You are on page 1of 257

Prof. Fatma K.

Gad

UNIT OPERATION ( PART I )

BY

DR. FATMA KHALIFA GAD

0
Prof.Fatma K. Gad

UNIT OPERATION
Introduction:

Chemical engineering has to do with industrial processes in which row


materials are changed or separated into useful products.

The unit operations are as applicable to many physical processes as to chemical


ones. For example, the process used to manufacture common salts consists of
the following sequence of the unit operations, transportation’s of solids and
liquids, transfer of heat, evaporations, crystallization, drying and screening. No
chemical reaction appears in these steps. On the other hand, the cracking of
petroleum with or without the aid of a catalyst is a typical chemical reaction
conducted on an enormous scale.

Here the unit operations, transportation of fluid and solids, distillation, and
various mechanical separations are vital and the cracking reaction could not be
utilized without them.

A number of scientific principles and techniques are basic to the treatment of


the unit operation. some are elementary physical and chemical lows, and others
are special techniques particularly useful in chemical engineering.

The following physical and chemical lows will be used frequently later.

1. Material and energy balances

Conservation of mass (energy) requires that the material (energy) entering


any process must be either accumulate or leave the process. There can be
neither loss nor gain during the process.

2. Molar units

A mole of any pure substance is defined as the quantity of that substance


whose mass is numerically equal to its molecular weight.

The mole is a mass unit and it can be used in material balances just like
mass in pounds or grams (mole).

The sum of the mole fraction is unity.

1
Prof.Fatma K. Gad

3. Ideal gas low

The low applies strictly to actual gases only at low pressures

PV = nRT

( V1 / V2 ) = ( P2 / P1 ) * ( T1 / T2 )

4 Partial pressure

All partial pressures in a given mixture add to the total pressure. This
applies to mixtures of both ideal and non ideal gasses. A generally useful
consequence of the ideal gas low is:

Volume percent = Pressure percent = mol percent.

For example, since air contains 79% N2 and 21% O2 by volume, 1 ft3 of air
under a pressure of P atm may be considered to be a mixture of 0.21 ft3 of
O2 , 0.79 ft3 of N2, both measured at P atm and the temperature of the
mixture.

It may be also considered to be 1ft 3 of O2 at 0.21 atm and 1ft3 of N2 at 0.79


atm, so that 0.21% of the pressure is from O2 and 0.79% from N2.

Finally 1 mole of air contains 0.21 mole of O2 and 0.79 mole of N2 at all
temperatures and pressures.

5. Equilibria

Equilibrium states represent end points of naturally occurring processes.


Equilibrium can not be shifted without making some change in the
conditions governing the system. Consequently the phase concept may be
defined as a homogeneous substance viewed independently of its shape
and size.

Gibbs phase rule:

F ( No. of independent variables ) = C ( No. of components ) – P ( No. of


phase ) + 2.

There are three phases, liquid, gas and solid.

2
Prof.Fatma K. Gad

The utility of the phase concept comes from the fact that at equilibrium the
temperature pressure and concentration are independent of the geometric
shape and size of the phases.

6. Process kinetics

Kinetics may be divided into chemical or physical parts. Physical kinetics


are that important in unit operations.

The rate of most physical processes is proportional to a quantity called a


driving force or driving potential ( F).

F = Resistance (R) * (d G) quantity can be heat, mass .../(dt) time

MASS TRANSFER OPERATION


Introduction:

When a concentration gradient exists within a fluid consisting of two or more


components, there is a tendency for each constituent to F low in such a
direction as to reduce the concentration gradient, this process is known as mass
transfer.

In a fluid flowing under stream line conditions in a direction at right angles to


the concentration gradient, the transfer is effected as a result of the random
motion of the molecules.

In a turbulent fluid, this mechanism is supplemented by trans force of material


by eddy currents.

Mass transfer can take place in either gas, liquid or solid phases, or in both two
phases simultaneously as follow:

1. Evaporation, transfer mass from liquid to gas.

2. Absorption; transfer mass from gas to liquid.

3. In a distillation column; the less volatile component diffuses (transfer) in


the gas phase to the liquid surface and the more volatile material diffuses at
an approximately equal molar rate in the opposite direction.

3
Prof.Fatma K. Gad

4. In liquid extraction; the solute diffuses through the raffinate phase to the
interface and then into the extract phase.

5. Crystallization; Solute diffuse through the mother liquid to the crystals


and deposits on the solid surface.

6. Drying: Liquid water diffuses through the solid toward the surface of the
solid vaporized and diffuses as vapor into the gas.

7. Humidification; There is no diffusion through the liquid phase because


the liquid phase is pure and no concentration gradient through it can exist.
But the vapor diffuses to or from the liquid gas interface in or out of the
gas phase.

So there are certain characteristics common to all the diffusion operations.


In every case a fluid is brought into intimate contact with another phase, a
concentration gradient is established and a transfer of matter from one
phase to the other results. The concentration gradient tends to move the
component in each direction as to equalize concentration and destroys the
gradient. For example the substance moves from a phase of high
concentration to one of low concentration.

The concentration difference is thus a driving force for the mass transfer.
Diffusion can also be caused by a pressure or temperature gradients
applied over the mixture.

In all mass transfer operation, diffusion occurs in at least one phase and
often in both phase. Example in gas absorption solute diffuses through the
gas phase to the interface and through the liquid phase from the interface.

Diffusion is not restricted, molecular transfer through stagnant layer of


solid or liquid. It also takes place in fluid phase by physical mixing and by
the eddies or turbulent flow.

Usually the diffusion process is accompanied by bulk flow of the mixture.


Thus we have to distinguish between two phenomena.

1. Molecular diffusion or spontaneous approach to uniformly a


molecular scale in stagnant liquid.

2. That which may be superimposed upon this process as a result


of bulk flow.

4
Prof.Fatma K. Gad

MOLECULAR DIFFUSION

The rate of transfer of A in a mixture of two components A and B will therefore


be determined not only by the rate of diffusion of A but also by the behavior of
B. The molar rate of transfer of A per unit area due to molecular motion is
given by Fick’s low.

NA = - DAB * ( dCA / dy ) (1)

Where NA The molar rate of diffusion per unit area ( lb mole / ft2 hr).

DAB The diffusivity of A in B ft/hr.

CA The molar concentration of A.

y The distance in the direction of diffusion.


(Equation (1) is one form of Fick’s low where NA refer to the heat flux
( q ) and D is analogous to the conductivity and concentration gradient (dC/dy)
to temperature gradient (dT/dy).

If the total pressure and hence the total molar concentration is every where
constant, (dCA/dy) and (dCB/dy) must be equal and opposite and therefore A
and B tend to diffuse in opposite directions.

In a distillation processes, where the two components have equal molar latent
heats. Condensation of one mole of the less volatile material releases just
sufficient heat for the vaporization of one mole of the more volatile component
and therefore equimolecular counter-diffusion take place with two components
diffusing at equal and opposite rate as determined by Fick’s law.

In many process B will neither remain stationary nor will it diffuse at an equal
and opposite molar rate to A. Exact calculations relating to this type of
problem are difficult. An example of this nature is the mass transfer in a
distillation column when the two components have unequal molar latent heat.
When the fluid is turbulent, Eddy diffusion takes place in addition to molecular
diffusion on the rate of diffusion is increased, so

5
Prof.Fatma K. Gad

NA = - ( DAB + ED ) * ( dCA / dy ) (2)

Where ED is known as the eddy diffusivity which will increase as the turbulence
is increased.

DIFFUSION IN THE GAS PHASE

1. Equimolecular counter diffusion

Suppose two vapors A and B are diffusing at equal and opposite rates and P A
and PB are their partial pressures at any point in the system.

If A and B are ideal gas

PAV = nART & PBV = nBRT

Where nA and nB are the number of moles of A and B in volume V so that.

PA = ( nA / V ) RT = CART = ( cA / MA ) RT (3)

PB = ( nB / V ) RT = CBRT = ( cB / MB ) RT

Where CA = PA / RT (4)

CA mole concentration and CA mass concentration MA, MB are molecular


weights.
If total pressure ( P ) is constant at the system; therefore,

P = PA + PB = RT [CA + CB ] = RT [ CA /MA + CB / MB ]( 5 )

If CT = CA + CB. Therefore,

6
Prof.Fatma K. Gad

P = RT CT (6)

Where CT = [ CA + CB ] is the sum of the molar concentrations thus,

dPA / dy = - dPB / dy

dCA / dy = - dCB / dy
(7)
and

dCA / dy = - ( dCB / dy ) * ( MA / MB )

From equations (6) and equation (1) therefore,

NA = - ( DAB / RT ) * ( dPA / dy )

and NB = - ( DBA / RT ) * ( dPB / dy )

In equimolecular counter diffusion, the total pressure at any point in the


system will remain constant as a result of diffusion and therefore, equal
number of molecules will diffuse in each direction.

Where NA = -NB

So that DAB = DBA = D ( from equation 7 )

NA = - D * ( dCA / dy ) = - ( D / RT ) * ( dPA / dy ) ( 8 )

7
Prof.Fatma K. Gad

and NB = +D * ( dCA / dy ) = + ( D / RT ) * ( dPA / dy )

The above equations which give the transfer rate of a point value of a
concentration or partial pressure gradient can be integrated directly. Thus if
the conditions at two different points in the system are denoted by the
suffixes 1 and 2, integration of equation (8) gives.

NA = - D * ( CA2 – CA1 ) / ( y2 – y1 ) (9)

NA = - (D / RT) * ( PA2 – PA1 ) / ( y2 - y1 ) ( 9’)

II Diffusion through a stationary gas:

If a surface is introduced on which A is absorbed but B in not, a partial


pressure gradient will be set up, causing A to diffuse towards and B away
from the surface. As a result a total pressure gradient will be produced
causing a bulk motion of A and B towards the surface, in addition to the
transfer by diffusion. Since there is no net motion of B, the bulk rate of flow
must exactly balance its transfer by diffusion.

Thus the bulk rate of flow of B = - NB

NB = - D (dCA / dy)

The bulk flow of B is accompanied by bulk flow of A and since the rate of
the number of molecules of A at a point to those of B is CA/CB

Surface absorbing A Diffusion A


Bulk flow of A
Diffusion of B
Bulk flow of B

Fig. (1) : Mass transfer through stationary gas.

Bulk flow of NA = - NB ( CA / CB )

8
Prof.Fatma K. Gad

NA = - D * ( dCA / dy ) * ( CA / ( CT - CA )) ( 9’’ )

The total rate of transfer of A is obtained by assuming the transfer by


diffusion and by bulk flow.

Thus, by adding equation 8 and 9’’ the total transfer N’A is given by:

N’A = - D *( dCA / dy ) - D * ( dCA / dy ) * ( CA / ( CT - CA ))

N’A = - D * ( dCA / dy ) * ( CT / ( CT - CA ))

N’A = - D * ( CT / CB ) * ( dCA / dy ) (10)

The previous relation is known as Stefan’s law. Integration of equation


(10) between low positions 1,2.

N’A = + DCT / (y2 - y1) * Ln ((CT – CA2 ) / (CT - CA1)) (11)

N’A = - D / ( y2 - y1 ) * ( CT / ( CT - CA )m ) * ( CA2 - CA1 )

N’A = - D / ( y2 - y1 ) * ( CT / CB)m *( CA2 – CA1 ) (11’)

N’A = - D / (RT (y2 - y1))*(P / PBM) * (PA2 - PA1) (11’’)

PBM is the log mean of PB1 and PB2 so,

PBM = ( PB2 - PB1 ) / Ln ( PB2 / PB1 )

9
Prof.Fatma K. Gad

Equation (11) can be simplified when the concentration of the diffusing


component A is small. Under these conditions C A small compared with CT.
Equation (11) gives.

N’A = D CT / ( y2 - y1 ) * Ln [ 1 - ( CA2 - CA1 ) / ( CT - CA1))]

N’A = D CT / (y2-y1)*[-(CA2-CA1)/(CT-CA1))–½(CA2-CA1)/(CT-CA1)2]

For small valves of CA this reduces to

N’A  - D / ( y2 - y1 ) * ( CA2 - CA )

This is identical to equation (9) for equimolecular counter diffusion. Thus


the effects of bulk flow can be neglected at low concentrations

Comparison of rates of mass transfer in equimolecular counter diffusion and


in diffusion through a stationary gas:-

For equimolecular counter diffusion from equation (9)


NA = - D [( CA2 - CA1 )/( y2-y1 )] = + kL [ CA1 - CA2 ] = + kG [ PA1 – PA2 ]

Where kG = D / ( RT (y2 - y1)) & kL = (D / (y2 - y1)) (12)

Are mass transfer coefficient for equimolecular counter diffusion.

For diffusion through a stationary gas B from equation (T1)

N’A = -D (( CA - CA1 ) / ( y2 - y1 )) * ( CT / CBM ) = k’L(CA1-CA2)

N’A = k’G ( PA1 – PA2 )

Where

10
Prof.Fatma K. Gad

kG = D / ( RT ( y2 - y1 )) * ( PT / PBM )
(13)
kL = D / ( y2 - y1 ) * ( CT / CBM )

are mass transfer coefficient for diffusion through a stationary gas:-

Thus stationary equimolecular

N’A / NA = CT / CBM = k’G / (14)


kG

Gas diffusivities:
The value of gas diffusivity (D) for the transfer one gas in one there may not
be known and experiment determination may not be practicable. Many
attempts have been made to express D in terms of other physical properties
and the following empirical equation of Maxwell’s , modified by Gilliland,
gives satisfactory agreement with the observed figures;

D = [4.3*T1.5*((1/MA)+(1/MB))0.5]/[P*(VA1/3+VB1/3)] (15)

Where D is the diffusivity in Cm2/sec


T absolute temp (oK)
MA & MB the molecular weight of A and B
P Total pressure in (N/m2)
VA / VB the molecular volumes of A and B.
[the vol. in m3 of 1 kmole of the material in the form of liquid at its B.P]

The diffusivity of component A in a mixture of gas B and C can be


calculated from the following equation.

D = (1-yA) / (yA/DAA+yB/DAB+yC/DAC ..etc ) (16)

when yA, yB, yC are mole fraction of A, B and C components.

11
Prof.Fatma K. Gad

DAA. DAb and DAC calculated through equation (15)

Example (1)

Calculate the diffusivity of CO2 in a gas mixture (20% (wt %) CO2) 70 %


air, and 10% H2O vapor] assume T = 20 oC and pressure = 1 atm.
where VCO2 ,Vair and VH2O are 22.2, 29.9 and 33.3 respectively.

(0.108 cm2/sec).

Example (2)

Ammonia is absorbed in water from a mixture with an using a column


operating at 101.3 KN/m2 and 2.95 oK. The resistance to transfer can be
regarded as lying entirely within the gas phase. At a point in the column, the
partial pressure of the ammonia is 7.0 kN/m2. The back pressure at the water
interface is negligible and the resistance to transfer can be regarded as lying
in a stationary gas filter 1 mm thick. If the diffusivity of ammonia in air is
2.36 x 10-5 m2/s; What is the transfer rate per unit area at that point in the
column? How would the rate of transfer be affected if the ammonia air
mixture were compressed to double the pressure?

Solution :

concentration of ammonia in the gas (C1) = ( P1 / P2 ) ( T2 / T1 ) (N2 / V2 )

C1= (1/22.4) (7/101.3) ( 273/295) - 0.00285 k mol / m3

P/PBM = CT / CBM = 101.3 Ln( 101.3/ 94.3 ) / Ln ( 101.3-94.3) = 1.036

PB2 = ( 101.3 - 0 ) PB1 = ( 101.3 - 7 )

NA = -[ D / ( y2 - y1 )] ( CT / CBM ) ( CA2 - CA1 )

= - ( 2.36 * 10-5 / 1*10-3 ) * 1.036 * .00285= 6.97*10-5 kmol/m2s

If the pressure is increased to 202.6 kg/m2 the driving force is doubled,


CT/CBM is essentially unaltered, and the diffusivity being inversely
proportional to the pressure equation (15).

The mass transfer rate therefore remain the same.

12
Prof.Fatma K. Gad

Atm = 1.0333 kg/cm2 = 1.0333 x 104 kg/m2.


Newton = Kg m/sec2 = 10 dyne (gm cm/sec)
Atm = (101.33) K. NEWTON (KN) / m2

DIFFUSION IN THE LIQUID PHASE

The rate of diffusion of a material in the liquid phase is represented by the same
basic equation as for gas phase diffusion, i.e.

NA = -DL (CA/dy)

Where DL is the liquid phase diffusivity.

Simple integrated forms are available only if the total concentration (CT) and
(DL) are constant.
The integrated form of the above equation for equimolecular counter diffusion
of a material A is then of the same as for gaseous diffusion,

NA = - DL ( CA2-CA1 ) / ( y2 – y1 ) (17)

Where, CA2, CA1 are the molar concentrations.


For diffusion through a stagnant liquid B

NA = -DL ( CA2 - CA1 ) / ( y2 – y1 ) CT / CBM (18)

An empirical equation from which liquid phase diffusivities in dilute solutions


can be calculated approximately is,

DL = 7.7 X 10-16 T / (  (V1/3 – Vo 1/3)) (19)

Where DL is the diffusivity in m2/sec


T is the temperature in oK
 is the viscosity in N s / m2
V is the molecular volumes

13
Prof.Fatma K. Gad

Equation (19) do not apply to electrolytes or to concentrated solutions.

Transfer of material between phases


The mass transfer rate between two fluid phases will depend on the physical
properties of the two phases, the concentration difference, the interfacial area
and the degree of turbulence.

A number of mechanisms have been suggested to represent conditions in the


region of the phase boundary. The earliest of this is the two film theory
propounded by Whitman.

The two film theory

Since the solute is diffusing from the gas phase into the liquid, that must be
concentrate gradient in the direction of diffusion within each phase.

In the gas phase the laminar diffusion resistance which concentrate gradient
overcomes lines lay within a thin laminar film at the interface between the two
phases. The film resistance is the same as the true molecular and eddy diffusion
resistance actually present.

A similar situation exist, while the liquid phase from the following figure.

Fig. (2) Two film theory.

The average concentration of A in the gas phase is P AG and this falls to PAi at
the interface due to the effective film thickness YG.

14
Prof.Fatma K. Gad

In the liquid phase the concentration falls from G Ai at the interface to GAL the
average concentration in the mean bulk of the liquid..

The concentrations at the interface are in equilibrium, and the resistance to


transfer is centered in the thin films on either sides. There is no resistance to
the transfer of the diffusing solute occurs in the interface itself.

In a steady state conditions, the rate of transfer of material through the first film
will be the same as the through the second film.

i.e. NA = (DP / R T ∆YG) * [ (PAG - PAi) / PBM ] (20)

NA = (DLCT / ∆YL CBM) * ( CAi - CAL ) (21)

Since the film thickness is rarely known, the putting.

kG = DP / R T ∆YG PBM and KL = DL C / ∆YL CBM (22)

Where, kG & kL are the gas film and liquid film transfer coefficients. They are
direct measure of the rate of transfer permit area of interface with a driving face
of unity.

Since NA = kG (PAG – PAi ) = kL ( CA – CAL ) (23)

kG / kL = ( CAi – CAL ) / ( PAG - PAi ) (24)

These conditions may be illustrated graphically as shown below. Where, ABF


is equilibrium curve. Point D(CL & PG) represent condition Ce in the liquid in
equilibrium with PG. Point A (Ce , PG) represent a concentration CL in the
liquid in equilibrium with (PG). Point B[Ci & Pi] represent the concentration
Pi in gas Ci in liquid in equilibrium at interfaces.

15
Prof.Fatma K. Gad

Fig. (3) Graphical illustration for equation 24.

Point F (CL , Pe) represent Pe in gas in equilibrium with CL in liquid.

Then the driving force coming transfer in the gas phase,= (PG – Pi) = DE.
and the driving force causing transfer in the liquid phase = (C i-CL) = BE.
Then PG – Pi/Ci – CL = kL/kG
And the concentrates at the interface (point B) are found by drawing a line
through D of slop; - kG /kL.

To cut the equilibrium curve in B. However the above relation has little
practical utility since it is necessary to know both, kG & kL as well as the
equilibrium relation ship between Pi and Ci in order to calculate NA.

It is convenient therefore to employ overall coefficient KL , KG based on the


overall driving force (P-Pe) and (Ce – CL) where Pe is the P.P of solute over a
solution of composition CL and Ce is the concentration of solution in
equilibrium with the solute partial pressure PG in the gas. Thus

NA = KG (PG – Pe)
(25)
=KL (Ce – CL)

RELATION BETWEEN FILM AND OVERALL COEFFICIENT

NA = kG ( PG - Pi ) = kL ( Ci – CL ) = KG ( PG – Pe ) = KL ( Ce - CL )

16
Prof.Fatma K. Gad

1/KG = 1/kG [(PG-Pe)/(PG-Pi)]

= 1/kG [(PG-Pi)/(PG-Pi)]+1/kG[Pi-Pe)/(PG-Pi)]

but 1/kG = 1/kL [( PG - Pi ) / ( Ci - CL )]

1/KG = 1/kG + 1/kL [ ( PG - Pi ) / ( Ci - CL ) ] [ ( Pi - Pe ) / ( PG - Pi ) ]

=1/kG + 1/kL [ ( Pi - Pe ) / ( Ci - CL ) ]

but (Pi - Pe) / (Ci - CL)is the average slope, slope of equilibrium curve and when
solution obeys Henry’s law, we have

H = dp / dc = ( Pi - Pe ) / ( Ci - CL )

So 1/KG = 1/kG + H/kL

Also 1/KL = 1/kL + 1/HkG (26)

And 1/KG = H/KL

Now for very soluble gas, the equilibrium curve lines close to the C axis and
point E approaches very close to point F. The driving force over the gas film
DE is then approximately equal to the overall driving force DF. So that k G is
approximately equal to KG.

If the gas is sparingly soluble, the equilibrium curve rises very steeply to that
the choices force EB. In the liquid film because approximately equal to the
overall driving force AD.

So that kL is approximately equal to KL. If the gas is moderately soluble, then


the film offer an appreciable resistance and the point B at the interface must be

17
Prof.Fatma K. Gad

located by drawing a line through D of slope

- kL / kG = ( PG – Pi ) / ( Ci - CL )

Examples:-

A sample rectifying column consists of a tube, arranged vertically and supplied


at the bottom with a mixture of benzene and toluene as vapor. At the top a
condenser returns some of the product as a reflux which flows in a thin film
down the inner wall of the tube. The tube is insolated and heat losses can be
neglected. At once point in the column, the vapor contains 70 mol% B and the
adjacent liquid reflux contains 59 mol% B. The temperature at this point is
365k. Assuming the diffusion resistance to vapor layer 0.2 mm thick, calculate
the rate of interchange of benzene (B) and toluene (T) between vapor and
liquid. The molar latent heat of the two components can be taken as equal. The
vapor pressure of toluene at 365k is 54.o KN/m2 and the diffusivity of the
vapors is 0.051 cm2/sec.

Solution:

Let subscripts 1 and 2 refer to the liquid surface and vapor side of the stagnant
layer respectively. The example of equimolecular counter diffusion where,

NA = - D / R T X ( PA2 - PA1 )

So for Toluene Vapor pressure at 365 oK = 54 KN/m2

From Raoult’s law;

P.P = (V.P) X

(PT1) P.PT1 = (1-59) * 54 = 22-14 KN/m2

PT2 = (1-0.70) * 101.3 = 30-39 KN/m2

NT = -(0.51 * 10-4) (30-39-22.14)/8.314 * 365 * 0.0002 = -6.93 x 10-5 kn/m2

For Benzene:-

18
Prof.Fatma K. Gad

PB1 = 101.3 – 22.14 = 79.16 KN/m2

PB2 = 101.3 – 30.39 = 70.91 KN/m2

NB = - (0.051 * 10-4) (70.91-79.16)/8.314 * 365 * 0.0002


= 6.93 * 10-5 Kmol/m2s

Thus the rate of interchange of B and T is equal but opposite in direction.

Example II.

A tube of small diameter was filled with acetone to 1.1 cm from the top and
maintained at a temperature of 20 oC, a gently current if air be supplied.
After 5 hr the level of acetone had fallen to 2.05 cm from top.

Total pressure = 750 mm Hg , V.P of acetone = 180 mm Hg


at 20o C, L = 0.79 gm/cm3 , M. wt = 58.
Calculate diffusivity of acetone ( R=82.06 cm3 atm / g mol ok )

Solution:-

NA = [ DP / ( R T Y )] Ln ( PB2 / PB1 ) (1)

NA = (ρL / Mwt) dL / d
= (L / Mwt) dL/d = D/RT (P/PBm) (P/L)

from (1) D = (/Mwt) (R T/P).ln(PB2/PB1) ∫L(dL/d

=(L/ 2 Mwt) (R T / P) . Ln ( PB2 / PB1 ) ((L22 – L12) /) cm2/sec

=[82.06*293*0.79/(58*(750/760)*Ln750/(750-180)]*(2.052)-
(1.12)/2*5*3600

= 0.101 cm2/sec

19
Prof.Fatma K. Gad

Example III.

Calculate the rate of diffusion of acetic acid (A) across a film of water (B)
solution oil an thick at 17oC, when the concentration on apposite sides of the
film are respectively 9 and 3 wt %. The diffusivity of A in the solution may
be taken as 0.95 x 10-5 cm2/sec. [M wt of CH3 COOH = 60)  of 9% solution
at 17oC = 1.012 gm/cm3

 of 3% solution at 17oC = 1.003 gm/cm3

Solution:-

Since concentration are given in wt % one must convert into mole/cm2.

Concentration in mole/cm3 = wt % (gm acid/gm sol) * P of sol (m/m3) * 1/m


wt of acid)

9% Acetic acid sot = 0.09 * 1.019/60 = 0.001515 gm mole / cm3

3% Acidic acid sot = 0.03 * 1.003/60 = 0.000595.

CA1 = 0.001515 gmole/cm3

CA2 = 0.000505 gmole/cm3

CB1 = 0.91 * 1.012/18 = 0.0455 gmole/cm3

CB2 = 0.47 * 1.003 / 18 = 0.0540 gmole/cm3

C1 = CA1 + CB1 = 0.001515 + 0.0455 = 0.047 mole/cm3

C2 = CA2 + CB2 = 0.000505 + 0.054 = 0.0545 mole/cm3

Ctotal = ( C1 + C2 ) / 2 = (0.047+0.0543)/2 = 0.0508 mole/cm3


(but for gasses Ptotal = PA1 + PB1 = PA2 + PB2

CBm =( CB2 – CB1 ) / Ln ( CB2 / CB1 )


=0.054–0.0455/Ln0.054/0.0455=0.0497 gmole/cm3

NA = ( DL / Y ) * ( Ctotal / CBm ) * ( CA1 - CA2 )

=0.95 * 10-5/0.1 * 0.0497 0.058 (0.001515 – 0.000505)

20
Prof.Fatma K. Gad

= 9.85 * 10-8 gmole/sec. cm2

GAS OBSORPTION
This chapter deals with the mass transfer operations known as gas absorption
and stripping or desorption.

In gas absorption a soluble vapor is absorbed by means of a liquid in which the


solute gas is more or less soluble, from its mixture with an inert gas. The
washing of ammonia from a mixture of ammonia and air by means of liquid
water is a typical example.

The solute is recovered from the liquid by distillation, and the absorbing liquid
can be either discarded or re used. Sometimes a solute is removed from a liquid
by bringing the liquid into contact with an inert gas, such an operation, the
reverse of gas absorption is desorption or gas stripping.

Design of packed towers:

A common apparatus used in gas absorption and certain other operations is the
packed tower, Fig (4).

The device consists of a cylindrical column, or tower equipped with a gas inlet
and distributing space at the bottom; a liquid inlet and distributor at the top,
liquid and gas outlets at the bottom and top, respectively; and a supported mass
of inert solid shapes, called tower packing or tower filling. The inlet liquid may
be pure solvent or a dilute solution of solute in the solvent and is called the
weak liquor, is distributed over the top of the packing by distributor and in ideal
operation, uniformly wets the surfaces of the packing.

The solute containing gas, or rich gas, enters the distributing space below the
packing and flows upward through the interstices in the packing counter current
to the flow of the liquid. The packing provides a large area of contact between
the liquid and gas and encourages intimate contact between the phases. The
solute in the rich gas is absorbed by the fresh liquid entering the tower, and
dilute or lean gas leaves the top. The liquid is enriched in solute as it flows
down the tower, and concentrated liquid, called strong liquor, leaves the bottom
of the tower through the liquid outlet.

Many kinds of tower packing have been invented as shown in Fig (5). Packing
are divided into those which are (I) dumped at random into the tower and those

21
Prof.Fatma K. Gad

which must be (II) stacked by hand.

I) Dumped packing consist of units ¼ to 2 inch and are used most in the
smaller columns and give high contact between liquid and gas.

II) Tthe units in stacked packing are 2 to about 8 inch in size, they are used
only in the larger towers.

The principal requirements of a tower packing are:-

1) It must be chemically inert to the fluids in the towers.


2) It must be strong without excessive weight.
3) It must contain adequate passages for both streams without excessive liquid
holdup or pressure drop.
4) It must provide good contact between liquid and gas.

Fig (1) Tower packing; a)Beel saddle b)Intalox saddle


c)Rashing ring d)Lassing ring e)Pall ring
f)Tellerette g)Double spiral ring h)Triple spiral ring.

Thus most towers packing are made of cheap, inert light materials such as, clay,
porcelain, or graphite. Thin walled rings of steel or aluminum are some times
used. High void spaces and large passage for the fluids are achieved by making
the packing unit irregular or hollow, so that they interlock into open structures
w a porosity of 60% or more.

Contact between liquid and gas:

The requirement of good contact between liquid and gas is the hardest to meet,
especially in large towers. Ideally the liquid once distributed over the top of the
packing, flows in thin films all over the packing surface all the way down the

22
Prof.Fatma K. Gad

tower. Actually the film tends to grow thicker in some places and thinner in
others, so that the liquid collects into small rivulets and flows along localized
paths through the packing. Especially at low liquid rates much of the packing
surface may be dry or, at best, covered by a stagnant film of liquid. This effect
is known as Channeling, it is the chief reason for the poor performance of large
pack towers.

Channeling is most serve in towers packed with stacked packing, less severe in
dumped packing of crushed solid, and least severe in dumped packing of
regular units such as rings.

In towers of moderate size channeling can be minimized by having the diameter


of the tower at least 8 times the packing diameter. If the ratio of tower diameter
to packing diameter is less than 8:1, the liquid tend to flow out of the packing
and down the walls of the column.

Even in small towers filled with the packing that meet this requirement,
however, liquid distribution and channeling have a major effect on column
performance.

In toll towers filled with large packing the effect of channeling may be
pronounced, and distributors for the liquid are normally included every 10 or 15
ft in the packed section.

At low liquid rates, regardless of the initial liquid distribution, much of the
packing surface is not wetted by the flowing liquid. As the liquid rises, the
wetted fraction of the packing surface increases until at a critical liquid rate, is
usually high, the entire packing surface becomes wetted and effective. At
liquid rates higher than the critical the effect of channeling is not important.

Limiting flow rates, loading and flooding:-

In a tower with a definite flow of liquid, there is an upper limit to the rate of gas
flow. The gas velocity corresponding to this limits called the flooding velocity.

Three factor effects on flooding velocity:-

a) Pressure drop through the bed of packing.


b) Gas flow rate.

23
Prof.Fatma K. Gad

c) Hold up liquid and the visual appearance of the packing.

Fig (6) shows the relation between pressure drop and gas flow rate in a packed
tower. The pressure drop per unit packing depth comes from liquid friction; it
is plotted on logarithmic coordinates against the gas flow rate G y expressed in
mass of gas per hour per unit of cross sectional area, base on the empty tower.

Gy is therefore, related to the superficial gas velocity by the equation;

Gy = Vsy y (1)

Where

y is the density of the gas.


Vsy superficial gas velocity.

When the packing is dry the line so obtained is straight and has a slope of about
1.8.

If the packing is irrigated with a constant flow of liquid, at a superficial mass


velocity Gx the relation between pressure drop and gas flow rate follows a line
such as cbde in fig (3). At low and moderate velocity, the pressure drop is
proportional to the 1.8the power of the flow rate but is greater than that in dry
packing at the same gas velocity. A gas velocity increases, the line curves
upward starting at point C in fig (3). Then at higher velocity, as shown by the
line cde, the pressure drop increases sharply at nearly constant gas velocity.

Fig. (2) Pressure drop in packed tower © loading point (e) flooding point.

The rise may follow a smooth curve, as shown by the solid line cd, or may show

24
Prof.Fatma K. Gad

sharp breaks at point C and d as shown by the dotted line. As the pressure drop
increases along line bc, the amount of liquid held in the packing is constant and
independent of the gas velocity. The liquid moves downward through the
packing uninfluenced by the motion of the gas. At point C, called the
LOADING POINT, the gas flow begins to impede the downward motion of the
liquid. Local accumulations of liquid appear here and there in the packing.

As the gas rate rise further, the liquid holdup increases and the pressure drop
changes along line cde, increasing more rapidly with gas velocity than before.

At point e the FLOODING POINT, the top of the packing is covered with a
layer of liquid through with bubbles of gas issue. The liquid can no longer flow
down through the packing, and the layer grows until liquid is blown out of the
top of the tower with the gas.

The gas velocity in an operating tower must obviously be lower than the
velocity which will cause flooding. How much lower is a choice to be made by
the designer. The lower the velocity the lower the cost of power and the larger
the tower.

The higher the gas velocity, the larger the power cost and the smaller the tower.

Economically, the most favorable gas velocity depends on a balance between


the cost of the power and the fixed charges on the experiment.

The optimum velocity can be estimated, but it is usually taken about one-half
the flooding velocity.

Packed towers are also commonly designed on the basis of a definite pressure
drop per unit height of packing. For absorption towers the design value is
usually between 0.25 and 0.5 inch H2O per foot of packing, for distillation
columns it is in the range 0.5 to 0.8 inch H 2O per foot. In most towers packed
with rings or saddles loading usually beings at a pressure drop of about 0.5 inch
H2O per foot, and FLOODING OCCURS at a pressure drop between 2 and 3
inch H2O per foot.

Fig (7) gives corrections for estimating flooding velocities and pressure drops
in packed towers, it consists of a logarithmic plot of

Gy2 Fp [ 62.3 / x ] x0.2 / ( gc x y ) v.s ( Gx / Gy ) * ( y / x )0.5

Where,

25
Prof.Fatma K. Gad

Gx mass velocity of liquid lb/Ft2 sec


Gy mass velocity of gas lb/ft2 sec
Fp Packing factor Ft-1 ft-1
Px Density of liquid lb/ft3
Py Density of gas lb/ft3
x Viscosity of liquid cp.
gc Newton’s law proportionality factor, 32.174 ft/lb, S 2

Fig. (3) Generalized correlation for flooding and pressure drop in packed
column after Eckart.

Example (1)

A tower packed with 1 inch (25.4 mm) ceramic Rasching rings is to be built to
treat 25000ft3 (708m3) of entering gas/hrs (air). The ammonia content of the
entering gas is 2% by volume. Ammonia free water is used as absorbent. The
temp 20°C and pressure 1 atm. The ratio of gas flow to liquid flow is 1 lb of

26
Prof.Fatma K. Gad

gas per pound of liquid.

a) If the gas velocity is to be one half the flooding velocity, what should be the
diameter of the tower?

b) What is the pressure drop if the packed section is 20 ft (6.1m) high?

Solution:

The average M.wt of the entering gas = 29x0.98+0.02x17 = 28.76

y = 28.76 x 492/359(460+68) = 0,07465 lb ft 3

x = 62.3 lb/ft3

x = 1 cp

For 1 inch ceramic rings, Fp = 155 ft-1 (from tables)

.. ( Gx / Gy ) * ( y / x )0.5 = ( 0.07465/62.3 )0.5 = 0.0346

from fig (4)

Gy2 Fp ( 62.3 / x ) x0.2 / gc x y = 0.19 at flooding velocity

The mass velocity at flooding is

Gy = 0.19 * 32.174*0.07465*62.3/155*1 0.2 * 1 = 0.428 lb / ft2 S

a) The total gas flow is 25000 x 0.07465/3600 = 0.518 lb/S. If the actual
velocity is one – half the flooding velocity. The cross sectional area S of
the tower is

S = 0.518/(0.42812) = 2.42 ft2

The diameter of the tower is 2.42/0.7854 = 1.76 ft (53.6 cm).

b) At the flooding velocity, Gy = 0.428/2 = 0.214 lb/ft2S


Gx is unchanged and the abscissa value in Fig (4) = 0.0692. The ordinate
becomes 0.19/4 = 0.0475. For these conditions the pressure drop is about
0.45 inch H2O per foot of packed height; the total pressure drop is 20 * 0.45

27
Prof.Fatma K. Gad

=9 inch H2O. U.E.D

Principle of absorption:-

As shown in the previous section, the diameter of packed absorption tower


depends on;

1) The quantities of gas and liquid handled.


2) Their properties
3) The ratio of one stream to the other.

Where the height of the tower and hence the total volume of packing depends
on;
1) The rate of mass transfer per unit of packed volume.
2) The magnitude of the desired concentration change.

Therefore, the calculation of the tower height rest on:

A) Material balances
B) Enthalpy balances.
C) Estimates of driving force and mass transfer coefficients.

(A) Material Balances


In a differential contact plant such as the packed absorption tower, where there
is no sudden discrete change in composition. As it is known that in an
absorption process, the solute (A) in the gas phase is transferred to the liquid
phase,
Consequently a material balance for the solute “A” for a zone from bottom to
any section;

The solute leaving gas = solute entering liquid (2)

So, material balance over a differential section, where the flow rate of the two
phases through a differential length of column G, L, x and y.

d(Gy) = d (Lx) = dNA (3)

Where NA is the rate of transfer of component A in moles per hour.

28
Prof.Fatma K. Gad

At low concentration [(yA1  10%)] L and G nearly constant at any section in


the column, so

G(yA1 – yA) = L (xA1 – xA) (4)

Where

G moles of gas entering the column


L moles of liquid
yA1 xA1 mole fraction of A in gas and liquid at the bottom
yA xA mole fraction of A in gas and liquid at any section

Fig (4) Vapor – liquid absorption contactor

Total material balance where GL constant

G (yA1 – yA2) = L (xA1 – xA2) (5)

Where yA2 , xA2 mole fraction of component A at the top of the tower for gas
and liquid. The slope of equation 5

L/G= ( yA1 – yA2 ) / ( xA1-xA2 ) (6)

This equation is true only when G , L do not change from one section to another

29
Prof.Fatma K. Gad

Fig. (5) limiting gas-liquid ratio

However, in absorption operations the flow rates of liquid and gas L , G usually
change appreciably form one point to another in the column and the operating
lines are often strongly curved specially at high concentration of solute (A)

From the figure, it is clear that for a given gas flow, a reduction in liquid flow
decreases the slope of the operating line. Consequently the maximum possible
liquor concentration and the minimum possible liquid rate are obtained when
the operating line just touches the equilibrium line as show above.

At any actual tower, the liquid rate must be greater than this minimum if the
tower to be operated.

The limiting liquid–gas ratio (L/G)min can be computed by setting x A1 = x*


where x* is in equilibrium with yA1.

(L/G) ratio is important in economics of absorption in a counter current tower.

If the liquid gas ratio is large, the average distance between the operating and
equilibrium line is also large, the concentration difference is favorable through
out the tower and the tower is short.

If the solute gas is to be recovered, however the cost of recovery is high

30
Prof.Fatma K. Gad

because of the dilution of the strong liquor.

If on the other hand concentration costs are reduced by using less liquid, the
driving forces in the absorber are reduced and the tower become taller and more
expensive.

The optimum liquid rate is found by balancing recovery costs against fixed
costs of the equipment.

In general, the operating line should be approximately parallel to the


equilibrium line for economical operation.

So at the bottom of tower the driving force is in gas phase.

At bottom driving force = yA1 – y*A1 = yA1


At top driving force = yA2 – y*A2 = yA2

For all the tower

(y)Lm = (yA1 – yA2) / Ln ( yA1 / yA2 ) (7)

Where
y*A1 is in equilibrium with xA1
y*A2 is in equilibrium with xA2 as shown

Fig (6) absorption process

31
Prof.Fatma K. Gad

Since at low concentration where Henery law apply

y*A1 = m xA1 , y*A2 = m xA2

Where m Henrry’s constant.

Example 2:-

By mean of a plate column, Acetone is absorbed from its mixture with air in a
non-volatile absorption oil. The entering gas contains 30 mole percent acetone,
and the entering oil is acetone free. The acetone in the air 97% is to be
absorbed and the concentrated liquid at the bottom of the tower is to contain
10% acetone. Calculate the quantity of liquid (oil) used and plot the operating
line.

Solution:

to concentration of acetone in the leaving air is obtained by an acetone balance.


Since the concentration in the entering gas is more that 10% so that there is a
change in G and L across the column and consequently are shall use mole ratio
instead mole fraction as we shall see.

Basis of 100 mole of entering are them;

Gm (inert air without acetone) = 1-0.30) 100 = 0.70 * 100 = 70 mole

The acetone in the entering air (yA1) = 0.30 * 100 = 30 mole

Acetone leaves with the gas = 0.03 * 30 = 0.9 mole

 (yA2) = (0.9/0.9 + 70) = 0.0127 (mole fraction)

Acetone absorbed by oil = 30 –0.9 – 29.1 mole

 Quantity of oil needed to form 10% acetone = 29.1 * 90 /10


= 261.9 mole ………..(1)

So far the equation of the operating line

32
Prof.Fatma K. Gad

70 (0.0127/0.2127 – yn+1/1-yn+1) = 261.9 (0 – xn/1-xn)

The operating line passes through points (0) 0.0127) and (0.1 > 0.3).

Plot the line by chose value of x and calculate the values of y from the above
equation

When x 0.03 0.05 0.08


y 0.114 0.173 0.253 ……..(2) U.E.D

Temperature variation in packed towers:-

When rich gas is fed to an absorption tower, the temperature in the tower varies
appreciably from bottom to top. This temperature gradient affects the shape of
the equilibrium line. The rate of absorption is large at the gas inlet, and the heat
of condensation and solution of the absorbed constituent may be sufficient to
increase the temperature of the liquid considerably.

Since the partial pressure of the absorbed component increases with increase
the temperature, the concentration of the vapor in equilibrium of definite
composition also increases with temperature. Even if Henry’s law applies, the
equilibrium line may be strongly curved in this situation.

The effect on the equilibrium line of a temperature gradient in a tower is shown


in Fig (11). Line “OA” is the equilibrium line for isothermal operation, and line
“OB” is that where the temp at the bottom of the tower is larger than at the top.

33
Prof.Fatma K. Gad

Fig (7) effect of temperature gradient on equilibrium line.

If the temperature effect is sufficiently great, the equilibrium line may intersect
the operating line, and the process is in operative near the bottom of the tower.
Some times cooling coils or other cooling means are installed in the tower to
reduce this heat effect.

Evaluation of the height of packed tower:-

The height of a packed tower depends on the rate of absorption which in turn is
affected by the rate of transfer of mass through the liquid and gas phases. The
following treatment has two limitations;

1) No chemical relations occur between the absorbed component and


liquid.
2) Heat of solution is neglected.

Since the rate of solute (A) transfer from gas phase to liquid phase pass through
two diffusion resistances:

a) Gas resistance
b) liquid resistance.

Assuming that the resistance in the interface itself is Zero; consequently the gas
and liquid at the interface are in equilibrium.

(A) Height of column based on conditions in Gas film

To illustrate the condition occur during the steady operation of a counter

34
Prof.Fatma K. Gad

current gas liquid absorption tower. It is convenient to express the


concentration of the streams in term of moles of solute gas per mol of inert gas
in the gas phase, and as moles of solute gas per mol of solute free liquid in the
liquid phase as will seen later.

The actual area of interface between the two phases is not known, and the term
“a” is introduced as the interfacial area per unit volume of the column.

On this basis the general equation for mass transfer can be written as;

AGm dY = ALm dX (8)

From mass transfer section;

A Gm dY = NA ad(V) = kG a(PG – Pi) Adz (9)

since PG = ( Y / ( 1 + Y )) *  (Dalton’s law)


Where
NA mol of solute absorber / unit time and unit interfacial area.
a Surface area of interface per unit volume of column.
A Cross sectional area of column.
Z height of packed section.
Gm mole of inert gas /(unit time) (unit cross section)
Y Moles of solute gas A/moles of inert gas B in gas phase.
X moles of solute A / moles of inert solvent in liquid phase.
Consider any plane at which the molar ratios of the diffusing material in the gas
and liquid phases are Y and X. Then over a small height of Z, the moles of gas
leaving the gas phase will equal the moles taken up by the liquid as shown in
equations 8 , 9.

From these two equations 8 and 9 and Daltons law;

 Gm dY = kg a  [ (( Y / ( 1 + y )) - (( Yi / ( 1 + Yi )) ] dz

Hence the height of column Z required to achieve change in Y from Y1 at the


bottom to Y2 at the top of the column is given by;

z y1

35
Prof.Fatma K. Gad

 dz = Z = Gm / kg a   ( 1 + Y ) ( 1 + Yi ) dy / ( Y - Yi ) (10)
0 y2

Which for weak mixtures can be written as:


y1
Z = G m / kg a   dy / ( Y - Yi ) (11)
y2

In this analysis it has been assumed that k g is a constant throughout the column,
and provided the concentration changes are not too large.

B) Height of column based on conditions in liquid film:-

A similar analysis may be made in terms of the liquid film, thus

A Lm dX = kL a ( Ci - CL ) Adz (12)

Where the concentrations C are in terms of moles of solute per unit volume
of liquor, where
CT = (mole of solute + solvent) / volume of liquid.
X = moles of solute/moles of solvent = C/(CT-C)

So C = ( X / ( 1 + X )) * CT (13)

Now the transfer equation may be written;

Lm dX = kL a CT [ ( Xi / ( 1 + Xi )) – ( X / ( 1 + X ))] dz

z x1
 Z = Z = Lm / kL a CT  [( 1+Xi ) ( 1+X )/( Xi - X )] dx (14)
0 x2

For dilute concentrations;

x1
Z = Lm / kL a CT  dx / ( 36Xi –X )
x2
Prof.Fatma K. Gad

(15)

Where CT and kL have been taken as constant over the column;

Y and X are mol ratio.

C) Height based on overall coefficients:-

If the driving force based on the gas, concentration is written as (Y-Ye) and
the overall gas transfer coefficient as KG

Then the height of the tower for dilute concentrations becomes.

Y1
(16)
Z = Gm / KG a   (dy) / ( Y - Ye )
Y2
And in term of liquor;

X1
Z = Lm / KL a CT  dx / ( Xe – X ) (17)
X2

where Ye Mol ration at equilibrium with X


Xe mol ration at equilibrium with Y.
Consequently for dilute concentrations; as the mol fraction is approximately
equal to the molar ratio at dilute concentrations then considering the gas
film;

Y1
Z = Gm / Kg a   dY / ( Y - Ye )
Y2
y1 (18)
Z = Gm / Kg a   dy / ( y - ye )
y2
37
Prof.Fatma K. Gad

Considering the liquid film;

X1
Z = Lm / KL a CT  dX / ( X - Xe )
X2
x1
Z = Lm / KL a CT  dx / ( x - xe ) (19)
x2

Where x and y are mole fraction


X and Y are mole ratio.

Special case when equilibrium curve is straight line:-

If over the range of concentrations considered the equilibrium curve is straight


line, it is permissible to use an average value of the driving force over the
column.

(1) For dilute concentrations:-

The absorption is given by

NA A a dz = Gm A dy = KG a A  ( y - ye ) dz (20)

if y* = ye = m x + c ;

consequently ye1 = m xi + c , ye2 = m x2 + c

So that m = ( ye2 - ye1 ) / ( x2 - x ) slope of equilibrium (21)

 Lm (x-x1) = Gm (y-y1)  x = x1 + Gm/Lm (y-y1).

from equation 20

38
Prof.Fatma K. Gad

y1 z
 dy / ( y - ye ) =  (( kG a  ) / Gm ) dz (22)
y2 0

by substituting ye in form x

y1 y1
 dy/(y-ye)=  dy/{y-m[x1+(Gm/Lm)(y-y1)] – C} (23)
y2 y2

From equations 21 , 22 and 23.

y1
dy / ( y - ye ) = ( y1 – y2 ) / ( y – ye )Lm (24)
y2

Where ( y- ye )Lm is the logarithms mean of ( y - ye ) = ( y - y* ) = y

( y - ye )Lm = [(y - ye)1 – (y - ye)2 ] / Ln ((y - ye)1 / (y - ye)2) (25)

Where

ye1 = y*1 Which is in equilibrium with x1

( y1 - ye1 ) = ( y1 – y*1 ) = y1

ye2 = y*2 which is in equilibrium with x2

( y2 – ye2 ) = ( y2 - y*2 ) = y2

As mention in equation (7) before

39
Prof.Fatma K. Gad

For dilute solutions:-

Z = ( Gm / (KG a ) ) * ( y1 - y2 ) / (y)Lm (26)

In liquid term;

Z = Lm / KL a CT ( x1 - x2 ) / (x )Lm (27)

Equation 27 can be written as follows:

Since CA = mol of A total vol. , xA = moles of A/total moles , m


(molar)density) = total mol/total vol.

 CA = ρm xA

So equation 27 can be in the form

Z= Lm / KL a Pm ( x1 – x2 ) / ( x )Lm (27’)

In term of mol ration equations 26, 27 and 27 will be the same, since mole
fraction equal mol ration.

B) For concentrated solution and straight equilibrium line:-

Z a A NA = Gm (Y1-Y2) A = KG a A   (Y-Ye)Lm Z (28)

The valve of  may be found from the relation.

Y=Y/(1+Y) (29)

hence  = 1 / ( 1 + Y ), although the value of  will change slightly over the

40
Prof.Fatma K. Gad

column, an average value will generally be sufficiently accurate.

Tower height in terms of partial pressures for high concentrations

A material balance taken between the bottom and some plane where the partial
pressure in the gas phase is Pg and the concentration in the liquid is X will give;

Lm ( X1 - X ) = Gm [ ( Pg1 / (  - Pg1 )) - ( Pg / (  - Pg )) ] (30)

Over a small height of the column dz therefore

Z Pg1
 dz = Gm  PBm dPg / [kg a (- Pg )2 ( Pg - Pi )]
(31)
0 Pg2

Pg pressure of A in balk of gas phase. Where

PBm Logarithmic mean value of PB = (  - PA )


= ( PB2 - PB1) / Lm ( PB2 / PB1 )

Kg has a mean value for the whole tower, but it is not constant in
this equation.

The transfer units:-

The group dy / ( y - ye ) is present in equation (18); has been defined by Chilton


and Colburn as the number of overall gas transfer units (NOG) since as been
known.

Gm A dy = kg a A ( y - yi ) dz

y1
so  dy / ( y - yi ) = ( kg a / G ) dz
y2

41
Prof.Fatma K. Gad

Where kg is taken as constant over the column.

The group of the left hand side of this equation represents the integrated ratio of
the change in composition to the driving force tending to bring this about. This
group has been defined as the number of transfer units N G. The quantity (G / kg
a ) which is the reciprocal of right hand side, and has the dimensions, they
define as the height of transfer unit (HTU) H G and consequently the above
equation tern to;

Z = HG NG (32)

Where NG = dy/(y-yi) =Number of a gas film transfer units (33)

HG = G / kg a  = Height of a gas film transfer units. (34)

Using same analysis for liquid side;

Z = HL NL (35)

NL=  dx / ( x - xi ) = number of liquid film transfer unit (36)

HL = L / kg A CT = Height of liquid film transfer unit (37)

Using overall coefficients of transfers:-

a) For gas;

NOG = dy/(y-ye) (38)

42
Prof.Fatma K. Gad

HOG = Gm / (kg a ) (39)

b) For liquid

NOL =  dx / ( xe - x ) (40)

HOL = Lm / (Kg a CT) (41)

Where CT is the mean molar density of the liquid ( m ).

For special case where the equilibrium curve is a straight line and dilute
solution;

NOG =(( y1 - y2 ) / y )Lm

=(y1-y2)/(y1-y*1) - (y2-y*2) Ln (y1-y*1)/(y2-y*2) (42)

NOL = ( x1 - x2 ) / (x)Lm

= (x1-x2)/(x1-x*1) - (x2-x*2) Ln (x1-x*1)/(x2-x*2) (43)

Relation between overall and film transfer units:-

From the above definitions the following equation may be written.

dz / HOG = dy / ( y - ye ) , dz / HG = dy / ( y - yi ) (33')

dz / HOL = dx / ( xe - x ) , dz / HL = dx / ( xi - x ) (33'')

HOG / ( y - ye ) = HG / ( y - yi
)
43
Prof.Fatma K. Gad

HOG = HG [ 1 + ( yi - ye ) / ( y - yi )] (44)

If the equilibrium curve is straight line, then assuming equilibrium at the


interface.

yi - ye = m ( xi - x )

And

H0G = HG + [ m ( xi - x ) / ( y - yi ) ] HG

From 33' and 33'' : ( xi - x ) / ( y - yi ) = dx HL / dy HG = Gm HL / Lm HG

Since Gm dy = Lm dx from above

dz / dy = HG / ( y - yi )

dz / dx = HL / ( xi - x )

dx HL / dy Hg = dz ( xi - x ) / dz ( y - yi ).

HOG = HG + m (Gm / Lm )HL (45)

Similarly

HOL = HL + Lm / ( m Gm HG ) (46)

By dividing

44
Prof.Fatma K. Gad

HOG / HOL = m Gm / Lm = NOL / NOG (47)

The importance of liquid and gas flow rates and slope of the equilibrium
curve:-

For a packed tower operating with dilute concentration; mol fraction = mol ratio
so

Gm ( y - y2 ) = Lm ( x1 - x2 )

For x2 = 0

 x = xi = Gm / Lm ( y - y2 )

Since NOG =  dy / ( y - ye )

For dilute solution Henry’s law holds and ye = mx therefore,


y1
NOG =  dy / [( y - m Gm / Lm) ( y - y2 )]
y2
Finally

NOG={1/(1-m Gm/Lm) Ln [(1-m Gm/Lm) (y1/y2)+ m Gm/Lm)]} (48)

It is shown from equation 48 that the greater mGm/Lm is the greater will be the
value of NOG for a given ratio of (y1/y2).

Example 3:-

Gas, from a petroleum distillation column, has its concentration of H 2S reduce


from 0.03 k mol H2S per k mol of inert HCS gas to 1% of this value with a tri-
ethanolamine – water solvent in a counter current tower, operating at 300 k and
at 1 atm. The equilibrium relation for the solution may be taken as Ye=2X.
The solvent enters the tower free of H2S and leaves containing 0.013 k mol
H2S/kmol of solvent. If the flow of inert gas is 0.015 Kmol/Sm2 of tower cross
section, calculate;

45
Prof.Fatma K. Gad

a) The height of the absorber necessary.


b) The NOG required.

The overall coefficient for absorption Kg a may be taken as 0.04 kmol/sm3.


(Unit mol fraction driving force).

Solution:

Since law concentration of H2S, So

Gm ( y1 - y2 ) A = Kg a  ( y - ye )Lm AZ

or

Z = Gm / Kg a  ( y1 - y2 ) / ( y )Lm

 driving force at top of column = y2 = y2e = 0.0003

Driving force at bottom of column = y1-y1e = 0/03-0.02 = 0.004

 Logarithmic mean driving force = 0.004-0.0003/ln 0.004/0.0003


=0.00143

 z = 0.015/0.04 x 1 x (0.03 – 0.0003)/0.00143 = 7.79 = 7.8 m

HOG (O.H.T.U) = Gm/kg a P = 0.015/0.04 = 0.375 m

NOG (O.N.T.U) = Z/HOG = 7.79/0.375 = 20.7 = 21.

Also NOG = 0.0297/0.0019 = 21

Example 4

A 2000 f3/min gas mixture contains 1% by vol. SO2 the rest air at 68°F. We
want to reduce the concentration to 0.01 % by vol. Using a dilute alkaline in a
packed tower. It may be assured that the P.P of SO 2 in equilibrium with solvent
is always zero and Kg equal 0.002 G0.8 where G in lb/hr. ft2 , kg in lb mol/hr atm.
Calculate the size of tower and height of packing (z). Velocity of gas phase =

46
Prof.Fatma K. Gad

6.7 ft/sec , a = 30 ft2/ft3

Solution:-

kg = 0.002 G0.8 lb mol/hr atm

a = 30 Ft2/ft3

v = 6.7 ft/sec

Since the velocity = vol flow rate/cross sectional area

6.7 = 2000/60 A

Area “A” = 5 ft2

D = ( 5 )0.5 = 2.25 ft for square tower

 Z = Gm / Kg a  ( yA1 - yA2 ) / ( yA )Lm

since G = 2000*60*0.075/5 area = 1800 1bmas/hr ft 2

Gm = G/mint = 1800/298 = 62 1bmol/hr ft2

Kg = 0.002 x 402 = 0.8

y = yA1 = yA1* = 0.01 – 0 = 0.01

yA2 = yA2 – yA2* = 0.001 – 0 = 0.0001

(y)Lm = 0.01 – 0.0001/lm0.01/0.001 = 215 x 10-3

 Z = 62[0.01-0.001]/0.8x30x1x2.15x10-3 = 12 ft. U.E.D

Example 5:-

We want to scrub 2000 ft3/min air containing 0.02 by vol NH3 at 68°F in the
same previous tower using water. 98 % of the ammonia entering is to be
removed. Calculate;

1) The theoretical minimum water mass velocity.

47
Prof.Fatma K. Gad

2) Z of packing if the water mass velocity is 1.5 the minimum value.

The equilibrium P. pressure of Nh3 is given by the equation PNH3 = 0.76 X


NH3, Kg is = 1.2 lb mol/hr ft2 at a = 30 ft2/ft3.

Solution:-

Gm = 2000*60*0.075/5 * 29 = 62 lb mol/hr ft 2 enter the column.

yA1 = 0.02 yA2 = 0.02(1-0.98) = 0.0004

  = 1 ATM  PA1 =  yA1

 PA1 = yA1 = 0.02  PA = 0.76 xA

ZA1= 0.020/0.70 = 0.0263 , xA2=0

Lm = Gm ( y1 - y2 ) / ( x1 - x2 ) = 62(0.02-0.0004)/(0.0263 – 0)
= 45 lb mol/hr ft2

(y)Lm = [0.02–(0.76*0.01360] – [0.004-(0.76*0)] / Ln (n / m)= 0.002

 Z = 6.2/1.2 *30-*1 *(10.02 – 0.0004)/0.002 = 15 FT U.E.D

1 mol occupied 3.59 ft3 at S.T.P

1 mol occupied 22.4 ft3 at S.T.P.

M unit of air = 29

Z = NOG * HOG = NTU * HTU

For dilute concentration 10% mol fraction = mol ratio

48
Prof.Fatma K. Gad

Absorption in a single equilibrium stage:-

An equilibrium stage is a hypothetical devise in w the gas and liquid streams


are brought together for a sufficient time to reach equilibrium and the separated.

A single stage may be represented as shown in Fig (13).

Gas and liquid streams enter the mixer at molar flow rate Gm2 and Lm0 moles the
respectively, containing the component to be absorbed at mole fractions y2 , x01
and leave the separator at equilibrium.

Fig. (10) Mixer and separator

The outlet composition y1e , x1e must lie on the equilibrium line.

Composition changes over the stage may be computed by satisfaction of the


material balance over the stage and the equilibrium relation.
Thus;

Gm2 y2 + Lmo xo = Gm1 y1e + Lm1 x1e (1)

And ye = f(x2) (2)

49
Prof.Fatma K. Gad

For the special case of absorption of a single component from a dilute solution
in the gas and liquid flows are changed negligibly by absorption equation 1 , 2
becomes;

-Lm / Gm = ( y2 - y1e ) / ( xo – xie ) (3)

Assume Lm and Gm are constant.

In this case the absorption process may be determined graphically as shown in


Fig (11)

Fig. (11) Absorption in single equilibrium stage.

The composition changes until the equilibrium line is reduced a long the line
abc which has a slope equal to –Lm / Gm as required by equation (3).

It should be noted at this point that in practice complete equilibrium can never
be attained since infinite contact time would be necessary. Depending on the
prevailing mass transfer rates in the stages, the composition will move along the
line ac and stop at some point b of composition x1 , y1 short of reacting the
equilibrium line.

A stage efficiency “E” may be defined as the ratio of the composition change
actually obtained to that which would be attained at equilibrium, so;

50
Prof.Fatma K. Gad

E = (y2 - y1) / (y2 - ye) = (x1 - xo) / (xie - xo) = ab / ac (4)

In many of the stage wise contactors used in practice stage efficiencies close to
unity can be realized.

1.3-D-3 Countercurrent Multiple-Contact Stages


1. Derivation of general equation.

In the previous section we used single stage contact to transfer the solute A
between the V and L phases. In order to transfer more solute from, say, the V1
stream, the single-stage contact can be repeated by again contacting the V 1
stream leaving the first stage with fresh L0. This can be repeated using multiple
stages. However, this is wasteful of the L0 stream and gives a dilute product in
the outlet L1 streams. To conserve use of the L0 stream and to get a more
concentrated product, countercurrent multiple-stage contacting is generally
used. This is somewhat similar to countercurrent heat transfer in a heat
exchanger, where the outlet heated stream approaches more closely the
temperature of the inlet hot stream.

The process flow diagram for a countercurrent stage process is shown in Fig.
1.3-4. The inlet L stream is L0 and the inlet V stream is VN+1 , instead of V2 as
for a single-stage in Fig. 1.3-3. The outlet product streams are V1 and LN and
the total number of stages is N. The component A is being exchanged between
the V and L streams.

Figure 1.3-4: Countercurrent multiple stage process.

The V stream is composed mainly of component B and the L stream of


component C. Components B and C may or may not be somewhat miscible in
each other. The two-phase system can be gas-liquid, vapor-liquid, liquid-liquid,
or other. Making a total overall balance on all stages,

51
Prof.Fatma K. Gad

(1.3-21)

where VN+1 is mol/h entering, LN is mol/h leaving the process, and M is the total
flow. Note in Fig. 1.3-4 that any two streams leaving a stage are in equilibrium
with each other. For example, in stage n, Vn and Ln are in equilibrium. For an
overall component balance on A, B, or C,

(1.3-22)

where x and y are mole fractions. Flows in kg/h (lb m/h) and mass fraction can
also be used in these equations.

Making a total balance over the first n stages,

(1.3-21)

Making a component balance over the first n stages,

(1.3-22)

Solving for yn+1, in Eq. (1.3-22).

(1.3-23)

This is an important material-balance equation, often called an operating line. It


relates the concentration yn+l in the V stream with xn in the L stream passing it.
The terms V1, y1, L0, and x0 are constant and usually known.

2. Countercurrent contact with immiscible streams.

An important case where the solute A is being transferred occurs when the
solvent stream V contains components A and B with no C and the solvent
stream L contains A and C with no B. The two streams L and V are immiscible
in each other with only A being transferred. When Eq. (1.3-23) is plotted on an
xy plot (xA and yA of component A) as in Fig. 1.3-4, it is often curved, since the
slope Ln / Vn+1 of the operating line varies if the L and V streams vary from
stage to stage.

52
Prof.Fatma K. Gad

Figure 1.3-5. Number of stages in a countercurrent multiple-stage contact


process.

In Fig. 1.3-5 is plotted the equilibrium line that relates the compositions of two
streams leaving a stage in equilibrium with each other. To determine the
number of ideal stages required to bring about a given separation or reduction
of the concentration of A from yN+1 to y1, the calculation is often done
graphically. Starting at stage l, y1 and x0 are on the operating line, Eq. (1.3-23),
plotted in the figure. The vapor y1 leaving is in equilibrium with the leaving x 1
and both compositions are on the equilibrium line. Then y2 and x1 are on the
operating line and y2 is in equilibrium with x2, and so on. Each stage is
represented by a step drawn on Fig. 1.3-5. The steps are continued on the graph
until yN+1 is reached. Alternatively, we can start at yN+l and draw the steps going
to y1.

If the streams L and V are dilute in component A, the streams are approximately
constant and the slope Ln/Vn+1 of Eq. (1.3-23) is nearly constant. Hence, the
operating line is essentially a straight line on an xy plot. In distillation, where
only components A and B are present, Eq. (1.3-23) also holds for the operating
line. Cases where A, B, and C are appreciably soluble in each other often occur
in liquid-liquid extraction.

EXAMPLE 1.3-2. Absorption of Acetone in a Countercurrent Stage Tower

53
Prof.Fatma K. Gad

It is desired to absorb 90% of the acetone in a gas containing 1.0 mol % acetone
in air in a countercurrent stage tower. The total inlet gas flow to the tower is
30.0 kg mol/h, and the total inlet pure water flow to be used to absorb the
acetone is 90 kg mol H2O/h. The process is to operate iso-thermally at 300 K
and a total pressure of 101.3 kPa. The equilibrium relation for the acetone (A)
in the gas-liquid is yA = 2.53 xA.

Determine the number of theoretical stages required for this separation.

Figure 1.3-6: Theorerical stages for counter current absorption in example 1.3-
2.

Solution:

The process flow diagram is similar to Fig. 1.3-5. Given values are
yAN+1 = 0.01, xA0 = 0,
VN+l = 30.0 kg mol/h, and L0 = 90.0 kg mol/h.
Making an acetone material balance,
amount of entering acetone : yAN+1 VN+1 = 0.01 x 30 = 0.30 kg mpl / hr
entering air : (1-yAN+1) VN+1 = (1-0.01) 30 = 29.7 kg mol air / hr

54
Prof.Fatma K. Gad

acetone leaving in V1 : 0.1 x 0.30 = 0.03 kg mol/ hr


acetone leaving in LN : 0.90 (0.30) = 0.27 kg mol/h

V1 = 29.7 + 0.03 = 29.73 kg mol air + acetone/h


yA1 = 0.03 / 29.73 = 0.00101
LN = 90.0 + 0.27 = 90.27 kg mol water + acetone/h
xAN = 0.27 / 90.27 = 0.0030

Since the flow of liquid varies only slightly from L0 = 90.0 at the inlet to LN =
90.27 at the outlet and V from 30.0 to 29.73, the slope L n / Vn+1, of the
operating line in Eq. (1.3-23) is essentially constant. This line is plotted in Fig.
1.3-6 and the equilibrium relation yA = 2.53 xA is also plotted. Starting at point
yAl, xA0, the stages are drawn as shown. About 5.2 theoretical stages are
required.

1.3-D-4 Analytical Equations for Countercurrent Stage


Contact
When the flow rates V and L in a countercurrent process are essentially
constant, the operating-line equation (1.3-23) becomes straight. If the
equilibrium line is also a straight line over the concentration range, simplified
analytical expressions can be derived for the number of equilibrium stages in a
countercurrent stage process.

Referring again to Fig. 1.3-4, Eq. (1.3-14) is an overall component balance on


component A.

(1.3-24)

(1.3-25)

Making a component balance for A on the first n stages,

(1.3-26)

(1.3-27)

55
Prof.Fatma K. Gad

Equating Eq. (1.3-25) to (1.3-27),

(1.3-28)

Since the molar flows are constant,

Ln = LN = constant = L and
Vn+l = VN+l = constant = V.

Then Eq. (1.3-28) becomes

(1.3-29)

Since yn+l and xn+1 are in equilibrium, and the equilibrium line is straight,

yn+1 = m xn+l. Also yN+1 =m xN+l.

Substituting m xn+l for yn+1 and calling A = L/mV, Eq. (1.3-29) becomes

(1.3-30)

where A is an absorption factor and is constant.

All factors on the right-hand side of Eq. (1.3-30) are constant. This equation is a
linear first-order difference equation and can be solved by the calculus of finite-
difference methods. The final derived equations are as follows.
For transfer of solute A from phase L to V (stripping),

(1.3-31)

(1.3-32)

When A = 1,

56
Prof.Fatma K. Gad

(1.3-33)

For transfer of solute A from phase V to L (absorption),

(1.3-34)

(1.3-35)

When A = l,

(Kremser equation) (1.3-36)

Often the term A is called the absorption factor and S the stripping factor,
where S = l/A. These equations can be used with any consistent of units such as
mass flow and mass fraction or molar flow and mole fraction. Such series of
equations are often called Kremser equations and are convenient to use. If A
varies slightly from the inlet to the outlet, the geometric average of the two
values can be used, with the value of m at the dilute end being used for both
values of A

EXAMPLE 1.3-3. Number of Stages by Analytical Equation.

Repeat Example 1.3-2 but use the Kremser analytical equations for
countercurrent stage processes.

Solution:

At one end of the process at stage 1,


V1 = 29.73 kg mol/h, yAl =0.00101,
L0 = 90.0, and xA0 = 0.

57
Prof.Fatma K. Gad

Also, the equilibrium relation is


yA = 2.53 xA where m = 2.53. Then,
A1 = L / mV = L0 / m V1 = 90.0 / ( 2.53 x 29.73) = 1.20
At stage N;
VN+l = 30.0, yAN+1 = 0.01,
LN = 90.27, and xAN = 0.00300.
AN = LN / m VN+1 = 90.27 / (2.53 x 30.0) = 1.19
The geometric average A=√(A1 x AN)=√(1.2 x 1.19) = 1.195.
The acetone solute is transferred from the V to the L phase (absorption).
Substituting into Eq. (1.3-35),
N=log[{(0.01–2.53(0))/(0.00101–2.53(0))
(1–(1/1.195))}+(1/1.195)]/log(1/1.195)
=5.04 stages
This compares closely with 5.2 stages obtained using the graphical method.
desorption

This operation, also called stripping which is the reverse of absorption. And is
treated by the same methods. The only differentces are that:
1- The driving potential ∆x and ∆y are reversed and are given by ( x-xi) , (yi-y)
2-The operating line lies below the equilibrium line.

The mass transfer in absorption and stripping is almost completely in one


direction. Consequently the phase rates changes rapidly through the column if
an appreciable amount of material is transferred.

In absorption, the wet gas which rises through the column loses material to the
descending absorber oil but little of the relatively non available oil passes to the
gas phase. As a result, the gas phase decreases in amount as it rises and the
liquid phase increases as corresponding amount as it descends.

In stripping, the entering gas contains little or none of the components to be


stripped from the oil to the gas phase to an extent described by their respective
distribution coefficients. Stripping gas normally used are essentially in soluble
in the oil phase, and little transfer of the gas to the liquid phase occurs.

58
Prof.Fatma K. Gad

Consequently, the oil phase decreases and the gas phase increases in amount as
they pass through the column.

The assumption of constant phase rates or constant L/V ratio in absorption or


stripping are valid only in the special case where the amount of material
transferred is so small as to have a negligible effect on the phase rates.

Besides the effect on the mass rates, the unidirectional transfer of material
creates thermal effect which must be considered. The material transferred
undergoes a phase change and the heats of vaporization or condensation are
involved along with the smaller heats of solution.

In absorption, most of the heat released as the gas components condense into
the liquid phase must appear as sensible heat in the liquid since insufficient
material is being vaporized to utilize the released heat as heat of vaporization.
Consequently the temperature of the liquid phase rises as it passes down the
column. The raise in liquid temperature occurs despite the fact that the liquid is
at its bubble point. The vaporization of only a small amount of a light
component say methane will cause a relatively large rise in the bubble point.
The amount of sensible heat required to raise the liquid temperature is large
compared with the heat required to vaporize the small amount of volatile
material

In stripping, the effect is opposite, where the heat of vaporization of the light
components as they pass from the liquid to the gas phase is supplied from the
sensible heat content of the liquid phase.
Consequently the temperature of the liquid phase decreases as it passes down
the column.
In the special case where only very small amount of material is transferred, the
temperature changes will be small and possible negligible.

Design procedures for multi component absorption graphical

Because of the unidirectional nature of the transfer in absorption and stripping


columns and the resulting variation in phase rates and temperature graphical
solution methods are less useful than in the cases of binary mixture.

The change in the phase rates as they pass through the column causes the L/V
ratio to change and therefore, the operating line equation does not plot as a
straight line.

59
Prof.Fatma K. Gad

yn = ( Ln+1 / Vn ) xn+1 + (Vo yo – L1 x1 ) / Vn )

Fig. (12) Schematic diagram for a column.

Only in special case of very small transfer will be L/V ratio remain essentially
constant.. If the temperature changes from stage to stage the distribution
coefficient K= y/x for each component changes and the equilibrium curves are
not straight lines; only in special case of an isothermal column.
If the wet gas to an absorber is very dilute in the material being absorbed, the
operation utilized only those portions of the equilibrium curve very close to the
origin and these portions often can be drawn as straight line (Henry’s holds).

To illustrate the above statements, consider a three stage absorber which


operates at 100 psia and changes the following feed;
Component ye
C1 0.1
C2 0.15
C3 0.10
C4 0.05
 = 1.0
Which has an Ln+1 / Vo = 2

The wet gas and the absorber oil both at 80°F and the oil is completely stripper
of light HCS to get a quick estimate of the amount absorbed.

Let us assume that L / V = Ln+1 / Vo and t = 80°F, through out the column. Both
operating and equilibrium curves will be straight lines in this case and can be
plotted as shown.

The slope of the equilibrium line for each component is the value for each

60
Prof.Fatma K. Gad

component is the value of K at 100 psig and 80°F.

The slop of all the operating lines is the same and equal 2.

The position of each operation line with respect to the corresponding


equilibrium line must be such as to satisfy the known concentration in L n+1 and
Vo and also allow three stages to be stepped off between these concentrations.

Fig. (13) Graphical construction for multi-component when L/V and T are
constant.

Considering the construction for C3 as in the above fig. the point ( yo = 0.1 , x1 )
represents passing streams Vo and L1 on the operating line and it must lie some
where along the horizontal line y = 0.1

Similarly, the point (y3 , xn+1 = 0) represents passing streams V3 and Ln+1 on the
operating line and since the entering absorber oil has been completely stripped,
it must fall somewhere on the ordinate at x = 0. So, by trial and error that
positions can be formed where exactly three steps can be constructed between
the intersections of the operating line with the lines y= 0.1 and x = 0.0. Similar
trial and error construction can be made for the other components.

It can be seen that most of the methane is transferred in the top stage and by the

61
Prof.Fatma K. Gad

time the oil leaves the second stage, it is essentially saturated with methane.
Therefore, little change in methane concentration occurs across the bottom
stage and for all practical purposes the point (y0 , x1) on the operative line
coincides with the equilibrium line.
In the case of the butane most of the transfer occurs at the bottom of the column
and to construct three stages it must be assumed that (y 3 , xn+1) essentially
coincides with the origin as shown.

The above case can be solved by choosing a key component.

The key component is defined as that component absorbed in appreciable


amount whose equilibrium curve falls most nearly parallel to the operating line
(component having a value of K most nearly equal to L/V).

Thus components heavier than the key component are absorbed nearly
completely and approach equilibrium with oil into the top if the fresh solvent
contains none of these component.

Components more volatile than the key component will approach equilibrium
with the liquid phase at the rich end (bottom) of the column.

Varying the L/V ratio clearly changes the nature of the key component. The
optimum ratio is the one which will allow substantially complete absorption of
the desired component.

The theoretical plate can be calculated by fixing the oil gas ratio and fraction of
the key component (butane) absorbed. Starting at the ordinate scale at the left
(top of the column) and counting off one step for each composition of the gas
entering and liquid leaving the top plate.

For component heavier than the key component C5 , it is evident that no


appreciable change occurs in gas composition in top plate.

It follows that the absorption of the more volatile components takes place in the
upper part of the column.

The construction for a stripping column is analogous to that shown for absorber
except that the operating lines fall below their respective equilibrium line.

Graphical solutions are satisfactory only in the case of dilute concentration of


the components being transferred. For all other cases a numerical solution
should be obtained.

62
Prof.Fatma K. Gad

GAS ABSORPTION WITH PLATE ABSORBERS :

Two assumptions are made; Gas Liquid solvent


1
v1 = V1 y1 Lo = Xo lo
(1) Effective constant absorption for
Solute (n) = AE = 1.4 N
Effective constant absorption for
Solvent (k) = zero vn+1 = yn+1 Vn+1 LN=LN Xn

(2) Effective constant, solute to solvent

Akn = 1.4 / αkn (1)

(3) Definition of the two functions NK and N-!K as for solute follows:

NK = ( 1 - AEk )N+1 / ( 1 - AEk ) (2)

N-1K = ( 1 - AEk )N / ( 1 - AEk ) (3)

Where N-1K , NK for solvent = 1 and N is the number of trays.

(4) Amount of solvent evaporate and gone with gas outlet;

νki = νkN+1 / NK + ( N-1K / NK ) LKo (4)

(5) Amount of solvent


SkN = [1 - N-1K / NK] LKo + [1 – ( 1 / NK )] VkN+1 (5)

(6) Amount of solvent in Lo:

Lo = 1.4 VN+1 (6)


Kn

(7) No. of trays:

63
Prof.Fatma K. Gad

N = Ln [ ( r – AkE ) / ( 1 – r ) AkE ] / Ln AkE (7)

Where r recovery percent


n refere to the solute
k refere to the solvent

If the solvent is not present in the solvent lon = zero ( pure solvent )

Example :
Gas flow 11 mole acetone in 1 mole Lo = pure water
V
Recovery of acetone = 95 % = r
T=300 k ,
Absorption factor AE = 1.4 P=10 bar
Pressure = 10 bar
Temperature at top of the tower = 300 oK
Calculate flow rate of water.
Lw
Vn+1=10 mole/sec air
Vn+1= 1 mole/sec acetone

Solution:

1) Flow of solvent inlet ( equation 6 )

Lko = 1.4 VN+1 Pon / π = 1.4 * (10+1) * 0.331 / 10 = 0.51 mol/sec

2) No. Of plate equation 7 where E = n , AE = 1.4 , r = 0.95

N = Ln [( r-AE)/(r-1) AE] Ln AE = 5.53 no acetone in H2O

3) For material balance where Aair = 0 ( non condensible )

and N-1air = Nair = 1

Since w/Ac = Pok(300o) / Pon(300o) = 0.106

Ank = 1.4 / kn = 13.24 eq 1

4) N-1k = [ 1 - ( Ank )N ] / ( 1 – Ank ) = 1.307E+05 eq 3

Nk = [ 1 - ( Ank )N+1 ] / ( 1 – Ank ) = 1.73E+06 eq 2

64
Prof.Fatma K. Gad

5) from material balance:

The flow rate of inert gas air = 10 mole /sec , acetone = 1 mole/sec

Flow rate of acetone in gas outlet = 1 ( 1 – 0.95 ) = 0.05 moles

6) flow rate of vaporized solvent ( H2O ) vik eq 4

V1k =(vN+1k)/Nk+(N-1k/Nk)Lok
=(zero/1.73e6)+(1.303e5/1.93e6)0.51= 0.038

7) Outlet solvent (H2O) material balance

H2O in gas inlet = zero , H2O evaporate with gas outlet = 0.038
H2O oulet = 0.51 – 0.038 = 1.472

U.E.D.

65
Prof.Fatma K. Gad

2.3 CONTINUOUS HUMIDIFICATION PROCESSES

2.3A Introduction and Types of Equipment for


Humidification
1-Introduction to gas-liquid contactors. When a relatively warm liquid is
directly contacted with gas that is unsaturated, some of the liquid is vaporized.
The liquid temperature will drop mainly because of the latent heat of
evaporation. This direct contact of a gas with a pure liquid occurs most often in
contacting air with water. This is done for the following purposes: humidifying
air for control of the moisture content of air in drying or air conditioning;
dehumidifying air, where cold water condenses some water vapor from warm
air; and water cooling, where evaporation of water to the air cools warm water.

Since, the fundamentals of humidity and adiabatic humidification were


discussed. The performance and design of continuous air-water contactors is
considered. The emphasis is on cooling of water, since this is the most
important type of process in the process industries. There are many cases in
industry in which warm water is discharged from heat exchangers and
condensers when it would be more economical to cool and reuse it than to
discard it.

2-Towers for water cooling. In a typical water-cooling tower, warm water


flows countercurrently to an air stream. Typically, the warm water enters the
top of a packed tower and cascades down through the packing, leaving at the
bottom. Air enters at the bottom of the tower and flows upward through the
descending water. The tower packing often consists of slats of wood or plastic
or of a packed bed. The water is distributed by troughs and overflows to
cascade over slat gratings or packing that provide large interfacial areas of
contact between the water and air in the form of droplets and films of water.
The flow of air upward through the tower can be induced by the buoyancy of
the warm air in the tower (natural draft) or by the action of a fan. Detailed
descriptions of towers are given in other texts.

The water cannot be cooled below the wet bulb temperature. The driving force
for the evaporation of the water is approximately the vapor pressure of the
water less the vapor pressure it would have at the wet bulb temperature. The
water can be cooled only to the wet bulb temperature, and in practice it is

66
Prof.Fatma K. Gad

cooled to about 3 K or more above this. Only a small amount of water is lost by
evaporation in cooling water. Since the latent heat of vaporization of water is
about 2300 kJ/kg, a typical change of about 8 K in water temperature
corresponds to an evaporation loss of about 1.5%. Hence, the total flow of
water is usually assumed to be constant in calculations of tower size.

In humidification and dehumidification, intimate contact between the gas phase


and liquid phase is needed for large rates of mass transfer and heat transfer. The
gas-phase resistance controls the rate of transfer. Spray or packed towers are
used to give large interfacial areas and to promote turbulence in the gas phase.

VAPOR PRESSURE OF WATER AND HUMIDITY


Vapor Pressure of Water
1. Introduction. In a number of the unit operations and transport processes it
is necessary to make calculations involving the properties of mixtures of water
vapor and air. These calculations involve knowledge of the concentration of
water vapor in air under various conditions of temperature and pressure, the
thermal properties of these mixtures, and the changes occurring when this
mixture is brought into contact with water or with wet solids in drying.

Humidification involves the transfer of water from the liquid phase into a
gaseous mixture of air and water vapor. Dehumidification involves the reverse
transfer, whereby water vapor is transferred from the vapor state to the liquid
state. Humidification and dehumidification can also refer to vapor mixtures of
materials such as benzene, but most practical applications occur with water. To
better understand humidity, it is first necessary to discuss the vapor pressure of
water.
2. Vapor pressure of water and physical states. Pure water can
exist in three different physical states: solid ice, liquid, and vapor. The physical
state in which it exists depends on the pressure and temperature.

Figure illustrates the various physical states of water and the pressure-
temperature relationships at equilibrium. The regions of the solid, liquid, and
vapor states are shown. Along the line AB, the phases liquid and vapor coexist.
Along line AC, the phases ice and liquid coexist. Along line AD, ice and vapor
coexist. If ice at point (1) is heated at constant pressure, the temperature rises
and the physical condition is shown moving horizontally. As the line crosses

67
Prof.Fatma K. Gad

AC, the solid melts, and on crossing AB the liquid vaporizes. Moving from
point (3) to (4), ice sublimes (vaporizes) to a vapor without becoming a liquid.

Liquid and vapor coexist in equilibrium along the line AB, which is the vapor-
pressure line of water. Boiling occurs when the vapor pressure of the water is
equal to the total pressure above the water surface. For example, at 100°C
(212°F) the vapor pressure of water is 101.3 kPa (1.0 atm), and hence it will
boil at 1 atm pressure. At 65.6°C (150°F), from the steam tables in Appendix
A.2, the vapor pressure of water is 25.7 kPa (3.72 psia). Hence, at 25.7 kPa and
65.6°C, water will boil.

Phase diagram for water.

If a pan of water is held at 65.6°C in a room at 101.3 kPa abs


pressure, the vapor pressure of water will a gain be 25.7 kPa. This
illustrates an important property of the vapor pressure of water,
which is not influenced by the presence of an inert gas suchfas air;
i.e., the vapor pressure of water is essentially independent of the
total pressure of the system.

Humidity and Humidity Chart

1-Definition of humidity. The humidity H of an air-water vapor


mixture is defined as the kg of water vapor contained in 1 kg of dry
air. The humidity so defined depends only on the partial pressure pA
of water vapor in the air and on the total pressure P (assumed
throughout this chapter to be 101.325 kPa, 1.0 atm abs, or 760 mm
Hg). Using the molecular weight of water (A) as 18.02 and of air as
28.97, the humdiity H in kg H2O/kg dry air or in English units as lb
H2O/lb dry air is as follows:

H = (kg H 2 O / kg dry air)

68
Prof.Fatma K. Gad

= (p A /(P—p A ))(kg mol H 2 O/kg mol air)(18.02 kg H 2 O/kg mol H 2 O)


(1 / (28.97 kg air / kg mol air))

H = (18.02 / 28.97) (P A / (P – P A ))

Saturated air is air in which the water va por is in equilibrium with


liquid water at the given conditions of pressure and temperature. In
this mixture the partial pressure of the water vapor in the air -water
mixture is equal to the vapor pressure p AS of pure water at the given
temperature. Hence, the saturation humidity Hs is

H s = (18.02 / 28.97) (P AS / (P – P AS ))

2-Percentage humidity. The percentage humidity HP is defined as


100 times the actual humidity H of the air divided by the humidity
Hs if the air were saturated at the same temperature and pressure.

H P = 100 H / H S

3-Percentage relative humidity. The amount of saturation of an


air-water vapor mixture is also given as percentage relative
humidity HR using partial pressures.

H R = 100 (P A / P AS )

Note that H R # H P , since HP expressed in partial pressures by


combining the first three Equations.

HP = 100 (H / H S )
= (100) (18.02/28.97) (P A /(P–P A ))/(18.02/28.97) (P AS /(P–P S ))
= (P A / P AS ) ((P – P AS ) / ( P – P A ))

EXAMPLE.

Humidity from Vapor-Pressure Data

The air in a room is at 26 .7°C (80°F) and a pressure of 101.325 kPa


and contains water vapor with a partial pressure p A = 2.76 kPa.
Calculate the following.

Humidity, H.

69
Prof.Fatma K. Gad

Saturation humidity, Hs, and percentage humidity HP.


Percentage relative humidity, HR .

Solution:

From the steam tables at 26.7°C, the vapor pressure of water is p AS


= 3.50 kPa (0.507 psia). Also, p A = 2.76 kPa and P = 101.3 kPa
(14.7 psia).

For part (a),

H = (18.02/28.97) (P A /(P–P A )) = 18.02 (2.76)/(28.97 (101.3–2.76))


= 0.01742 kg H 2 O / kg air

For part (b), the saturation humidity is

H S = (18.02/28.97) (p AS /P -p AS ))
= 18.02(3.50) / (28.97(101.3 - 3.50))
= 0.02226 kg H 2 O / kg air

The percentage humidity,

H P = 100 (H / H S ) = 100 (0.01742 / 0.02226 = 78.3 %

For part (c), the percentage relative humidity is

H g = 100 (P A / P AS ) = 100 (2.76) / 3.5 = 78.9 %

4-Dew point of an air-water vapor mixture. The temperature at


which a given mixture of air and water vapor would be saturated is
called the dew-point temperature or simply the dew point. For
example, at 26.7°C (80°F), the saturation vapor pressure of water is
P A S = 3.50 kPa (0.507 psia). Hence, the dew point of a mixture
containing water vapor having a partial pressure of 3.50 kPa is
26.7°C. If an air-water vapor mixture is at 37.8°C (often called the
dry bulb temperature, since this is the actual temperature a dry
thermometer bulb would indicate in this mixture) and contains water
vapor of p A = 3.50 kPa, the mixture would not be saturated. On
cooling to 26.7°C, the air would be saturated, i.e., at the dew point.

70
Prof.Fatma K. Gad

On further cooling, some water vapor would condense, since the


partial pressure cannot be greater than the saturation vapor pressure.

5-Humid heat of an air-water vapor mixture. The humid heat cs


is the amount of heat in J (or kJ) required to raise the temperature of
1 kg of dry air plus the water vapor present by 1 K or 1=C. The heat
capacity of air and water vapor can be assumed constant over the
temperature ranges usually encountered at 1.005 kJ/kg dry air • K
and 1.88 kJ/kg water vapor • K, respectively. Hence, for SI and
English units,

cs kJ/kg dry air.K = 1.005 + 1.88/7 (SI)

cs btu/lbm dry air.oF = 0.24 + 0.45 H (English)

[In some cases cs will be given as (1.005 + 1.88H)103 J/kg.K.]

6-Humid volume of an air-water vapor mixture. The humid


volume vH is the total volume in m 3 of 1 kg of dry air plus the
vapor it contains at 101.325 kPa (1.0 atm) abs pressure and the
given gas temperature. Using the ideal gas law,

υ H m 3 /kg dry air = (22.41 / 273) T K ((1 / 28.97) + (1 / 18.02) H)


= (2.83 x 10 - 3 + 4.56 x 10 - 3 H)T K

υ H ft 3 /lb m dry air = (359 / 492) T o R ((1 / 28.97) + (1 / 18.02) H)


= (0.0252 + 0.0405 H) T o R

For a saturated air-water vapor mixture, H = Hs, and υ H is the


saturated volume.

7-Total enthalpy of an air-water vapor mixture. The total


enthalpy of 1 kg of air plus its water vapor is Hy J/kg or kJ/kg dry
air. If T0 is the datum temperature chosen for both components, the
total enthalpy is the sensible heat of the air-water vapor mixture
plus the latent heat  0 in J/kg or kJ/kg water vapor of the water
vapor at T 0 . Note that (T - T0)°C = (T - T0) K and that this enthalpy
is referred to liquid water.

H y kJ/kg dry air = c s (T - T 0 )+H 0 =-(1.005+1.88H)(T - T0°C) + H 0

71
Prof.Fatma K. Gad

H y , btu/lb m dry air = (0.24 + 0.45H) (T - T 0 °F) + H 0

If the total enthalpy is referred to a base temperature T 0 of 0°C


(32°F), the equation for tfy becomes

H y kJ/kg dry air = (1.005 + 1.88 H) (T°C - 0) + 2501.4 H (SI)

H y btu/lb m dry air = (0.24 + 0.45H) (T°F - 32) + 1075.4H (English)

8-Humidity chart of air-water vapor mixtures. A convenient


chart of the properties of air-water vapor mixtures at 1.0 atm abs
pressure is the humidity chart. In this figure the humidity H is
plotted versus the actual temperature of the air-water vapor mixture
(dry bulb temperature).

The curve marked 100% running upward to the right gives the
saturation humidity tfs as a function of temperature. In the Example,
for 26.7°C Hs was calculated as 0.02226 kg H 2 O/kg air. Plotting
this point of 26.7°C (80°F) and Hs = 0.02226 on Fig. 9.3 -2, it falls
on the 100% saturated line.
Any point below the saturation line represents unsaturated air -water
vapor mixtures. The curved lines below the 100% saturation line'and
running upward to the right represent unsaturated mixtures of
definite percentage humidity HP. Going downward vertically from
the saturation line at a given temperature, the line between 100%
saturation and zero humidity H (the bottom horizontal line) is
divided evenly into 10 increments of 10% each.

All the percentage humidity lines H P mentioned and the saturation


humidity line Hs can be calculated from the data of vapor pressure
of water.

EX A MPLE . Use of Humidity Chart

Air entering a dryer has a temperat ure (dry bulb temperature) of


60°C (140°F) and a dew point of 26.1'C (80°F). Using the humidity
chart, determine the actual humidity tf, percentage humidity H P ,
humid heatcs, and the humid volume vu in SI and English units.

Solution:

72
Prof.Fatma K. Gad

The dew point of 26.7°C is the temperature when the given mixture
is at 100% saturation. Starting at 26.7°C, Fig. 9.3 -2, and drawing a
vertical line until it intersects the line for 100% humidity, a
humidity of H = 0.0225 kg H 2 O/kg dry air is read off the plot. This
is the actual humidity of the air at 60°C. Stated in another way, if
air at 60°C and having a humidity H = 0.0225 is cooled, its dew
point will be 26.7°C. In English units, H = 0.0225 lb H 2 O/lb dry air.

Locating this point of H = 0.0225 and t = 60°C on the chart, the


percentage humidity H P is found to be 14%, by linear interpolation
vertically between the 10 and 20% lines. The humid heat for tf =
0.0225 is, from

c s = 1.005 + 1.88(0.0225)
= 1.047 kJ/kg dry air • K or 1.047 x 103 J/kg • K

c s = 0.24 + 0.45(0.0225)
= 0.250 btu/lb m dry air • °F (English)

The humid volume at 60°C (140°F), is


v H = (2.83 x 10 - 3 + 4.56 x 10 - 3 x 0.0225) (60 + 273)
= 0.977 m3/kg dry air
In English units,
v H = (0.0252 + 0.0405 x 0.0225)(460 + 140) = 15.67 ft 3 /lbm
dry air 9.3C

73
Prof.Fatma K. Gad

Humidity chart for mixtures of air and water vapor at a pressure of


101.325 kPa (760 mm Hg)

74
Prof.Fatma K. Gad

2.3B Theory and Calculation of Water-Cooling Towers


1. Temperature and concentration profiles at interface. In Fig. 2.5-1 the
temperature profile and the concentration profile in terms of humidity are
shown at the water-gas interface. Water vapor diffuses from the interface to the
bulk gas phase with a driving force in the gas phase of (H i - HG) kg H2O/kg dry
air. There is no driving force for mass transfer in the liquid phase, since water is
a pure liquid. The temperature driving force is T L - Ti in the liquid phase and Ti
- TG K or °C in the gas phase. Sensible heat flows from the bulk liquid to the
interface in the liquid. Sensible heat also flows from the interface to the gas
phase. Latent heat also leaves the interface in the water vapor, diffusing to the
gas phase. The sensible heat flow from the liquid to the interface equals the
sensible heat flow in the gas plus the latent heat flow in the gas.

The conditions in Fig. 2.3-1 occur at the upper part of the cooling tower. In the
lower part of the cooling tower the temperature of the bulk water is higher than
the wet bulb temperature of the air but may be below the dry bulb temperature.
Then the direction of the sensible heat flow in Fig. 2.3-1 is reversed.

2. Rate equations for heat and mass transfer.

Rigure 2.3.1: The conditions in the upper part of the cooling tower.

We shall consider a packed water-cooling tower with air flowing upward and
water countercurrently downward in the tower. The total interfacial area
between the air and water phases is unknown, since the surface area of the
packing is not equal to the interfacial area between the water droplets and the
air. Hence, we define a quantity a, defined as m2 of interfacial area per m3
volume of packed section, or m2/m3. This is combined with the gas-phase mass-

75
Prof.Fatma K. Gad

transfer coefficient kG in kg mol/s-m2-Pa or kg mol/s-m2-atm to give a


volumetric coefficient kGa in kg mol/s ■ m3 volume • Pa or kg mol/s • m3 • atm
(lb mol/h • ft3 • atm).

The process is carried out adiabatically and the various streams and conditions
are shown in Fig. 2.3-1,
where
L = water flow, kg water/s • m2 (lbm / h • ft2)
TL = temperature of water, °C or K (°F)
G = dry air flow, kg/s • m2 (lbm / h • ft2)
TG = temperature of air, °C or K (°F)
H = humidity of air, kg water/kg dry air (lb water/lb dry air)
Hy — enthalpy of air-water vapor mixture, J/kg dry air (Btu/lbm dry air)
The enthalpy Hy is

Hy = cs(T - T0) + Hλ0 = (1.005 + 1.88H)103(T - 0) + 2.501 x 106H (SI)


Hy = cs(T - T0) + H λ0 = (0.24 + 0.45H) (T - 32) + 1075.4H (English)

The base temperature selected is 0°C or 273 K (32°F). Note that(T - T0)°C = (T
— T0) K.
Making a total heat balance for the dashed-line box shown in Fig. 2.5-2, an
operating line is obtained.

G (Hy — Hyl) = LcL(TL — TLl) (2.3-1)

This assumes that L is essentially constant, since only a small amount is


evaporated. The heat capacity cL of the liquid is assumed constant at 4.187 x
103 J/kg - K (1.00btu/lbm • °F). When plotted on a chart of H y versus TL, this
Eq. (3.5-1) is a straight line with a slope of L cL/G. Making an overall heat
balance over both ends of the tower,

G (Hy2 - Hyl) = LcL (TL2 - TLl) (2.3-2)

Again making a heat balance for the dz column height and neglecting sensible
heat terms compared to the latent heat,

LcL dTL = G dHy (2.3-3)

The total sensible heat transfer from the bulk liquid to the interface is (refer to
Fig. (2.3-1)

LcL dTL = G dHy = hL a dz (TL – Ti) (2.3-4)

76
Prof.Fatma K. Gad

Figure 10.5-2. Continuous countercurrent adiabatic water cooling

where hLa is the liquid-phase volumetric heat-transfer coefficient in W/m3-K


(btu/h-ft3 ■ °F) and Ti is the interface temperature.

For adiabatic mass transfer the rate of heat transfer due to the latent heat in the
water vapor being transferred can be obtained

qλ / A = MB kG a P λ0 (Hi- HG)dz (2.3-5)

where qλ/A is in W/m2 (btu/h - ft2), MB = molecular weight of air, kGa is a


volumetric mass-transfer coefficient in the gas in kg mol/s - m3 • Pa, P = atm
pressure in Pa, λ0 is the latent heat of water in J/kg water, Hi, is the humidity of
the gas at the interface in kg water/kg dry air, and HG is the humidity of the gas
in the bulk gas phase in kg water/kg dry air. The rate of sensible heat transfer in
the gas is

qs / A = hG a (Ti - TG) dz (2.3-6)

where qs/A is in W/m2 and hca is a volumetric heat-transfer coefficient in the


gas in W/m3 • K.

Now from Fig. 2.3-1, Eq. (2.3-4) must equal the sum of Eqs. (2.3-5) and (2.3-6).

GdHy = MB kG a P λ0 (Hi - HG) dz + hG a (Ti- TG) dz (2.3-7)

And hG a / MB ky a ≈ cs (2.3-8)

Substituting PkG a for k a,

77
Prof.Fatma K. Gad

hG a / MB P kG a ≈ cs (2.3-9)

Substituting Eq. (2.3-9) into Eq. (2.3-7) and rearranging,

G dHy = MB kG a P dz
[(cs Ti + λ0 Hi) - (cs TG + λ0 HG)] (2.3-10)

Adding and subtracting cs T0 inside the brackets,

G dHy=MB kG a P dz (cs(Ti-T0)
+Hi λ0-[cs (TG-T0)+HG λ0] (2.3-11)

The terms inside the braces are (Hyi - Hy), and Eq. (2.3-11) becomes

G dHy = MB kG aP dz(Hyi - Hy) (2.3-12)

Integrating, the final equation to use to calculate the tower height is

(2.3-13)

If Eq. (2.3-4) is equated to Eq. (2.3-12) and the result rearranged,

-hL a / kG a MB P = (Hyi – hy) / (Ti – TL) (3.5-14)

2.3c Design of Water-Cooling Tower Using Film Mass-


Transfer Coefficients
The tower design is done using the following steps. Liquid temperature (°C)

1-The enthalpy of saturated air Hyi is plotted versus Ti on an H versus T plot as


shown in Fig. 2.3-3. This enthalpy is calculated using the saturation humidity
from the humidity chart for a given temperature, with 0°C (273 K) as a base
temperature. Calculated values are tabulated in Table 2.3-1.

2-Knowing the entering air conditions TG1 and Hl the enthalpy of this air Hyl is
calculated. The point Hyl and TL1 (desired leaving water temperature) is plotted
in Fig. 2.3-3 as one point on the operating line. The operating line is plotted

78
Prof.Fatma K. Gad

with a slope LcL/G and ends at point TL2, which is the entering water
temperature. This gives Hy2 . Alternatively, Hy2 can be calculated from Eq. (2.3-
2).

3-Knowing hLa and kGa, lines with a slope of –hL a / kG a MB P are plotted as
shown in Fig. 2.3-3. From Eq. (2.3-14) point P represents Hy and TL on the
operating line, and point M represents Hyi and Ti, the interface conditions.
Hence, line MS or Hyi - Hy represents the driving force in Eq. (2.3-13).

Figure 2.3-3: Temperature – Enthalpy diagram and operating line for water
cooling tower.

4. The driving force Hyi - Hy is computed for various values of TL between TLl
andTL2-Then by plotting l/(Hyl - Hy) versus Hy from Hyl to Hy2, a graphical
integration is performed to obtain the value of the integral in Eq. (2.3-13).
Finally, the height z is calculated from Eq. (2.3-13).

Table 2.3-1. Enthalpies of Saturated Air-Water Vapor Mixtures (0°C Base


Temperature)

79
Prof.Fatma K. Gad

2.3D Design of Water-Cooling Tower Using Overall


Mass-Transfer Coefficients
Often, only an overall mass-transfer coefficient KGa in kg mol/s-m3-Pa or kg
mol/s-m3 - atm is available, and Eq. (2.3-13) becomes

(2.3-15)

The value of H*y is determined by going vertically from the value of H y at point
P up to the equilibrium line to give H* y at point R, as shown in Fig. 2.3-3. In
many cases the experimental film coefficients k Ga and hLa are not available.
The few experimental data available indicate that h La is quite large and the
slope of the lines -hLal (kG a MB P) in Eq. (2.3-14) would be very large and the
value of Hyi would approach that of H*y in Fig. 2.3-3.

The tower design using the overall mass-transfer coefficient is done using the
following steps.

1-The enthalpy-temperature data from Table 2.3-1 are plotted as shown in Fig.
2.3-3.

2-The operating line is calculated as in steps 1 and 2 for the film coefficients
and plotted in Fig. 2.3-3.

3-In Fig. 2.3-3 point P represents Hy and TL on the operating line and point R
represents H*y on the equilibrium line. Hence, the vertical line RP or H* y - Hy
represents the driving force in Eq. (2.3-15).

80
Prof.Fatma K. Gad

4-The driving force H*y - Hy is computed for various values of TL between TL1
and TL2. Then by plotting l/(H*y — Hy) versus Hy from Hy1 to Hy2, a graphical
integration is performed to obtain the value of the integral in Eq. (2.3-15).
Finally, the height z is obtained from Eq. (2.3-15).

If experimental cooling data in an actual run in a cooling tower with known


height z are available, then using Eq. (2.3-15), the experimental value of KGa
can be obtained.

EXAMPLE 2.3-1. Design of Water-Cooling Tower Using Film


Coefficients

A packed countercurrent water-cooling tower using a


gas flow rate of G = 1.356 kg dry air/s-m2 and a
water flow rate of L = 1.356 kg water/s • m2 is to cool the water from
TL2 = 43.3°C (110°F) to
TLl = 29.4°C (85°F).
The entering air at 29.4°C has a wet bulb temperature of 23.9°C.
The mass-transfer coefficient kG a is estimated as 1.207 x 10-7 kg mol/s • m3 •
Pa and hL a / kc a MB P as 4.187 x 104 J/kg-K (10.0 btu/lbm-°F).
Calculate the height of packed tower z. The tower operates at a pressure of
1.013 x 105 Pa.

Solution:

Following the steps outlined,


The enthalpies from the saturated air-water vapor mixtures from Table 2.3-1 are
plotted in Fig. 2.3-4.
The inlet air at TG1 = 29.4°C has a wet bulb temperature of 23.9°C.
The humidity from the humidity chart is 77, = 0.0165 kg H20/kg dry air.
Noting that (29.4 - 0)°C = (29.4 - 0) K,
Hyl = (1.005 4- 1.88 x 0.0165)103(29.4 - 0) 4- 2.501 x 106(0.0165)
= 71.7 x 103 J/kg
The point Hyl = 71.7 x 103 and TLl = 29.4°C is plotted.
Then substituting into Eq. (2.3-2) and solving,

81
Prof.Fatma K. Gad

1.356(Hy2 - 71.7 x 103) = 1.356(4.187 x 103)(43.3 - 29.4)


Hy2 = 129.9 x 103 J/kg dry air (55.8 Btu/lbm. The point
Hy2 = 129.9 x 103 and
TL2 = 43.3°C is also plotted, giving the operating line.

Figure 2.3-4: Graphical solution of Example 2.3-1.

Lines with slope –hL a /kG a MB P = -41.87 x 103 J/kg-K are plotted giving Hyi
and Hy values, which are tabulated in Table 2.3-2 along with derived values as
shown.

Table 2.3-2. Enthalpy Values for Solution to Example 2.3-1 (enthalpy in J/kg
dry air)

82
Prof.Fatma K. Gad

Values of l/(Hyi — Hy) are plotted versus Hy and the area under the curve from
Hyl = 71.7 x 103 to Hy2 = 129.9 x 103 is Substituting into Eq. (2.3-13),

= 6.98 m (22.9 ft)

2.3-E Minimum Value of Air Flow


Often the air flow G is not fixed but must be set for the design of the cooling
tower. As shown in Fig. 3.5-5 for a minimum value of G, the operating line MN
is drawn through the point Hyi and TL) with a slope that touches the equilibrium
line at TL2, point N. If the equilibrium line is quite curved, line MN could
become tangent to the equilibrium line at a point farther down the equilibrium
line than point N. For the actual tower, a value of G greater than G min must be
used. Often, a value of G equal to 1.3 to 1.5 times Gmin is used.

3.5F Design of Water-Cooling Tower Using Height of a


Transfer Unit
Often another form of the film mass-transfer coefficient is used in Eq. (2.3-13):

(2.3-16)

83
Prof.Fatma K. Gad

Figure 2.3-5: Operating line construction for minimum gas flow.

HG = G / MB kG a P (2.3-17)

where HG is the height of a gas enthalpy transfer unit in m and the integral term
is called the number of transfer units. The term H G is often used since it is less
dependent upon flow rates than kG a.

Often another form of the overall mass-transfer coefficient KGa in kg mol/s • m3


• Pa or kg mol/s • m3 • atm is used and Eq. (2.3-15) becomes

(2.3-18)

where HoG is the height of an overall gas enthalpy transfer unit in m. The value
of H*y is determined by going vertically from the value of Hy up to the
equilibrium line as shown in Fig. 2.3-3. This method should be used only when
the equilibrium line is almost straight over the range used. However, the H oG is
often used even if the equilibrium line is somewhat curved because of the lack
of film mass-transfer coefficient data.

2.3G Temperature and Humidity of Air Stream in Tower


The procedures outlined above do not yield any information on the changes in

84
Prof.Fatma K. Gad

temperature and humidity of the air-water vapor stream through the tower. If
this information is of interest, a graphical method by Mickley is available. The
equation used for the graphical method is derived by first setting Eq. (2.3-6)
equal to Gcs dTG and combining it with Eqs. (2.3-12) and (2.3-9) to yield Eq.
(2.3-19).

dHy / dTG = (Hyi – Hy) / (Ti – TG) (2.3-19)

2.3H Dehumidification Tower


For the cooling or humidification tower discussed, the operating line lies below
the equilibrium line and water is cooled and air humidified. In a
dehumidification tower cool water is used to reduce the humidity and
temperature of the air that enters. In this case the operating line is above the
equilibrium line. Similar calculation methods are used (TI).

2.4 ESTIMATION OF MASS-TRANSFER


COEFFICIENTS FOR PACKED TOWERS

2.4A Experimental Determination of Film Coefficients


The individual film mass-transfer coefficients k`ya and k`xa depend generally
upon Schmidt number, Reynolds number, and the size and shape of the packing.
The interactions among these factors are quite complex. Hence, the correlations
for mass-transfer coefficients are highly empirical. The reliability of these
correlations is not too satisfactory. Deviations of up to 25% are not uncommon.
A main difficulty arises because an overall coefficient or resistance is measured
experimentally that represents the two film resistances in series. To obtain the
single-phase film coefficient, the experiment is so arranged that the other film
resistance is negligible or can be approximately calculated.

To measure the liquid film mass-transfer coefficient k`xa, a system for


absorption or desorption of very insoluble gases such as O 2 or CO2 in water is
used. The experiment gives K`xa, which equals k`xa, since the gas-phase
resistance is negligible.

85
Prof.Fatma K. Gad

To measure the gas-phase film coefficient k`ya, we desire to use a system such
that the solute is very soluble in the liquid and the liquid-phase resistance is
negligible. Most such systems as NH3-air-water have a liquid-phase resistance
of about 10%. By subtracting this known liquid phase resistance (obtained by
correcting k`xa data for absorption of CO2 or O2 to NH3 data for k`xa) from the
overall resistance in Eq. (1.4-24), we obtain the coefficient k`ya. Details of these
are discussed elsewhere.

2.4B Correlations for Film Coefficients


The experimental data for the gas film coefficient in dilute mixtures have been
correlated in terms of HG, where

HG = V / k`y a S (2.4-1)

The empirical equation is as follows:

HG = α Gyβ Gxγ NSc0.5 (2.4-2)

where Gy = kg total gas/s • m2; Gx = kg total liquid/s • m2; and a, p1, and y are
constants for a packing as given in Table 2.4-1. The temperature effect, which
is small, is included in the Schmidt number μ/ρD, where μ is the viscosity of the
gas mixture in kg/m-s, ρ the density in kg/m3, and D the diffusivity of solute A
in the gas in m2/s. The coefficients k`ya and HG can be shown to be independent
of pressure.

Equation (2.4-2) can be used to correct existing data for absorption of solute A
in a gas on a specific packing to absorption of solute E in the same system and
the same mass-flow rates. This is done by Eq. (2.4-3).

HG(E) = HG(A) [NSc(E) / NSc(A)]0.5 (2-4-3)

The correlations for liquid film coefficients in dilute mixtures show that HL is
independent of gas rate until loading occurs, as given by the following:

HL = θ (Gx / μL)η NSc0.5 (2.4-4)

where HL is in m, μL is liquid viscosity in kg/m.s, NSc is Schmidt number μ/ρD,

86
Prof.Fatma K. Gad

ρ is liquid density in kg/m3, and D is diffusivity of solute A in the liquid in m2/s.


Data are given in Table 3.8-2 for different packings. Equation (3.8-3) can be
used to correct existing data on a given packing and solute to another solute.

Table 2.4-1: Gas Film Height of Transfer Unit HG in Meters.

Packing Parameters Range of values


Type α β γ Gy Gx
Rashing rings
9.5 mm(3/8”) 0.62 0.45 -0.47 0.271-.678 0.678-.034
25.4mm(1”) 0.557 0.32 -0.51 0.271-.814 0.678-6.1
38.1mm(1.5”) 0.83 0.38 -0.66 0.271-0.95 0.678-.034
38.1mm(1.5”) 0.689 0.38 -0.4 0.271-0.95 2.034-6.1
80.8mm(2”) 0.894 0.41 -0.45 0.271-.085 0.678-6.1
Berl saddles
12.7mm(0.5”) 0.541 0.3 -0.74 0.271-0.95 0.678-.034
12.7mm(0.5”) 0.367 0.3 -0.24 0.271-0.95 2.034-6.1
25.4mm(1”) 0.461 0.36 -0.4 0.271-.085 0.542-6.1
38.1mm(2”) 0.652 0.32 -0.45 0.271- 0.542-6.1
1.356

Table 2.4-2: Liquid Film Hight of a Transfer Unit HL in Meters.

EXAMPLE 2.4-1. Prediction of Film Coefficients for Ammonia Absorption

Predict HG, HL, and K`ya for absorption of NH3 from water in a dilute solution
in a packed tower with 25.4-mm Raschig rings at 303 K (86°F) and 101.32 kPa

87
Prof.Fatma K. Gad

pressure. The flow rates are Gx = 2.543 kg/s • m2 and Gy = 0.339 kg/s • m2.

Solution:

The equilibrium relation in a dilute solution is 0.0151 = m(0.0126) or


y = 1.20x. Also, for air,
u = 1.86 x 10-5 kg/m-s. Density
 = 1.168 kg/m3.
The diffusivity of NH3 in air at 273 K from Table 6.2-1 is 1.98 x 10-5
m2/s.
Correcting to 303 K by Eq. (6.2-45), DAB = 2.379 x 10-5 m2/s. Hence,
NSc = μ / ρ D = 1.86x10-5 / (1.168) (2.379x10-5) = 0.669
Substituting into Eq. (2.4-2) using data from Table 2.4-1,
HG = α Gyβ Gxγ NSc0.5 = 0.557(0.339)0.32 (2.543)-0.51 (0.669)0.5 = 0.2 m
The viscosity of water = 1.1404 x 10 -3 kg/m-s at 15°C and 0.8007 x 10-3 at
30°C. The DAB of NH3 in water at 288 K (15°C) from Table 1.3-1 is 1.77 x 10-9
m2/s. Correcting this to 303 K (30°C),
DAB = (1.1404x10-3 / 0.8007x10-3) (303 / 288) (1.77x10-9) m2 / s
Then, using ρ = 996 kg/m3 for water,
NSc μ / ρ D = 0.8007x10-3 / (996) (2.652x10-9) = 303.1
Substituting into Eq. (2.4-4), using data from Table 2.4-1,
HL = θ (Gx / μL)η NSc0.5 = (2.35x10-3) (2.543 / 0.8007x10-3)0.22 (303.1)0.5
= 0.2412 m

Converting to k`ya using Eq. (2.1-37),

k`y a = V / HG S = (0.339 / 29) / 0.2 = 0.0584 ks mol/s m3 mo1 frac

For K`xa using Eq. (2.1-38),

k`x a = L / HL S = (2.543 / 18) / 0.2412 = 0.586 kg mol/s m3 mo1 frac

Substituting into Eq. (1.4-16) for dilute solutions,

1 / K`y a = 1 / k`y a + m’ / k`x a = (1 / 0.584) + (1.2 / 0.586)


= 17.12 + 2.048 = 19.168
K`y a = 0.0522 kg mol/s - m3 • mol frac

Note that the percent resistance in the gas film is (17.12/19.168) (100) = 89.3%.

88
Prof.Fatma K. Gad

( I ) DISTILLATION

Distillation is the most widely used method of separation of liquid mixtures into
their several components. On the engineering side, distillation column have to
be designed with a larger range in capacity than any other types of chemical
engineering equipment, with single column from 0.3 to 10 meter in diameter
and from 3 to 75 meter in height.

The demand on designers is not only to achieve the desired product quality at
minimum cost, but also to provide constant purity of product even though there
may be some variation in the feed composition.

A distillation unit must never be considered without its associated control unit
and usually it with operate in association with several other separate units.

In the following part, consideration is given to the theory if the process,


methods of distillation and calculation of the number of stages required for both
binary and multi component systems.

Vapor liquid equilibrium


The composition of the vapor in equilibrium with a liquid of given composition
is determined experimentally using an equilibrium still. The result are
conveniently shown on a temperature – composition diagram 16 - a , b, c).

In the normal case (1.a), the curve ABC shows the composition of the liquid
which boils at any given temperature and the curve ADC the corresponding
composition of the vapor at that temperature.

Thus, a liquid of composition x 1 will boil at T1 and the vapor in equilibrium is


indicated by point D of composition y1.

In Fig (16) there is a critical composition x g where the vapor has the same
composition as the liquid, so that no change occurs on boiling, such

critical mixtures are called azeotropes, and special methods are necessary to
effect separation. it is important to note that the composition of the vapor in
equilibrium with a given liquid will change with pressure.

89
Prof.Fatma K. Gad

Fig.(1)a-Benzene-Toluene b-Acetone-CS2 c-Acetone-Chloroform

Fig.(2) a-Benzene-Toluene b-Acetone-CS2 c-Acetone-Chloroform

Vapor composition a function of liquid composition at constant pressure.

For distillation purposes it is more convenient to plot y against x at constant


pressure as shown in fig (17 a,b,c). It must be remembered that the temperature
varies along each of the curves.

Partial vaporization and partial condensation

Suppose a mixture of benzene (B) and Toluene (T) to be heated in a vessel,


closed in such a way that the pressure remains atm, but no material can escape.
If x of the more volatile component (B) is plotted as abscissa, and the
temperature at which the mixture boils as ordinate, then the boiling curve is
obtained as shown by ABCJ fig(18). The corresponding dew point curve ADEJ

90
Prof.Fatma K. Gad

shows the temperature at which a vapor of composition y starts to condense.

If a mixture of composition x2 is at t3 below its boiling point T2 as shown by


point G on the diagram, then on heating at constant P the following changes
will occur.
1. When the T reaches T2, the liquid will boil (point B), and some vapor of
composition y2 shown by point E.

2. On further heating the composition of the liquid will change, because of the
loss of the more volatile component to the vapor. The boiling point will
therefore, rise to T’. At this point (T’1) the liquid will have a composition
represented by point L and the vapor by point No. and therefore, the change
in the proportion of liquid to vapor, the ratio being ;

Liquid / Vapor = MN / ML

3. On further heating to T1, all of the liquid is vaporized to give vapor D of the
same composition y1 as the originally.

It is seen that partial vaporization of the liquid gives a vapor richer in the more
volatile component that the liquid. If the vapor initially formed, as for instant
of point E, is at once removed by condensation, the liquid of composition x 3 is
obtained, represented by point C. The step BEC may be regarded as
representing an ideal stage, since the liquid passes from composition x 2 to a
liquid of composition x3, which is a greater enrichment in the more volatile
component than can be obtained by any other degree of vaporization.

Starting with superheated vapor represented by point H, on cooling to D


condensation will commence, and the first drop of liquid will have a
composition K. Further cooling to T’ will give liquid L and vapor N. Thus
partial condensation brings about enrichment of the vapor in the more volatile
component in the same manner as partial vaporization.

The industrial distillation column is essence a series of units in which these two
process of partial vaporization and partial condensation are affected
simultaneously.

91
Prof.Fatma K. Gad

Fig. (3) Effect of partial vaporization and condensation at B. P.

Partial pressure and Dalton’s Raoults and Henry’s laws


The partial pressure (PA) of component A in a mixture of vapor is the pressure
that would be exerted by component A at the same temperature, if present in the
same volumetric concentration as in the mixture.

Dalton’s law of partial pressure:-

The total pressure is equal to the summation of the partial pressures.

P = PA

Then, in an ideal mixture, the partial pressure is proportional to the mol fraction
of the consistent in the vapor phase;

(1)
PA = P y1

Henry’s law:-

92
Prof.Fatma K. Gad

At law concentration, a simple relation for the partial pressure developed by a


liquid solute A in a solvent.

PA = H x1 (2)

H is Henry’s constant.

According to this law, the p. pressure is directly proportional to the mol fraction
in liquid phase (xA). It is found that this holds only for dilute solution. Where,
H is a constant depending on temperature, on the solvent and to a minor degree
on pressure.

Raoult’s law.:-

When the solution is nearly pure A, the partial pressure of component A is


proportional to its concentration,

PA = P*A xA (3)

Where, PA is vapor pressure.

The above relation is usually found to be true only for high values of xA or
correspondingly law values of xB. If A is regarded as the solute and B as the
solvent, then Henry’s law applies when xA is small, and Raoult’s law when xB
is small.

If the mixture follows Raoults law, then P* of a mixture can be obtained


graphically from a knowledge of the V.P of the two components as in figure
(19). Where OA represents the P.P of A in a mixture and CB the PP of B and
the total pressure being shown by the line BA. Thus, in a mixture of
composition D the PP of A is given by DE, PB by DE and the total pressure (P)
by DG.

93
Prof.Fatma K. Gad

Fig. (4) Partial pressure of ideal mixture.

For a non ideal mixture fig (20), it will be found that over the range OC the
mixture follows Henry’s law and over BA it follows Raoults law.

If it is known that the mixture follows Raoult’s law, then the values of y A for
various values of xA may be calculated from the knowledge of the P (v.p) of the
true components at various temperatures thus,

PA = xA PoA , PA = P yA

So that yA = PoA xA / P , yB = PoB xB / P = PoB ( 1 - xA ) / P

Where yA + yB = 1

PoA xA / P + PoB ( 1 - xA ) / P = 1

Therefore, xA = ( P - PoB ) / ( PoA - PoB ) (4)

94
Prof.Fatma K. Gad

Fig. (5) Partial pressure of non-ideal mixture

Example:-

For a mixture of nC7 (A) and Toluene at 373° k and 1 atm (101.3 KN/m2)

P°A = 106 kN/M2

P°B = 73.7

xA = 101.3 – 73.7/106 – 73.7 = 0.856

 yA = P°A xA / P

yA = 106 * 0.856 / 101.3 = 0.896.

Vapor pressure can be determined at any temperature from Antonie equation

Log P° = A - B / (t + C) (5)

Where A, B and C are constants t is temperature in centigrade

95
Prof.Fatma K. Gad

Determination of boiling point diagram from Raoult’s law:-

It is necessary only to apply equations;

P°A xA + P°B ( 1 - xA ) = P (5)

yA = ( P°A / P ) xA & ( 1 - yA ) = P°B / P ( 1 - xA )

At several temperatures between the BP of pure component A and B and plot


curves of xA is T and yA vs the equilibrium curve which is simply a plot of x A
vs yA as shown in fig (21, 22).

Fig .(6) Boiling point construction diagram

96
Prof.Fatma K. Gad

Fig. (7) Equilibrium curve

Example:-

The vapor pressure of B and t are shown in the following table. Assuming that
mixture of these materials follow Raoult’s law, calculate and plot the B.P and
equilibrium curve for the system at a 1 atm.

Solution:-

Several temperatures between 80.6 (B , P of benzene) and 110.6° C (B.P of


toluene) are chosen and the corresponding data for P° A and P°B are substituted
A equation (6) .

Ex. At 85°C , P°A = 877 , P°B = 345 mm Hg

 877xA + 345 ( 1 - xA ) = 760  xA = 0.780 and consequently.

yA = 877/760 xA = ( 877 * 0.78 ) / 760 = 0.9

Then we plot xA vs yA and (xA , yA) vs T

v.p of benzene and toluene.

t°C B VP T
80.1 760 -

97
Prof.Fatma K. Gad

85 877 345
90 1016 405
95 1168 475
100 1344 557
105 1532 645
110 1748 743
110.6 1800 760

Relative volatility ()

The action between the composition of the vapor yA and if the liquid xA is
equilibrium can be also expressed in another way , which is particularly useful
in distillation calculations. If the ratio of the partial pressure to the mol fraction
in the liquid is defined as the volatility;

Volatility of A = PA / xA , Volatility of B = PB / xB

The ratio of these two volatilities is known as the relative volatility () given by

 = (PA xB) / (PB xA)

Substituting P yA for PA , P yB for PB

 = (yA xB) / (yB xA) (7)

yA / yB =  xA / xB

since yB = 1- yA , xB = 1 – xA.

yA =  xA / ( 1 + (  - 1 ) x A )
(8)
(8)
xA = yA / (  - (  - 1 ) y A )

98
Prof.Fatma K. Gad

For separation to be achieved  must be not equal 1. And as  increases above


unity the separation becomes much easier.

For an ideal system the volatility is numerically equal to the vapor pressure of
the pure component, thus the relative volatility () may be expressed as

 = P°A / P°B (9)

This is follows by applying PA / PB = yA / yB so that

  = PA xB / PB xA = P°A xA xB / P°B xB xA = P°A / P°B

 vary somewhat with temperature; as temperature falls,  rises.

In the distillation column, if the relative volatility at the top and bottom are less
than 15 percent apart an average value of  must be taken over the whole
column. If they differ by more that his figure, the equilibrium curve must be
constructed incrementally by calculating the relative volatility at several point
along the column.

Another frequently used relation between y , x as following.

yA = K. xA (10)

Where K is called equilibrium constant equal (P°A / P)

(II) The methods of distillation two components mixtures


For a binary mixture with a normal y-x fig (2-a) the vapor is always richer in
the more volatile component than the liquid from which it is formed.

There are four main methods used in distillation,

99
Prof.Fatma K. Gad

a) Differential distillation.
b) Flash or equilibrium distillation
c) Rectification
d) Batch distillation.

(I) (a) Differential distillation


The simplest example of batch distillation is a single stage, differential
distillation, starting with a still pot, initially full, heated at constant rate. In this
process the vapor formed on boiling the liquid is removed at once from the
system. Since this vapor is richer in the more volatile component than the
liquid, it follows that the liquid remaining becomes steadily weaker in this
component. Thus, whilst the vapor formed over a short period is in equilibrium
with the liquid, the total vapor formed is not in equilibrium with the residual
liquid.

Let S be the number of moles of material in the still and x be the mol fraction of
component A. Suppose an amount of ds containing a mol fraction y of A be
vaporized. Then a material balance on component A gives

y dS = d(Sx)
= Sdx + xdS

s x
 dS / S =  dx / ( y - x )
s0 x0

x
Ln (S ( intial ) / S ( final )) =  (dx / ( y - x ) (a-1)
x0

This equation is called Rayleigh equation. The integral on the right hand side
can be solved graphically if the equilibrium relationship between y and x is
available.

In some cases a direct integration is possible. Thus if over the range concerned
the equilibrium relationship is a straight line of the form :
y = mx+c.

100
Prof.Fatma K. Gad

so Ln ( S / So) = [ 1 / ( m-1 )] Ln (m-1) x + c / ( m - 1 ) xo + c

Si / So = [( y - x ) / ( yo - xo )]( 1 / ( m - 1 )) generally

( S / So )( m -1 ) = ( y - x ) / ( yo - xo ) (a-2)

Alternatively if the relative volatility may be assumed constant over the range
concerned, then :

y =  x / ( 1 + (  - 1 ) x ) can be substituted as follows


x
Ln ( S1(intial) / So(final) ) =  dx / {  x / [ 1 + (  - 1 ) x ]}
xo

or equal
xo
Ln ( So / S ) =  dx / [  x / ( 1 + (  - 1 ) x ) - x ]
x
xo
 Ln (So/S) =  [1+(-1) x] dx / [  x - x -  x2 + x2 ]
x

xo
=  [ ( 1 +  x - x ) / (  - 1 ) ( 1 - x ) x ] dx
x
xo
Ln ( So / S ) = 1 / (  - 1 )  [ ( 1 +  x - x ) / ( 1 - x ) x ]dx
x

xo
= 1 / (  - 1 )  [( 1 / x ) + (  / ( 1 - x )) ]dx
x

Ln (So/S) = 1 / (-1) [Ln (xo/x) -  Ln (1-xo)/(1-x)] (a-3)

Also

 Ln ( So / S ) – Ln ( So / S ) = Ln ( xo / x ) +  Ln ( 1 - x ) / (1-xo )

101
Prof.Fatma K. Gad

Ln ( S / So ) + Ln ( x / xo ) =  Ln (( 1 - x ) / ( 1 - xo )+ Ln ( S / So )

Ln (S x / (So xo)) =  Ln ( S (1-x)/( So ( 1 - xo )) (a-4)


for binary mixture of A and B

Ln A(original) / A(final) = AB Ln B(original) / B(final)

for multi component


since  xi = 1

Ln ( S xi ) / ( So xio ) = iB Ln ( S xB ) / ( So xBo ) (a-5)

As this process consists of only a single stage a complete separation is


impossible unless the relative volatility is infinite.

Example
A liquid containing 50 mol % benzene and 25 mol % toluene, 25 mol % o-
xylene is differentially distillate at 1 atm with vaporization of 32.5 mol% of the
charge. Assume Roult’s law applied compute the distillate and residue
composition.

BT = 2.49 , xT = 0.364


Solution.

Ln ( S xi ) / ( So xio ) = iT Ln ( S xT ) / ( So xTO )


For benzene
Ln 100 * 0.5/67.5 * xBo = 2.49 Ln 100* 0.25/67.5 xTO (1)

Ln 100 * 0.25/67.5 xo = 0.364 Ln 100* 0.25/67.3 xTO (2)

 xBo + xTo + xxo = 1 (3)

By solving the three equation we yet xBO = 0.385 , xTo


= 0.285 , xxo = 0.331
Making material balance for each component.
S xi = So xio + D xid
xBd = 0.792
xTd = 0.178

102
Prof.Fatma K. Gad

xxd = 0.07S
II (B) Flash or equilibrium distillation

This method is frequently carried out as a continuous process. the feed is


passed into the still where part is vaporized, the vapor remaining in contact with
the liquid. The mixture of vapor and liquid leaves the still and is separated so
that the vapor is in equilibrium with the liquid.

An analysis of the process can be given as below; for a binary mixture of A and
B.

Let V be the number of moles of vapor formed and y mol fraction of A


F is moles of feed of mol fraction xF of A.
S moles of liquid and x mol fraction of A.

Then material balance on component A gives.

F xF = V y + S x
F=V+S
V / F = ( xF – x ) / ( y - x )

y = ( F/V ) xF – [ ( F / V ) - 1 ] x (b-1)

Equation (b-1) represent a straight line of slope – ( F - V ) / V


= - S / V passing through the point ( x , xF )

The value of x and y required must satisfy not only the above equation
(operating line) but also the appropriate equilibrium data as shown in the
following figure.

So that equilibrium equation for any component

yi (for component i in the distillate V) = Ki xi ( in liquids)


yi in vapor (v) = Ki xi (in liquids)

Note that the richest vapor its composition will vary between its babble and
dew points depending on the relative amount of S/V.

103
Prof.Fatma K. Gad

Fig. (8) Diagram for solution of flash distillation.

Since the separation obtained in ordinary cases is poor (x , x F) or So; such


operations generally used only on mixtures of large () or in conjunction with
other distillation operation.

All the previous equations apply equally well to the use of equilibrium
condensation. Where a vapor feed is partially condensed to yield the two
products.

Heat requirement is calculated from a heat balance on the system.

Q = Shs + VHv – Fhf - H (b-2)

Where
hf ,hs , Hv are the enthalpies of feed, vapor and liquid respectively
H is heat of solution (equal zero for ideal)
Since for binary mixture of A and B

hs = [ CLA MA x + CLB MB ( 1 - x )] ( tL - to )

HV = y ( CLA MA ( tg - to ) + A MA] + (1-y) [CLB MB ( tg - to ) + B MB]


Where CLA , CLB are the heat capacities of component A , B
A , B are the latest heat of vaporization of A, B
MA , MB are the molecular weights of A, B
tL , tg are temp of liquid and vapor

104
Prof.Fatma K. Gad

to is datum temp. usually is taken to that of feed.

So that tF = Zero in equation (b-2)

So for mixture which form ideal solution

( yi )v = Ki ( xi )w

F xif = V yi + S xi (for each component)

F = V+ S

S / V = ( yi – xiF ) / ( xiF – xi )

S/V = (ki xi–xiF)/(xiF–xi)=(yi–xiF)/(xiF–(yi/ki)) (b-3)

This provides the following:

1) for vaporization Where  yi = 1

yi = ( xiF ( 1 + (S / V) ) / ( 1 + (S / V ki) ) (b-4)

2) for condensation Where  xi = 1

xi = ( xiF ( (S / V) +1 ) / ( ki +(S / V) ) (b-5)

In such case S/V would have to be assumed and calculate yi , xi and the correct
value of S/V is that which gives  yi = 1 or  xi = 1.

This is useful in calculating the bubble point and dew point.

Bubble and dew point of mixture.


Bubble point, is the temp. at the first bubble of vapor formed by heating the
mixture at constant pressure where;

 yi = 1
105
Prof.Fatma K. Gad

(b-6)

Dew point is the temp. at the first droplet of liquid condenses when we cool the
vapor mixture at constant pressure where;

 xi = 1 (b-7)

Example
A liquid mixture containing 50 % (mol) C7 and 50% C8 at 80°F is to be
continuously flash vaporized at 1 atm to vaporize 60 mol % of the feed. What
will be the composition of the vapor and liquid, the temperature in the separator
and the heat requirement.
Vapor pressure data of C7 , C8 as follows,

t°c 98.4 105 110 115 120 125.6

V.pH 760 940 1050 1200 1250 1540

V.po 333 417 430 561 650 760

Heat capacity of CH , C8 in the range of 80 - 233 °C are

CLH = 0.507, CLO = 0.505 BTU /K °F

MH = 100 , Mo = 114

H AT 235 °f = 133 , o = 134 BTU/LB

And then plot S / V function of


(1) temp in the separator
(2) composition of liquid leaving the separator.

Solution

Slope = S/V = - 40/60 = -2/3

ToC VPH VPo P-Po PH – Po xH PH/P yH


98.4 760 333 427 427 1-0 1 1
105 940 412 343 523 0.655 1.23 0.81
110 1050 984 276 566 0.487 1.38 0.674

106
Prof.Fatma K. Gad

115 1200 561 149 639 0.312 1.57 0.492


120 1350 650 110 700 0.1571 1.77 2.279
125.8 1540 760 00 780 0 2.02 0

Therefore, yv = 0.575 , xs = 0.387 for S / V = 40/60

And temp = 113°C = 235.4 °F

Since yi = xiF ( 1 + ( S / V )) / (( 1 + ( S / V Ki ))

xi = xiF (( S / V ) + 1 ) / ( Ki + ( S / V ))

By heat balance taking to = 80 °F

  Hsolution = 0 , HF = 0

hS = [CLH MH x + CLO MO ( 1 - x ) ] ( t - to )
= [0.57 * 100 * 0.387 + 0.505 * 114 * 0.613][235-80]
= 8.550 BTU/lb

HV = yH [CLH MH ( tg - to ) + H MH] + (1-y)[CLO MO ( Tg - TO ) + O HO )


=00.575[0.507*100(235.9-80)+133*100]
+0.425[0.505*114(2354.80+134*114]
= 22500 BTU/lb mol.

Q = 60 * 22500 + 40 * 8550 = 1692000 BTU/LB MOL feed.

Example 2:

Find the bubble point of the following liquid, where the pressure is 1 atom.

Composition 140°F 130 °F 125 °F


xi Ki Ki xi Ki Ki xi Ki Ki xi
C4 0.05 5.6 0.28 5.0 0.25 4.7 0.235
C5 0.17 2.0 0.30 1.8 2.305 1.7 0.289
C6 0.65 0.77 0.5 0.67 0.435 0.6 0.39
C7 0.1 0.29 0.029 0.24 0.24 0.22 0.022

107
Prof.Fatma K. Gad

C8 0.03 0.1 0.003 0.07 0.002 0.06 0.002


1.016 0.938

Solution

Since at the bubble point yi = 1


So bubble point lay between 130 and 125

 tBP = 125 +(130 – 125) (1-0.938)/(1.016 – 0.938) = 129 °F or

tBP = 130 – (130 – 125) (1.016 – 1)/(1.016/0.938) = 129 °F.

Also yi at bubble point = (ki xi) (ki xi at bubble point)/(ki xi at any temp)

Example 3:

y Ki at 140°F K at 142°F
C4 0.145 5.6 5.7
C5 0.285 2.0 2.1
C6 0.530 2.0 0.8
C7 0.035 0.29 0.3
C8 0.005 0.1 0.11

Solution

Since  yi / ki = xi

so at 14°f yi / ki = 1.028


at 102 yi / ki = 0.988

Since at dew point xi = 1


So dew point of the mixture lies between 140 , 142 °F
So at xi = 1 =  yi / ki = 140 +2 * 0.028/0.04 = 141.4°F. U.E.D
 = M wt / 22.4 or M wt / 359

FOR KNOWING THE FEED STATE:

1) Mole fraction of feed = z ;

108
Prof.Fatma K. Gad

Assume that feed is vapor z = y , then


Assume once again that feed is liquid z = x
Get dew and bubble point for Σ x , Σ y
If
Σ x > 1 system below dew point
Σ y < 1 system below bubble point
Σ x < 1 system above dew point
Σ y > 1 system below bubble point

If system Σ x >1 and Σ y <1 then the system all liquid

2) Bubble point pressure Σ y = 1 = 1 / π Σ Po x π = Σ Po x ;

Dew point pressure Σ x = 1 = y / k = π Σ y / Po 1/π = Σ y/ Po

3)Calculate Po at T by knowing two values of V.P. at known temperature

Ln Po = C + D / T mmHg
Where
D = T1 T2 / ( T1 – T2 ) Ln ( Po2 / Po1 )
C = Ln P2 – ( D / T2 )

CALCULATION OF FLASH UNITS :-

There are three procedures of calculation for simple phase separation unit
(flash unit) as shown in the fig.

U k f1

Ukin

Ukf2

Some definitions :-

First define an overall split fraction, (k )

Φ = V/F (1)

ξn = Ө / (1 + Ө) (2)

109
Prof.Fatma K. Gad

Ө = kn Φ / (1 – Φ) (3)

( ξk ) = (Vk / Fk) (4)

For each non-key component k

( ξk ) = (αk/n) (ξn) / (1 + (αk/n -1) ξn ) (5)

Where component ( n ) defined as key component.

So the variable n ,  , P , T and Q ( heat supplied to flash unit ) can be


specified. Where

Lk = (1 – ξk) fk
(6)

 Lk = L ,  Vk = V and Fk = Lk + Vk ( 1 )

( Vk / V ) = k ( Lk / L ) (7)

Where k is equilibrium constant (2)

So for a specified feed, the number of variables less the number of equations
equals to 2, which is the degree of freedom.

So this means that we can completely specify the condition of the flash unit by
selecting two of the variables, ( we defer specification on Q ); we will get the
following procedures;

Case 1 :  specified and T or P specified

Case II : T and P specified ( isothermal flash )

Case iii ;  specified and T or P specified.

110
Prof.Fatma K. Gad

Example to show the difference between the cases:

Consider the mixture with the components, flow rates, boiling point and
Antoine coefficient

Componen Fk mole/hr B.P ok Ak Bk Ck


t
Benzene 30 353 15.9008 2788.51 -52.34
Toluene 50 383 16.0137 3096.52 -53.67
O-xylene 40 418 16.1156 3395.57 -59.44

Solution

Key component ( n ) choose Toluene ( 2 )

Case I : Fixed 2 = 0.9 and P = 1 bar. ( calculate temp. )

Assuming k/n constant over the temperature range.

( 1 ) Specified 2 = 0.9 , P = 1 bar and guess T = 390 ok

( 2 ) Calculate relative volatilities k/n = Pok / Pon


1/2 = 2.305 , 3/2 = 0.381

( 3 ) Calculate recoveries of non key component,


k = k/n n / ( 1 + ( k/n – 1 ) n )

for each component 1 , 3 : 1 = 0.954 , 3 = 0.774

( 4 ) Solve for mass balance and evaluate mole fraction;

Vk = fk k and Lk = ( 1 - k )fk
Yk = Vk / Vi and xk = Lk / Li

So

V1 = 30 * 0.954 = 28.62 L1 = 30 -28.62 = 1.38


xi = 1.38 / ( 1.38 + 9.04 )
V2 = 50 * 0.9 = 45 L2 = 50 – 45 = 5 x2 = 0.324
V3 = 40 * 0.774 = 30.96 L3 = 40 – 30.96 = 9.04 x3 = 0.586

111
Prof.Fatma K. Gad

( 5 ) Evaluate the bubble point :

P = ( ’ / k/n ) * Pok ( at flash T )


Where ’ =  i/n xI

So for Po3 = P k/n / ’ = 750 * 0.381 / 0.752 = 380


mm Hg

The temperature at which Po3 = 380 mm Hg = 393 ok close to the suggested one.

Case II : P and T specified :-

( 1 ) Temp. 390 ok , P = 1 atm , guess 2 = 0.9

( 2 ) From Antoin equation determine vapor pressure at 390 ok ,

Ln Pok = Ak – Bk / ( Ck + T )

k/n = Pok / Pon

1/2 = 3682.8 / 904.1 , 2/3 = 344.7 / 904.1 = 0.381


( 3 ) Recoveries of k component:

1 = k/n n / (1 + (k/n – 1) n) = 2.305 * 0.9 / (1+(2.305–1)*0.9) = 0.954

3 = 0.381 * 0.9 / ( 1 + ( 1.619 – 1 ) * 0.9 ) = 0.774

( 4 ) Mass balance :

V1 = 30 * ( 0.954 ) = 28.62 L1 = 1.38 x1 = 0.89


V2 = 50 * ( 0.91 ) = 45 L2 = 5 x2 = 0.624
V3 = 40 * ( 0.774 ) = 30.96 L3 = 9.04 x3 = 0.586

( 5 ) Check bubble point :

Use component ( 3 ) ’ = P 3/2 / Po3 = 750 * 0.381 / 344.7 = 0.82

But ’ =  xf = 0.752 which is far from 0.82

So we re-guess 2 = 0.8 instead of 0.9 and repeat the procedure;

112
Prof.Fatma K. Gad

So 1 = 0.902 , 3 = 0.604

L1 = 2.94 , x1 = 0.102
L2 = 10 , x2 = 0.347
L3 = 15.84 , x3 = 0.55

( 6 ) ’ = 0.792 So by referring to component 3 ,


’ = P 3/2 / Po3 = 0.829 (close enough)

Case III : Where  and P is known :

( 1 )  = V / F = 0.8 , P = 1 bar guess T = 390 ok

( 2 ) Evaluate equilibrium constant ( k ) then relative volatility

1/2 = 2.305 , k1 = 2.778


2/2 = 1.0 , k2 = 1.205
3/2 = 0.381 , k3 = 0.460

( 3 )  quantities = kn  / ( 1 -  ) = Vn / Ln

and n =  / ( 1 +  ) ( n = 2 toluene )

So  = ( 1.205 ) * ( 0.8 ) / ( 1 – 0.8 ) = 4.82

n = 0.828

( 4 ) Non key recoveries :

1 = ( 2.305 ) * ( 0.828 ) / [ 1 + ( 1.305 ) * ( 0.828 ) ] = 0.917 ,


3 = 0.647

( 5 ) Mass balance :

V1 = 27.5 , V2 = 41.4 , V3 = 75.9


L1 = 2.5 , L2 = 8.6 , L3 = 14.1
x1 = 0.0999 , x2 = 0.341 , x3 = 0.560

( 6 ) ’ =  k/n x2 = 0.782

113
Prof.Fatma K. Gad

So Po3 = P k/n / ’ = 365.4 mm Hg at 391.9 close to 390.

III ( C ) BATCH DISTILLATION


In these processes the whole of a batch is run into the boiler of the still and, on
heating the vapor is passed into fractionation column as indicated in fig (24).
As with continuous distillation the composition of the top product will depend
on the still composition and the number of plates in the column and on the
reflex ration used.

As the process proceeded, the liquid remaining in the still will become steadily
weaker in the M.V.C

Thus the still may be changed with S1 moles of a mixture containing xS1 of
M.V.C. Initially with reflex ration R1 the top product has a composition x d1. If
after a contain interval of time the composition of the top product starts to fall,
then if the reflex is increased to new value R 2, it will be possible to obtain the
some composition again. This method required a continuous increase in R to
maintain a constant quality of the top product.

Fig. (III-1) Batch distillation column

114
Prof.Fatma K. Gad

An alternative method is to work with a constant R and allow the composition


of the top product to fall.

Batch distillation is useful for handling small quantities of multi component,


since it is possible to obtain the different components at reasonable degree of
purity one after the other.

To obtain the maximum recovery of a valuable component, the charge


remaining in the still after first distillation may be added to the next batch.

1) Operation at constant product composition:-

Suppose a column of four ideal plates will be used to separate a mixture of


EtOH and H2O. Initially there are in the still S1 moles of liquor of mol
fraction xS1 with respect to the M.V.C (EtOH). The top product is to
contain xd and this needed a R1. Suppose the distillation is to be continued
till there are S2 moles in the still with xS2. Then for the same no. of plates
the reflex ratio will be increased to R2.

Therefore, S1 – S2 = Db
S1 xS1 – S2 xS2 = Db xd

D (product) = S1 [ ( xS1 – xS2) / ( xd – xS2) ]= ( a / b ) S1 (III-1)

Where a and b are as shown in Fig (III-2).


If  is the intercept on the y axis for any operating line then,

 = xd / ( R + 1 ) or R = ( xd /  ) - 1

On comparing the operation at constant R with that of constant x d is the


difference in the amount of steam used in the distillation for a given quantity D b

115
Prof.Fatma K. Gad

Fig. (III2) Graphical solution of batch distillation.

If R is assumed to be adjusted to keep x d constant, than at any moment R is give


by

R = dLp / dDp and the total reflex flowing down the column is given by

Lb R2
 dLb =  R dDb (2)
0 R1

to provide reflex dLb requires the removal of heat  dLb in the condenser, where
 is the latent heat per mol. Thus the heat to be supplied in the reboiler Q R to
provide this reflex during the total distillation in given by

Lb R2
QR =   DLb =   R dDb
0 R1
(III-3)

116
Prof.Fatma K. Gad

The Rm can be calculated from Underwood and Fenske equations.

2. Operation at constant reflux ratio

If the same column as before is operated at a constant R, the concentration of


the MVC in the top product will continuously fall. Over a small interval of
time dt, xd becomes xd – dxd. If in this time the amount of product obtained is
dDb ; material balance on MVC gives.

MVC in the top product = dDb [ xd - dxd / 2] = xd dDb


And xd dDb = - d( S xS ) = - S dxS – xS dS
But dDb = - dS , S dxS = dS ( xd - xS )

s2 xS2
And  dS / S =  dxS / ( xd - xS )
s1 xS1

Finally
xS1
Ln ( S1 / S2 ) =  dxS / ( xd - xS ) (4)
xS2

The right hand side of this equation can be integrated by plotting (1/(x d-xS)) vs
xS. This will enable the ratio of the initial to final quantity in the still to be
found for any desired charge in xS and hence the amount of distillate Db.

The heat to be supplied to provide the reflux will be

QR =  R Db (5)

117
Prof.Fatma K. Gad

Fig Batch distillation at constant reflux ratio.

Example:(I)

A mixture of EtOH and H2O with 0.55 mol fraction of a alcohol is distilled to
give a top product of 0.75 mol fraction EtOH. The column has 4 ideal plates
and the distillation is stopped when the reflex ratio has to be increased beyond
4.0.
Find the amount of distillate and heat required / mol product.

Solution

Assuming various values of R between 0.85 and 4, we get the corresponding


values of the intercept  and the concentration in the still as shown in the table
below.

R  xS Db
0.85 0.405 0.55 0
1.0 0.375 0.50 20.0

118
Prof.Fatma K. Gad

1.5 0.3 0.37 47.4


2.0 0.25 0.20 63.8
3.0 0.187 0.075 70.5
4.0 0.15 0.05 71.9

The amount of product for R = 4 calculated as

Db = 100 [0.55 – 0.05/0.75 – 0.05] = 100 * 0.5/0.7 = 71.4 kmol

R=4
The value  R dDb from the plot = 96 kmol
R=0.85

Taking and average latent heat per k mol for the EtOH – H2O mixture as 4000
kJ.
The heat to be supplied
QR = (96 * 4000)/1000 = 380 MJ
QR / Db = 380/71.4 = 5.32 MJ U.E.D

Example (ii)

If the same batch as above is distilled with constant R of 2.1. What will be the
heat required and the average composition of the distillate, if the distillation is
stopped when the composition in the still has fallen to 0.105 mol fraction of
MVC.

Solution:

The initial composition of the top product will be 0.78 as shown in the
following table and the final composition will be 0.74. values of xd, xS, xd-xS
and 1 / (xd-xS ) for various values of xS and a constant reflex ratio are given in
the following table.

xS xd xd-xS 1/(xd-xS)
0.55 0.780 0.230 4.35
0.5 0.775 0.275 3.65
0.425 0.770 0.345 2.90
0.310 0.766 0.450 2.22
0.225 0.750 0.525 1.91
0.105 0.740 0.635 1.58

119
Prof.Fatma K. Gad

Fig of Example 1

Values of xS and 1 / ( xd - xS ) are plotted as shown

0.55
For which  dxS / ( xd - xS ) = 1.1
0.105

Ln ( S1 / S2 ) = 1.1 , ( S1 / S2 ) = 3.0

Hence Db = S1 – S2 = 100 – 100/3 = 66.7

Amount of EtOH in the product = x1 S1 – x2 S2


= 0.55*100–0.105*33.3 = 51.5 k mole.
Average composition of product is = 51.5/66.7 = 0.77 and

The heat required = 4000 * 2.1 * 656.7 = 560 380 KJ

The heat required per kmol product = 360380/66.7 = 8400 KJ U.E.D

120
Prof.Fatma K. Gad

Fig. for example 2

DISTILLATION WITH REFLUX AND McCABE-


THIELE METHOD

3.4A Introduction to Distillation with Reflux


Rectification (fractionation) or stage distillation with reflux, from a simplified
point of view, can be considered to be a process in which a series of flash-
vaporization stages are arranged in a series in such a manner that the vapor and

121
Prof.Fatma K. Gad

liquid products from each stage flow countercurrently to each other. The liquid
in a stage is conducted or flows to the stage below and the vapor from a stage
flows upward to the stage above. Hence, in each stage a vapor stream V and a
liquid stream L enter, are mixed and equilibrated, and a vapor and a liquid
stream leave in equilibrium. This process flow diagram was shown in Fig. 3.3-1
for a single stage and an example given in Example 3.2-1 for a benzene-toluene
mixture.

For the countercurrent contact with multiple stages, the material-balance or


operating-line equation (3.2-1) was derived which relates the concentrations of
the vapor and liquid streams passing each other in each stage. In a distillation
column the stages (referred to as sieve plates or trays) in a distillation tower are
arranged vertically, as shown schematically in Fig. 3.4-1.

The feed enters the column in Fig. 3.4-1 somewhere in the middle of the
column. If the feed is liquid, it flows down to a sieve tray or stage. Vapor enters
the tray and bubbles through the liquid on this tray as the entering liquid flows
across. The vapor and liquid leaving the tray are essentially in equilibrium. The
vapor continues up to the next tray or stage, where it is again contacted with a
downflowing liquid. In this case the concentration of the more volatile
component (the lower-boiling component A) is being increased in the vapor
from each stage going upward and decreased in the liquid from each stage
going downward. The final vapor product coming overhead is condensed in a
condenser and a portion of the liquid product (distillate) is removed, which
contains a high concentration of A. The remaining liquid from the condenser is
returned (refluxed) as a liquid to the top tray.

The liquid leaving the bottom tray enters a reboiler, where it is partially
vaporized, and the remaining liquid, which is lean in A or rich in B, is
withdrawn
as liquid product. The vapor from the reboiler is sent back to the bottom stage
or
tray. Only three trays are shown in the tower of Fig. 3.4-1. In most cases the
number of trays is much greater. In the sieve tray the vapor enters through an
opening and bubbles up through the liquid to give intimate contact of the liquid
and vapor on the tray. In a theoretical tray the vapor and liquid leaving are in
equilibrium. The reboiler can be considered as a theoretical stage or tray.

122
Prof.Fatma K. Gad

Figure 4.4-1: Process flow of fractionating tower containing sieve trays.

3.4B McCabe-Thiele Method of Calculation for Number


of Theoretical Stages
1. Introduction and assumptions. A mathematical-graphical method for
determining the number of theoretical trays or stages needed for a given
separation of a binary mixture of A and B has been developed by McCabe and
Thiele. The method uses material balances around certain parts of the tower,
which give operating lines, and the xy equilibrium curve for the system.

The main assumption made in the McCabe—Thiele method is that there must
be equimolar overflow through the tower between the feed inlet and the top tray
and the feed inlet and bottom tray. This can be shown in Fig. 3.4-2, where
liquid and vapor streams enter a tray, are equilibrated, and leave. A total
material balance gives

Vn+1 + Ln-1 = Vn + Ln (3.4-1)

A component balance on A gives

Vn+1 yn+1 + Ln-1 xn-1 = Vn yn + Ln xn (3.4-2)

where Vn+1 is mol / h of vapor from tray n+1, Ln is mol/h liquid from tray n, yn+l
is mole fraction of A in Vn+1 and so on. The compositions yn and xn are in
equilibrium and the temperature of the tray n is T n. If Tn is taken as a datum, it
can be shown by a heat balance that the sensible heat differences in the four
streams are quite small if heats of solution are negligible. Hence, only the latent
heats in stream Vn+1 and Vn are important. Since molar latent heats for

123
Prof.Fatma K. Gad

chemically similar compounds are almost the same, V n+l = Vn and Ln = Ln-1.
Therefore, we have constant molal overflow in the tower.

2. Equations for enriching section. In Fig. 3.4-3 a continuous distillation


column is shown with feed being introduced to the column at an intermediate
point and an overhead distillate product and a bottoms product being
withdrawn. The upper part of the tower above the feed entrance is called the
enriching section, since the entering feed of binary components A and B is
enriched in. this section, so that the distillate is richer in A than the feed. The
tower is at steady state.

An overall material balance around the entire column in Fig. 3.4-3 states that
the entering feed of F mol / h must equal the distillate D in mol/h plus the
bottoms W in mol/h.

F=D+W (3.4-3)

A total material balance on component A gives

F xF = D x D + W x w (3.4-4)

Figure 3.4-2. Vapor and liquid flows entering and leaving a tray.

124
Prof.Fatma K. Gad

Figure 3.4-3: Distillation column showing material balance section for McCabe
thiele method.

In Fig. 3.4-4a the distillation tower section above the feed, the enriching
section, is shown schematically. The vapor from the top tray having a
composition yl passes to the condenser, where it is condensed so that the
resulting liquid is at the boiling point. The reflux stream L mol/h and distillate
D mol/h have the same composition, so yl = xD. Since equimolal overflow is
assumed, L1 = L2 = Ln and V1 = V2 = Vn = Vn+l.

Making a total material balance over the dashed-line section in Fig. 4.4-4a,

Vn+l = Ln + D (3.4-5)

Figure 3.4-4: Material balance and operating line for enriching section:
(a)Schematic of tower (b)Operating and equilibrium lines.

125
Prof.Fatma K. Gad

Making a balance on component A,

Vn+1 yn+1 = Ln xn + D xD (3.4-6)

Solving for yn+1( the enriching-section operating line is

yn+1 = (Ln / Vn+1) xn + D xD / Vn+1 (3.4-7)

Since Vn+1 = Ln + D, Ln / Vn+l = R / (R + 1) and Eq. (3.4-7) becomes

yn+1 = (R / (R + 1)) xn + xD / (R + 1) (3.4-8)

where R = L / D = reflux ratio = constant. Equation (3.4-7) is a straight line on


a plot of vapor composition versus liquid composition. It relates the
compositions of two streams passing each other and is plotted in Fig. 3.4-4b.
The slope is L / Vn+l or R/(R+1), as given in Eq. (4.4-8). It intersects the y = x
line (45° diagonal line) at x = xD. The intercept of the operating line at x = 0 is y
= xD/(R+1).

The theoretical stages are determined by starting at x D and stepping off the first
plate to x1. Then y2 is the composition of the vapor passing the liquid x 1. In a
similar manner, the other theoretical trays are stepped off down the tower in the
enriching section to the feed tray.

3. Equations for stripping section. Making a total material balance over the
dashed-line section in Fig. 3.4-5a for the stripping section of the tower below
the feed entrance,

Vm+1 = Lm – W (3.4-9)

Making a balance on component A,

Vm+l ym+1 = L,m xm – W xW (3.4-10)

Solving for ym +the stripping-section operating line is

ym+1 = (Lm / Vm+1) xm – W xW / Vm+1 (3.4-11)

Again, since equimolal flow is assumed, Lm = LN = constant and Vm+1 = VN =


constant. Equation (3.4-11) is a straight line when plotted as y versus x in Fig.

126
Prof.Fatma K. Gad

4.4-5b, with a slope of L/Vm+l. It intersects the y = x line at x = xw. The intercept
at x = 0 is y = - W xw / Vm+l.

Figure 3.4-5. Material balance and operating line for stripping section:
(a) schematic of tower, (b) operating and equilibrium lines.

Again the theoretical stages for the stripping section are determined by starting
at xw, going up to yw, and then across to the operating line, etc.

4. Effect of feed conditions. The condition of the feed stream F entering the
tower determines the relation between the vapor V m in the stripping section and
Vn in the enriching section and also between L m and Ln. If the feed is part liquid
and part vapor, the vapor will add to Vm to give Vn.

For convenience, we represent the condition of the feed by the quantity q,


which is defined as

heat needed to vaporize 1 mol of feed at entering conditions


q= (3.4-12)
molar latent heat of vaporization of feed

If the feed enters at its boiling point, the numerator of Eq. (3.4-12), is the same
as the denominator and q = 1.0. Equation (3.4-12) can also be written in terms
of enthalpies.

q = Hv – Hf / Hv – HL (3.4-I3)

127
Prof.Fatma K. Gad

where Hv is the enthalpy of the feed at the dew point, H L the enthalpy of the
feed at the boiling point (bubble point), and HF the enthalpy of the feed at its
entrance conditions. If the feed enters as vapor at the dew point, q = 0. For cold
liquid feed q > 1.0, for superheated vapor q < 0, and for the feed being part
liquid and part vapor, q is the fraction of feed that is liquid.

We can look at q also as the number of moles of saturated liquid produced on


the feed plate by each mole of feed added to the tower. In Fig. 3.4-6 a diagram
shows the relationship between flows above and below the feed extrance. From
the definition of q, the following equations hold:

Lm = Ln + q F (3.4-14)

Vn=Vm + (1-q) F (3.4-15)

Figure 3.4-6. Relationship between flows above and below the feed entrance.

The point of intersection of the enriching and the stripping operating-line


equations on an x y plot can be derived as follows. Rewriting Eqs. (3.4-6) and
(4.4-10) as follows without the tray subscripts:

Vn y = Ln x + D xD (3.4-16)

Vm y = Lm x – W xw (3.4-17)

where the y and x values are the point of intersection of the two operating lines.
Subtracting Eq. (3.4-16) from (3.4-17),

(Vm - Vn) y = (Lm - Ln) x - (D xD + W xw) (3.4-18)

Substituting Eqs. (3.4-4), (3.4-14), and (3.4-15) into Eq. (3.4-18) and
rearranging,

128
Prof.Fatma K. Gad

y = (q / (q – 1)) x – (xF / (q – 1)) (3.4-19)

This equation is the q-line equation and is the locus of the intersection of the
two operating lines. Setting y = x in Eq. (3.4-19), the intersection of the q-line
equation with the 45° line is y = x = x f, where xF is the overall composition of
the feed.

In Fig. 3.4-7 the q line is plotted for various feed conditions given below the
figure. The slope of the q line is q/(q — 1). For example, for the liquid below
the boiling point, q > 1, and the slope is > 1.0, as shown. The enriching and
operating lines are plotted for the case of a feed of part liquid and part vapor
and the two lines intersect on the q line. A convenient way to locate the
stripping operating line is to first plot the enriching operating line and the q
line. Then draw the stripping line between the intersection of the q line and
enriching operating line and the point y = x = x w.

5. Location of the feed tray in a tower and number of trays. To


determine the number of theoretical trays needed in a tower, the stripping and
operating lines are drawn to intersect on the q line as shown in Fig. 3.4-8.
Starting at the top at xD,, the trays are stepped off. For trays 2 and 3, the steps
can go to the enriching operating line, as shown in Fig. 3.4-8a. At step 4 the
step goes to the stripping line. A total of about 4.6 theoretical steps are needed.
The feed enters on tray 4.

For the correct method, the shift is made on step 2 to the stripping line, as
shown in Fig. 3.4-8b. A total of only about 3.7 steps are needed with the feed
on tray 2. To keep the number of trays to a minimum, the shift from the
enriching to the stripping operating line should be made at the first opportunity
after passing the operating-line intersection.

129
Prof.Fatma K. Gad

Figure 3.4-7: Location of the q line for various feed conditions:


Liquid below boiling point (q > 1) Liquid at boiling point (q = 1)
Liquid + vapor (0 < q < 1) Saturated vapor (q = 0)

In Fig. 3.4-8b the feed is part liquid and part vapor since 0 < q < 1. Hence, in
adding the feed to tray 2, the vapor portion of the feed is separated and added
beneath plate 2 and the liquid added to the liquid from above entering tray 2. If
the feed is all liquid, it should be added to the liquid flowing to tray 2 from the
tray above. If the feed is all vapor, it should be added below tray 2 and joins the
vapor rising from the plate below.

Since a reboiler is considered a theoretical step when the vapor y w is in


equilibrium with xw as in Fig. 3.4-5b, the number of theoretical trays in a tower
is equal to the number of theoretical steps minus one.

EXAMPLE 3.4-1. Rectification of a Benzene-Toluene Mixture

A liquid mixture of benzene-toluene is to be distilled in a fractionating tower at


101.3 kPa pressure. The feed of 100 kg mol/h is liquid and it contains 45 mol %
benzene and 55 mol % toluene and enters at 327.6 K (130°F). A distillate
containing 95 mol % benzene and 5 mol % toluene and a bottoms containing 10
mol % benzene and 90 mol % toluene are to be obtained. The reflux ratio is 4:
1. The average heat capacity of the feed is 159 kl/kg mol • K (38 btu/lb mol •
°F) and the average latent heat 32 099 kI/kg mol (13 800 btu/lb mol).
Equilibrium data for this system are given in Table 4.1-1 and in Fig. 4.1-1.
Calculate the kg moles per hour distillate, kg moles per hour bottoms, and the
number of theoretical trays needed.

Solution:

The given data are

F = 100 kg mol/h,
xF = 0.45,
xD = 0.95,
xw = 0.10, and
R = L/D = 4.
For the overall material balance substituting into Eq. (3.4-3),

F=D+W (3.4-3)

130
Prof.Fatma K. Gad

100 = D + W

Substituting into Eq. (3.4-4) and solving for D and W,

F xf = D x D + W x w (3.4-4)

100 (0.45) = D (0.95) + (100 - D) (0.10)


D = 41.2 kg mol/h W = 58.8 kg mol/h
For the enriching operating line, using Eq. (4.4-8),
Yn+1 = (R/(R+1)) xn+xD/(R+1) = (4/(4+1)) xn + 0.95/(4+1) = 0.8 xn + 0.19

FIGURE 3.4-8. Method of stepping off number of theoretical trays and location
of feed plate:[a) improper location of feed on tray 4,(b) proper location of feed
on tray 2 to give minimum number of steps.

The equilibrium data from Table 3.1-1 and the enriching operating line above
are plotted in Fig. 3.4-9.

Next, the value of q is calculated. From the boiling-point diagram, Fig. 3.1-1,
for xF = 0.45, the boiling point of the feed is 93.5°C or 366.7 K (200.3°F). From
Eq. (3.4-13),

q = (Hv –Hf) / (Hv – HL) (3.3-13)

The value of Hv — HL = latent heat = 32 099 kJ/kg mol. The numerator of Eq.
(4.4-13) is

131
Prof.Fatma K. Gad

Hy — HF = [Hy — HL) + (HL — HF) (3.4-20)

Also, HL - HF = cpL(TB – TF) (3.4-21)

where the heat capacity of the liquid feed

cpL = 159 kJ/kg mol • K,


TB = 366.7 K (boiling point of feed), and
TF = 327.6 K (inlet feed temperature).

q = ((Hv – HL) + cpL (TB – TF)) / (Hv – HL) (3.4-22)

Figure 3.4-9. McCabe-Thiele diagram for distillation of benzene-toluene.

Substituting the known values into Eq. (3.4-22),

q = (32099 + 159(366.7 - 327.6)) / 32099 = 1.195 (SI)


q = (13 800 + 38(200.3 - 130) / 13 800 = 1.195 (English)

From Eq. (3.4-19), the slope of the q line is

q / (q – 1) = 1.195 / (1.195-1) = 6.12

The q line is plotted in Fig. 3.4-9 starting at the pointy = x F = 0.45 with a slope

132
Prof.Fatma K. Gad

of 6.12.

The stripping operating line is drawn connecting the point y = x = x w = 0.10


with the intersection of the q line and the enriching operating line. Starting at
the point y = x = xD, the theoretical steps are drawn in as shown in Fig. 4.4-9.
The number of theoretical steps is 7.6 or 7.6 steps minus a reboiler, which gives
6.6 theoretical trays. The feed is introduced on tray 5 from the top.

3.4C Total and Minimum Reflux Ratio for McCabe-


Thiele Method
1. Total reflux. In distillation of a binary mixture A and B the feed conditions,
distillate composition, and bottoms composition are usually specified and the
number of theoretical trays are" to be calculated. However, the number of
theoretical trays needed depends upon the operating lines. To fix the operating
lines, the reflux ratio R = L/D at the top of the column must be set.

One of the limiting values of reflux ratio is that of total reflux, or R = ∞. Since
R = L/D and, by Eq. (3.4-5),

Vn+1 = Ln + D (3.4-5)

then Ln is very large, as is the vapor flow Vn. This means that the slope R/(R +
1) of the enriching operating line becomes 1.0 and the operating lines of both
sections of the column coincide with the 45° diagonal line, as shown in Fig.
3.4-10.

The number of theoretical trays required is obtained as before by stepping off


the trays from the distillate to the bottoms. This gives the minimum number of
trays that can possibly be used to obtain the given separation. In actual practice,
this condition can be realized by returning all the overhead condensed vapor V1
from the top of the tower back to the tower as reflux, i.e., total reflux. Also, all
the liquid in the bottoms is reboiled. Hence, all the products distillate and
bottoms are reduced to zero flow, as is the fresh feed to the tower.

This condition of total reflux can also be interpreted as requiring infinite sizes
of condenser, reboiler, and tower diameter for a given feed rate.

If the relative volatility α of the binary mixture is approximately constant, the


following analytical expression by Fenske can be used to calculate the
minimum number of theoretical steps Nm when a total condenser is used.

133
Prof.Fatma K. Gad

Nm = log [(xD / (1 – xD)) – ((1 – xw) / xw)] / log αav (3.4-23)

For small variations in α, αav = (α1 αw)1/2, where α, is the relative volatility of the
overhead vapor and αw is the relative volatility of the bottoms liquid.

Figure 3.4-10: Total reflux and minimum number of trays by McCabe Thiele.

2. Minimum reflux ratio. The minimum reflux ratio can be defined as the
reflux ratio Rm that will require an infinite number of trays for the given
separation desired of xD and xw This corresponds to the minimum vapor flow in
the tower, and hence the minimum reboiler and condenser sizes. This case is
shown in Fig. 3.4-11. If R is decreased, the slope of the enriching operating line
R/(R + 1) is decreased, and the intersection of this line and the stripping line
with the q line moves farther from the 45° line and closer to the equilibrium
line. As a result, the number of

134
Prof.Fatma K. Gad

Figure 3.4-11. Minimum reflux ratio and infinite number of trays by McCabe-
Thiele method.

steps required to give a fixed x D and xw increases. When the two operating lines
touch the equilibrium line, a "pinch point" at y' and x' occurs where the number
of steps required becomes infinite. The slope of the enriching operating line is
as follows from Fig. 3.4-11, since the line passes through the points x', y', and
xD (y = xD).

Rm / (Rm +1) = (xD – y') / (xD * x') (3.4-24)

In some cases, where the equilibrium line has an inflection in it as shown in


Fig. 3.4-12, the operating line at minimum reflux will be tangent to the
equilibrium line.

3. Operating and optimum reflux ratio. For the case of total reflux, the
number of plates is a minimum, but the tower diameter is infinite. This
corresponds to an infinite cost of tower and steam and cooling water. This is
one limit in the tower operation. Also, for minimum reflux, the number of trays
is infinite, which again gives an infinite cost. These are the two limits in
operation of the tower.

The actual operating reflux ratio to use is in between these two limits. To select
the proper value of R requires a complete economic balance on the fixed costs
of the tower and operating costs. The optimum reflux ratio to use for lowest
total cost per year is between the minimum R m and total reflux. This has been
shown for many cases to be at an operating reflux ratio between 2R m to l.5Rm.

EXAMPLE 3.4-2. Minimum Reflux Ratio and Total Reflux in Rectification

135
Prof.Fatma K. Gad

For the rectification in Example 3.4-1, where a benzene-toluene feed is being


distilled to give a distillate composition of
xD = 0.95 and a bottoms composition of
xw = 0.10,
Calculate the following.
(a)Minimum reflux ratio Rm.
(b)Minimum number of theoretical plates at total reflux.

Solution:

For part (a) the equilibrium line is plotted in Fig. 3.4-13 and the q-line equation
is also shown for xF = 0.45. Using the same xD and xw as in Example 3.4-1, the
enriching operating line for minimum reflux is plotted as a dashed line and
intersects the equilibrium line at the same point at which the q line intersects.
Reading off the values of. x' = 0.49 and y' = 0.702, substituting into Eq. (3.4-
24), and solving for Rm,

Rm / (Rm + 1) = (xD – y') / (xD – x') = (0.95 - 0.702) / (0.95 - 0.49)

Hence, the minimum reflux ratio Rm = 1.17.

Figure 3.4-12. Minimum reflux ratio and infinite number of trays when
operating line is tangent to equilibrium line.

For the case of total reflux in part (b), the theoretical steps are drawn as shown

136
Prof.Fatma K. Gad

in Fig. 3.4-13. The minimum number of theoretical steps is 5.8, which gives 4.8
theoretical trays plus a reboiler.

3.4D Special Cases for Rectification Using McCabe—


Thiele Method
1-Stripping-column distillation. In some cases the feed to be distilled is not
supplied to an intermediate point in a column but is added to the top of the
stripping column as shown in Fig. 3.4-14a. The feed is usually a saturated
liquid at the boiling point and the overhead product VD is the vapor rising from
the top plate, which goes to a condenser with no reflux or liquid returned back
to the tower.

The bottoms product W usually has a high concentration of the less volatile
component B. Hence, the column operates as a stripping tower with the vapor
removing the more volatile A from the liquid as it flows downward. Assuming
constant molar flow rates, a material balance of the more volatile component A
around the dashed line in Fig. 3.4-14a gives, on rearrangement,

ym+1 = (Lm / (Vm+1) xm) – (W xw / Vm+1) (3.4-25)

This stripping-line equation is the same as the stripping-line equation for a


complete tower given as Eq. (3.4-11). It intersects the y = x line at x = x w, and
the slope is constant at L/Vm+l.

If the feed is saturated liquid, then Lm = F. If the feed is cold liquid below the
boiling point, the q line should be used and q > 1.

Lm = qF (3.4-26)

137
Prof.Fatma K. Gad

FIGURE 3.4-13. Graphical solution for minimum reflux ratio Rm and total
reflux for Example

Figure 3.4-14: Material balance and operating line for stripping tower:
(a) flows in tower. (b) operating and equilibrium line.

138
Prof.Fatma K. Gad

In Fig. 3.4-14 the stripping operating-line equation (3.4-25) is plotted and the q
line, Eq. (3.4-19), is also shown for q = 1.0. Starting at.x F, the steps are drawn
down the tower.

EXAMPLE 3.4-3. Number of Trays in Stripping Tower

A liquid feed at the boiling point of 400 kg mol/h containing 70 mol % benzene
(A) and 30 mol % toluene (B) is fed to a stripping tower at 101.3 kPa pressure.
The bottoms product flow is to be 60 kg mol/h containing only 10 mol % A
and.the rest B. Calculate the kg mol/h overhead vapor, its composition, and the
number of theoretical steps required.

Solution:

Referring to Fig. 3.4-14a, the known values are F = 400 kg mol/h, x F = 0.70, W
= 60 kg mol/h, and xw = 0.10. The equilibrium data from Table 3.1-1 are plotted
in Fig. 3.4-15. Making an overall material balance,

F = W + VD

400 = 60 + VD

Solving, VD = 340 kg mol/h. Making a component A balance and solving,

F xF = W xw + VD yD
400 (0.70) = 60 (0.10) + 340 (yD)
yD = 0.806

139
Prof.Fatma K. Gad

Figure 3.4-15: Stripping tower for example 3.4-3.

For a saturated liquid, the q line is vertical and is plotted in Fig. 4.4-15. The
operating line is plotted through the point y = x w - 0.10 and the intersection of
yD = 0.806 with the q line. Alternatively, Eq. (3.4-25) can be used with a slope
of Lm / Vm+1 = 400 / 340. Stepping off the trays from the top, 5.3 theoretical
steps or 4.3 theoretical trays plus a reboiler are needed.

2-Enriching-column distillation. Enriching towers are also sometimes used,


where the feed enters the bottom of the tower as a vapor. The overhead
distillate is produced in the same manner as in a complete fractionating tower
and is usually quite rich in the more volatile component A. The liquid bottoms
is usually comparable to the feed in composition, being slightly leaner in
component A. If the feed is saturated vapor, the vapor in the tower V n = F.
Enriching-line equation (3.4-7) holds, as does the q-line equation (3.4-19).

3-Rectification with direct steam injection. Generally, the heat to a


distillation tower is applied to one side of a heat exchanger (reboiler) and the
steam does not directly contact the boiling solution, as shown in Fig. 3.4-5.
However, when an aqueous solution of more volatile A and water B is being
distilled, the heat required may be provided by the use of open steam injected
directly at the bottom of the tower. The reboiler exchanger is then not needed.

The steam is injected as small bubbles into the liquid in the tower bottom, as
shown in Fig. 3.4-16a. The vapor leaving the liquid is then in equilibrium with
the liquid if sufficient contact is obtained. Making an overall balance on the
tower and a balance on-A,

F+S=D+W (3.4-27)

F xF + S ys = D x D + W x w (3.4-28)

where

S = mol/h of steam and


ys = 0 = mole fraction of A in steam.

The enriching operating-line equation is the same as for indirect steam.

140
Prof.Fatma K. Gad

For the stripping-line equation, an overall balance and a balance on component


A are as follows:

Lm + S =Vm+l + W (3.4-29)

Lm xm + S(0) = Vm+l yn+l + W xw (3.4-30)

Solving for ym+1 in Eq. (3.4-30),

ym+1 = (Lm / Vm+1) xm – W xw / Vm+1 (3.4-31)

For saturated steam entering, S = Vm+l and hence, by Eq. (3.4-29), Lm = W.


Substituting into Eq. (3.4-31), the stripping operating line is

ym+1 = W / S xm – W / S xw (3.4-32)

When y = 0, x = xw. Hence, the stripping line passes through the point y = 0, x =
xw, as shown in Fig. 3.4-16b, and is continued to the x axis. Also, for the
intersection of the stripping line with the 45° line, when y = x in Eq. (3.4-32),x
= Wxw / (W — S).

For a given reflux ratio and overhead distillate composition, the use of open
steam rather than closed requires an extra fraction of a stage, since the bottom
step starts below the y = x line (Fig. 3.4-16b). The advantage of open steam lies
in simpler construction of the heater, which is a sparger.

141
Prof.Fatma K. Gad

Figure 3.4-16. Use of direct steam injection:


(a) schematic of tower, (b) operating and equilibrium lines.

4. Rectification tower with side stream. In certain situations, intermediate


product or side streams are removed from sections of the tower between the
distillate and the bottoms. The side stream may be vapor or liquid and may be
removed at a point above the feed entrance or below depending on the
composition desired.

The flows for a column with a liquid side stream removed above the feed inlet
are shown in Fig. 3.4-17. The top enriching operating line above the liquid side
stream and the stripping operating line below the feed are found in the usual
way. The equation of the q line is also unaffected by the side stream and is
found as before. The liquid side stream alters the liquid rate below it, and hence
the material balance or operating line in the middle portion between the feed
and liquid side stream plates.

Making a total material balance on the top portion of the tower as shown in the
dashed-line box in Fig. 3.4-17,

Vs+l = Ls+ O + D (3.4-33)

142
Prof Fatma K. Gad

where O is mol/h saturated liquid removed as a side stream. Since the liquid
side stream is saturated,

Ln = Ls + O (3.4-34)

VS+1 = Vn+1 (3.4-35)

Making a balance on the most volatile component,

Vs+1 ys+1 = Ls xs + O xO + D xD (3.4-36)

Solving for yS+1 the operating line for the region between the side stream and
the feed is

ys+1 = (Ls / Vs+1) xs + (O xO + D xD) / Vs+1 (3.4-37)

The slope of this line is Ls / Vs+l. The line can be located as shown in Fig. 3.4-
18 by the q line, which determines the intersection of the stripping line and Eq.
(3.4-37), or it may be fixed by the specification of x D, the side-stream
composition. The step on the McCabe-Thiele diagram must actually be at the
intersection of the two operating lines at xQ in an actual tower. If this does not
occur, the reflux ratio can be altered slightly to change the steps.

Figure 3.4-17. Process flow for a rectification tower with a liquid side stream.

143
Prof Fatma K. Gad

Figure 4.4-18. McCabe-Thiele plot for a tower with a liquid side stream above
the feed.

5. Partial condensers. In a few cases it may be desired to remove the


overhead distillate product as a vapor instead of a liquid. This can also occur
when the low boiling point of the distillate makes complete condensation
difficult. The liquid condensate in a partial condenser is returned to the tower as
reflux and the vapor removed as product as shown in Fig. 3.4-19.

If the time of contact between the vapor product and the liquid is sufficient, the
partial condenser is a theoretical stage. Then the composition x R of the liquid
reflux is in equilibrium with the vapor composition y D, where yD = xD. If the
cooling in the condenser is rapid and the vapor and liquid do not reach
equilibrium, only a partial stage separation is obtained.

144
Prof Fatma K. Gad

3.5 DISTILLATION AND ABSORPTION TRAY


EFFICIENCIES

3.5A Introduction
In all the previous discussions on theoretical trays or stages in distillation, we
assumed that the vapor leaving a tray was in equilibrium with the liquid
leaving. However, if the time of contact and the degree of mixing on the tray is
insufficient, the streams will not be in equilibrium. As a result the efficiency of
the stage or tray is not 100%. This means that we must use more actual trays for
a given separation than the theoretical number of trays determined by
calculation. The discussions in this section apply to both absorption and
distillation tray towers.

Three types of tray or plate efficiency are used:

Overall tray efficiency EO, Murphree tray efficiency EM, and point or local tray
efficiency EMP (sometimes called Murphree point efficiency). These will be
considered individually.

Figure 3.4-19: Partial condenser where the vapor and liquid leave in

145
Prof Fatma K. Gad

equilibrium:
(a) Process flow diagram. (b) McCabe Thiele plot.

3.5B Types of Tray Efficiencies


1. Overall tray efficiency. The overall tray or plate efficiency E0 concerns
the entire tower and is simple to use but is the least fundamental. It is defined as
the ratio of the number of theoretical or ideal trays needed in an entire tower to
the number of actual trays used.

EO = number of ideal trays / number of actual trays (3.5-1)

For example, if eight theoretical steps are needed and the overall efficiency is
60%, the number of theoretical trays is eight minus a reboiler, or seven trays.
The actual number of trays is 7 / 0.60, or 11.7 trays.

Two empirical correlations for absorption and distillation overall tray


efficiencies in commercial towers are available for standard tray designs (O1).
For hydrocarbon distillation these values range from about 50 to 85% and for
hydrocarbon absorption from about 10 to 50%. These correlations should only
be used for approximate estimates.

2. Murphree tray efficiency. The Murphree tray efficiency EM is defined as


follows:

EM = (ym – ym+1) / (y*n – y n+1) (3.5-2)

where yn is the average actual concentration of the mixed vapor leaving the tray
n shown in Fig. 11.5-1, yn+1 , the average actual concentration of the mixed
vapor entering tray n, and y*n the concentration of the vapor that would be in
equilibrium with the liquid of concentration x n leaving the tray to the down
comer.

The liquid entering the tray has a concentration of x n+!, and as it travels across
the tray, its concentration drops to xn at the outlet. Hence, there is a
concentration gradient in the liquid as it flows across the tray. The vapor
entering the tray contacts liquid of different concentrations, and the outlet vapor
will not be uniform in concentration.

3. Point efficiency. The point or local efficiency EMP on a tray is defined as

146
Prof Fatma K. Gad

EMP = (y'n – yn+1) / (y*n – y'n+1) (3.5-3)

where y'n is the concentration of the vapor at a specific point in plate n as


shown in Fig. 3.5-1, y'n+l the concentration of the vapor entering the plate n at
the same point, and y'*n the concentration of the vapor that would be in
equilibrium with x'n at the same point. Since y'n cannot be greater than y*n, the
local efficiency cannot be greater than 1.00 or 100%.

In small-diameter towers the vapor flow sufficiently agitates the liquid so that it
is uniform on the tray. Then the concentration of the liquid leaving is the same
as that on the tray. Then yn = yn, yn+1 = yn+1, and y'*n = y*n. The point efficiency
then equals the Murphree tray efficiency or EM = EMP.

In large-diameter columns incomplete mixing of the liquid occurs on the trays.


Some vapor will contact the entering liquid x n-1, which is richer in component A
than xn. This will give a richer vapor at this point than at the exit point, where x„
leaves. Hence, the tray efficiency EM will be greater than the point efficiency
EMP. The value of EM can be related to EMP by the integration of EMP over the
entire tray.

4.5C Relationship Between Efficiencies


The relationship between EMP and EM can be derived mathematically if the
amount of liquid mixing is specified and the amount of vapor mixing is also set.
Derivations for three different sets of assumptions are given by Robinson and
Gilliland. However, experimental data are usually needed to obtain amounts of
mixing. Semitheoretical methods to predict EMP and EM are summarized in
detail by Van Winkle.

147
Prof Fatma K. Gad

Figure 3.5-1: Vapor and liquid compositions on a sieve tray efficiency.

Figure 3.5-2: Use of Murphy plate efficiency to determine actual number of


trays.

When the Murphree tray efficiency EM is known or can be predicted, the overall
tray efficiency EO can be related to EM by several methods. In the first method
an analytical expression is as follows when the slope m of the equilibrium line
is constant and also the slope L/V of the operating line:

EO = log[ 1 + EM (m V / L – 1] / log(m V/L) (3.5-4)

If the equilibrium and operating lines of the tower are not straight, a graphical
method in the McCabe-Thiele diagram can be used to determine the actual
number of trays when the Murphree tray efficiency is known. In Fig. 3.5-2 a
diagram is given for an actual plate as compared with an ideal plate. The
triangle acd represents an ideal plate and the smaller triangle abe the actual
plate. For the case shown, the Murphree efficiency EM — 0.60 = ba/ca. The
dashed line going through point b is drawn so that ba/ca for each tray is 0.60.
The trays are stepped off using this efficiency, and the total number of steps

148
Prof Fatma K. Gad

gives the actual number of trays needed. The reboiler is considered to be one
theoretical tray, so the true equilibrium curve is used for this tray as shown. In
Fig. 4.5-2, 6.0 actual trays plus a reboiler are obtained.

Binary Nonequimolal Overflow Distillation

3.6.A INTRODUCTION
In chapter 8 we discussed distillation under the assumption of equimolal
overflow, i.e., vapor and liquid molar flow rates in any section of the distillation
column are the same on all trays in that section. In this chapter we will not
make that assumption. Instead, energy balances on all trays will be used to
calculate the liquid and vapor flow rates on all trays. This nonequimolal
overflow (NEMO) model of the column is more rigorous and should be used in
those columns where there are

1-Large temperature differences from the top of the column to the bottom
(where sensible heat effects are important).

2-Significant differences in the molar heats of vaporization of the components.

3-Large heats of mixing.

The graphical calculation procedure that includes energy balances is called the
Ponchon method. The graphical solution uses an Hxy diagram on which several
A points are drawn, using the same graphical mixing rule that was used in
nonadiabatic flash calculations and in liquid-liquid extraction calculations.

It might be useful to summarize at this point both the equimolal overflow


McCabe-Thiele method, and the nonequimolal overflow Ponchon method so
that you can compare them directly. Both methods use VLE data and feed
conditions to calculate a minimum reflux ratio. Then component balances
(together with energy balances in the Ponchon method) are used with VLE
relationships to step up the column.

3.6.B Review of McCabe-Thiele Method Limitations

149
Prof Fatma K. Gad

Limited to binary systems only.

Requires equimolal overflow (gives straight operating lines).

Design Method
1-Find the minimum reflux ratio. Usually, this is obtained from the intersection
of the feed thermal condition q line with the VLE equilibrium curve.

2-Set the actual reflux ratio. Typically, this is some preset multiple of the
minimum reflux ratio, for example 1.1 times the minimum.

3-Draw lines representing component balances. These are the straight operating
lines for each section of the column.

4-Step off stages. Use alternately the operating line and the VLE curve. Use the
stripping operating line from xB up to where it intersects the q line. Then use
the rectifying operating line.

3.6.C Summary of Ponchon Method


1-Limitations: Limited to binary systems.

2-Design Procedure

(a)-Find the minimum reflux ratio. Usually, this is done by extending the VLE
tie-line on an Hxy diagram that goes through the feed point (z, hp) to a vertical
line drawn at the distillate composition xD. As shown in Figure 13-1, the
coordinates of the point of intersection are (xD,h%'n). The minimum reflux
ratio is calculated from

RRmin = (hΔmin – HNT) / (HNT – hD) (3.6-1)

where
HNT = enthalpy of vapor leaving the top tray
hD = enthalpy of liquid distillate product

150
Prof Fatma K. Gad

(b)-Set the actual reflux ratio. Typically, this is 1.1 times the minimum. Then
calculate the vertical coordinate hΔ of a Δ point:

hΔ = HNT + RR (HNT - hD) (3.6-2)

(c)-Locate the Δ and ΔB points using component and energy balances. Draw a
straight line from the Δ point through the feed point to a vertical line though the
bottoms composition xB. (See Figure 3.6-2.) The point of intersection is called
the ΔB point.

(d)-Beginning at the bottom of the column, use the ΔB point to construct


operating lines in the stripping section and the VLE tie-lines for phase
equilibrium. When a VLE tie-line is used that crosses the line that joins the Δ
and ΔB points, this is the optimum feed tray. Now use the Δ point to construct
operating lines in the rectifying section.

figure 3.6-1: Minimum reflux ratio on Ponchon diagram.

3.6-D PONCHON DIAGRAMS FOR SIMPLE


COLUMNS

Rectifying Section
The standard simple column has one feed, a total condenser, and a partial
reboiler, and makes two products (distillate and bottoms). (See Figure 3.6-3.)

151
Prof Fatma K. Gad

Consider the top part of the column. We cut the column above the nth tray, as
shown in Figure 3.6-4, and do mass, component, and energy balances around
this system.

We assume that we know Qc, D, xD, hD, Ln+1, xn+1, and hn+l. We wish to calculate
Vn, _yn, and Hn. The problem is mathematically similar to the mixing problem
discussed in chapter 11. The only difference is that one of the streams is
flowing out instead of in. This can be taken care of simply by using a negative
sign, as illustrated in Figure 3.6-5.

Instead of using  for summation, it is convenient to use the term Δ for


difference. Δ, a pseudo flow rate, is the difference between the vapor and liquid
streams passing each other in the rectifying section of the column. It is defined
by
Δ = Vn - Ln+l (3.6-3)

A total mass balance, a light component balance, and an energy balance around
the system under consideration give

152
Prof Fatma K. Gad

Fgure : 3.6-2(above) Stepping off stages on Ponchon diagram.


Fgure : 3.6-3(below) Distillation column nomenclature.

153
Prof Fatma K. Gad

Figure 3.6-4: Top section of the distillation column.

Total: Vn - Ln+l = D (3.6-4)

Component: Vn yn - Ln+1 xn+1 = D xD (3.6-5)

Energy: Vn Hn - Ln+1 hn+1 = -Qc + D hD


= D (hD – (Ac / D)) (3.6-6)

The right-hand side of equation 13-6 is written in the form shown because we
want to define a pseudo enthalpy hΔ and a pseudo composition xA.

Comparing equations 3.6-3 and 3.6-4, we see that A = D. Therefore, we define


hΔ and xΔ such that they give us the three equations

Vn - Ln+l = Δ (3.6-7)

Vn yn - Ln+1 xn+l = ΔxΔ (3.6-8)

Vn Hn - Ln+l hn+l = ΔhΔ (3.6-9)

Qc

Vn

(-Ln+1) D

Figure 3.6-5 System equivalent to the top of a distillation column.

154
Prof Fatma K. Gad

The coordinates of the Δ point are

XΔ = xD (3.6-10)

hΔ = hD – (Qc / D) (3.6-11)

Chap. 13 Binary Nonequimolal Overflow Distillation

Figure 3.6-6: use of Δ point.

The graphical mixing rule tells us that if we mix the Vn and Ln+1, streams
together (with Ln+1 taken to be negative) to get the Δ stream, the coordinates of
this Δ stream will lie on a straight line joining the points Vn and Ln+1. Thus, the
Δ point can be used to do component and energy balances. The method is
similar to the use of operating lines in the McCabe-Thiele procedure to do
component balances.

Figure 3.6-6 illustrates the use of a Δ point. If we know the point Ln+1 and the Δ
point on an Hxy diagram, we can simply draw a straight line joining the two.
The intersection of this line with the saturated vapor line gives us the point Vn.
Thus, yn and Hn can be read off the graph, and then the flow rate Vn can be
calculated by means of a component balance.

The straight line that we drew can be considered an operating line. It passes
through the coordinates (xn+1, hn+1) and (yn, Hn). The Δ point permits us to relate
the compositions and enthalpies of passing streams in the column. In effect, we

155
Prof Fatma K. Gad

are solving simultaneously mass, component, and energy balances graphically.


3.6-E Condenser
Balances around the condenser and reflux drum, sketched in Figure 3.6-7, give

VNT = R + D (3.6-12)

VNT(hD - HNT) = Qc (3.6-13)

where
R = reflux flow rate
D = distillate flow rate
Solving for R yields

R = (- Qc – D (HNT – hD)) / (HNT – hD) (3.6-14)

The reflux ratio RR = R / D can then be expressed as

RR = ((hD – Qc / D) – HNT) / (HNT – hD) (3.6-15)

Thus, the reflux ratio is the ratio of the distances a and b shown in Figure 3.6-8.
Note that the point hD lies on the saturated liquid line if there is no subcooling.
If the distillate and the reflux are subcooled, the point will lie below the
saturated liquid line.

3.6-F Stripping Section


Similar balances in the stripping section lead to the ΔB point which is used to
relate the compositions of passing streams in that section. The coordinates of
the ΔB point are

composition xΔB = xB (3.6-17)

Enthalpy hΔB = hB – QR / B (3.6-18)

156
Prof Fatma K. Gad

Figure 3.6-7: System around the condenser of a distillation column.

Figure 3.6-8: Reflux ratio from Δ point.

where
QR = heat transfer rate in the reboiler
B = bottoms product flow rate
hB = enthalpy of liquid bottoms product

3.6-G Whole Column


Mass, component, and energy balances around the entire system as sketched in
Figure 3.6-9(a) are

Total : F=B+D (3.6-19)

Component : z F = B x B + D xD (3.6-20)

Energy : hF F = B(hB – (QR / B)) + D (hD – (Qc / D)) (3.6-21)

These three equations tell us that mixing a stream B with enthalpy (hB - QR / B)
and composition xB together with a stream D with enthalpy (hD — Qc / D) and
composition xD will give a stream F which has an enthalpy hF and composition
z.

Thus, a straight line from the Δ point to the ΔB point must go through the feed
point, as shown in Figure 3.6-9(b). The point Fcan lie in the two-phase region,
in the subcooled-liquid region, or in the super-heated-vapor region. This overall
balance line replaces the q line used in the McCabe-Thiele method.

Example 3.6.1.

157
Prof Fatma K. Gad

A methanol-water separation is achieved in a distillation column operating at


atmospheric pressure. Feed flow rate is 3,000 kg/hr, and feed composition is 60
weight percent methanol. The feed enters the column at a pressure of 1
atmosphere and a temperature of 75°C. Determine the number of trays required
and the optimum feed tray to make 95 weight percent methanol distillate and 5
weight percent methanol bottoms when a reflux ratio of 0.75 is used. Assume
theoretical trays, a total condenser, a partial reboiler, and saturated reflux.

Figure 3.6-9: Overall column (a) System. (b) HXY diagram.

Solution:

Total mass and component balances give the product flow rates.

Total F =D + B

158
Prof Fatma K. Gad

Component z F — x D D + x B B

(0.60)(3,000) = (0.95)(D) + (0.05)(B)

D = 1,833 kg/hr

B = 1,167 kg/hr

Since the reflux ratio is given, the Δ point can be located on the Hxy diagram for methanol-water at atmospheric
pressure (see Figure 3.6-10) by using equation 3.6-2:

h Δ = H N T + RR(H N T - h D )

= 310 + 0.75 (310 - 50) = 505 cal/g

The feed point is plotted at z = 60 weight percent and a temperature of 75°C.

Note that the feed is partially vaporized. A straight line from the Δ point through the feed point intersects the vertical
line through x B at —500 cal/g. This is the Δ B point.

Stages are stepped off by starting at x B = 5 weight percent on the saturated liquid line in Figure 3.6-10. The vapor
leaving the base of the column has a composition y B = 31 weight percent. An operating line is drawn from y B to the Δ B
point (the dashed straight line), and the composition x 1 = 19 weight percent is read off the point of intersection with the
saturated liquid curve. Then the VLE tie-line is used to get y l = 55 weight percent.

This stepping procedure is continued for five steps (partial reboiler plus four trays). Between trays 4 and 5 the feed line is
crossed. The optimum feed tray is tray 4.

To get x 5 from y4, the Δ point is used, not the Δ B point. Stepping is continued until x D is reached. Eight trays plus the
partial reboiler are required.

3.6-H LIMITING CONDITIONS


In the McCabe-Thiele method, the minimum reflux ratio and the minimum
number of stages can be easily visualized and graphically calculated on an xy

159
Prof Fatma K. Gad

diagram. These limiting conditions represent useful information about the


constraints on the separation. In the Ponchon method, the same limiting
conditions can be determined with essentially the same ease, and they yield the
same useful information.

160
Prof Fatma K. Gad

Figure 3.6-10: Solution to example 3.6-1 shown on an Hxy diagram.

161
Prof Fatma K. Gad

3.6-9 Minimum Reflux Ratio

If we extend the VLE tie-lines to intersect with a vertical line through x D , the highest intersection point establishes the minimum
Δ point. As sketched in Figure 3.6-ll(a), the highest point usually occurs for the tie-line through the feed point. This is equivalent
to establishing the minimum reflux ratio in the McCabe-Thiele method by finding the rectifying operating line that intersects the
feed q line on the VLE curve. However, there are some systems in which the highest point of intersection on the vertical line
through x D may occur on the extension of a tie-line that is not the one that goes through the feed point.

Once the minimum Δ point is established, the minimum reflux ratio can be calculated from equation 3.6-1.

3.6-J Minimum Number of Trays

The minimum number of trays occurs at an infinite reflux ratio. If the reflux ratio is infinite, the condenser heat removal must be
infinite (or D must be zero). Therefore, the Δ point must be located at infinity. The operating lines on a Ponchon diagram are
vertical lines at an infinite reflux ratio (total reflux).

Figure 3.6—11(b) sketches the stepping-off of the minimum number of stages. Vertical operating lines and the VLE tie-lines are
used. For the system shown, four trays plus'a reboiler are required.

Figure :3.6-11: Limiting conditions.


(a) Minimum reflux ratio. (b) Total reflux.

3.6.K COMPLEX COLUMNS

162
Prof Fatma K. Gad

So far we have applied the Ponchon method to just the "plain vanilla" distillation column with a single feed and liquid distillate and
bottoms products. However, there are no inherent limitations to the method that would restrict us to just these simple columns.

Complex columns may have multiple feeds, multiple products, and multiple heat input or heat removal locations. The Ponchon
method becomes a little more complex, but the procedure is essentially the same. Total mass, component, and energy balances are
written around each section of the column, giving a different Δ point for each section. Several important cases will be considered.

Partial Condenser

Many columns produce a distillate product as a vapor from the reflux drum, as illustrated in Figure 3.6-12. The flow rate of this
vapor product is D, its composition is y D , and its enthalpy is H D . The D point on the Hxy diagram is located on the
saturated vapor line.

The Δ point is located at (y D , H D - Q C /D). Note that the liquid in the reflux drum is in equilibrium with the vapor
product. This liquid has a composition x R which is located at the other end of the tie-line from y D . The reflux drum gives an
additional stage of separation because of the partial condenser.

Occasionally one sees a distillation column where the product is removed as a vapor directly from the overhead vapor line, before
the condenser. In this case the condenser is a total condenser, producing liquid reflux at the same composition as the vapor product.
The reflux drum does not provide an additional stage. The Ponchon diagram looks almost the same as the liquid distillate case,
except that Q c is reduced because the distillate product is not condensed.

Other variations on the theme are also possible. For example, some columns take both vapor and liquid products. This obviously
changes the Ponchon diagram. It is instructive to sketch the diagram for this kind of system.

Dual Feeds

Many columns have more than a single feed stream. If the compositions of the two feed streams are identical, they can be
combined and fed onto a single tray. However, if the compositions are not the same (as is usual), they should be fed
onto different feed trays in order to consume the minimum amount of energy.

163
Prof Fatma K. Gad

Figure 3.6-12: Partial condenser.

Figure 3.6-13 shows a column with feed F, introduced on a tray low in the column and
feed F 2 introduced on a tray higher in the column. The compositions are z x and z2,
where z2 is greater than z t .

The column now has three different sections: a rectifying section at the top above the
top feed, a middle section between the feed trays, and a stripping section below the
lower feed. Therefore, we need three Δ points to analyze it. Let us call these points Δ,
ΔM, and ΔB. Clearly, the Δ and ΔB points will be located at the same coordinates as in
a conventional column:

Δ point = (x D , h D – (Qc /D)

ΔB point = (x B , h B – (Q R / B))

To find the location of the A M point, we write the balance equations and rearrange them so that they have the form

Vn - L N + 1 = Δ M = D - F2 (3.6-22)

V n y n - L n+l x n+1 = Δ M X Δ m = D x D - F 2 z 2 (3.6-23)

164
Prof Fatma K. Gad

V n H n - L n+l h n+l = Δ M h Δ M = -Q c + D h D - F 2 h F2 (3.6-24)

The pseudo flow rate A M , the pseudo composition X A M , and the pseudo enthalpy h AM define the coordinates of the A M point that
is used to step from tray to tray in the middle section of the column. The coordinates of the A M point are

x Δ M = (D x D – F 2 z 2 ) / (D – F 2 ) (3.6-25)

h Δ M = - Q c + Dh D - F 2 h F2 (D – F 2 ) (3.6-26)

This procedure is used for any complex column to locate the Δ points for each section. Remember that the coordinates of the Δ
points are pseudo compositions, so there is no restriction that they must have values between 0 and 1 only.

Figure 3.6-13: Two feeds.

Intermediate Reboilers and Condensers and Sidestreams


The procedure just described is used for each section of the column. The following
example should help to show the details of the method.

Example 3.6.2.

165
Prof Fatma K. Gad

30,000 lbm/hr of feed at 80°C is fed into an atmospheric distillation column. Feed composition is 45 weight percent methanol and
55 weight percent water. The distillate product from a total condenser is 98 weight percent methanol, and the bottoms product from
the base of the column is 4 weight percent methanol. The partial reboiler in the base of the column has a heat duty of 5.65x106
Btu/hr. There is another intermediate reboiler up in the stripping seetion which uses lower pressure steam and has a heat duty of
9.13x106 Btu/hr. intermediate reboiler is located where the liquid composition on the tray is approximately 10 weight percent
methanol. Feed is introduced on the optimum feed tray. Without assuming equimolai overflow. Calculate the number of trays, the
condenser duty, and the thermal condition of the feed.

Solution:

Since equimolai ovferflow cannot be assumed, we must use a Ponchon Hxy diagram for the design of the column. The two
reboilers will result in three different column sections and three Δ points: Δ above the feed, ΔM between the feed and the
intermediate reboiler, and ΔB below the intermediate reboiler. Figure 3.6-14 shows the general layout of the column. Figure 3.6-
15 gives the Hxy diagram for methanol and water at atmospheric pressure.

The ΔB point will be located at (x B , h B — Q B /B). We first, calculate product flow rates D and B from overall total and
component balance:

F=B+D

z F = xB B + xD D 30,000 = B + D

(0.45)(30,000) = (0.04)(B) + (0.98)(D)

D = 13,085 lbm/hr B = 16,915 lbm/hr

(Q B / B) = (5.65x106 Btu/hr) / (16915 lbm / hr)

h B at 4 weight percent on the saturated liquid line = 80 cal/g

h B – (QB / B) = 80 - (334 Btu/lbm)(lbm/454 g)(252 cal/Btu)

= -106 cal/g
The ΔB point is thus located at 4 weight percent and - 106 cal/g. It is used to draw the operating lines in the lower section of the
column. The operating line connecting ΔB and y B intersects the saturated liquid line (see Figure.3.6-15) at x 1 = 12 weight
percent. Therefore, we should locate the intermediate reboiler at tray 1.

166
Prof Fatma K. Gad

Figure 3.6-14: Intermediate reboiler (example 3.6-2).

For the section above the intermediate reboiler, the mass, component, and energy equations are written in standard format to
determine the location of the Δ M point:

Ln + i - Vn = B = Δ M (3.6-27)

L n+1 x n+i – Vn yn = B x B = Δ M x Δ M (3.6-28)

Ln +l hn +1 - V n H n = B h B - Q R - Q m = Δ M h Xu (3.6-29)

Thus, the coordinates of the AM point are

XΔ m = x B = 4 wt. % .

h Δ M = hB - (Q R + Q int ) / B

= 80 cal/g–[(5.65+9.13) (106 Btu/hr)/(16915 lbm/hr)](lbm/454g)(252 cal/Btu)

= - 405 cal/g

167
Prof Fatma K. Gad

We use the Δ M point to draw operating lines on Figure 3.6-15 to get x 2 and x 3 . The optimum feed tray is tray 3, since we cross
the straight line connecting the feed point to the Δ M point.

Figure 3.6-15: Solution to example 3.6-2 shown on an Hxy diagram.

The Δ point for the top section of the column can be calculated by determining Q c from an overall energy balance. Alternatively,
we can get the Δ point graphically by drawing a straight line from the Δ M point through the feed point at (z,h F ) to a vertical line
through x D .

Stepping is continued up the column using the Δ point. A total of six trays plus the total reboiler are needed.

The Δ point is located at an enthalpy of 830 cal/g. We can solve for Q c from

hA = hD – Qc / D

We have

830 cal/g = 50 cal/g – [Qc Btu/hr/13085 lbm/hr] (252 cal/Btu) (lbm/454 g)

168
Prof Fatma K. Gad

Qc = - 18.4x106 Btu/hr

The feed is 45 weight percent methanol at 80°C and is in the two-phase region. We can read the compositions of the liquid and
vapor phases at 80°C from the ends of the VLE tie-line:

x F = 37 wt. % y F = 76 wt. %

Let F L and F v be the flow rates of the liquid and vapor feed, respectively. Then

30,000 = F L + F v

(30,000)(0.45) = (F L(0.37) + (F v)(0.76)

F L = 23,840 lbm/hr

Therefore, the feed thermal condition is

(FL / F) = 23840 / 30000 = 0.79 (79 % liquid)

EXAMPLE 3.6-3.

Distillation Using Enthalpy-Concentration Method

A liquid mixture of benzene-toluene is being distilled using the same conditions as in


Example 3.4-1 except that a reflux ratio of 1.5 times the minimum reflux ratio is to be
used. The value of Rm= 1.17 from Example 3.4-2 will be used. Use enthalpy balances
to calculate the flow rates of the liquid and vapor at various points in the tower and
plot the curved operating lines. Determine the number of theoretical stages needed.

Solution:

The given data are as follows:


F = 100 kg mol/h, xF = 0.45,
xD = 0.95, xw = 0.10,
R = 1.5 Rm = 1.5(1.17) = 1.755,
D = 41.2 kg mol/h, W = 58.8 kg mol/h.
The feed enters at 54.4°C and q = 1.195.

169
Prof Fatma K. Gad

The flows at the top of the tower are calculated as follows.

L / D = 1.755;
L= 1.755(41.2) = 72.3;
V1 = L + D = 72.3 + 41.2 = 113.5

The saturation temperature at the top of the tower for y1 = xD = 0.95 is 82.3°C from
Fig. 3.1-1. Using Eq. (3.6-2),

H1 = 0.95[30820+96.3(82.3-80.1)]+(1-0.95)[34224+138.2(82.3-80.1)]
= 31 206

This value of 31206 could also have been obtained from the enthalpy-concentration
plot, Fig. 3.6-1. The boiling point of the distillate D is obtained from Fig. 3.1-1 and is
81.1°C. Using Eq. (3.6-5),

hD = 0.95(138.2)(81.1 - 80.1) + (1 - 0.95)(167.5)(81.1 - 80.1) = 139

Again this value could have been obtained from Fig. 3.6-1.

Following the procedure outlined for the enriching section for step 1, a value of x n
=0.55 is selected. Assuming a straight operating line for Eq. (3.6-8), an approximate
value of yn+1 is obtained.

yn+1 = (72.3 / 113.5) xn + (41.2 / 113.5) (0.95) = 0.637(xn) + 0.345

= 0.637(0.55) + 0.345 = 0.695

Starting with step 2 and using Fig. 3.6-1, for

xn = 0.55, hn = 1590 and for


yn+1 = 0.695, Hn+1 = 33 240.

Substituting into Eq. (3.6-12) and solving,

Vn+1 (33240) = (Vn+1-41.2) 1590+113.5(31206) - 72.3(139) Vn +1= 109.5

Using Eq. (3.6-6),

109.5 = Ln + 41.2 or Ln = 68.3

For step 3, substituting into Eq. (3.6-8),

170
Prof Fatma K. Gad

yn+1 = (68.3 / 109.5) (0.55) + (41.2 / 109.5) (0.95) = 0.700

This calculated value of yn+1 = 0.700 is sufficiently close to the approximate value of
0.695 so that no further trials are needed.

Selecting another value for xn = 0.70 and using Eq. (3.6-8), an approximate value of
yn+I is calculated.

yn+1 = (72.3 / 113.5) (0.70) + (41.2 / 113.5) (0.95) = 0.791

Using Fig. 3.6-1 for

xn = 0.70,
hn = 1000, and for
yn+1 = 0.791,
Hn+1 = 32 500.

Substituting into Eq. (3.6-12) and solving,

Vn+1 (32 500) = (Vn+1- 41.2)1000 + 1 13.5(31 206) - 72.3(139)


Vn+l = 110.8

Using Eq. (3.6-6),

Ln = 110.8 - 41.2 = 69.6

Substituting into Eq. 3.6-8),

yn+1 = (69.6 / 110.8) (0.70) + (41.2 / 110.8) (0.95) = 0.793

In Fig. 3.6-3, the points for the curved operating line in the enriching section are
plotted. This line is approximately straight and is very slightly above that for constant
molal overflow.

171
Prof Fatma K. Gad

Figure 3.6-3: plot of curved operating line using enthalpy concentration method for
example 4.6-2. Solid lines are for enthalpy concentration method and dashed lines for
constant molal over flow.

Using Eq. (3.6-10), the condenser duty is calculated.

qc = 113.5(31 206) - 72.3(139) - 41.2(139) = 3 526 100 kJ/h

To obtain the reboiler duty qR, values for hw and hF are needed.

Using Fig. 3.6-1 for

xw = 0.10,
hw = 4350.

The feed is at 54.5°C. Using Eq. (3.6-5),

hF = 0.45(138.2)(54.5 - 80.1) + (1 - 0.45)(167.5)(54.5 - 80.1) = -3929

Using Eq. (3.6-17),

qR = 41.2(139) + 58.8(4350) + 3 526 100 - 100( - 3929) = 4 180 500kJ/h

Using Fig. 3.4-5 and making a material balance below the bottom tray and around the

172
Prof Fatma K. Gad

reboiler,

Ln = W + Vw (3.6-18)

Rewriting Eq. (4.6-16) for this bottom section,

Vw Hw = (Vw+ W) hN + qR – W hw (3.6-19)

From the equilibrium diagram, Fig. 3.1-2, for


xw = 0.10,
yw = 0.207, which is the vapor composition leaving the reboiler.

For equimolal overflow in the stripping section using Eqs. (3.4-14) and (3.4-15),

Lm = Ln + q F = 72.3 + 1.195(100) = 19l. 8 (3.4-14)

Vm+1 = Vn+l – (l – q) F = 113.5 - (1 - 1.195)100 = 133.0 (3.4-15)

Selecting ym+1 = yw = 0.207, and using Eq. (3.6-15), an approximate value of xm = xN


is obtained.

ym+1 = (Lm / Vm+1) xm – (W xw / Vm+1) (3.6-15)

0.207 = (191.8 / 133.0) (xN) – (58.8 / 133.0) (0.10)

Solving, xN = 0.174.

From Fig. (4.6-1) for

xN = 0.174,

hN = 3800, and for

yw = 0.207,

Hw = 37 000.

Substituting into Eq. (3.6-19),


Vw(37 000) = (Vw+ 58.8)(3800) + 4 180 500 - 58.8(4350)

Solving, Vw = 125.0.

173
Prof Fatma K. Gad

Using Eq. (4.6-18), LN = 183.8.

Substituting into Eq. (3.6-15) and solving for xN,

0.207 = (183.8 / 125.0) (xN) – (58.8 / 125.0) (0.10)


xN = 0.173

This value of 0.173 is quite close to the approximate value of 0.174.

Assuming a value of ym+1 = 0.55 and using Eq. (3.6-15), an approximate value of xm
is obtained.

ym+1 = 0.55 = (191.8 / 133.0) (xm) – (58.8 / 133.0) (0.10)


xm = 0.412

From Fig. (11.6-1) for

xm = 0.412,
hm = 2300, and for
ym+1 = 0.55,
Hm = 34 400.

Substituting into Eq. (3.6-16),

Vm+1 (34 400) = (Vm + 1 4- 58.8X2300) + 4 180 500 - 58.8(4350)

Solving, Vm+1 = 126.5.


Using Eq. (4.6-13),

Lm = W + Vm+1 = 58.8 + 126.5 = 185.3

Substituting into Eq. (3.6-15) and solving for xm,

ym+1 = 0.55 = (185.3 / 126.5) xm – (58.8 / 126.5) (0.1)


xm = 0.407

This value of 0.407 is sufficiently close to the approximate value of 0.412 so that no
further trials are needed. The two points calculated for the stripping section are
plotted in Fig. 3.6-3. This stripping line is also approximately
straight and is very slightly above the operating line for constant molal overflow.

Using the operating line for the enthalpy balance method, the number of theoretical

174
Prof Fatma K. Gad

steps is 3.4. For the equimolal method 9.9 steps are obtained. This difference would
be larger if the reflux ratio of 1.5 times Rm were decreased to, say, 1.2 or 1.3. At
larger reflux ratios, this difference in number of steps would be less.

Note that in Example 3.6-2 in the stripping section the vapor flow increases slightly
from 125.0 to 126.5 in going from the reboiler to near the feed tray. These values are
lower than the value of 133.0 obtained assuming equimolal overflow. Similar
conclusions hold for the enriching section. The enthalpy-concentration method is
useful in calculating the internal vapor and liquid flows at any point in the column.
These data are then used in sizing the trays. Also, calculations of qc and q R are used in
designing the condenser and reboiler. This method is very applicable for design using
a computer solution for binary and multicomponent mixtures to make tray to tray
mass and enthalpy balances for the whole tower. A more restrictive Ponchon-Savarit
graphical method for only binary mixtures is available (K3, T2).

3.7 DISTILLATION OF MULTICOMPONENT MIXTURES

3.7A Introduction to Multicomponent Distillation


In industry many of the distillation processes involve the separation of more than two
components. The general principles of design of multi*component distillation towers
are the same in many respects as those described for binary systems. There is one
mass balance for each component in the multi-component mixture. Enthalpy or heat
balances are made which are similar to those for the binary case. Equilibrium data are
used to calculate boiling points and dew points. The concepts of minimum reflux and
total reflux as limiting cases are also used.

1. Number of distillation towers needed. In binary distillation one tower was


used to separate the two components A and B into relatively pure components with A
in the overhead and B in the bottoms. However, in a multi-component mixture of n
components, n-1 fractionators will be required for separation. For example, for a
three-component system of components A, B, and C, where A is the most volatile and
C the least volatile, two columns will be needed, as shown in Fig. 3.7-1. The feed of
A, B, and C is distilled in column 1 and A and B are removed in the overhead and C
in the bottoms. Since the separation in this column is between B and C, the bottoms
containing C will contain a small amount of B and often a negligible amount of A
(often called trace component). The amount of the trace component A in the bottoms
can often be neglected if the relative volatilities are reasonably large. In column 2 the
feed of A and B is distilled with A in the distillate containing a small amount of
component B and a much smaller amount of C. The bottoms containing B will also be

175
Prof Fatma K. Gad

contaminated with a small amount of C and A. Alternately, column 1 could be used to


remove A overhead with B plus C being fed to column 2 for separation of B and C. '

2. Design calculation methods. In multicomponent distillation, as in binary, ideal


stages or trays are assumed in the stage-to-stage calculations. Using equilibrium data,
equilibrium calculations are used to obtain the boiling point and equilibrium vapor
composition from a given liquid composition or the dew point and liquid composition
from a given vapor composition. Material balances and heat balances similar to those
described in Section 3.6 are then used to calculate the flows to and from the adjacent
stages. These stage-to-stage design calculations involve trial-and-error calculations,
and high-speed digital computers are generally used to provide rigorous solutions.

FlGURE 3.7-1. Separation of a ternary system of A, B, and C.

In a design the conditions of the feed are generally known or specified (temperature,
pressure, composition, flow rate). Then, in most cases, the calculation procedure
follows either of two general methods. In the first method, the desired separation or
split between two of the components is specified and the number of theoretical trays
are calculated for a selected reflux ratio. It is clear that with more than two
components in the feed the complete compositions of the distillate and bottoms are
not then known and trial-and-error procedures must be used. In the second method,
the number of stages in the enriching section and in the stripping section and the
reflux ratio are specified or assumed and the separation of the components is
calculated using assumed liquid and vapor flows and temperatures for the first trial.
This approach is often preferred for computer calculations. In the trial-and-error
procedures, the design method of Thiele and Geddes, which is a reliable procedure, is
often used to calculate resulting distillate and bottoms compositions and tray
temperatures and compositions. Various combinations and variations of the above
rigorous calculation methods are available in the literature and are not considered
further.

176
Prof Fatma K. Gad

The variables in the design of a distillation column are all interrelated, and there are
only a certain number of these which may be fixed in the design. For a more detailed
discussion of the specification of these variables, see Kwauk.

3. Shortcut calculation methods. In the remainder of this chapter, shortcut


calculation methods for the approximate solution of multicomponent distillation are
considered. These methods are quite useful to study a large number of cases rapidly to
help orient the designer, to determine approximate optimum conditions, or to provide
information for a cost estimate. Before discussing these methods, equilibrium
relationships and calculation methods of bubble point, dew point, and flash
vaporization for multicomponent systems are covered.

3.7B Equilibrium Data in Multicomponent Distillation


For multicomponent systems which can be considered ideal, Raoult's law can be used
to determine the composition of the vapor in equilibrium with the liquid. For example,
for a system composed of four components, A, B, C, and D,

PA = pA xA,
PB = PB xB,
PC = PC xC,
PD = PD xD (3.7-1)

yA = pA / P = (PA / P) xA,
yB = (PB / P) xB,
yC = (PC / P) xC,
yD = (PD / P) xD (3.7-2)

In hydrocarbon systems, because of nonidealities, the equilibrium data are often


represented by

yA = KA xA,
y B = KB x B ,
yC = KC xC,
y D = KD x D (3.7-3)

where KA is the vapor-liquid equilibrium constant or distribution coefficient for


component A. These K values for light hydrocarbon systems (methane to decane)
have been determined semiempirically and each K is a function of temperature and
pressure. Convenient K factor charts are available by Depriester and Hadden and

177
Prof Fatma K. Gad

Grayson. For light hydrocarbon systems K is generally assumed not to be a function


of composition, which is sufficiently accurate for most engineering calculations. Note
that for an ideal system, KA = P/P, and so on. As an example, data for the
hydrocarbons n-butane, n-pentane, n-hexane, and n-heptane are plotted in Fig. 3.7-2
at 405.3 kPa (4.0 atm) absolute.

Figure 3.7-2. Equilibrium K values for light hydrocarbon systems at 405.3 kPa (4.0
atm) absolute.

The relative volatility αi for each individual component in a multicomponent mixture


can be defined in a manner similar to that for a binary mixture. If component C in a
mixture of A, B, C, and D is selected as the base component,

αi = Ki / KC,
αA = KA / KC ,
αB = KB / KC,
αC = KC / KC = 1,
αD = KD / KC (3.7-4)

178
Prof Fatma K. Gad

The values of Ki will be a stronger function of temperature than the α i values since the
Ki lines in Fig. 3.7-2 all increase with temperature in a similar manner.

3.7C Boiling Point, Dew Point, and Flash Distillation


1-Boiling point. At a specified pressure, the boiling point or bubble point of a given
multicomponent mixture must satisfy the relation ∑ yi = 1.0. For a mixture of A, B, C,
and D with C as the base component,

∑yi = ∑Ki xi = Kc∑αi xi = 1.0 (3.7-5)

The calculation is a trial-and-error process, as follows. First a temperature is assumed


and the values of αi are calculated from the values of Ki at this temperature. Then the
value of Kc is calculated from Kc = 1.0/αi xi. The temperature corresponding to the
calculated value of Kc = 1.0/ ∑ αi, xi. The temperature corresponding to the
calculated value of Kc is compared to the assumed temperature. If the values differ,
the calculated temperature is used for the next iteration. After the final temperature is
known, the vapor composition is calculated from

yi = αi xi / ∑(αi xi) (3.7-6)

2. Dew point. For the dew point calculation, which is also trial and error,

∑ xi = ∑ (yi / Ki) = (1 / KC) ∑ (yi / αi) = 1 (3.7-7)

The value of Kc is calculated from Kc = ∑ (yi / αi). After the final temperature is
known, the liquid composition is calculated from

xi = (yi / αi) / ∑ (yi / αi) (3.7-8)

EXAMPLE 3.7-1. Boiling Point of a Multicomponent Liquid

A liquid feed to a distillation tower at 405.3 kPa abs is fed to a distillation tower. The
composition in mol fractions is as follows:
n-butane (xA = 0.40),
n-pentane (xB = 0.25),
N-hexane(xc = 0.20),
n-heptane(xD = 0.15).
Calculate the boiling point and the vapor in equilibrium with the liquid.

179
Prof Fatma K. Gad

Solution:

1-First a temperature of 65°C is assumed and the values of K obtained from Fig. 3.7-
2. Using component C (n-hexane) as the base component, the following values are
calculated using Eq. (3.7-5) for.the first trial:

The calculated value of Kc is 0.2745, which corresponds to 69°C, Fig. 3.7-2. Using
69°C for trial 2, a temperature of 70°C was obtained. Using 70°C for trial 3, the
calculations shown in the table give a final calculated value of 70°C, which is the
bubble point. Values of yi are calculated from Eq. (3.7-6).

3. Flash distillation of multicomponent mixture. For flash distillation, the process


flow diagram is shown in Fig. 3.3-1. Defining/ = V/F as the fraction of the feed
vaporized and (1 - f) = L/F as the fraction of the feed remaining as liquid and making
a component / balance as in Eq. (3.3-6), the following is obtained:

yi = ((ƒ – 1) / ƒ) xi + xi F / ƒ (3.7-9)

where yi is the composition of component i in the vapor, which is in equilibrium with


xi in the liquid after vaporization. Also, for equilibrium, y1 = K1 x1 = Kc αi xi, where αi
= Ki / KC.Then Eq. (4.7-9) becomes

yi = Kc ai xi = ((ƒ – 1) / ƒ) xi + xi F / ƒ (3.7-10)

Solving for xi and summing for all components,

∑ xi = ∑ (xiF / (ƒ (Kc αi – 1) + 1 )) (3.7-11)

180
Prof Fatma K. Gad

This equation is solved by trial and error by first assuming a temperature if the
fraction f vaporized has been set. When the ∑ x i values add up to 1.0, the proper
temperature has been chosen. The composition of the vapor yi can be obtained from y i
= Kc αi xi; or by a material balance.

3.7D Key Components in Multicomponent Distillation


Fractionation of a multicomponent mixture in a distillation tower will allow
separation only between two components. For a mixture of A, B, C, D, and so on, a
separation in one tower can only be made between A and B, or B and C, and so on.
The components separated are called the light key, which is the more volatile
(identified by the subscript L), and the heavy key (H). The components more volatile
than the light key are called light components and will be present in the bottoms in
small amounts. The components less volatile than the heavy key are called heavy
components and are present in the distillate in small amounts. The two key
components are present in significant amounts in both the distillate and bottoms.

3.7E Total Reflux for Multicomponent Distillation


1. Minimum stages for total reflux. Just as in binary distillation, the minimum
number of theoretical stages or steps, Nm, can be determined for multicomponent
distillation for total reflux. The Fenske equation (3.7-12) also applies to any two
components in a multicomponent system. When applied to the heavy key H and the
light key L, it becomes

Nm = log [(xLD D / xND (xHW W/ xLw W] / log (αLav) (3.7-12)

where xLD is mole fraction of light key in distillate, xLW is mole fraction in bottoms,
x[ID is mole fraction of heavy key in distillate, and x llw is mole fraction in bottoms. The
average value of αL of the light key is calculated from the αLD at the top temperature
(dew point) of the tower and αLav at the bottoms temperature.

αLav = √(αLD αLw) (3-7-13)

Note that the distillate dew-point and bottoms boiling-point estimation is partially
trial and error, since the distribution of the other components in the distillate and
bottoms is not known and can affect these values.

181
Prof Fatma K. Gad

2. Distribution of other components. To determine the distribution or


concentration of other components in the distillate and the bottoms at total reflux, Eq.
(3.7-12) can be rearranged and written for any other component i as follows:

xiD D / xiw W = (αiav)Nm (xHD D / xHw W) (3.7-14)

These concentrations of the other components determined at total reflux can be used
as approximations with finite and minimum reflux ratios. More accurate methods for
finite and minimum reflux are available elsewhere.

EXAMPLE 3.7-2. Calculation of Top and Bottom Temperatures and Total Reflux

The liquid feed of 100 mol/h at the boiling point given in Example 3.7-1 is fed to a
distillation tower at 405.3 kPa and is to be fractionated so that 90% of the n-pentane
(B) is recovered in the distillate and 90% of the n-hexane (C) in the bottoms.
Calculate the following.
(a)Moles per hour and composition of distillate and bottoms.
(b)Top temperature (dew point) of distillate and boiling point of bottoms.
©Minimum stages for total reflux and distribution of other components in the
distillate and bottoms.

Solution:

(a) Material balances are made for each component, with component n-pentane (B)
being the light key (L) and n-hexane (C) the heavy key (H). For the overall balance,

F=D+W (3.7-15)

For component B, the light key,

xBF F = 0.25(100) = 25.0 = yBD D + xBW W (3.7-16)

Since 90% of B is in the distillate,

yBD D = (0.90X25) = 22.5.

Hence, xBw W = 2.5.

For component C, the heavy key,

xCF F = 0.20(100) = 20.0 = yCDD + xCW W (3.7-17)

182
Prof Fatma K. Gad

Also, 90% of C is in the bottoms and

xCW W = 0.90(20) = 18.0. Then,


yCD D = 2.0.

For the first trial, it is assumed that no component D (heavier than the heavy key C) is
in the distillate and no light A in the bottoms. Hence,
moles A in distillate = yAD D = 0.40(100) = 40.0. Also,
moles D in bottoms = xDW W = 0.15(100) = 15.0.
These values are tabulated below.

(b)The dew point of the distillate (top temperature) in part (b), a value of 67°C will be
estimated for the first trial. The K values are read from Fig. 3.7-2 and the a values
calculated. Using Eqs. (3.7-7) and (3.7-8), the following values are calculated:

The calculated value of Kc is 0.2627, which corresponds very closely to 67°C, which
is the final temperature of the dew point.
For the bubble point of the bottoms, a temperature of 135°C is assumed for trial 1 and
Eqs. (3.7-5) and (3.7-6) used for the calculations. A second trial using 132°C gives
the final temperature as shown below.

The calculated value of Kc is 1.144, which'is close to the value at 132°C.

183
Prof Fatma K. Gad

(c) The proper a values of the light key L (n-pentane) to use in Eq. (3.7-13) are as
follows:

αLD = 2.50 (t = 67°C at the column top)

αLW = 2.04 (t = 132°C at the column bottom)

αLav = √(αLD αLW = 72.50(2.04) = 2.258 (3.7-13)

Substituting into Eq. (3.7-12),

NM = log [(0.349 x 64.5/0.031 x 64.5)(0-507 x 35.5/0.070 x 35.5)]/log (2.258)

= 5.404 theoretical stages (4.404 theoretical trays)

The distribution or compositions of the other components can be calculated using Eq.
(3.7-14). For component A, the average a value to use is

αAav = √(αAD αAW) = √(6.73 x 4.348) = 5.409

(xAD D / xAW W)= (αAav)Nm (xHD D / xHW W)

= (5.409)5.404 (0.031(64.5)) / (0.507(35.5)) = 1017 (3.7-14)

Making an overall balance on A,

F xFA = 40.0 = D xDA + W xWA (3.7-18)

Substituting DxDA = 1017 W xwA from Eq. (3.7-14) into (3.7-18) and solving,

xWA W = 0.039, xDA D = 39.961

For the distribution of component D, αDav = √(0.385 x 0.530) = 0.452.

(D xDD / W xDW) = (αAav)Nm (D xHD / W xHW)

= (0.452)5.404 (0.031 x 645) / (0.507 x 355) = 0.001521

F xDF = 15.0 = D xDD + W xDW

Solving, D xDD = 0.023, xDW, W = 14.977.

184
Prof Fatma K. Gad

The revised distillate and bottoms compositions are as follows.

Hence, the moles of D in the distillate are quite small, as are the moles of A in the
bottoms.

Using the new distillate composition, a recalculation of the dew point assuming 67°C
gives a calculated value of Kc = 0.2637. This is very close to that of 0.2627 obtained
when the trace amount of D in the distillate was assumed as zero. Hence, the dew
point is 67°C. Repeating the bubble-point calculation for the bottoms assuming
132°C, a calculated value of Kc = 1.138, which is close to the value at 132°C. Hence,
the bubble point remains at 132°C. If either the bubble- or dew-point temperatures
had changed, the new values would then be used in a recalculation of N m.

3.7F Shortcut Method for Minimum Reflux Ratio for Multi-


component Distillation

As in the case for binary distillation, the minimum reflux ratio R m is that reflux ratio
that will require an infinite number of trays for the given separation of the key
components.

For binary distillation only one "pinch point" occurs where the number of steps
become infinite, and this is usually at the feed tray. For multicomponent distillation
two pinch points or zones of constant composition occur: one in the section above the
feed plate and another below the feed tray. The rigorous plate-by-plate stepwise
procedure to calculate is trial and error and can be extremely tedious for hand

185
Prof Fatma K. Gad

calculations.

Underwood's shortcut method to calculate R m (Ul, U2) uses constant average a values
and also assumes constant flows in both sections of the tower. This method provides a
reasonably accurate value. The two equations to be solved to determine the minimum
reflux ratio are

1 – q = ∑((αi xiF) / (αi – θ)) (3.7-19)

Rm + l = ∑((αi xiD) / (αi – θ)) (3.7-20)

The values of xiD for each component in the distillate in Eq. (3.7-20) are supposed to
be the values at the minimum reflex. However, as an approximation, the values
obtained using the Fenski total reflux equation are used. Since each a; may vary with
temperature, the average value of a; to use in the preceeding equations is
approximated by using a; at the average temperature of the top and bottom of the
tower. Some have used the average α which is used in the Fenske equation or the a at
the entering feed temperature.

To solve for Rm, the value of 9 in Eq. (3.7-19) is first obtained by trial and error. This
value of 0 lies between the α value of the light key and a value of the heavy key,
which is 1.0. Using this value of 0 in Eq. (3.7-20), the value of Rm is obtained
directly. When distributed components appear between the key components, modified
methods described by others can be used.

3.7G Shortcut Method for Number of Stages at Operating


Reflux Ratio

1.Number of stages at operating reflux ratio. The determination of the minimum


number of stages for total reflux in Section 3.7E and the minimum reflux ratio in
Section 3.7F are useful for setting the allowable ranges of number of stages and flow
conditions. These ranges are helpful for selecting the particular operating conditions
for a design calculation. The relatively complex rigorous procedures for doing a
stage-by-stage calculation at any operating reflux ratio have been discussed in Section
3.7A.

An important shortcut method to determine the theoretical number of stages required


for an operating reflux ratio R is the empirical correlation of Erbar and Maddox (El)
given in Fig. 3.7-3. This correlation is somewhat similar to a correlation by Gilliland
(Gl) and should be considered as an approximate method. In Fig. 3.7-3 the operating

186
Prof Fatma K. Gad

reflux ratio R (for flow rates at the column top) is correlated with the minimum R m
obtained using the Underwood method, the minimum number of stages Nm obtained
by the Fenske method, and the number of stages N at the operating R.

2.Estimate of feed plate location. Kirkbride has devised an approximate method


to estimate the number of theoretical stages above and below the feed which can be
used to estimate the feed stage location. This empirical relation is as follows:
Log(Ne / Ns) =
0.206 log [ (xHF / xLF) (W / D) (xLW / xHD)2] (3.7-21)

where Ne is the number of theoretical stages above the feed plate and Ns the number
of theoretical stages below the feed plate.

EXAMPLE 3.7-3. Minimum Reflux Ratio and Number of Stages at Operating Reflux
Ratio

Using the conditions and results given in Example 3.7-2, calculate the following.

(a)Minimum reflux ratio using the Underwood method.

(b)Number of theoretical stages at an operating reflux ratio R of 1.5 R m using the


Erbar-Maddox correlation.

(c)Location of feed tray using the method of Kirkbride.

Solution;

The temperature to use for determining the values of αi is the average between the top
of 67°C and the bottom of 132°C and is (67 + 132)/2 or 99.5°C. The K, values
obtained from Fig. 3.7-2 and the αi values and distillate and feed compositions to use
in Eqs. (3.7-19) and (3.7-20) are as follows:

187
Prof Fatma K. Gad

Figure 3.7-3: Erbar Maddox correlation between reflux ratio and number of stages
(Rm based on Underwood method).

Substituting into Eq. (3.7-19) with q = 1.0 for feed at the boiling point,

1–q=1–1=0
=(5.2 x 0.4)/(5.2–θ)
+(2.3x0.25)/(2.3–θ)+(1x0.2)/(1–θ)
+(0.467x0.15)/(0.467–θ) (3.7-22)

This is trial and error, so a value of 0 — 1.210 will be used for the first trial (0 must
be between 2.30 and 1.00). This and other trials are shown below.

The final value of θ = 1.0296 is substituted into Eq. (3.7-20) to solve for Rm.

Rm+1 = (5.20(0.6197) / (5.20 - 1.2096))


+ (2.30(0.3489) / (2.30 - 1.2096))
+ (1.00(0.031) / (1.00 - 1.2096))

188
Prof Fatma K. Gad

+ (0.467(0.0004) / (0.467 - 1.2096))

Solving, Rm = 0.395.

(b) The following values are calculated.


R = l.5 Rm = 1.5(0.395) = 0.593,
R/(R+1) = 0.593/(0.593 + 1.0) = 0.3723,
R/(Rm+1) = 0.395/(0.395 + 1.0) = 0.2832.
From Fig. 3.7-3, Nm/N = 0.49. Hence,
Nm/N = 0.49
From example 3.7-2, Nm = 5.40.
Solving, N = 11.0 theoretical stages in the tower.
This gives 11.0 — 1.0 (reboiler) or 10.0 theoretical trays.

(c)For the location of the feed tray in part(c) using Eq. (3.7-21),

Log(Ne/Ns) = 0.206 log [(0.20/0.25)(35.516/64.484)(0.0704/0.031) 2] = 0.07344

Hence, Ne/Ns = 1.184.


Also, Ne = 1.184 (N – Ne)
So Ne = 6
Ne + Ns = N = 11.0 stages.

Solving, Ns — 5.0 and Ne = 6.0. This means that the feed tray is 6.0 trays from the
top.

III ( C ) BATCH DISTILLATION


In these processes the whole of a batch is run into the boiler of the still and, on
heating the vapor is passed into fractionation column as indicated in fig (24). As with
continuous distillation the composition of the top product will depend on the still
composition and the number of plates in the column and on the reflex ration used.

As the process proceeded, the liquid remaining in the still will become steadily
weaker in the M.V.C

Thus the still may be changed with S1 moles of a mixture containing xS1 of M.V.C.
Initially with reflex ration R1 the top product has a composition x d1. If after a contain
interval of time the composition of the top product starts to fall, then if the reflex is

189
Prof Fatma K. Gad

increased to new value R2, it will be possible to obtain the some composition again.
This method required a continuous increase in R to maintain a constant quality of the
top product.

Fig. (III-1) Batch distillation column

An alternative method is to work with a constant R and allow the composition of the
top product to fall.

Batch distillation is useful for handling small quantities of multi component, since it
is possible to obtain the different components at reasonable degree of purity one after
the other.

To obtain the maximum recovery of a valuable component, the charge remaining in


the still after first distillation may be added to the next batch.

1) Operation at constant product composition:-


Suppose a column of four ideal plates will be used to separate a mixture of EtOH
and H2O. Initially there are in the still S1 moles of liquor of mol fraction xS1 with
respect to the M.V.C (EtOH). The top product is to contain x d and this needed a
R1. Suppose the distillation is to be continued till there are S 2 moles in the still
with xS2. Then for the same no. of plates the reflex ratio will be increased to R2.

190
Prof Fatma K. Gad

Therefore, S1 – S2 = Db
S1 xS1 – S2 xS2 = Db xd

D (product) = S1 [ ( xS1 – xS2) / ( xd – xS2) ]= ( a / b ) S1 (III-1)

Where a and b are as shown in Fig (III-2).


If  is the intercept on the y axis for any operating line then,

 = xd / ( R + 1 ) or R = ( xd /  ) - 1

On comparing the operation at constant R with that of constant x d is the difference in


the amount of steam used in the distillation for a given quantity D b

If R is assumed to be adjusted to keep xd constant, than at any moment R is give by

R = dLp / dDp and the total reflex flowing down the column is given by

Lb R2
 dLb =  R dDb
0 R1

191
Prof Fatma K. Gad

Fig. (III2) Graphical solution of batch distillation.

to provide reflex dLb requires the removal of heat  dLb in the condenser, where  is
the latent heat per mol. Thus the heat to be supplied in the reboiler Q R to provide this
reflex during the total distillation in given by

Lb R2
QR =   DLb =   R dDb
0 R1

The Rm can be calculated from Underwood and Fenske equations.

2. Operation at constant reflux ratio


If the same column as before is operated at a constant R, the concentration of the
MVC in the top product will continuously fall. Over a small interval of time dt, x d
becomes xd – dxd. If in this time the amount of product obtained is dD b ; material
balance on MVC gives.
MVC in the top product = dDb [ xd - dxd / 2] = xd dDb

192
Prof Fatma K. Gad

And xd dDb = - d( S xS ) = - S dxS – xS dS


But dDb = - dS , S dxS = dS ( xd - xS )

s2 xS2
And  dS / S =  dxS / ( xd - xS )
s1 xS1

xS1
Ln ( S1 / S2 ) =  dxS / ( xd - xS )
xS2

The right hand side of this equation can be integrated by plotting (1/(x d-xS)) vs xS.
This will enable the ratio of the initial to final quantity in the still to be found for any
desired charge in xS and hence the amount of distillate Db.

The heat to be supplied to provide the reflux will be

QR =  R Db

Fig Batch distillation at constant reflux ratio.

193
Prof Fatma K. Gad

Example:(I)

A mixture of EtOH and H2O with 0.55 mol fraction of a alcohol is distilled to give a
top product of 0.75 mol fraction EtOH. The column has 4 ideal plates and the
distillation is stopped when the reflex ratio has to be increased beyond 4.0.
Find the amount of distillate and heat required / mol product.

Solution

Assuming various values of R between 0.85 and 4, we get the corresponding values of
the intercept  and the concentration in the still as shown in the table below.

R  xS Db
0.85 0.405 0.55 0
1.0 0.375 0.50 20.0
1.5 0.3 0.37 47.4
2.0 0.25 0.20 63.8
3.0 0.187 0.075 70.5
4.0 0.15 0.05 71.9

The amount of product for R = 4 calculated as

Db = 100 [0.55 – 0.05/0.75 – 0.05] = 100 * 0.5/0.7 = 71.4 kmol

R=4
The value  R dDb from the plot = 96 kmol
R=0.85

Taking and average latent heat per k mol for the EtOH – H2O mixture as 4000 kJ.

The heat to be supplied


QR = (96 * 4000)/1000 = 380 MJ
QR / Db = 380/71.4 = 5.32 MJ U.E.D

Example (ii)

If the same batch as above is distilled with constant R of 2.1. What will be the heat
required and the average composition of the distillate, if the distillation is stopped
when the composition in the still has fallen to 0.105 mol fraction of MVC.

Solution:

194
Prof Fatma K. Gad

The initial composition of the top product will be 0.78 as shown in the following table
and the final composition will be 0.74. values of x d, xS, xd-xS and 1 / (xd-xS ) for
various values of xS and a constant reflex ratio are given in the following table.

xS xd xd-xS 1/(xd-xS)
0.55 0.780 0.230 4.35
0.5 0.775 0.275 3.65
0.425 0.770 0.345 2.90
0.310 0.766 0.450 2.22
0.225 0.750 0.525 1.91
0.105 0.740 0.635 1.58

Fig of Example 1

Values of xS and 1 / ( xd - xS ) are plotted as shown

0.55
For which  dxS / ( xd - xS ) = 1.1
0.105

Ln ( S1 / S2 ) = 1.1 , ( S1 / S2 ) = 3.0

Hence Db = S1 – S2 = 100 – 100/3 = 66.7

195
Prof Fatma K. Gad

Amount of EtOH in the product = x1 S1 – x2 S2


= 0.55*100–0.105*33.3 = 51.5 k mole.

Average composition of product is = 51.5/66.7 = 0.77 and

The heat required = 4000 * 2.1 * 656.7 = 560 380 KJ

The heat required per kmol product = 360380/66.7 = 8400 KJ U.E.D

Fig. for example 2

AZEOTROPIC DISTILLATION
In the processes so far considered, the vapor becomes steadily richer in the M.V.C on

196
Prof Fatma K. Gad

successive plates.

There are two types of mixtures where this steady increase in the concentration of the
M.V.C either does not take place or else take place so slowly that an uneconomical
number of plates is required.

For instance if a mixture of EtoH and H2O is distilled, then the composition of the
alcohol steadily increases until it reaches 96% by wt, then the composition of the
vapor equals that of the liquid and no further enrichment occurs. This mixture is
called an Azeotropic and cannot be separated by straight forward distillation.

In this case the equilibrium curve crosses the diagonal, indicating the existence of an
Azeotropic.

i.e. H2O + HNO3 & H2O + HCl & H2O + many alcohol’s.

The principle of Azeotropic distillation lies in the addition of a new substance which
forms an Azeotropic with one or more of the components in the mixture, and as a
result is present on most of the plates of the column in appreciable concentration.

EXTRACTIVE DISTILLATION
When the relative volatility of a binary mixture is very low, in which case continuous
distillation of the mixture to give nearly pure components will require high (1) Reflux
ratio (2) High heat requirements (3) A tower of large cross section (4) Large No. of
trays. Examples of such mixtures;

197
Prof Fatma K. Gad

n – heptane – Methyl cyclo hexane & Toluene – Iso Octane.

The principle of Extraction Distillation lies in the addition of a substance which has
relatively less volatility compared with the components to be separated, and it is
therefore feed continuously near the top of the column. This extractive agent
(solvent) runs down the column as reflux and is present in appreciable concentration
on all the plates.

Since for a non-ideal binary mixture, the partial pressure may be expressed in the
form;

PA = A PoA xA

Where, A is the activity co-efficient for component ‘A.


Consequently,

 = ( PA xB ) /( PB XA ) = ( A PoA ) /( B PoA )

The solvent added to the mixture in extractive distillation differentially affects the
activities of the two components and hence ().

Such process depends upon the difference in departure from ideality between the
solvent and the components to be separated. As shown in the figure for a mixture of
Toluene – Iso Octane is seen to be very difficult to separate by adding Phenol as a
solvent 55% and 83% the relative volatility of Iso Octane increases, with as much as
83 Mol% Phenol in the liquid.

The solvent to be used will be selected on the basis of;


(1) Selectivity
(2) Volatility
(3) Ease of Separation from top and bottom products
(4) Its cost
Extractive distillation is usually a more desirable process than Azeotropic distillation,
since no large quantities of solvent must be vaporized. Also, a greater choice of
added component (solvent) is possible to use.

198
Prof Fatma K. Gad

Extractive Distillation of Toluene – Iso Octane with Phenol

PLATE COLUMNS
Plate column for the most important category and their features and performance will
be considered. The number of theoretical stages required to effect a required
separation, and the corresponding rates for the liquid and vapor phases, can be
determined by the procedure described earlier. To translate these quantities to form
an actual design the following factors must be considered;

1) The type of plate or tray.

2) The vapor velocity, which is the major factor in determining the diameter of the
column.

3) The plate spacing, which is the major factor fixing the height of the column when
the number of plates is known.

TYPES OF TRAYS

199
Prof Fatma K. Gad

I) TRAYS HAVING SEPARATE LIQUID DOWN-COMERS:


1) The bubble cup tray as shown in Fig., which is widely used in petroleum and
chemical industries.

2) Sieve or perforated tray

3) The Flexi-tray (Valve) is a recent development which has already made quite
a place for itself and appears to have a promising future.

4) The Uniflux tray is shown in the figure.

BUBBLE CAP SIEVE TRAY

FLEXI TRAY UNIFLUX TRAY

II) TRAYS HAVING NO LIQUID DOWN-COMERS:


1) The Turbo grid tray Fig. (a)
2) The Ripple tray Fig. (b) as shown in figure below.

200
Prof Fatma K. Gad

FIG. (A) FIG. (B)

Liquid level is maintained on each tray by a dynamic balance of the fluid flows.

OBJECTIONABLE PHENOMENA:
Areas of objectionable phenomena surround the region of satisfactory operation of
each tray. Changes in the design of a tray can shift the satisfactory operating range
with respect to any one or more than one of the objectionable areas.

1) CONING:
Coning occurs at low liquid rates. The vapor pushes the liquids back from the
slots or holes and passes upward with poor liquid contact. This causes poor tray
efficiency.

2) PULSATING VAPOR FLOW:


It occurs when the vapor rate is so small, so liquid can dump to the tray below.

3) DUMPING LIQUID:
It occurs at high liquid rates and low vapor loads.

4) BLOWING & EXCESSIVE ENTRAINING:


It occurs at very high vapor rates. The vapor can literally tear holes in the liquid,
giving poor contact between phases. Excessive entrainment of large droplets and
slugs of liquid up to the tray above usually accompany blowing.

5) FLOODING:
It occurs when the pressure drop across a tray is so high that the liquid cannot
flow down the column as fast as required. Flooding is a sharp line of distinction
between satisfactory and non-operating.

SATISFACTORY OPERATING REGION FOR TRAYS


The region of satisfactory operation is that range of liquid and vapor rates where;

201
Prof Fatma K. Gad

1) The contacting efficiency of the tray is at or near maximum.


2) The column is mechanically able to handle the liquid and vapor loads in a steady sate
manner.

As shown in the Figure ‘I’, areas where undesirable phenomena occur bound the
region.

Both pulsating vapor flow and dumping can be referred to jointly as weeping.

At very high vapor rate, the vapor bubbles carry Operating Region liquid as spray or
droplets to the plate above giving excessive entrainment.

LIQUID FLOW ARRANGEMENT ON A TRAY


Such arrangement depends largely on the ratio of liquid to vapor flow. Three layouts
are shown in the figure.

a) CROSS FLOW:
Normally, allow good length of liquid path giving good opportunity for mass
transfer.
b) REVERSE:
Down-comers much reduced in area, and very long liquid path suitable for low
liquid – vapor ration.

c) DOUBLE PASS:
Liquid flow splits into two directions, so that this system will handle high liquid.

202
Prof Fatma K. Gad

Vapor ratios.

FACTORS DETERMINING COLUMN PERFORMANCE


The performance of a column that can be judged in relation to two separate but related
criteria:

1) If in a column the vapor leaving a tray is in equilibrium with the liquid leaving,
this gives a theoretical tray and provides a standard of performance.
2) The relative performance of, say two column of the same diameter must be
considered in relation to their capacity for liquid and vapor flow.

The main features influencing the performance are;

1) Liquid and vapor velocities.


2) Physical properties of vapor and liquid.
3) Extent of entrainment of liquid by rising vapor streams.
4) The hydraulics of liquid and vapor across and through the tray.

(I) Bubble cap trays


Bubble cap trays are now rarely used for new installation on account of:

1) Their high cost.

203
Prof Fatma K. Gad

2) Their high-pressure drop.


3) The difficulties arise in large columns because of the tray hydraulic gradient,
which set up across the tray.

However Bubble cup tray are capable of dealing with very low liquid rates and are
therefore useful for operation at low reflux ratios.

View of Bubble Cup

TRAY SPACING:
The necessary tray spacing to avoid entrainment of the liquid by the vapor and to
enable down-comer to handle the flow, so the liquid does not rise up the down-comer.

The head of liquid in the down-comer (Z\p) is made up of the head of liquid on the
tray (hW + hoW) plus the head of the lost by the vapor (h p) plus the head lost at the
bottom of the down-comer (hd) plus the liquid gradient (hg) so;

Z\p = hp + hd + hg + hw + how
= (hrc + hr + hs) + hd + hg + hw + how

An approximate relation has been given by Bolles and Kirkbride is ;

Z\p = 2 (hw + how) + hd + hrc

Then fore safe operation, the tray spacing is given as the following equation to
overcome the foam in down comer.

204
Prof Fatma K. Gad

Zp = 2 Z\ p

DIAMETER OF BUBBLE TOWERS:


Since the allowable vapor rate is a function of the densities PV and PL at tower
conditions and of constant ((K)) which is related to service and to tray spacing; so

W / a = k * ( v * (L - v ))0.5 lb / ft2 . hr


U = k * (( L - v ) / v )0.5 ft / sec

Where ‘K’ is constant can be calculated from Nelson Fig (16-13)

II SIEVE TRAYS
Studies on Sieve Trays have been shown that they can offer several advantages over
the bubble cap tray.

1) Simpler and cheaper construction.

2) The general form of the flow on sieve tray is indicated in the figure, it is seen that the
tray forms a typical cross flow system with perforations in the tray taking the place of
the more complex bubble caps.

205
Prof Fatma K. Gad

Sieve tray

It is important to notice the key differences in operation between these two types of
tray; With the Sieve Tray the vapor passes vertically through the holes into the liquid
on the tray, whereas with bubble cup the vapor issues in an approximately horizontal
direction from the slots.

Height of liquid in Down-comer (Z\p):-


It is given in the same manner as bubble cup.

Z\p = ho + (hw + how) + hw + how + hg + hd

Where, ‘ho’ is the drop in pressure through the holes.


‘’ the aeration factor (0.8 – 0.5) depend on hw & how.
Approximation Z\p = ho + 2 (hw + how) + hd.
Zp = 2 Z\p

III : VALVE TRAYS


The Valve Tray, which may be regarded as intermediate between the bubble and the
sieve trays, offers advantages over both.

The feature of the tray is that lift able caps as variable orifices which adjust
themselves to changes in vapor flow.

Advantages claimed for valve trays include;

206
Prof Fatma K. Gad

1) Operation at the same capacity and efficiency as sieve tray.

2) A low pressure drop which, is fairly constant over a large portion of the operating
range.

3) A relatively simple construction which leads to a cost of only 20% higher than a
comparable sieve tray.

EXAMPLE-1:

Calculate the diameter of topping tower under the following condition;


lb Gasoline 3,415 lb 3415/110 = 31 Moles
lb Hot Reflux 15,000 lb 15500/110 = 141 Moles
lb Steam 567 lb567/18 = 31.5 Moles
TOTAL 19,482 lb 203.5 Moles

The pressure of the column = 780 mm and T = 287 oF at the Top.


So the volume at the top of the tower is;

V = ( 203.5 ) * ( 379 ) * (( 287 + 460 ) /520 ) * ( 760 / 780 )


= 107600 ft3

v = ( 19482 / 107600 ) = 0.181 lb /ft3

Assume PL = 42.7 lb/ft3 from Fig. (5-14) Nelson.


So, PV (PL – PV) = 0.181 (42.7 – 0.181) = 7.7
‘K’ at a tray spacing of 22” = 735 Fig.(16-13) Nelson.
 Mass Velocity = W / a = 735 * ( 7.7 )0.5 = 2040
 Area (a) = 19582 / 2040 = 9.55
0.5
Diameter = ( 9.55 / 0.785 ) = 3.5 ft U.E.D

EXAMPLE-2:

Calculate the diameter of Vacuum Tower under the following operating condition.

Trays at 30”, Pressure at Top = 30 mm & Top T = 490 oF.


Wax Distillate & Reflux 106,000 lb 331 Moles
Gas Oil 12,000 lb 49 Moles
Steam 1,400 lb 78 Moles
TOTAL 119,400 lb 459 Moles

207
Prof Fatma K. Gad

Volume = 459*359*(760/30)*(490+460)/(460+60) = 8,050,000 ft3/hr.


v = ( 119400 / 8050000 ) = 0.0148 lb/ft 3
Assume PL = 50 lb/ft3 Fig. (5-14)
 V (L – V) = 0.0148 (50 – 0.0148) = 0.74
‘K’ at 30” spacing tray = 600 Fig. (16-13)
 W/a = 600 ( 0.74 ) 0.5
= 516 lb/ft2 hr.
a = 119400 / 516 = 231 ft2

D = ( 231 / 0.782 )0.5 = 17.25 ft U.E.D

CHAPTER 4

SINGLE-STAGE LIQUID-LIQUID EXTRACTION


PROCESSES

4-1 Introduction to Extraction Processes


In order to separate one or more of the components in a mixture, the mixture is
contacted with another phase. The two-phase pair can be gas-liquid, which was dis-
cussed in Chapter 3; vapor-liquid which was covered in Chapter 4; liquid-liquid; or
fluid-solid. In this section liquid-liquid extraction separation processes are considered
first. Alternative terms are liquid extraction or solvent extraction.

In distillation the liquid is partially vaporized to create another phase, which is a


vapor. The separation of the components depends on the relative vapor pressures of

208
Prof Fatma K. Gad

the substances. The vapor and liquid phases are similar chemically. In liquid-liquid
extraction the two phases are chemically quite different, which leads to a separation
of the components according to physical and chemical properties.

Solvent extraction can sometimes be used as an alternative to separation by distil-


lation or evaporation. For example, acetic acid can be removed from water by
distillation or by solvent extraction using an organic solvent. The resulting organic
solvent-acetic acid solution is then distilled. Choice of distillation or solvent
extraction would depend on relative costs. In another example high-molecular-weight
fatty acids can be separated from vegetable oils by extraction with liquid propane or
by high-vacuum distillation, which is more expensive.

In the pharmaceutical industry products such as penicillin occur in fermentation


mixtures that are quite complex, and liquid extraction can be used to separate the
penicillin. Many metal separations are being done commercially by extraction of
aqueous solutions, such as copper-iron, uranium-vanadium, and tantalum-columbium.

4.2 EQUIPMENT FOR LIQUID-LIQUID EXTRACTION

4.2A Introduction and Equipment Types


As in the separation processes of absorption and distillation, the two phases in liquid-
liquid extraction must be brought into intimate contact with a high degree of
turbulence in order to obtain high mass-transfer rates. After this contact of the two
phases, they must be separated. In absorption and in distillation, this separation is
rapid and easy because of the large difference in density between the gas or vapor
phase and the liquid phase. In solvent extraction the density difference between the
two phases is not large and separation is more difficult.

There are two main classes of solvent-extraction equipment, vessels in which mecha-
nical agitation is provided for mixing, and vessels in which the mixing is done by the
flow of the fluids themselves. The extraction equipment can be operated batch wise or
operated continuously as in absorption and in distillation.

4.2B Mixer-Settlers for Extraction

209
Prof Fatma K. Gad

To provide efficient mass transfer, a mechanical mixer is often used to provide


intimate contact of the two liquid phases. One phase is usually dispersed into the
other in the form of small droplets. Sufficient time of contact should be provided for
the extraction to take place. Small droplets produce large interfacial areas and faster
extraction. However, the droplets must not be so small that the subsequent settling
time in the settler is too large.

The design and power requirements of baffled agitators or mixers. In Fig. 4.2-1a a
typical mixer-settler is shown, where the mixer or agitator is entirely separate from
the settler. The feed of aqueous phase and organic phase are mixed in the mixer, and
then the mixed phases are separated in the settler. In Fig. 4.2-lb a combined mixer-
settler is shown, which is sometimes used in extraction of uranium salts or copper
salts from aqueous solutions. Both types of mixer-settlers can be used in series for
countercurrent or multiple-stage extraction.

4.2C Plate and Agitated Tower Contactors for Extraction


As discussed in Section 2-3 for plate absorption and distillation towers, somewhat
similar types of devices are used for liquid-liquid contacting. A perforated-plate or
sieve-tray extraction tower is shown wherein the rising droplets of the light solvent
liquid are dispersed. The dispersed droplets coalesce below each tray and are then re-
formed on each tray by passing through the perforations. The heavy aqueous liquid
flows across each plate, where it is contacted by the rising droplets and then passes
through the down comer to the plate below.

Figure 4.2-1. Typical- mixer-settlers for extraction: (a) separate mixer-settler, (b)
combined mixer-settler.

210
Prof Fatma K. Gad

Figure 4.2-2. Extraction towers:


(a) perforated-plate or sieve-tray tower,
(b) agitated extraction tower.

In Fig. 4.2-2b an agitated extraction tower is shown. A series of paddle agitators


mounted on a central rotating shaft provides the agitation for the two phases. Each
agitator is separated from the next agitator by a calming section of wire mesh to
encourage coalescence of the droplets and phase separation. This apparatus is
essentially a series of mixer-settlers one above the other. Another type is the Karr
reciprocating-plate column, which contains a series of sieve trays with a large open
area of 60% where the plates are moved up and down. This is one of the few types of
extraction towers that can be scaled up with reasonable accuracy.

4.2D Packed and Spray Extraction Towers


Packed and spray tower extractors give differential contacts, where mixing and
settling proceed continuously and simultaneously. In the plate-type towers or mixer-
settler contactors, the extraction and settling proceeds in definite stages. The heavy
liquid enters the top of the spray tower, fills the tower as the continuous phase, and
flows out through the bottom. The light liquid enters through a nozzle distributor at
the bottom which disperses or sprays the droplets upward. The light liquid coalesces
at the top and flows out. In some cases the heavy liquid is sprayed downward into a
rising light continuous phase.

A more effective type of tower is to pack the column with packing such as Raschig

211
Prof Fatma K. Gad

rings or Berl saddles, which cause the droplets to coalesce and redisperse at frequent
intervals throughout the tower. Detailed discussions of flooding and construction of
packed towers are given elsewhere.

4.3 CHOICE OF SOLVENT


Finding the best solvent is the most critical aspect of developing a liquid-liquid
extraction process. The solvent should have a high selectivity for the extracted solute.
The selectivity p of a solvent is similar to relative volatility and is given by

β = (x1s / x2s) / (x1R / x2R) (4.3-1)

where

x1s = weight fraction of component 1 (solute) in the solvent phase


x2s = weight fraction of component 2 in the solvent phase
x1R = weight fraction of component 1 (solute) in the raffinate phase
x2R = weight fraction of component 2 in the raffinate phase

Selectivity is thus the ratio of the compositions in the solvent phase of the two
components being separated to the ratio of the compositions in the raffinate phase of
the same components. If the ratio is high, it means that the solvent preferentially
soaks up the solute. Note that p goes to unity at the plait point.

Of course, the solvent should be easily recoverable from the extracted solute. Usually,
this is done by distillation. The solvent and raffinate phases should have as large a
density difference as possible to help settling.

It is also desirable to find a solvent that is inexpensive, nontoxic, non-corrosive, and


without safety or pollution problems. Satisfying these last properties simultaneously
is rarely possible.

4.4 Equilibrium Relations in Extraction


1-Phase rule. Generally in a liquid-liquid system we have three components, A, B,
and C, and two phases in equilibrium. Substituting into the phase rule, the number of
degrees of freedom is 3. The variables are temperature, pressure, and four

212
Prof Fatma K. Gad

concentrations. Four concentrations occur because only two of the three mass fraction
concentrations in a phase can be specified. The third must make the total mass
fractions total to 1.0, xA + xB + xC= 1.0. If pressure and temperature are set, which is
the usual case, then, at equilibrium, setting one concentration in either phase fixes the
system.

2-Triangular coordinates and equilibrium data. Equilateral triangular coordinates are


often used to represent the equilibrium data of a three-component system, since there
are three axes. This is shown in Fig. 4.4-1. Each of the three corners represents a pure
component, A, B, or C. The point M represents a mixture of A, B, and C. The
perpendicular distance from the point M to the base AB represents the mass fraction
xc of C in the mixture at M, the distance to base C B the mass fraction xA of A, and the
distance to base AC the mass fraction xB of B. Thus,

xA + xB + xC = 0.40 + 0.20 + 0.40 = 1.0 (4.4-1)

213
Prof Fatma K. Gad

Figure 4.4-1. Coordinates for a triangular diagram.

A common phase diagram where a pair of components A and B are partially miscible
is shown in Fig. 4.4-2. Typical examples are methyl isobutyl ketone (A)-water (B)-
acetone (C), water (A)-chloroform (B)-acetone (C), and benzene (A)-water (B)-acetic
acid (C). Referring to Fig. 4.4-2, liquid C dissolves completely in A or in B. Liquid A
is only slightly soluble in B and B slightly soluble in A. The two-phase region is
included inside below the curved envelope. An original mixture of composition M
will separate into two phases A and B which are on the equilibrium tie line through
point M. Other tie lines are also shown. The two phases are identical at point P, the
Plait point.

214
Prof Fatma K. Gad

Figure 4.4-2: Liquid – liquid phase diagram where component A and B are partially
miscible.

3. Equilibrium data on rectangular coordinates. Since triangular diagrams have


some disadvantages because of the special coordinates, a more useful method of
plotting the three component data is to use rectangular coordinates. This is shown in
Fig. 4.4-3 for the system acetic acid (A)-water (B)-isopropyl ether solvent (C). The
solvent pair B and C are partially miscible. The concentration of the component C is
plotted on the vertical axis and that of A on the horizontal axis. The concentration of
component B is obtained by difference from Eqs. (4.4-2) or (4.4-3).

xB = 1.0 - xA – xc (4.4-2)

yB = 1.0 -yA- yc (4.4-3)

The two-phase region in Fig. 4.4-3 is inside the envelope and the one-phase region
outside. A tie line gi is shown connecting the water-rich layer i, called the raffinate
layer, and the ether-rich solvent layer g, called the extract layer. The raffinate
composition is designated by x and the extract by y. Hence, the mass fraction of C is
designated as yc in the extract layer and as xc in the raffinate layer. To construct the tie
line gi using the equilibrium yA – xA plot below the phase diagram, vertical lines to g
and i are drawn.

215
Prof Fatma K. Gad

Figure 4.4-3: Acitic acid (A)-Water (B)-isopropyl ether (C).


Liquid – Liquid phase diagram at 293 K (20 oC).

EXAMPLE 4.4-1. Material Balance for Equilibrium Layers

An original mixture weighing 100 kg and containing 30 kg of isopropyl ether (C), 10


kg of acetic acid (A), and 60 kg water (B) is equilibrated and the equilibrium phases
separated. What are the compositions of the two equilibrium phases?

Solution:

The composition of the original mixture is

xc = 0.30,
xA = 0.10, and
xB = 0.60.

This composition of

216
Prof Fatma K. Gad

xc = 0.30 and
xA = 0.10
is plotted as point h on Fig. 4.4-3.
The tie line gi is drawn through point h by trial and error.
The composition of the extract (ether) layer at g is
yA = 0.04,
yc= 0.94, and
yB = 1.00 - 0.04 - 0.94 = 0.02 mass fraction.
The raffinate (water) layer composition at i is
xA = 0.12,
xc = 0.02, and
xB= 1.00 — 0.12- 0.02 = 0.86.
Another common type of phase diagram is shown in Fig. 4.4-4, where the solvent
pairs B and C and also A and C are partially miscible. Examples are the system
styrene (A)-ethylbenzene (B)-diethylene glycol (C) and the system chlorobenzene
(A)-methyl ethyl ketone (B)-water (C).

Figure 5.4-4: Phase diagram where the solvent pairs B – C and A – C are partially
miscible.

4.5 Single-Stage Equilibrium Extraction


1. Derivation of lever-arm rule for graphical addition. This will be derived for
use in the rectangular extraction-phase-diagram charts. In Fig. 4.5-1a two streams, L
kg and V kg, containing components A, B, and C, are mixed (added) to give a result-
ing mixture stream M kg total mass. Writing an overall mass balance and a balance on
A,

V + L= M (4.5-1)

V yA + L xA = M xAM (4.5-2)

217
Prof Fatma K. Gad

where xAM is the mass fraction of A in the M stream. Writing a balance for component
C,

Vyc + L xC = M xCM (4.5-3)

Combining Eqs. (4.5-1) and (4.5-2),

L / V = (yA – xAM) / (xAM — xA) (4.5-4)

Combining Eqs. (4.5-1) and (4.5-3),

L / V = (yc – xCM) / (xCM – xc) (4.5-5)

Equating Eqs. (4.5-3) and (4.5-4) and rearranging,

(xC – xCM) / ( xA – xAM) = (xCM – yc) / (xAM – yA) (4.5-6)

This shows that points L, M, and V must lie on a straight line. By using the properties
of similar right triangles,

L (kg) / V (kg) = VM / LM (4.5-7)

This is the lever-arm rule and states that kg L / kg V is equal to the length of line VM
/ length of line LM. Also,

L (kg) / M (kg) = VM / LV (4.5-8)

These same equations also hold for kg mol and mol frac, lbm, and so on.

Figure 4.5-5. Graphical addition and lever-arm rule : (a) process flow, (b) graphical
addition.

218
Prof Fatma K. Gad

EXAMPLE 4.5-1. Amounts of Phases in Solvent Extraction

The compositions of the two equilibrium layers in Example 4.4-1 are for the extract
layer (V)
yA = 0.04, yB = 0.02, and
yc = 0.94,
and for the raffinate layer (L)
xA =0.12, xB = 0.86, and
xc = 0.02.
The original mixture contained 100 kg and
xAM = 0.10. Determine the amounts of V and L.

Solution:

Substituting into Eq. (4.5-1),

V + L = M= 100

Substituting into Eq. (4.5-2), where M = 100 kg and xAM = 0.10,

V(0.04) + L(0.12) = 100(0.10)

Solving the two equations simultaneously,

L = 75.0 and V = 25.0.

Alternatively, using the lever-arm rule, the distance hg in Fig. 4.5-2 is measured as
4.2 units and gi as 5.8 units. Then by Eq. ,

L / M = L / 100 = 4.2 / 5.8

Solving,

L = 72.5 kg and

V = 27.5 kg,

which is a reasonably close check on the material-balance method.

2. Single-stage equilibrium extraction. We now study the separation of A from a

219
Prof Fatma K. Gad

mixture of A and B by a solvent C in a single equilibrium stage. The process is shown


in Fig. 4.5-2a, where the solvent, as stream V2, enters and also the stream LO. The
streams are mixed and equilibrated and the exit streams L 1 and V1 leave in
equilibrium with each other.

Figure 4.5-2: Single stage equilibrium liquid – liquid extraction:


(a) process flow diagram. (b) plot on phase diagram.

The equations for this process are the same as those given for a single equilibrium
stage where y represents the composition of the V streams and x the L streams.

L0 + V2 = L1 + V1 = M (4.5-9)

L0 xA0 + V2 yA2 = Ll xAl + V1 yA1 = M xAM (4.5-10)

L0 xC0 + V2 yC2 = L1 xC1 + V1 yCl = M xCM (4.5-11)

Since xA + xB + xC = 1.0, an equation for B is not needed. To solve the three


equations, the equilibrium-phase diagram in Fig. 4.5-2b is used. Since the amounts
and compositions of L0 and V2 are known, we can calculate values of M, xAM, and xCM
from Eqs. (4.5-9)-(4.5-11). The points L0, V2, and M can be plotted as shown in Fig.
4.5-2b. Then using trial and error a tie line is drawn through the point M, which
locates the compositions of L1 and V1 The amounts of L1 and V1 can be determined by
substitution into Eqs. (4.5-9)—(4.5-11) or by using the lever-arm rule.

220
Prof Fatma K. Gad

4.6 MULTIPLE STAGES WITH CROSSFLOW OF


SOLVENT
If the process liquid stream from the first stage is fed into a second extractor and
mixed with more fresh solvent, as shown in Figure 4-6-1, we have what is called
cross-flow extraction. In cross-flow extraction, the procedure described in

Figure 4.6-1(above): Multistage extractor with cross flow of solvent.


Figure 4.6-2(below): Mix point for cross flow extractor.

the previous section for a single stage is simply repeated again for each stage, using
the raffinate phase from the upstream stage as the feed to each stage. The graphical
solution is sketched in Figure 4.6-1 for a three-stage cross-flow extractor. For each
stage, the point M is calculated from the flow rates of the feed and fresh solvent
streams to that particular stage.

221
Prof Fatma K. Gad

4.7 CONTINUOUS MULTISTAGE COUNTERCURRENT


EXTRACTION
Introduction

Single-stage equilibrium contact was used to transfer the solute A from one liquid to
the other liquid phase. To transfer more solute, the single-stage contact can be
repeated by contacting the exit Ll stream with fresh solvent V2 in Fig. 4.5-2. In this
way a greater percentage removal of the solute A is obtained. However, this is
wasteful of the solvent stream and also gives a dilute product of A in the outlet
solvent extract streams. In order to use less solvent and to obtain a more concentrated
exit extract stream, countercurrent multistage contacting is often used.

Figure 4.7-1: Continuous Multistage Countercurrent Extraction.

Many of the fundamental equations of countercurrent gas absorption and of rec-


tification are the same or similar to those used in countercurrent extraction. Because
of the frequently high solubility of the two liquid phases in each other, the
equilibrium relationships in extraction are more complicated than in absorption and
distillation.

4.7A Continuous Multistage Countercurrent Extraction

222
Prof Fatma K. Gad

1. Countercurrent process and overall balance. The process flow for this
extraction process is the same as given previously in Fig. 3.3-2 and is shown in Fig.
4.7-1. The feed stream containing the solute A to be extracted enters at one end of the
process and the solvent stream enters at the other end. The extract and raffinate
streams flow counter-currently from stage to stage, and the final products are the
extract stream V1 leaving stage 1 and the raffinate stream LN leaving stage N.

Figure 4.7-2: Countercurrent multistage Extraction process flow diagram.

Making an overall balance on all the N stages,

L0 + VN+l = LN + V1 = M (4.7-1)

where M represents total kg/h (lb„yh) and is a constant,L0 the inlet feed flow rate in
kg/h, VN+l the inlet solvent flow rate in kg/h, V l the exit extract stream, and LN the exit
raffinate stream. Making an overall component balance on component C,

L0 xC0 + VN+1 yCN+1 = Ln xCN + V1 yC1 = M xCM (4.7-2)

Combining Eqs. (5.7-1) and (5.7-2) and rearranging,

xCM = [(L0 xC0+VN+1 yCN+1 )/(L0 + VN+1)]


= [(LN xCN + V yC1)/(LN + V1)] (4.7-3)

Asimilar balance on component A gives

xAM = [(L0 xCA0+VN+1 yAN+1 )/(L0+VN+1)]


= [(LN xAN + V yA1)/(LN + V1)] (4.7-4)

Equations (4.7-3) and (4.7-4) can be used to calculate the coordinates of point M on
the phase diagram that ties together the two entering streams L 0 and VN+1 and the two
exit streams V1 and LN. Usually, the flows and compositions of L0 and VN+l are known
and the desired exit composition x AN is set. If we plot pointsL0, VN+1, and M as in Fig.
5.7-2, a straight line must connect these three points. Then L N, M, and V1 must lie on

223
Prof Fatma K. Gad

one line. Also, LN and Vl must also lie on the phase envelope, as shown. These
balances also hold for lbm and mass fraction, kg mol and mol fractions, and so on.

EXAMPLE 4.7-1. Material Balance for Countercurrent Stage Process

Pure solvent isopropyl ether at the rate of V N+1 = 600 kg/h is being used to extract an
aqueous solution of L0 = 200 kg/h containing 30 wt % acetic acid (A) by
countercurrent multistage extraction. The desired exit acetic acid concentration in the
aqueous phase is 4%. Calculate the compositions and amounts of the ether extract V 1
and the aqueous raffinate LN. Use equilibrium data from Appendix A.3.

Solution:

The given values are :


VN+t = 600 kg/h xA0 = 0.30
xB0 = 0.70 xC0 = 0.0 and
xAN = 0.04
and L0 are plotted.

Also, since LN is on the phase boundary, it can be plotted at xAN — 0.04. For the
mixture point M, substituting into Eqs. (4.7-3) and (4.7-4),

Figure 4.7-3: use of the mixture point M for overall material balance in countercurrent
solvent extraction.

xCM = [(L0 xC0+VN+1 yCN+1 )/(L0 + VN+1)]


= [(200 (0) + 600 (1.0))/(200 + 600)] = 0.75

xAM = [(L0 xCA0+VN+1 yAN+1 )/(L0+VN+1)]


= [(200 (0.3) + 600 (0))/(200 + 600)] = 0.075

224
Prof Fatma K. Gad

Using these coordinates, the point M is plotted in Fig. 4.7-3. We locate V1 by drawing
a line from LN through M and extending it until it intersects the phase boundary.
This gives

yAl = 0.08 and

yCl = 0.90. For LN a value of

xCN = 0.017 is obtained.

By substituting into Eqs. (4.7-1) and (4.7-2) and solving,


LN = 136 kg/h and
V1 = 664 kg/h.

Figure 4.7-4: Method to perform overall material balance for example 4.7-1.

2. Stage-to-stage calculations for countercurrent extraction. The next step


after an overall balance has been made is to go stage by stage to determine the
concentrations at each stage and the total number of stages N needed to reach LN in
Fig. 4.7-2.

Making a total balance on stage 1,

L0 + V2 = Ll + Vl (4.7-5)

Making a similar balance on stage n,

Ln-l + Vn+1 = Ln + Vn (4.7-6)

225
Prof Fatma K. Gad

Rearranging Eq. (4.7-5) to obtain the difference A in flows,

L0 - Vl = L1 - V2 = Δ (4.7-7)

This value of Δ in kg/h is constant and for all stages,

Δ = L0 – V1 = Ln - Vn+1 = Ln - VN+l = • • • (4.7-8)

This also holds for a balance on component A, B, or C.

ΔxΔ = L0 x0 – V1 y1 = Ln xn - Vn+1 yn+1


= LN xN - VN +1 yN+1 = • • • (4.7-9)

Combining Eqs. (4.7-8) and (4.7-9) and solving for xΔ,

xΔ = [(L0 x0 – Vl y1) / (L0 – V1)] =


[(Ln xn – Vn+l yn+l) / (Ln – Vn+1)] =
[(LN xN – VN+l yN+l) / (LN – VN+1)] (4.7-10)

where xA is the x coordinate of point A.


Equations (5.7-7) and (5.7-8) can be written as

L0 = Δ + Vl Ln = Δ + Vn+l LN = Δ + VN+l (4.7-11)

From Eq. (4.7-11), we see that L0 is on a line through Δ and V1, Ln is on a line
through Δ and Vn+l, and so on. This means Δ is a point common to all streams passing
each other, such as L0 and V1, Ln and Vn+l, LN and VN+l, and so on. The coordinates to
locate this Δ operating point are given for x CA and xAΔ in Eq. (4.7-10). Since the end
points VN+l, LN or V1 and L0 are known, xΔ can be calculated and point Δ located.
Alternatively, the Δ point is located graphically in Fig. 4.7-5 as the intersection of
lines L0 V1 and LN VN+l.

In order to step off the number of stages using Eq. (4.7-11) we start at L0 and draw the
line L0 Δ, which locates V1 on the phase boundary. Next a tie line through V1 locates
Ll. which is in equilibrium with V1 Then line L1 Δ is drawn giving V2. The tie line V2
L2 is drawn. This stepwise procedure is repeated until the desired raffinate
composition LN is reached. The number of stages N is obtained to perform the
extraction.

226
Prof Fatma K. Gad

Figure 4.7-5. Operating point Δ and number of theoretical stages needed for
countercurrent extraction.

EXAMPLE 4.7-2. Number of Stages in Countercurrent Extraction

Pure isopropyl ether of 450 kg/h is being used to extract an aqueous solution of 150
kg/h with 30 wt % acetic acid (A) by countercurrent multistage extraction. The exit
acid concentration in the aqueous phase is 10 wt %. Calculate the number of stages
required.

Solution:

The known values are


Vn+1 =450, yAN+1 =0,
yCN+1 = l.0 L0 = 150,
xA0 = 0.30, xB0 = 0.70,
xC0 = 0, and xAN = 0.10.

The points VN+l, L0, and LN are plotted in Fig. 4.7-5. For the mixture point M,
substituting into Eqs. (4.7-3) and (4.7-4),xCM = 0.75 and xAM = 0.075. The point M is
plotted and V1 is located at the intersection of line L N M with the phase boundary to
give yA1 = 0.072 and yCl = 0.895. This construction is not shown. (See Example 4.7-1
for construction of lines.)

The lines L0Vl and LNVN+l are drawn and the intersection is the operating point Δ as
shown. Alternatively, the coordinates of Δ can be calculated from Eq. (4.7-10) to

227
Prof Fatma K. Gad

locate point Δ. Starting at L0 we draw line L0 Δ, which locates Vl. Then a tie line
through V1 locates Lx in equilibrium with Vt. (The tie-line data are obtained from an
enlarged plot such as the bottom of Fig. 4.5-3. Line L1 Δ is next drawn locating V2. A
tie line through V2 gives L2. A line L2 Δ gives V3. A final tie line gives L3, which has
gone beyond the desired LN. Hence, about 2.5 theoretical stages are needed.

3. Minimum solvent rate. If a solvent rate VN+1, is selected at too low a value a
limiting case will be reached with a line through A and a tie line being the same. Then
an infinite number of stages will be needed to reach the desired separation. The
minimum amount of solvent is reached. For actual operation a greater amount of
solvent must be used.

Figure 4.7-6. Graphical solution for countercurrent extraction in Example 4.7-2

The procedure to obtain this minimum is as follows. A tie line is drawn through point
L0 (Fig. 4.7-5) to intersect the extension of line LN VN+1. Other tie lines to the left of
this tie line are drawn including one through L N to intersect the line LN VN+l. The
intersection of a tie line on line LN VN+l which is nearest to VN+l represents the Δmin
point for minimum solvent. The actual position of Δ used must be closer to V N+l than
Δmin for a finite number of stages. This means that more solvent must be used.
Usually, the tie line through L0 represents the Δmin.

228
Prof Fatma K. Gad

4.7B Countercurrent-Stage Extraction with Immiscible


Liquids
If the solvent stream VN+l contains components A and C and the feed stream L 0
contains A and B and components B and C are relatively immiscible in each other, the
stage calculations are made more easily. The solute A is relatively dilute and is being
transferred from L0 to VN+1.

Referring to Fig. 4.7-1 and making an overall balance for A over the whole system
and then over the first n stages,

L` (x0 / (1 – x0))+V`(yN+1 / (1 – yN+1))


= L` (xN / (1 – xN)) +V`(y1 / (1 – y1)) (4.7-12)

L` (x0 / (1 – x0))+V`(yn+1 / (1 – yn+1))


= L` (xn / (1 – xn)) +V`(y1 / (1 – y1)) (4.7-13)

where
L` = kg inert B/h,
V = kg inert C/h,
y = mass fraction A in V stream, and
x = mass fraction A in L stream.

This Eq. (4.7-13) is an operating-line equation whose slope = L/V. If y and x are quite
dilute, the line will be straight when plotted on an x y diagram.

The number of stages are stepped off as shown previously in cases in distillation and
absorption.

If the equilibrium line is relatively dilute, then since the operating line is essentially
straight, and can be used to calculate the number of stages.

EXAMPLE 4.7-3. Extraction of Nicotine with Immiscible Liquids.

An inlet water solution of 100 kg/h containing 0.010 wt fraction nicotine (A) in water
is stripped with a kerosene stream of 200 kg/h containing 0.0005 wt fraction nicotine
in a countercurrent stage tower. The water and kerosene are essentially immiscible in
each other. It is desired to reduce the concentration of the exit water to 0.0010 wt
fraction nicotine. Determine the theoretical number of stages needed. The equilibrium
data are as follows (C5), with x the weight fraction of nicotine in the water solution
and y in the kerosene.

229
Prof Fatma K. Gad

x 0.00101 0.00246 0.005 0.0074 0.0098 0.020


6 8 2
y 0.00080 0.00195 .00045 0.0068 0.0090 0.018
6 9 4 2 4 5

Solution:

The given values are


L0 = 100 kg/h, x0 =0.010,
VN+l =200 kg/h, yN+1 = 0.0005,
xN = 0.0010.
The inert streams are

L` = L(l - x) = L0 (1 - x0) = 100(1 - 0.010) = 99.0 kg water/hr

V` = V(l - y) = VN+1 (l - yN+1) = 200(1 - 0.0005) = 199.9 kg kerosene/hr

Making an overall balance on A using Eq. (4.7-12) and solving, y1 = 0.00497. These
end points on the operating line are plotted in Fig. 4.7-7. Since the solutions are quite
dilute, the line is straight. The equilibrium line is also shown. The number of stages
are stepped off, giving N = 3.8 theoretical stages.

Figure 4.7-7. Solution for extraction with immiscible liquids in Example 4.7-3.

230
Prof Fatma K. Gad

4.7-C CO-CURRENT CONTACT WITH IMMISCIBLE


SOLVENTS
In this case, as illustrated in the Figure, there is no need to use Triangular diagram.

Suppose the initial solution contains A Kg of solvent A with a mass ration X F of


solute. Then the selective solvent to be added will be a mass B of solvent B.

On mixing and separating, a Raffinate is obtained with solvent A containing a mass


X1 of solute and an extract with solvent B containing a mass ration Y 1 of solute. A
material balance on the solute gives;

AXF = AX1 + BY1


 Y1 / ( X1 - XF ) = - A/B

This process can be illustrated by letting the point (F) represent the feed solution and
drawing a line (FE1) of slope (- A/B), to cut the equilibrium curve in E1. This will
then give composition Y1 of the extract and X1 of the Raffinate. If a further stage is
then carried out by the addition of solvent B to the stream AX 1, then Point E2 is found
on the equilibrium curve by drawing GE2 of slope – (A/B)2.

Point E2 then gives the composition X2 and Y2 of the Final Extract and Raffinate.
This system can be used for any number of stages, any assumed variation in the
proportion of solvent B to Raffinate from stage to stage.

Calculation of No. of stages for Co-Current using Immiscible Solvents.

If the distribution law is followed, then the equilibrium curve becomes a straight line

231
Prof Fatma K. Gad

of equation Y = m X.
The material balance on the solute may then be rewritten as;

AXF = AX1 + BY1 = ZX1 + B m X1 = (A + B m)X1

If a further mass B of B is added to Raffinate AX1 to give extract of composition Y2


and Raffinate X2 in a second stage, then.

A * X1 = A * X2 + B * m * X2 = (A + B * m)X2

X2 = [ A / ( A + B * m ) ] X1 = [ A / ( A + B * m ) ]2 * Xf

 Xn = [ A / ( A + B m ) ] n * Xf

where
n = Log ( xn / xf ) / Log [ A / ( A + B * m ) ]

EXAMPLE – 4:

Acetic Acid in an aqueous solution, containing 0.15 mass fraction acid is to be


extracted in a single stage using MIK as solvent. The amount of solvent used is 1.745
lb solvent/lb solution (Feed). Find the composition of extract and raffinate and the
quantities of them for each lb feed
X lb Acid/lb H2O 0.5 0.1 0.15 0.2
Y lb Acid/lb Solvent 0.04 0.075 0.12 0.17

SOLUTION:

Material balance on the solute


Feed Raffinate
A(X1 – X2) = B(Y2 – Y1) X1 X2
-A / B = Y2 / ( X1 – X2 ) Y1=0 Y2
Solvent extract
Where Y1 = 0 (Solvent free of solute)
Slope of operating line = - A / B = (1-0.15) / 1.745 = - 0.487
X1 mass ratio of solute (Acid) = X1 / (1 – X1) = 0.15 / (1 – 0.15) = 0.1768
Drawing a line with slope = - 0.487 and Point (X1 & Y1) = (0.178 & 0)
From the drawing we get Point (X2 & Y2) as shown
X2 = 0.074 & Y2 = 0.05
Mass Fraction X2 = X2 / (1 + X2 ) = 0.074 / (1 + 0.074) = 0.0688
Raffinate = H2O + Acid = 0.85 + (0.074 x 0.85) = 0.9129

232
Prof Fatma K. Gad

Extract = 1.745 + 0.05 x 1.745 = 1.832

Y Equilibrium line
(X2,Y2)

Operating line

(X1,Y1)
Lb acid / Lb H2O
U.E.D.

4.8 A - MINIMUM SOLVENT RATE


In the preceding discussion we assumed that the fresh solvent flow rate S 0 was given.
However, in the normal design situation the solvent rate must be specified by the
designer. Usually, the smaller the solvent flow rate the lower the energy costs because
less solvent must be circulated between the extractor and the stripping column and the
concentration of solute in the "fat solvent" from the extractor is higher. Hence,
separation costs are reduced.

However, the smaller the solvent flow the higher the capital costs, because more
stages are required in the liquid-liquid extractor. These engineering trade-offs
between the number of stages and the solvent flow rate are analogous to the situation
in distillation, where trade-offs must be made between the reflux ratio and the number
of trays.

A commonly used heuristic in distillation was setting the reflux ratio to 1.2 times the
minimum reflux ratio. In liquid-liquid extraction, there is similarly a minimum solvent
flow rate. Once this minimum is found, the actual solvent flow rate is set somewhat
higher. Since liquid-liquid extraction stages are usually more expensive than
distillation trays, the ratio of actual to minimum solvent flow rates is usually higher in
extraction (typically 2).

The minimum solvent flow rate S0min occurs when an LLE tie-line is coincident with
an "operating line" drawn to the Δ point. The minimum solvent rate is usually found
by extending the LLE tie-line that passes through the feed point. The intersection of
this straight line with the straight line through the points S 0 and RN gives the
minimum Δ, Δmin. As shown in Figure 4-8-1, Δmin can lie either to the left or to the
right of the phase diagram.

233
Prof Fatma K. Gad

In some unusual LLE systems, there are other tie-lines (not through the feed point)
of the tie-
that intersect the S0-RN line farther to the left. The extensions

lines that occur farthest to the left dictate the point


Δmin. Figure 4.8-1 illustrates this situation for
two cases: when Δmin lies to the right of the
diagram, and when it lies to the left. For both
cases, the leftmost intersection of a tie-line with
the line through the points R and gives Δmin. N

The point S1min is found on the straight line from


the feed point to Δmin, and a straight line is
drawn from it to the point R which has been N

specified. The point Mmin is then given by the


intersection of this line with the straight line
connecting the points F and S . S can then be 0 0
min

calculated using the known compositions s1min


and s3min of M : min

Figure 4.8-1: Minimum solvent flow rate.

234
Prof Fatma K. Gad

Chap. 9 Liquid-
Liquid Extraction

Figure 5.8-2: Minimum Δ point


(a) Δmin lies to the right of the phase diagram.
(b) Δmin lies to the left of the phase diagram
S0min = F (z1 – x1min) / (x1min – x1So) (4.8-1)

4.8 B - DESIGN PROCEDURE

235
Prof Fatma K. Gad

The foregoing graphical construction may sound somewhat complicated on first


reading. However, it is fairly straightforward once you get the basic idea. It parallels
the construction used for a single stripping or rectifying column in binary distillation.
Let us summarize the specific sequence of steps to design a countercurrent liquid-
liquid extractor for a ternary system.

1. Parameters specified.

(a) Feed flow rate and compositions (F, z1, z3).


(b) Composition of solute in the raffinate stream leaving the last stage (x 1RN).
(c) LLE data for the ternary system.
(d)Economic ratio of actual to minimum solvent flow rates (S0/ S0min).
(e) Composition of fresh solvent (x1S0, x3S0).

2. Calculate the minimum solvent flow rate S0min.

(a) Extend the tie-line through the point F.


(b) Draw a line through the points S0 and RN.
(c) Locate S1min on the tie-line through F.
(d) Draw a line from S1min to RN.
(e) The intersection of this line with a line through F and S0 gives the point Mmin.
(f) Read off the composition (x1min) at Mmin.
(g) Calculate S0min from equation 4.8-1.

3. Set the actual fresh solvent flow rate S0.

4. Calculate the point M.

5. Draw a line from RN through M to find S1 on the solubility curve.

6. Locate the Δ point by drawing two lines, one through S0 and RN, and the other
through S1 and F.

7. Step off stages, alternately using the LLE tie-lines and the A point, as follows:

LLE Δ LLE A

S1 Rl S2 R2 S3 •••

Example 4.8.1.

236
Prof Fatma K. Gad

Design a countercurrent liquid-liquid extractor


to separate acetone from methyl isobutyl ketone
(mik) using water as a solvent. The column is to
operate at 25°c.
The organic feed rate is 10,000 kg/hr, and the
composition is 45 weight percent acetone and 55
weight percent ketone. The fresh solvent is pure
water, and twice the minimum solvent rate is to
be used. The concentration of acetone in the
organic raffinate leaving the top of the extractor
is to be 2.5 weight percent acetone.

Solution:

In this system, acetone is the solute to be


separated from mik using water as the solvent.
Thus, component 1 is acetone and component 3
is water.

The point rn is plotted on the


Minimum solvent rate:
solubility curve at x1RN = 2.5 weight percent
acetone. The point s0 is plotted at the right
lower corner of the diagram, since x3S0 = 100
237
Prof Fatma K. Gad

weight percent water. the point F is plotted on


the vertical axis at z1 = 45 weight percent
acetone.

The LLE tie-line that passes through f is used to


determine the points Δmin and s1min. a straight
line is drawn from rn to s1min, and another from
F to S0. The intersection of these lines is the
point Mmin, giving a composition x1min = 24
weight percent acetone. The minimum solvent
flow rate is calculated:

S0min = F (z1 – x1min) / ( x1min – x1S0) =


(100000) (0.45 – 0.24) / (0.24 – 0)
= 8750 kg / hr minimum solvent flow rate.
The actual solvent rate is twice the minimum:

S0 = 2 (8750) = 17,500 kg/hr

Calculate the point M


M = F + S0 = 10,000 + 17,500 = 27,500 kg/hr
X1M = (F z1 + S0 x1S0) / M= [(10,000)(45) + (17,500)(0)] / 27,500
= 16.4 weight percent acetone

238
Prof Fatma K. Gad

Determine S1: Draw a straight line from RN through M to the solubility

curve. The composition of the solvent phase on stage 1 is found from Figure 4.9-2
to be x1S1 = 19 weight percent acetone.

336
Chap
. 9 Liquid-Liquid Extraction

Figure 4.8-3: Minimum solvent rate for example 4.8-1


The flow rate of S 1 can be calculated from a total mass balance and an acetone
component balance around the entire system:

F + S0 = RN + S1

z1 F + x1S0 = x1RN + x1S1 S1

10,000 + 17,500 = RN + S 1

(45)(10,000) + (0)(17,500) = (2.5) RN + (19) S 1

S1= 23,100 kg/hr RN = 4,400 kg/hr

Note that the original organic feed rate was 10,000 kg/hr, but the organic raffinate
leaving the unit is only 4,400 kg/hr. The recovery of acetone is found as follows:

Acetone in Feed = (10,000)(0.45) = 4,500 kg/hr acetone

Acetone in Raffinate (Rn) = (4,400)(0.025) = 110 kg/hr acetone

Acetone in Fat Solvent (S1) = (23,100)(0.19)


= 4,390 kg/hr acetone recovered

239
Prof Fatma K. Gad

Recovery of acetone = 4,390/4,500 = 0.976

Thus, 97.6 percent of the acetone fed to the unit is captured by the solvent.

Step Off Stages

As shown in Figure 4.9-3, LLE tie-lines and the Δ point are used alternately to step
off five stages. On the fifth stage, the concentration of the raffinate phase has been
reduced to the desired value of 2.5 weight percent acetone. Thus, five equilibrium
stages are required to achieve the separation at a solvent rate twice the minimum.

The problems discussed have been design problems, i.e., problems concerned with
determining the number of stages of a system.
Just as in distillation, however, there are also many interesting rating problems in
liquid-liquid extraction. For example, we may want to find the flow rate of fresh
solvent required to attain a specified recovery of solute in a column with a fixed
number of stages.
These rating problems are usually solved iteratively. For the previous example, a fresh
solvent flow rate would be guessed, and the number of stages in the actual column
would be stepped off to see if the desired recovery was achieved. If not, a new value
of fresh solvent flow rate would be guessed. Some of the problems at the end of the
chapter are rating problems.

Figure 4.8-3: Stage to stage calculations for example 4.8-1.

240
Prof Fatma K. Gad

CHAPTER 5

MEMBRANE SEPARATION PROCESSES

5.1 INTRODUCTION AND TYPES OF MEMBRANE SEPARATION


PROCESSES

5.1A Introduction

Separations by the use of membranes are becoming increasingly important in the


process industries. In this relatively new unit operation, the membrane acts as a semi-
permeable barrier and separation occurs by the membrane controlling the rate of
movement of various molecules between two liquid phases, two gas phases, or a
liquid and gas phase. The two fluid phases are usually miscible and the membrane
barrier prevents actual, ordinary hydrodynamic flow. A classification of the main
types of membrane separation is as follows.

241
Prof Fatma K. Gad

5.I Classification of Membrane Processes

1 . Gas diffusion in porous solid.

In this type a gas phase is present on both sides of the membrane, which is a micro-
porous solid. The rates of molecular diffusion of the various gas molecules depend on
the pore sizes and the molecular weights. This type of diffusion in the molecular,
transition, and Knudsen regions.

2.Gas permeation in a membrane.

The membrane in this process is usually a polymer such as rubber, polyamide, and so
on, and is not a porous solid. The solute gas first dissolves in the membrane and then
diffuses in the solid to the other gas phase. This was discussed in detail for solutes
following Fick's law and is considered again for the case where resistances are
present.

Examples are hydrogen diffusing through rubber and helium being separated from
natural gas by permeation through a fluorocarbon polymer. Separation of a gas
mixture occurs since each type of molecule diffuses at a different rate through the
membrane.
3.Liquid permeation or dialysis.

In this case the small solutes in one liquid phase diffuse readily because of
concentration differences through a porous membrane to the second liquid phase (or
vapour phase). Passage of large molecules through the membrane is more difficult.
This membrane process has been applied in chemical processing separations such as
separation of H2S04 from nickel and copper sulfates in aqueous solutions, food
processing, and artificial kidneys and is covered in detail in Section 5.2. In electro-
dialysis, separation of ions occurs by imposing an emf difference across the
membrane.

4.Reverse osmosis.

A membrane, which impedes the passage of a low-molecular-weight solute, is placed


between a solute-solvent solution and a pure solvent. The solvent diffuses into the
solution by osmosis. In reverse osmosis a reverse pressure difference is imposed
which causes the flow of solvent to reverse as in the desalination of seawater. This
process also is used to separate other low-molecular-weight solutes, such as salts,

242
Prof Fatma K. Gad

sugars, and simple acids from a solvent (usually water). This process is covered in
detail in Sections 13.9 and 13.10.

5.Ultrafiltration membrane process.

In this process, pressure is used to obtain a separation of molecules by a


semipermeable polymeric membrane (M2). The membrane discriminates on the basis
of molecular size, shape, or chemical structure and separates relatively high molecular
weight solutes such as proteins, polymers, colloidal materials such as minerals, and so
on. The osmotic pressure is usually negligible because of the high molecular weights.

6.Gel permeation chromatography.

The porous gel retards diffusion of the high-molecular-weight solutes. The driving
force is concentration. This process is quite useful in analyzing complex chemical
solutions and purification of very specialized and/or valuable components.

5.2 GAS PERMEATION MEMBRANE PROCESSES

5.2 Series Resistances in Membrane Processes

In membrane processes with two gas phases and a solid membrane. The
equilibrium relation between the solid and gas phases is given by

(5.2-1)

where S is the solubility of A in m3 (STP)/atm • m3 solid, and H is the equilibrium


relation in kg mol/m3 ■ atm. This is similar to Henry's law. The flux equations in each
phase are as follows:

(5.2-2)

243
Prof Fatma K. Gad

The permeability P M in kg mol/s - m • atm is given by

(5.2-3)

Eliminating the interfacial concentrations as before,

(5.2-4)

In the case where pure A ( p A l ) is on the left side of the membrane, there is no
diffusional resistance in the gas phase and kcl can be considered to be infinite. Note
that k G l = k c l / R T .

An example of gas permeation in a membrane is use of a polymeric membrane as an


oxygenator for a heart-lung machine to oxygenate blood. In this biomedical appli-
cation, pure O2 gas is on one side of a thin membrane and blood is on the other side.
Oxygen diffuses through the membrane into the blood and CO2 diffuses in a reverse
direction into the gas stream.

Types of Membranes and Permeabilities for Separation of Gases

1 . Types of membranes. Early membranes were limited in their use because of low-
selectivities in separating two gases and quite low permeation fluxes. This low-flux
problem was due to the fact that the membranes had to be relatively thick (1 mil or
1/1000 of an inch or greater) in order to avoid tiny holes which reduced the separation
by allowing viscous or Knudsen flow of the feed. Development of silicone polymers
(1 mil thickness) increased the permeability by factors of 10 to 20 or so.

Some newer asymmetric membranes include a very thin but dense skin on one side of
the membrane supported by a porous substructure (Rl). The dense skin has a thickness
of about 1000 A and the porous support thickness is about 25-100 /n.m. The flux
increase of these membranes is thousands of times higher than the 1-mil-thick original
membranes. Some typical materials of present membranes are a composite of
polysulfone coated with silicone rubber, cellulose acetate and modified cellulose
acetates, aromatic polyamides or aromatic polyimides, and silicone-polycarbonate
copolymer on a porous support.

2 . Permeability of membranes. The accurate prediction of permeabilities of gases


in membranes is generally not possible, and experimental values are needed.
Experimental data for common gases in some typical membranes are given in Table

244
Prof Fatma K. Gad

5.3-1. Note that there are wide differences among the permeabilities of various gases
in a given membrane. Silicone rubber exhibits very high permeabilities for the gases
in the table.

For the effect of temperature T in K, the In P'A is approximately a linear function of


1/T and increases with T . However, operation at high temperatures can often degrade
the membranes. When a mixture of gases is present, reductions of permability of an
individual component of up to 10% or so can often occur. In a few cases much larger
reductions have been observed (Rl). Hence, when using a mixture of gases,
experimental data should be obtained to determine if there is any interaction between
the gases. The presence of water vapor can also have similar effects on the
permeabilities and can also possibly damage the membranes.

5.3 Types of Equipment for Gas Permeation Membrane Processes

1. Flat membranes. Flat membranes are mainly used for experimental use to
characterize the permeability of the membrane. The modules are easy to fabricate and
use and the areas of the membranes are well defined. In some cases modules are
stacked together like a multilayer sandwich or plate-and-frame filter press. The major
drawback of this type is the very small membrane area per unit separator volume.

Small commercial flat membranes are used-for producing oxygen-enriched air for
individual medical applications.

2.Spiral-wound membranes. This configuration maintains the simplicity of fabricating


flat membranes while increasing markedly the membrane area per unit separator
volume up to 100 ft2/ft3 (328 m2/m3) while decreasing pressure drops (Rl). The
assembly consists of a sandwich of four sheets wrapped around a central core of a
perforated collecting tube. The four sheets consist of a top sheet of an open separator
grid for the feed channel, a membrane, a porous felt backing for the permeate channel,
and another membrane as shown in Fig. 5.3-1. The spiral-wound element is 100 to
200 mm in diameter and is about 1 to 1.5 m long in the axial direction. The flat sheets
before rolling are about 1 to 1.5 m by about 2 to 2.5 m. The space between the
membranes (open grid for feed) is about 1 mm and the thickness of the porous
backing (for permeate) is about 0.2 mm.

Table 5.3-1: Permeabilities of various gases in membranes.

245
Prof Fatma K. Gad

The whole spiral-wound element is located inside a metal shell. The feed gas enters at
the left end of the shell, enters the feed channel, and flows through this - channel in
the axial direction of the spiral to the right end of the assembly (Fig. 5.3-1). Then the
exit residue gas leaves the shell at this point. The feed stream, which is in the feed
channel, permeates perpendicularly through the membrane. This permeate then flows
through the permeate channel in a direction perpendicular to the feed stream toward
the perforated collecting tube, where it leaves the apparatus at one end. This is
illustrated in Fig. 5.3-2, where the local gas flow paths are shown for a small element
of the assembly.

3.Hollow-fiber membranes. The membranes are in the shape of very small diameter
hollow fibers. The inside diameter of the fibers is in the range of 100 to 500 µm and
the outside 200 to 1000 µm with the length up to 3 to 5 m. The module resembles a
shell-and-tube heat exchanger. Thousands of fine tubes are bound together at each
end into a tube sheet that is surrounded by a metal shell having a diameter of 0.1 to
0.2 m, so that the membrane area per unit volume is up to 10 000 m2/m3 as in Fig. 5.3-
3.

Typically, the high-pressure feed enters into the shell side at one end and leaves at the
other end. The hollow fibers are closed at one end of the tube bundles. The permeate
gas inside the fibers flows countercurrently to the shell-side flow and is collected in a
chamber where the open ends of the fibers terminate. Then the permeate exits the
device.

246
Prof Fatma K. Gad

Figure 5.3-1: Spiral wound elements and assembly.

Figure 5.3-2: Local gas flow path for spiral wound separator.

247
Prof Fatma K. Gad

Figure 5.3-3: Hollow finer separator assembly.

5.4 Introduction to Types of Flow in Gas Permeation

1.Types of flow and diffusion gradients.

In a membrane process high-pressure feed gas is supplied to one side of the


membrane and permeates normal to the membrane. The permeate leaves in a direction
normal to the membrane, accumulating on the low-pressure side. Because of the very
high diffusion coefficient in gases, concentration gradients in the gas phase in the
direction normal to the surface of the membrane are quite small. Hence, gas film
resistances compared to the membrane resistance can be neglected. This means that
the concentration in the gas phase in a direction perpendicular to the membrane is
essentially uniform whether or not the gas stream is flowing parallel to the surface or
is not flowing.

If the gas stream is flowing parallel to the membrane in essentially plug flow, a
concentration gradient occurs in this.direction. Hence, several cases can occur in the
operation of a membrane module. The permeate side of the membrane can be operated
so that the phase is completely mixed (uniform concentration) or where the phase is in

248
Prof Fatma K. Gad

plug flow. The high-pressure feed side can also be completely mixed or in plug flow.
Countercurrent or cocurrent flow can be used when both sides are in plug flow.
Hence, separate theoretical models must be derived for these different types of
operation.

2.Assumptions used and ideal flow patterns.

In deriving theoretical models for gas separation by membranes, isothermal


conditions and negligible pressure drop in the feed stream and permeate stream are
generally assumed. It is also assumed that the effects of total pressure and/or
composition of the gas are negligible and that the permeability of each component is
constant (i.e., no interactions between different components).

Since there are a number of idealized flow patterns, the important types are
summarized in Fig. 5.3-1. In Fig. 5.3-1a complete mixing is assumed for the feed
chamber and the permeate chamber. Similar to a continuous-stirred tank, the reject or
residue and the product or permeate compositions are equal to their respective
uniform compositions in the chambers.

Figure 5.4-1: Ideal flow patterns in a membrane separator for gases:


a)complete mixing. b) ctoss flow.
c)countercurrent flow. D)concurrent flow.

An ideal cross-flow pattern is given in Fig. 5.4-1b, where the feed stream is in plug
flow and the permeate flows in a normal direction away from the membrane without

249
mixing. Since the feed composition varies along its flow path, the local permeate
concentration also varies along the membrane path.
In Fig. 5.4-1c both the feed stream and permeate stream are in plug flow
countercurrent to each other. The composition of each stream varies along its flow
path. Cocurrent flow of the feed and permeate streams is shown in Fig. 5.4-1d.

5.5 COMPLETE-MIXING MODEL FOR GAS


SEPARATION BY MEMBRANES

5.5 Basic Equations Used

In Fig. 5.5-1 a detailed process flow diagram is shown for complete mixing. When
a separator element is operated at a low recovery (i.e., where the permeate flow rate is
a small fraction of the entering feed rate), there is a minimal change in composition.
Then the results derived using the complete-mixing model provide reasonable esti-
mates of permeate purity. This case was derived by Weller and Steiner (W4). The
overall material balance (Fig. 5.5-1) is as follows:

qf = qo + qp (5.5-1)

where is total feed flow rate in cm3(STP)/s; q0 is outlet reject flow rate, cm3(STP)/s;
and q p is outlet permeate flow rate, cm3(STP)/s. The cut or fraction of feed
permeated,  , is given as

 = qp / qf (5.5-2)

The rate of diffusion or permeation of species A (in a binary of A and B ) is given


below which uses cm3(STP)/s as rate of permeation rather than flux in kg mol/s • cm2.

(5.5-3)

where P'A is permeability of A in the membrane, cm3(STP) • cm/(s • cm2 • cm Hg);


q A is flow rate of A in permeate, cm3(STP)/s; A m is membrane area, cm2; t is
membrane thickness, c m / p h is total pressure in the high-pressure (feed) side, cm Hg;
p, is total pressure in the low-pressure or permeate side, cm Hg; x a is mole fraction of
A in reject side; x j is mole fraction of A in feed; and y p is mole fraction of A in A
similar equation can be written for component B . permeate. Note that phx0 is the
partial pressure of A in the reject gas phase.
Figure 5.5-1: Process flow for complete mixing case.

A similar equation can be written for component B.

(5.5-4)
where P ' B is permeability of B , cm3(STP) • cm/(s- cm2 • cm Hg). Dividing Eq.
(5.5-3) by (5.5-4)

(5.5-5)

This equation relates y p , the permeate composition, to x D , the reject composition,


and the ideal separation factor * is defined as

(5.5-6)
Making an overall material balance on component A

qf xf = qo xo + qp yp (5.5-7)

Dividing by qf and solving for the outlet reject composition,


(5.5-8)

Substituting q p =  qf from Eq. (5.5-2) into Eq. (5.5-3) and solving for the membrane
area, Am,

(5.5-9)

Solution of Equations for Design of Complete-Mixing Case

For design of a system there are seven variables in the complete-mixing model
(H1), x f , x o , y p ,  ,  * , P l / p h , and A m , four of which are independent variables.
Two commonly occurring cases are considered here.

Case 1. This is the simplest case where x f , x o ,  * , and p l / p h are given and y p ,
 , and A m are to be determined by solution of the equations. By use of the quadratic
equation formula, Eq. (5.5-5) is solved for the permeate composition y p in terms of
xQ.

(5.5-10)

Hence, to solve this case, y p is first calculated using Eq. (5.5-10). Then the fraction
of feed permeated, 9 , is calculated using Eq. (5.5-8) and the membrane area, Am from
Eq. (5.5-9).

EXAMPLE 5.5-1. Design of a Membrane Unit for Complete Mixing


A membrane is to be used to separate a gaseous mixture, of A and B whose feed
flow rate is q f = 1 x 10 cm3(STP)/s and feed composition of A is x f = 0.50 mole
fraction. The desired composition of the reject is x a = 0.25. The membrane thickness
t = 2.54 x 10-3 cm, the pressure on the feed side is p h = 80 cm Hg and on the permeate
side is p x = 20 cm Hg-The permeabilities are P'A = 50 x 10~10 cm3(STP) • cm/(s • cm2
• cm Hg) and P'B = 5 x 10-10. Assuming the complete-mixing model, calculate the
permeate composition, y p , the fraction permeated, 9 , and the membrane area, A m .

Solution:

Substituting into Eq. (5.5-6),

Using Eq. (5.5-10),


Using the material balance equation (5.5-8),

Solving,  = 0.706. Also, using Eq. (5.5-9),

Case 2 . In this case xf,  ,  * , and p l / p h are given and y p , x 0 , and


Am are to be determined. Equation (5.5-5) cannot be solved f o r y p since
xo is unknown. Hence, x o from Eq. (5.5-8) is substituted into Eq. (5.5-5)
and the resulting equation solved for y p using the quadratic equation to
give

(5.5-11)
Where

After solving foryp, the value of x0 is calculated from Eq. (5.5-8) and
A m from Eq. (5.5-9).

EXAMPLE 5.5-2. Membrane Design for Separation of Air

It is desired to determine the membrane area needed to separate an air


stream using a membrane 1 mil thick with an oxygen permeability of P'A

DR. FATMA KHALIFA GAD 254


= 500 x 10-10 cm3(STP)-cm/(s-cm2-cmHg). An * = 10 for oxygen
permeability divided by nitrogen permeability (S6) will be used. The
feed rate is q f = 1 x 106 cm3(STP)/s and the fraction cut  = 0.20. The
pressure selected for use are p h = 190 cm Hg and p t = 19 cm Hg. Again,
assuming the complete-mixing model, calculate the permeate
composition, the reject composition, and the area.

Solution:

Using Eq. (13.4-11) for a feed composition of x f = 0.209,

Substituting into Eq. (5.5-8),

DR. FATMA KHALIFA GAD 255


Finally, using Eq. (5.5-9) to find the area

DR. FATMA KHALIFA GAD 256