Non-ideal Gas and Liquid Mixtures:

Fugacity and Activity Coefficients

Chemical Potential
For a closed system:
d (nG )  (nV )dP  (nS )dT (1)
For an open multi-compnent system: G t  nG  f (T , P, n1 , n2 ...nN )
N
  (nG )    (nG )    (nG ) 
d (nG )    dP    dT     dni (2)
 P T ,n  T  P ,n i 1  ni 
P ,T ,n i
j

Comparing (1) and (2), in general:

  (nG ) 
d (nG )  (nV )dP  (nS )dT     dni (3)
i  ni  P ,T , n  i
j

Thus: d (nG )  (nV )dP  (nS )dT   Gi dni (4)

i

  (nG ) 
Note that by definition:    i  chemical potential
 ni T , P ,n j i
 Schema for Non-ideal System VLE Description

Raoult’s Law as starting point

yiP= xi Pis

Retain the form and modify both sides by

corrective factors (CF) that capture non-ideality of each phase
(CFivap) yiP= (CFiliq) xi Pis

Derive both these corrective factors

CFivap,CFiliq from a common
indicator of Non-deal behaviour (FUGACITY)

Derive Fugacity using chemical potential

which describes phase equilibria, and develop expressions for
Fugacity from PVT data or EOS and composition (yi, and xi)
Correcting for gas mixture non-ideal
behaviour
Fugacity and Fugacity Coefficient
dGiig  Vi ig dP  RTdP / P  RTd ln P (T  const )
dGi (real fluid )  RTd ln f i ( f  fugacity )
Fugacity  " Fictitious pressure "
It is equal to the pressure of an ideal gas which
has the same chemical potential as the real gas.
Contd..
For exam ple, nitrogen gas ( N 2 ) at 0°C and
a pressure of P = 100 atm has a fugacity= 97 .03atm . 

T his m eans that the chem ical potential o f real

nitrogen at a pressure of 100 atm is less than if
nitrogen w ere an ideal gas

T he value of the chem ical potential is that w hich

nitrogen as an ideal gas w ould have at a pressure
of 97.03 atm .
Fugacity and Fugacity Coefficient
ig ig
dGi  Vi dP  RTdP / P  RTd ln P (T  const )
dGi (real fluid )  RTd ln f i ( f  fugacity )
Fugacity  " Fictitious pressure "
It is equal to the pressure of an ideal gas which
has the same chemical potential as the real gas.
ig R fi
dG  dG   RTd ln  RTd ln i
dGi (13)
P
Fugacity Coefficient (i ): i  fi / P
ig R
For an ideal gas : fi  P  Gi  0; i  1
R
On integrating eqn(12) : Gi  RT ln i
 V R  (G R / RT ) 
By eqn.(10):  
RT  P T , x
R ig
R V  V  V 
Or: d (G / RT )  dP    dP
RT  RT 
R 1  ZRT RT  dP
d (G / RT )      ( Z  1) (14)
RT  P P  P
On integrating eqn(13)from 0 to P :
GiR GiR GiR P dP
RT P

RT P 0

RT
 
0
( Z  1)
P
P
dP
Thus : ln i  (GiR 
/ RT )  ( Z i  1) (15)
0
P
Eqn.(15) is suitable for volume explicit EOS
Fugacity Coefficient Calculation from EOS
For a pure gas ‘i’ obeying Virial EOS:
Bii P
Zi  1 
RT
Bii P
Hence, ln i 
RT
Computation of Gas Fugacity Coefficient (pure component)
by generalized Lee-Kessler Charts:
Pr
d Pr
O

ln   ( Z  1)
Pr
Pr Pr
O d Pr 1 d Pr
ln   ( Z  1)
  Z 
O
Pr O
Pr

ln   ln  0   ln  1
Pr Pr
0 d Pr
O 1 d Pr
1
ln   ( Z  1)
 ln   Z

O
Pr Pr
O

ln  0 and ln  1 are given as function of Tr and Pr

 see Tables13 to16 in Appendix VII of e-book 
 
 or Smith/Van Ness E.12-E.15 
Fugacity Coefficients from Cubic EOS
An alternate form for ln  suitable for
pressure explicit (cubic EOSs) is :
1 V
ln   ( Z  1)   PdV
RT V 
For SRK EOS :
a Z B bP
ln   Z  1  ln( Z  B)  ln  ; B 
bRT  Z  RT

For PR EOS :
a  Z  B (1  2)  bP
ln   Z  1  ln( Z  B )  ln  ;B 
2 2bRT  Z  B (1  2)  RT
Fugacity of a Pure Liquid:

dG = VidP = RTdlnfi
P-------

Psat -------

Ref.
point

Poynting factor
l sat
sat sat Vi ( P  Pi )
f i (T , P )  i Pi exp[ ]
RT
 Estimation of  i
sat

dG iV  RTd ln f iV
dG iL  RTd ln f i L
On integration between vapour and liquid states:
V
f
G iV  G iL  RT ln i L
fi
At equilibrium:G iV  G iL  0 (see derivation for Antoine Eq.)
Thus: f iV  f i L  f i sat
For a pure species coexisting liqu id and vapor phases are in
equilibrium when they have the same temperature, pressure,
and fugacity.
Alternately:
fiV / Pi sat  f i L / Pi sat  fi sat / Pi sat
iV  iL  isat
sat
sat fi
where :  i  sat
Pi
sat
BPi
V Lsat
If Virial EOS applies then,     i  ii
RT
For high pressures use either V or P explicit
expression for fugacity coefficient as shown earlier.
Fugacity/Fugacity Coefficient of Species in Mixture:

i ig  Giig  RT ln yi
d i ig  dGiig  RTd ln yi
But , dGiig  Vi ig dP  RTd ln P
Thus : d i ig  RTd ln P  RTd ln yi  RTd ln( yi P )
d i ig  RTd ln pi
By extension : d i (real fluid )  dGi  RTd ln fˆi
M R  M  M ig  nM R  nM  nM ig
It follows that:
  (nM R )    (nM )    (nM ig ) 
     
 ni  T ,P ,n j i  ni  T ,P ,n j i  ni  T ,P ,n j i
Or: M iR  M i  M iig
Thus : GiR  Gi  Giig  i  iig (16)
But:d iig = RTd ln pi ; and d i  RTd ln fˆi
On integrating between the real and ideal states:
fˆi fˆi
ig
i  i  RT ln  RT ln  RT ln ˆi
pi yi P
ˆ
f
Where, ˆi  species fugacity coefficient in a mixture  i
yi P
Thus, using eqn. (15):
ˆ
f
GiR  Gi  Giig  RT ln i  RT ln ˆi (17)
yi P
On comparing eqns.(9) and (17):

ˆ   ( nG R
/ RT ) 
ln i    (18)
  ni  T , P ,n j  i
Although most commonly applied to gases,
the fugacity coefficient in eqn. 15 may also
be used for liquids, in which case the mole
fraction yi is replaced by xi .
Contd…

R
G
Since ln ˆi is a partial of , it follows that:
RT
GR
  yi ln ˆi
RT i

Further, by Gibbs-Duhem relation for partial

molar properties:

ˆ  0 (const. T and P)
 i i
x
i
d ln 
Species fugacity coefficients by EOS:
R
  ( nG / RT 
ln ˆi   
 ni  P ,T , n j i
where, G R  G  G ig
R dP P
Also, (G / RT )  ( Z  1)
0  P
Thus, using eqn. (16). the general expression is:
ˆ
P dP P dP
ln i  (nZ  n)
  ( Z i  1) 
0 P 0 P
 Expression for Z varies with EOS
Contd…

Or, alternately:
V
 P  RT 
ˆ
RT ln i       dV  RT ln Z
V   n V 
  i T ,V , n j i

Species Fugacity Expressions by Virial EOS

BP
Z  1 ; where, Bmixture   yi y j Bij
RT i j

ˆ P dP
Using ln i   (nZ  n) it may be shown that
0 P
for a binary gas mixture obeying virial EOS:
ˆ P 2 ˆ P 2
ln 1  ( B11  y2 12 ); ln 2  ( B22  y1 12 )
RT RT
12  2 B12  B11  B22
 Expressions for fugacity coefficient for ‘kth species in a multi-
component mixture:
For Virial EOS :
P 1
ln k 
RT
[ Bkk 
2
 y y (2
i j
i j ik   ij )]

 ik  2 Bik  Bii  Bkk ;  ij  2 Bij  Bii  B jj ;  ii   jj   kk  0

For SRK EOS :
bk am  bk ak  Z  Bm
ln k  ( Z  1)  ln( Z  Bm )  (  2  ln( ); m  mixture
bm bm RT  bm am  Z

For PR EOS :
bk am  bk ak   Z  Bm (1  2) 
ln k  ( Z  1)  ln( Z  Bm )   2  ln  
bm b
2 2bm RT  m a   Z  Bm (1  2) 
m 

bm P
For both cubic EOSs above, Bm 
RT
Mixing Rules for Cubic EOS
Mixing rules for binary mixtures:
am   yi y j aii a jj (1  kij )
i j

kij  binary interaction parameter ;where kii  k jj  0

bm   yi bi
i

m   yii
i

For a binary (assuming all k ij  0) :

am  y12 a11  2 y1 y2 (a11a22 )1/ 2  y22 a22
bm  y1b1  y2b2
m  y11  y22
Ideal Liquid Solutions:
For any liquid mixture:
d i  RTd ln fˆi ; hence, i   i (T )  RT ln fˆi
For an ideal liquid mixture:
iid  Giid  i (T )  RT ln fˆi id
But, in general, iid  Giid  Gi  RT ln xi

Also, Gi   i (T )  RT ln f i

Thus, iid  Giid  i (T )  RT ln xi f i

It follows: fˆi id  xi fi  Lewis  Randall Rule

For real xixture: GE  G  Gid  GiE  Gi  Giid
Molar Excess Property Relations
M E  M  M id
Thus: H E  H  H id ; S E  S  S id ; G E  G  G id
Further: H E  H  x H i i  H mix
i

E  
S S

x S
i
i i R  i
xi ln xi 

 S mix  R  x ln x
i
i i

E  
G G

 x G  RT 
i
i i
i
xi ln xi 

 Gmix  RT  x ln x
i
i i
Molar Excess Gibbs Energy and the Activity Coefficient (for
liquid mixtures):

G iE  Gi  G idi  RT ln fˆi  RT ln fˆi id  RT ln( fˆi / fˆi id ); fˆi id  xi f i

id E fˆi fˆ
Gi  G i  G i  RT ln  ( Activity coefficient ) 
i
i

xi f i xf i i

E  (nG E / RT )
Thus, G  RT ln  i
i ln  i  (25)
ni T , P ,n j 1

i.e., G E / RT   xi ln  i
i

By Gibbs  Duhem relation : xi d ln  i  0 (const T & P )

lim  i   i & lim  i  1
xi 0 xi 1

Obtain suitable models for GE/RT and hence get γi

Effect of T & P on Excess Gibbs free energy
For an equimolar mixture of benzene and cyclohexane
at 298K and 1 bar:
V E  0.65cm3 / mol
H E  800cm3 / mol
Thus it follows:
  (G E / RT )  VE 0.65 5 1
     2.62 x10 bar
 P T , x RT 83.14  298
  (G E / RT )  HE 800 3 1
    2
  2
 1.08 x10 K
 T  P,x RT 83.14  298

G E ( for T  1K )  G E ( for P  40bar )

Activity Coefficients from Low pressure VLE Data
• Carry out VLE experiments at low pressure

• Measure vapour phase compositions at

various starting and liquid phase mole fractions

• Use modified Raoult’s Law for calculating activity

coefficients as function of liquid phase
composition

• Fit activity coefficient data to appropriate activity

coefficient model for GE/RT

• Note that G E / RT   xi ln  i
The activity coefficient model is derived next (see following slide)
Obtaining the Activity Coefficient Model

It follows : G E / x1 x2 RT  A21 x1  A12 x2

or G E / RT  x1 x2 ( A21 x1  A12 x2 )
 (nG E / RT )
Using ln  i 
ni T , P ,n j 1

Margules Equations:
ln  1  x22 [ A21  2( A21  A12 ) x1 ]
ln  2  x12 [ A12  2( A12  A21 ) x2 ]
Also : ln  1 ( x1  0)  A21 & ln  2 ( x2  0)  A12
Other Activity Coefficient Models
 Applicability of Activity Coefficient Models
System Type Models
Species similar is size and shape One-constant Margules
Moderately non-ideal mixtures Two-constant Margules, Van Laar,
Regular Solution
Strongly non-ideal mixtures Wilson, NRTL, UNIQUAC
(Alcohols+Hydrocarbons)
Solutions with miscibility gap NRTL, UNIQUAC
 Regular Solution (RS) Model (for moderately non-
ideal solutions)
2 2 2
G E / RT   x1V1  x2V2  1 2 1   2   ln  1  V1 22 1   2  ; ln  2  V2 12 1   2 

where, i = volume fraction =

xiVi
Vm
Vi = molar volume of pure ith species, Vm = ΣxiVi , and
δi = solubility parameter

i 
U   H
i
vap
i
vap
 RT 
Vi Vi

For Multi-component Solutions

2 x jV j

R T ln  i  Vi 1    ;     j j ; j 
j x V
j
j j
 Method of Group Contribution

UNIQUAC (Universal Quasi-chemical) based

Activity Coefficients:
ln  i  ln  iC (combinatorial )  ln  iR ( residual )
END