UNIT-III

CORRELATION AND PREDICTION OF PHASE EQUILIBRIA

1. Wohl’s three-suffix Equations:-

Wohl’s proposed a equation for expressing excess free energy contained terms for compositions,
effective molar volume & effective volumetric fraction of the separate constituents of solution.

  q  
ln  1  z 22  A  2 B 1  A  Z 1 ;
  q2  
  q  
ln  2  z12  B  2 A 2  B  Z 2     (1)
  q1  
Z1 & Z2 – effective volume fraction of separate components in the solution
q1 & q2 - effective molal volume
x1 x2
z1  , z2 
q  q 
x1  x 2  2  x 2  x1  1 
 q1   q2 
Eqn(1) is known as Wohl’s three-suffix equation. It involves three parameters A, B & (q1 / q2) are the
characteristics of binary system.

2. Margules Equation:-

When the term q1 / q2 =1, we get the following expression known as Margules three suffix equation.
By substituting Z1 & Z2 values, eqn (1) becomes,

ln  1  x 22 A  2B  Ax1   (2 B  A) x 22  2 A  B x 23
ln  2  x12 B  2 A  B x 2   (2 A  B) x12  2B  Ax13      (2)
Constant A – terminal value of ln  1 at x1= 0
Constant B – terminal value of ln  2 at x2= 0
 The 3-suffix magules equation represent the VLE data of systems like acetone-methanol,
acetone-chloroform.
When A=B in eqn.(2),
ln  1  Ax 22 ; ln  2  Ax12    (3)
Eqn.(3) is called Margules two suffix equation.
 It represents accurately and sufficiently the activity coefficients of simple liquid mixtures
(i.e.) mixtures of molecules, which are similar in shape, size and chemical nature.
 The constant A may be positive or negative. It depends on temperature.
 The VLE data of argon-oxygen, benzene-cyclohexane etc., are well represented by Margules
equation.

3. Van Laar Equation:-

Let (q1 / q2) = (A/B) in Eqn.(1). The resulting two parameter equation is known as the Van Laar
Equation.
Ax 22
ln  1  Az 22 
 
A x x
B 1 2
2


1
 Bx12
ln  2  Bz 2
     (4)
1
x  B Ax 
1 2
2

Constant A – terminal value of ln  1 at x1= 0

Constant B – terminal value of ln  2 at x2= 0
When A= B, the Van Laar equations simplify to the Margules equation (3). Equation (4) is very
convenient for evaluation of constant A & B.
2 2
 x ln  2   x ln  1 
A  ln  1 1  2  B  ln  2 1  1     (5)
 x1 ln  1   x 2 ln  2 

 Applicable for simple solutions preferably non-polar liquids.

 Applicable for more complex mixtures
 Van Laar equations are widely used for VLE calculations because of their flexibility and
mathematical simplicity.
 Activity coefficients in benzene-isooctane system, n-propanol water system etc., are accurately
represented by this equation.
Choice of Equation:
 The selection of proper equation for VLE data correlation depends on the molecular complexity
of the system and the precision of the experimental data.
 When an equation selected that fits experimental well, the constants for the constant pressure
conditions will be different from those applicable for constant temperature conditions.
 The effect of pressure on the constants is usually negligibly small, whereas the effect of
temperature is appreciable and cannot be neglected. However in VLE calculations the effect of
temperature on activity coefficient is usually ignored.
 The Margules three-suffix equation is suited for symmetrical systems.(constant A & B are
same). The Van Laar equations can be used for unsymmetrical solutions, where the ratio A/B
does not exceed 2. They cannot represent maxima or minima in the lnγ curve. Margules three
suffix equation should be used in such cases.
 For choice of appropriate equation, a rule of thumb is used: When the ratio of molar volumes
close to unity, the margules equation is preferred. When the ratio is different from unity, as is
the case when water is one of the constituents, the van Laar equation is found to be satisfactory.
 For example, the chloroform-ethyl alcohol system, which shows a maximum and a minimum
on the lnγ curves and whose ratio between pure component molar volume is 1.38, is accurately
represented by the Margules equation. For n-propanol-water system, this ratio is 4.16 and the
van Laar equations are found to represent the behavior accurately.
 In equations having only two constants, determination of γ1 and γ2 at a single known
composition permits the evaluation of the constants and the complete γ curve by the following
equation.
yP
 i  i S    (1)
xi Pi
 The data required are a single set of equilibrium vapour-liquid composition values and the
vapour pressures of the pure components. When azeotropes is formed, only the azeotropic
composition need be known, because it represents the composition of both the liquid and the
vapour phases. The activity coefficients can be evaluated by putting x = y
P
 i  S    (2)
Pi
4. Wilson Equation:
 Many equations cannot be derived from the Wohl’s general equation. Among such equations,
the Wilson equation, the NRTL equation & the UNIQUAC equation are the important from
practical point of view. All these are based on the concept of local compositions, which are
different from overall mixture compositions.
2
  Wilson proposed an equation for γ in a binary mixture,

 Wilson equations have two adjustable positive parameters Λ12 and Λ21. These are related to the
pure component molar volumes and to the characteristics energy differences by

where V1 & V2 - the molar volumes of pure liquids

λ’s - energies of interaction between the molecules designated in the subscripts.
 The differences in the characteristic energies (aij) are assumed to be temperature independent &
this introduces no serious error in practical calculations.
 It provides a good representation for of VLE of a variety of miscible mixtures & polar solutions
like alcohol in non-polar solvents for which the Margules and Van Laar equations are
inadequate.
 It is recommended only for liquid systems that are completely miscible, or for partially miscible
systems in the region where only one liquid phase exists.
 Disadvantages: (i) not suitable for system showing maxima or minima on the lnγ Vs x-curves.
(ii) not suitable for system exhibiting limited miscibility.

5. Non-random two-liquid (NRTL) Equation:

The NRTL model, proposed by Renon and Prausnitz (1968), also based on the local composition
concept. The activity coefficients are

The adjustable parameters are evaluated as

Constants b12 and b21 - similar to constants representing characteristic energy difference in Wilson
Equation
α12 - independent of composition and temperature
- related to non-randomness in the mixture
 If α12= 0, the mixture is completely random & NRTL equation reduces to Margules equation.
 This is applicable to partially miscible as well as totally miscible systems.
3
  For strongly non-ideal solutions and especially partially miscible systems, the NRTL
equations provide a good representation.

6. Universal quasi-chemical(UNIQUAC) equation:-

 Abrams & Prausnitz (1975) extended the quasi-chemical theory of liquid mixtures to solutions
containing molecules of different molecules of different sizes. This extension is called the
UNIQUAC theory.
 The UNIQUAC model consists of two parts
1) Combinatorial part – describes the prominent entropic contribution – determined by sizes and
shapes of the molecules require only pure component data.
2) Residual part – primarily due to intermolecular forces responsible for enthalpy of mixing-
depends on the intermolecular forces & involves two adjustable binary parameters.
The UNIQUAC equations for activity coefficients are

z – coordination number,
r, q & q' – pure component molecular structure constants
r – molecular size
q- surface area
 For fluids other than water or lower alcohols, q = q'
 For alcohols, the suface of interaction q' is smaller than geometric surface q.
 The adjustable binary parameters τ12 & τ21 are related to the characteristic energies Δu as follows.

4
 It satisfies a large number of non-electrolyte mixtures containing non-polar fluids such as
hydrocarbons, alcohols, nitriles, ketones, aldehydes, organic acids, etc., and water including
partially miscible mixtures.
 The main advantages of this equation are its wide applicability and simplicity arising primarily from
two adjustable parameters.

7. Universal functional activity coefficient (UNIFAC) method:-

 Activity coefficient estimated through group contributions.

 Liquid treated as solution of different structural group from pure component data which molecules
are formed.
 Based on UNIQUAC model.
 The UNIQUAC model consists of two parts
i. Combinatorial part –molecular size contribution – estimated from pure component properties
sizes and shapes of the molecules.
ii. Residual part –interaction contribution- group areas & group contributions are needed.
ln  i  ln Ci ln  iR
 UNIFAC method is successful for design of distillation column involving even azeotropic and
extractive distillation.

Problem1: Liquids A and B form an azeotrope containing 46.1 mole % A at 101.3 kPa and 345K. At
345k, the vapour pressure of A is 84.8kPa and that of B is 78.2kPa. Calculate the Van Laar constant.
Solution:-
Material A – component 1 & Material B – component 2
The activity coefficients at the azeotropic concentration can be evaluated as

P 101.3
1  
P1S 84.8
= 1.195
P 101.3
2    1.295
P2S 78.2
At x1 =0.461 & x2 = 0.539
2 2
 x ln  2   0.539 ln 1.1295 
A  ln  1 1  2   ln 1.1951 
 x1 ln  1   0.461 ln 1.195 
= 1.2955
2 2
 x ln  1   0.461  ln 1.195 
B  ln  2 1  1   ln 1.2951 
 x 2 ln  2   0.539  ln 1.295 

= 0.6530

Problem2: The azeotope of the ethanol-benzene system has a composition of 44.8 mole % ethanol with
a boiling point of 341.4K at101.3 kPa . At this temperature the vapour pressure of benzene is 68.9 kPa is
84.8kPa and the vapour pressure of ethanol is 67.4kPa. What are the activity coefficients in a solution
containing 10% alcohol.
Solution:-
P 101.3
1  S

P1 68.9 5
 = 1.4702

P 101.3
2  
P2S 67.4 = 1.5030

At x1 =0.552 & x2 = 0.448

2
 x ln  2 
A  ln  1 1  2 
 x1 ln  1 
= 1.3305
2
 x ln  1 
B  ln  2 1  1 
 x 2 ln  2 
= 1.9106

For 10% alcohol, x1 = 0.9 & x2 = 0.1

1.3305  0.1
2
Ax 22
ln  1  
A B x  x 
1 2
2
  1.3305 
 

0.9  0.1
2

  1.906  
= 0.0252
1.9106  0.9 
2
Bx12
ln  2  
   
x1  B x2
A
2
  1.9106  
 0.9   0.1
2

  1.3305  
=1.4210
  1  1.0255
 2 4.1412

Problem3: Water (1) –hydrazine(2) system forms an azeotrope containing 58.5% (mol) hydrazine at
393K and 101.3 kPa. Calculate the equilibrium vapour composition for a solution containing 20% (mol)
hydrazine. The relative volatility of water with reference to hydrazine is 1.6 and may be assumed to
remain constant in the temperature range involved. The vapour pressure of hydrazine at 393K is
124.76kPa.

Solution:

Vapour pressure of water at 393K = 1.6 X V.P. of hydrazine at 393K

= 1.6 X 124.76 =199.62 kPa
 P1  199.62kPa P2  124.76kPa
S S

For azeotrope,
x1  0.415, x 2  0.585
P 101.3
1  S
 = 1.5075
P1 199.62
P 101.3 = 1.5030
2  S

P2 124.76
 0.585   0.2083 
2 2
 x ln  2 
A  ln  1 1  2   ln 0.50751  
 x1 ln  1   0.415   0.6783 
= - 1.3927

6
  0.415   0.6783 
2 2
 x ln  1 
B  ln  2 1  1   ln 0.8120 1  
 x 2 ln  2   0.585   0.2083 
= - 2.2822
For solution containing 20 mol% hydrazine,

x1  0.8, x 2  0.2
 1.3927  0.2 
2
Ax 22
ln  1  
A B x  x 
1 2
2
   1.3927 
 

0.8  0.2 
2

   2.2822  
= - 0.1176
  1  0.8891
 2.2822  0.8
2
Bx12
ln  2  
x  B Ax 
1 2
2
   2.2822  
 0.8   0.2 
2

   1.3927  
= - 1.1485
 2 0.3171
wkt, y i P   i x i Pi
S

 1 x1P1S  2 x2 P2S
y1  , y2 
P P
To evaluate vapour compositions using the equations, we should know the values at the new
equilibrium temperature. Taking ratio of last two equations,

y 2 1  y1  2 x 2 P2S  x
   2 2
y1 y1  1 x1 P1S
 1 x1 12
y  x
1 2  1 2 2
y1  1 x1 12
1 1
y1  
 x 0.3171  0.2
1 2 2 1
 1 x1 12 0.8891  0.8  1.6
= 0.9472

y2 = 1-0.9472 =0.0528

The composition of vapour in equilibrium with the liquid containing 20% hydrazine is 5.28%
hydrazine and 94.72% water.

Problem4: A stream of isopropanol-water mixture is flashed into a separation chamber at 353K and
91.2kPa. A particular analysis of the liquid product showed an isopropanol content of 4.7%(mol), a valve
that deviated from norm. It is suspected that an air leak into the separator might have caused this. Do you
agree? The vapour pressure of pure propanol and water are 91.11kPa and 47.36kPa respectively and
vanlaar constants are A =2.470 and B = 1.094.
Solution:

For x1  0.047, x 2  0.953

A =2.470
B = 1.094.
7
 P = 91.2 kPa
 P1S  91.11kPa P2S  47.36kPa

Ax 22   1  7.388
ln  1 
A B x  x 
1 2
2

Bx12  2 1.011
ln  2 
x  B Ax 
1 2
2

The total pressure corresponding to this equilibrium

P   1 x1P1S   2 x2 P2S
= 0.047 x 7.388 x 91.11 +1.011 x 0.953 x 47.36 = 77.27 kPa
This is less than total pressure. This error must have been caused by air-leak.

Problem5: Ethyl alcohol and hexane form an azeotrope at 33.2% (mol) ethanol. It boils at 331.9 K at
101.3 KPa. At 331.9 K, the vapour pressures are 44.25 KPa for ethanol and 72.24 KPa for hexane.
Determine
(i) The Van laar constants
(ii) The vapor composition for a solution containing 50%(mol) hexane boiling at
331.9 K.
(iii) The total pressure for the conditions in Part (ii)

Solution:

For x1  0.668, x 2  0.332

P = 101.3 kPa
 P1S  72.24kPa P2S  44.25kPa

P 101.3
1  S

P1 72.24

= 1.4023

P 101.3
2  S

P2 44.25
= 2.289
(i) The Van laar constants:

0.332  ln 2.289 
2 2
 x ln  2  
A  ln  1 1  2   ln 1.40231  
 x1 ln  1   0.668  ln1.4023 
= 1.6623
 0.668  ln 1.4023 
2 2
 x ln  1 
B  ln  2 1  1   ln 2.289 1 
 x 2 ln  2   0.332  ln 2.289  
= 2.7475

(ii) For 50%(mol) hexane :

x1  0.5, x 2  0.5
8
 Ax 22  1 1.9065
ln  1 
A B x  x 
1 2
2

Bx12   2  1.4776
ln  2 
x  B Ax 
1 2
2

wkt, y i P   i xi Pi S

 1 x1P1S  2 x2 P2S
y1  , y2 
P P
 xP 1.9065  0.5  72.24
S
y1  
1 1 1
P 101.3
= 0.6773
y 2  1  y1
= 1- 0.6773
= 0.3227

(iii) The total pressure for the conditions in Part (ii):-

P   1 x1P1S   2 x2 P2S
 1.9065  0.5  72.24  1.4776  0.5  44.25
P = 101.56kPa

Problem6: Construct P-x-y diagram for cyclohexane(1)-benzene(2) system at 313K given that at 313K the
vapour pressures are P1S =24.62 kPa and P2S =24.41kPa. The liquid phase activity coefficients are given by
ln  1  0.458 x 22 ln  2  0.458 x12
Solution:

For x1  0.4, x 2  0.6

 P1S  24.62kPa P2S  24.41kPa

ln  1  0.458  0.6 2  0.16488

 1  1.1793
ln  2  0.458  0.4 2  0.0733
 2  1.0760
The total pressure is determined as
P   1 x1P1S   2 x2 P2S
= 1.1793 x 0.4 x 24.62 +1.0760 x 0.6 x 24.41
= 27.37 kPa

The vapour composition is found out from

 x PS
y1  1 1 1
P
1.1793  0.4  24.62

27.37
= 0.4243
The above calculation is repeated for various x1 values. The results are tabulated below:
9
 x1 0 0.2 0.4 0.6 0.8 1.0

γ1 1.5809 1.3406 1.1793 1.0760 1.0185 1.000

γ2 1.000 1.0185 1.0760 1.1793 1.3406 1.5809

P 24.41 26.49 27.37 27.41 26.61 24.62

y1 0 0.2492 0.4243 0.5799 0.7540 1.0

The results are plotted by taking P on the Y-axis and x1 & y1 on the x-axis.

Problem7: The following values refer to the Wilson parameters for the system acetone(1)-water(2):

a12  1225.31J / mol , a 21  6051.01J / mol V 1 74.05  10 6 m 3 / mol , V2  18.07  10 6 m 3 / mol.

The vapour pressures are given by

2795.817 3799.887
ln P1s  14.39155  , ln P2s  16.26205  ,
T  43.198 T  46.854

Where P is in kPa and T is in K. Calculate the equilibrium pressure and composition of

a) Vapour in equilibrium with a liquid of composition x1 =0.43 at 349 K.

b) The liquid in equilibrium with a vapour of concentration y1=0.8 at 349 K.

Solution: Using Antoine equations, at 349K, the vapour pressures are calculated as

P1s  190.37 kPa, P2s  39.87 kPa,

The parameters in the Wilson equation are calculated as

18.07  1225.31 
12  exp    0.1600
74.05  8.314  349 

74.05  6051.01 
 21  exp    0.5092
18.07  8.314  349 

a) Vapour in equilibrium with a liquid of composition x1 =0.43 at 349 K.

10
  1  1.5820 &  2  1.4661

The pressure is calculated as

P   1 x1 P1S   2 x 2 P2s

 1.5820  0.43  190.37  1.4661  0.57  39.87

= 162.82 kPa

The vapour composition is found out from the relation

y1 P   1 x1 P1S

1.5820  0.43  190.37

y i   1 x1 P1
S

P 162.82

=0.7954

b) The liquid in equilibrium with a vapour of concentration y1=0.8 at 349 K.

Since liquid composition is not known, the activity coefficients cannot be calculated. Assume
 1   2  1.
Wkt, y i P  xi Pi S
For component 1 & 2,
yP y P
x1  1 S x2  2 S
P1 & P2
Since x1  x 2  1,
y1 P y 2 P
S
 S 1
To Calculate Pressure P1 P2
1
P
 y1   y2 
 S  S 
 P1   P2 
1
 0.80 0.20 
P  
 190.37 39.87 

= 108.48KPa
To calculate x1,
y P 0.80  108.48
x1  1 S 
P1 190.37
= 0.4559
x 2  0.5441
Activity coefficients can now be determined for this composition using Wilson equations.

11
  1  1.513 &  2  1.5191

Now the pressure is recalculated by incorporating the activity coefficient values. The equation

1
P
 y1   y2 
 S  S 
becomes   1 P1    2P2 
Substituting the values, we get P=164.48kPa. Liquid compositions are now recalculated,

y1 P 0.80  164.48
x1  
P1
S
1.513  190.37
= 0.4568
So that x2 =0.5432.
Activity coefficient under this condition are
 1  1.5109 &  2  1.5208
1
P
 y1   y2 
 S  S 
  1 P1    2P2 
= 164.48
This is same as the previous value calculated for P.
 P  164.48kPa & x1  0.4568

VAPOUR-LIQUID EQUILIBRIA INVOLVING HIGH PRESSURES & MULTICOMPONENT

SYSTEMS

Vaporisation Equilibrium Constants:

 For high pressure VLE calculations, it is convenient to express the phase equilibrium relations in terms
of vaporisation equilibrium constants or K factors.
 It is defined as the ratio of mole fraction in the vapour phase y to that in the liquid x or K = (y / x). Since
the fugacities in the liquid and vapour are equal.

f i 0,V  iV y i  xi  iL f i 0, L

y i  iL f i 0, L
K i 
K factor, xi  iV f i 0,V

K factor is also obtained from eqn. by assuming poynting factor of unity,

12
 y i iS  iL Pi S  i  iL Pi S
K i  
xi i P iP

 S 
i   i 
where  i 
 If the poynting factor is not unity, it should also be included.
 The K factor contains all thermodynamic information needed for VLE calculations and it is complex
functions of T, P and the vapour phase and liquid phase compositions.
 iS  y  PS
  1 K i i  i i
 If the vapour phase is assumed to behave as an ideal gas 
i  & xi P .
 i  1& y PS
K i i  i
 If the liquid phase behaves as an ideal solution xi P
 For ideal solutions, K factors depend only on T & P and are readily correlated as function of these two
variables.

Bubble-point Equilibria:

 The bubble-point temperature is the one at which the first bubble of vapour is produced from the liquid
on heating at constant pressure.
 At the bubble point the liquid has same composition as the original mixture.
 In problems, where bubble point is to be determined, the xi are known. Assume a temperature and get
the Ki values at this temperature. Calculate yi using yi = Kixi. If the assumed temperature is correct then,
 y i   K i xi  1
 Otherwise, repeat the calculations with another temperature. To find the bubble-point pressure, a similar
procedure as above is adopted by assuming various values of pressure until  K i xi  1

Dew-point Equilibria:

 The dew-point temperature is the one at which the first drop of condensate is formed on cooling a
vapour at constant pressure.
 The vapour in equilibrium with the liquid at the dew point has the same composition as the original
mixture.
 To find the dew-point temperature, a temperature is assumed arbitrarily and Ki is determined. Then,
y
xi  i
Ki
y
 x i   Ki  1
At the dew point, i

 Otherwise, repeat the calculations with another temperature till this equation is satisfied. To find the
dew-point pressure, a similar procedure as above is adopted by assuming pressure instead of
temperature.

Flash Vaporisation:

It can be stated as: Given a mixture of known overall composition zi at temperature T and pressure P, The
overall material balance for the system is

F=V+L

where ‘F’ is the total number of moles of the initial mixture. The component ‘i’ balance for the system is
13
 Fzi = Vyi + Lxi ------ (1)

Since yi = Kixi, it can be eliminated from above eqn. to get the following:

Fz i F zi
xi  
K iV  L V K i  L  V     (2)
x  1,
K
zi

V
     (3)
 
i

i 
L F
Since we have V
The above equation can be used for the calculation of T, P or fractional vapourisation in an iterative
procedure. Alternatively xi can be eliminated from eqn(1) which leads to
Fz i F zi
yi      ( 4)
V   L  V 
1  L 
 
 Ki   VK i 
 yi  1, z V
  i   F      (5)
1  L 
Since we have  VK i 
Eqn(5) can also utilized in an iterative procedure to estimate T,P or the fraction of the initial mixture
that is vapourised.

CONSISTENCY TEST FOR VLE DATA

 Many practical cases like distillation calculations are dependent on VLE data & such data should be
reasonably accurate if the results are to be reliable.
 As the VLE measurements are prone to inaccuracies, it is essential that we have some means for
checking the consistency of measured results.
 Thermodynamics provides tests for consistency of experimental VLE data.
 Almost all these tests are based on the Gibbs-Duhem equations written in terms of activity
coefficients.
1. Using slope of lnγ curve
2. Using data at the midpoint
3. Redlich-Kister Method
4. Using the Coexistence Equation
5. Using the Partial Pressure Data.

1. Using slope of lnγ curve:-

The Gibbs-Duhem equation in terms of activity coefficient provides a simple test for thermodynamic
consistency.

------ (1)

14
 Test for thermodynamic consistency of VLE data

 Plot the logarithm of the activity coefficient against mole fraction x1 of component 1 in a binary
solution as shown in fig.(a) and measures the slopes of the tangents drawn to the resulting curves at
any selected composition x1.
 Eqn.(1) tells us that if the GDE is to be satisfied, both slopes must have opposite sign. Otherwise, the
data are inconsistent.
 If the slopes are of opposite sign, substitute the values in eqn.(1) & if it is satisfied reasonably well,
then the data is consistent at the selected composition.
 For a complete test, the slopes determined at other compositions are substituted into Eqn.(1) to see
whether the equality is satisfied or not.

In addition to above observations, the following generalizations with the help of GDE,
1. If one of the lnγ curves has a maximum (or minimum) at certain concentration, the other curve
should have a minimum (or maximum) at the same composition.
2. If there is no maximum or minimum point, then both curves must be positive or both must be
negative over the entire range. Or in other words, if one component has γ values always
greater than unity & has no maximum, the γ values of the other component must likewise be
greater than unity. This is a consequence of the fact that Raoult’s law is to be obeyed by the
component as its mole fraction tends to unity.

From the above discussions, it is clear that fig.(b) represents plots of consistent data, whereas fig.(c) &
(d) are plots of thermodynamically inconsistent data. In fig.(c), though there is maximum on one curve
and minimum on the other, these are shown at different compositions. In fig.(d), the slopes have the
same sign and the data are thermodynamically inaccurate.

2. Using data at the midpoint:-

 For testing the thermodynamic consistency of VLE data, the integrated forms of GDE like the van Laar
or the Margules equation are sometimes found very convenient.
Consider the van Laar equation,

15
 Ax 22
ln  1  Az 
2
2
A B x  x 
1 2
2

Bx12
ln  2  Bz12       (2)
  Ax 
x1  B 2
2

Put x1 = x2 = 0.5 & we see that

ln  1 ln  2 AB
 
B A  A  B 2
ln  1  B & ln 2 A
Case 1. Assume that A=B. Then 4 4

Case 2. Assume that A=2B. Then

ln  1  2
9
 
B & ln 2 2 A
9
 
Case 3. Assume that A=3B. Then
ln  1  3
16
 
B & ln 2 3
16
A  
In all cases, the lnγ1 at the mid-point is approximately one-fourth the van Laar constant B and lnγ2
approximately one-fourth the constant A. we know that the van Laar constant A is lnγ1 as x1 tends to
zero & B is lnγ2 as x1 tends to one. While checking the consistency, the ln  value at this point will be
approximately one-fourth terminal value of the other ln  curve. (i.e.) the curve, which is highest at the
end-point will be lowest at the mid-point and vice-versa, as shown in fig.(a) & (b).

The mid-point values are approximately one-fourth the terminal values of the other curves in both
the figures. However, fig.(c) reveals an inconsistent data, as the curve which is highest at the end-point
is also the highest at the mid-point.

Test for thermodynamic consistency using activity coefficient data at the mid-point
3. Redlich-Kister Method:-

 This method is also known as the zero-area method can be applied to test the consistency of
experimental data when the activity coefficient values over the entire concentration range are available.
 It is based on the excess free energy of mixing which is the difference between the free energy of mixing
of a real solution and that of an ideal solution.
 The excess free energy of mixing can be written as

16 ------ (2)
 ------ (3)

Substitute (1) in (3) gives,

------ (4)

This can be integrated between the limits x1 = 0 and x1 = 1. Since only pure components are involved at
these two end states & no mixing effects occur, ∆GE = 0 at these points.

------ (5)

 From the experimental values of activity coefficients, ln  1  2  values are calculated and plotted against
x1 taken on the x-axis. The net area of the diagram should be equal to zero, if the data are
thermodynamically consistent. (i.e.) the area above the x-axis will be equal to the area below it as shown
in fig.

Test for thermodynamic consistency by Redlich-Kister method

4. Using the Coexistence Equation:-

 The coexistence equation can be used for testing the consistency of VLE data.
 If the vapour in equilibrium with a binary mixture behaves as an ideal gas, equation
y i P   i xi Pi S
can be used to describe the equilibrium.
Rearranging this equation, the activity coefficient can be written as

17
 y1 P y2 P
1  S
, 2 
x1 P1 x2 P2S
Substitute this in GDE (1) written in the following form,
x1 d ln  1  1  x1 d ln  2    (6)
y1 P y2 P
x1d ln   1  x1 d ln
x1 P1S x2 P2S
P1S & P2S are constants at constant temperature, the above equation can be simplified as
y1  x1
d ln P  dy1
y1 (1  y1 )
which on rearrangement gives,

dP P y1  x1 
    (7 )
dy1 y1 (1  y1 )
Equation (7) is known as coexistence equation.
 It can be used to calculate any one of the three variables P, x or y if experimentally values of the other
two variables are available.
 If all the three variables are experimentally determined, then Eqn.(7) can be used to test the consistency
of the measured data.

5. Using the Partial Pressure Data:-

At low pressures, the fugacity of a gas equals the pressure therefore the GDE in terms of fugacity can be
written as
d ln p 1 d ln p 2
x1   x2
dx1 dx1
x1 d p 1 x d p2
 2    (8)
p 1 dx1 p 2 dx1

The partial pressures of both components are plotted against mole fraction x1 as in fig. The slopes,
 d p1   d p 2   x1  d p 1   x 2  d p 2 
      &  
 dx  &  dx       
 1   1  are determined at any selected composition.  p 1  dx1   p 2  dx1  are calculated.
Then according to Eqn. (8) the absolute values of these quantities should be the same if the data are
thermodynamically consistent.
18
Problem: The following results were obtained by experimental VLE measurements on the system, ethanol
(1) - benzene (2) at 101.3 KPa. Test whether the data are thermodynamically consistent or not.

x1 0.003 0.449 0.7 0.9

y1 0.432 0.449 0.52 0.719

P1s kPa 65.31 63.98 66.64 81.31

P2 s kPa 68.64 68.64 69.31 72.24

Solution:
Assuming that the gas phase behaves ideally, the activity coefficients are calculated as
yP y P
1  1 S , 2  2 S
x1 P1 x2 P2
The values so calculated are listed below:

ln  1  2  values are plotted against x1. The net area is found out.
Since this is not equal to zero, the given experimental measurements do not satisfy the Redlich-Kister
1
1
0 ln  2 dx1  0
criterion for consistency.
1
1
  ln dx1  0
0
2

The given VLE data for ethanol (1) - benzene (2) are thermodynamically inconsistent.

19