with a pore
structure which can accept substantial amounts
of suitable impregnant while allowing access of
the process gas to the complex pore structure.
Type HGR impregnated activated wee specifically
designed for this service. Under normal
operating conditions, a properly designed HGR
bed will remove mercury for many years. HGR
carbon has been used successfully since 1975 to
protect aluminum exchangers from corrosion due
to low level mercury contamination in LNG and,
more recently, olefin plants.
A south Texas pipeline company had a failure of
their aluminum cold box, which through
metallurgical analysis, was traced to mercury
corrosion. Calgon Carbon worked with the
pipeline company and, through on-site analysis,
determined there was over 50ug/Nm3 of mercury in
the natural gas. Using the operating conditions
of Table V, Calgon Carbon assisted in designing
a mercury removal system and determining the
best location of the system in the plant. The
mercury removal system has been operating since
1984 removing mercury to non-detectable levels.

CATALYST AND EQUIPMENT PROTECTION


Process and plant engineers have found that using
activated carbon is an excellent way to protect
expensive catalysts and process equipment from other
low concentration poisons which can deactivate
catalysts or damage equipment. Carbon is highly costeffective and relatively simple to use. Some of the
applications for utilizing carbon as a guard bed
include:
Removal of organic compounds that can foul
downstream equipment
Arsine control in acetylene, ethane-propane,
and other hydrocarbon streams
Aromatics removal to prevent "charring"
of Claus catalyst
Compressor oil removal for high purity
gases
Acid gas removal to protect sensitive
electrical equipment
Purification of process air.
Calgon Carbon utilizes impregnated and non-impregnated
carbons to control a wide variety of contaminants to
very low effluent levels. Since these carbons are
extremely versatile and can operate at existing plant
conditions, capital investments and operating costs are
minimal. All of the multitude of applications cannot be
discussed here; however, a few examples are presented.

One equipment protection application involved a tertiary


oil project which utilized triethylene glycol
(TEG)dehydration prior to the cryogenic separation
plant. Plant personnel and the design engineers became
concerned that the TEG would plate out in the cryogenic
processes and foul equipment. They installed Calgon
Carbon's pelletized activated carbon upstream for
removal of the TEG. The estimated influent TEG
concentration to the carbon beds was 9 ppm with an
effluent objective of 0.02 ppb. Data collected from the
site indicates high loading of the TEG (15 wt.%) on the
activated carbon and excellent removal efficiencies. See
Table VI.
With the advent of sophisticated computer and motor
control systems, the need has arisen to protect this
equipment from acid gas corrosion. The Instrument
Society of America has issued guidelines for computer
and motor control room air quality. The
concentrations listed below are for a G1 (mild)
environment:

COMPONENT MAXIMUM CONCENTRATION (ppb)


Hydrogen Sulfide
Chlorine
SO2, SO3
NOX
HF
Ozone
Failures of sensitive circuitry in computers,
laboratory equipment, and other electronic
equipment can lead to long downtimes and high
maintenance costs.

3
1
10
50
500
2

Calgon Carbon recommends an all-fiberglass system


utilizing a specially impregnated carbon
specifically developed for acid gas removal. Air
contaminated with corrosive gases is filtered at
the fan inlet to remove particulate matter. The
air is then delivered at a controlled rate and
pressure to a deep bed of Calgon Carbon IVP
Granular Activated Carbon to remove the corrosive
gases (see Figure III). Treated air is then
passed through a 30% and a 99% ASHRAE filter to
remove micron size particulates. As shown, the
carbon system can be placed on the make-up air
and/or the recirculating air depending upon the
level of severity. Sulfur plants are an obvious
source of hydrogen sulfide, SO2, and SO3.
Calgon Carbon's IVP Carbon is impregnated with
sodium hydroxide to enhance adsorption of acid
gases. Figure IV shows the chemical reactions
which occur. Since the gases are chemically
bonded to the carbon surface, there is no
potential for desorption of the acid gases and
the carbon can effectively remove hydrogen
sulfide to non-detectable levels even under
varying influent conditions.
SUMMARY
Granular activated carbon has been employed in
many gas conditioning and related processes. To
obtain optimum results, each application should
be individually evaluated by a carbon expert to
determine the appropriate carbon and design
parameters. A properly designed carbon system
can reduce downtime, improve gas quality,
protect equipment and catalysts, and reduce
overall operating costs.

TABLE I
TYPICAL PROPERTIES
OF
ACTIVATED CARBONS
BITUMINOUS SUB-BITUMINOUS
Iodine number
Bulk density as
packed (#/ft3)
Volume Activity
% Ash
% Phosphorus*
Abrasion Number

1,000
26
26,000
6.7
<0.05
80

LIGNITE

1,000

600

25

23

25,000
12.3
1-5
75

13,800
20.1
<0.05
60

* Component of ash which can cause foaming in amine


solutions

TABLE

II

OPERATING COST ELEMENTS BEFORE AND


AFTER CARBON SYSTEM INSTALLATION
PRE-CARBON
SYSTEM

AMINE CIRCULATION
(GPM)
ANTIFOAM CONSUMPTION
(DRUMS/YEAR)
FILTER CARTRIDGES
(NO. USED/DAY)

POST CARBON
SYSTEM

300
4-5
30-40

REDUCTION

265

35
1 -2

12-15

3
20

AMINE MAKE-UP
(GAL/MONTH)

1,400

1,000

400

STEAM USAGE
(LBS./MONTH)

18,000

16, 000

2,000

UNIT COST
DEA -$0.40/LB.
STEAM -$4.44/1,000 LBS.
FILTER CARTRIDGES -$7.37
ANTIFOAM -$4,000/DRUM

TABLE III

FIRST YEAR SAVINGS


AMINE MAKE-UP

$20,000

STEAM

77,800

FILTER CARTRIDGES

54,000

ANTIFOAM

12,000

CARBON USAGE

-16,000
TOTAL

$147,800

TABLE IV
H2S REMOVAL FROM NATURAL
GAS LOUISIANA CHEMICAL PLANT
DESIGN BASIS
Gas Flow (SCFM)
Gas Temp. (F)
Gas Pressure (PSIG)
H2S Conc.

Unit 1
6900

Unit 2
6800

Unit 3
3200

49

39

45

190

190

210

<---Varies between .1--->


and 10 ppm

FCA CARBON SYSTEM


Pounds of FCA/Vessel
2 Vessels Each Unit
H2S Concentration
Out of Vessels
Regeneration
About 6 months

8450

8450

6520

<---Non-detectable to--->
.22 ppm
<---Steam + Trace O2 --->

TABLE V
MERCURY REMOVAL FROM NATURAL GAS
SOUTH TEXAS PIPELINE COMPANY
DESIGN CONDITIONS
30 MMSCFD
1100 PSIG
110F
*Mercury content over 50 ug/Nm3
HGR CARBON SYSTEM (installed 1984)
Carbon Steel adsorber
1500 pounds Type HGR
*Mercury level after carbon 0.001 ug/Nm3

* Analyses are using a Jerome Model 301 Gold Film


Mercury Analyzer

TABLE VI
WEST TEXAS TERTIARY OIL RECOVERY
TEG REMOVAL
CARBON DIOXIDE INJECTION FACILITY

DESIGN CONDITIONS
2 Trains
93 MMSCFD, each train
370 psia
100F
GAS COMPOSITION
COMPOUND
Carbon Dioxide
Hydrogen Sulfide
Methane
C2+
TEG Contamination
CARBON SYSTEM
8.5' diameter
10' bed depth
16,000 pounds of carbon per
bed

MOLE PERCENT
77
1
12
10
9 ppm