Local Composition Models

Engr. Paul Gabriel E. Lerona

Topic 13
Adamson University
[email protected]

June 18, 2022

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 1 / 45
Overview I
1 References
Seader Table 2.9
Perry’s Chapter 2

2 Local Composition Models

Limitations
Local Composition Models

3 Wilson Equation

4 NRTL Equation

5 UNIQUAC Model

6 Local Composition Models

7 UNIFAC Model
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 2 / 45
 References

References

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 3 / 45
 References

References I

1 SVA
Chapter 12.2 Local-Composition Models
Appendix H UNIFAC Method
2 Perry’s
Chapter 2 Prediction and Correlation of Physical Properties
3 Elliott
Chapter 12 Local Composition Activity Models
4 Koretsky
Chapter 7.4 Fugacity in the Liquid Phase
5 Sandler
Chapter 9.5 Several Correlative Liquid Mixture Activity Coefficient
Models
Chapter 9.11 Choosing the Appropriate Thermodynamic Model

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 4 / 45
 References Seader Table 2.9

Seader Table 2.9

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 5 / 45
 References Seader Table 2.9

Seader Table 2.9 I

Name Equation for Species 1 Equation for Species 2

Margules log γ1 = Ax22 log γ2 = Ax12

log γ1 = x22 Ā12 + 2x1 Ā21 − Ā12 log γ2 = x12 Ā21 + 2x2 Ā12 − Ā21

 

Margules (two-constant)

A12 A21
van Laar(two-constant) ln γ1 = [1+(x1 A12 )/(x2 A21 )]2
ln γ2 = [1+(x2 A21 )/(x1 A12 )]2

ln γ1 = − ln (x1 + Λ12 x2 ) ln γ2 = − ln (x2 + Λ21 x1 )

Wilson (two-constant)
   
Λ12 Λ21 Λ12 Λ21
+ x2 − − x1 −
x1 + Λ12 x2 x2 + Λ21 x1 x1 + Λ12 x2 x2 + Λ21 x1

x22 τ21 G21

2 x12 τ12 G12 x12 τ12 G12
2 x22 τ21 G21
ln γ1 = + ln γ2 = +
NRTL (three-constant) (x1 + x2 G21 )2 (x2 + x1 G12 )2 (x2 + x1 G12 )2 (x1 + x2 G21 )2
Gij = exp (−αij τij ) Gij = exp (−αij τij )

Ψ1 Z̄ θ1 Ψ2 Z̄ θ2
ln γ1 = ln + q1 ln ln γ2 = ln + q2 ln
x1  2 Ψ1 x2  2 Ψ2
r1 r2
UNIQUAC (two-constant) + Ψ2 l1 − l2 − q1 ln (θ1 + θ2 T21 ) + Ψ1 l2 − l1 − q2 ln (θ2 + θ1 T12 )
r2 r1
   
T21 T12 T12 T21
+ θ2 q1 − + θ1 q2 −
θ1 + θ2 T21 θ2 + θ1 T12 θ2 + θ1 T12 θ1 + θ2 T21

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 6 / 45
 References Perry’s Chapter 2

Perry’s Chapter 2

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 7 / 45
 References Perry’s Chapter 2

Perry’s Chapter 2 I
1 Physical constants: critical properties, normal melting and boiling
point, acentric factor, radius of gyration, and dipole moment
2 Vapor pressure: liquid and solid
3 Thermal properties: enthalpy and Gibbs’ energy of formation and
ideal gas entropy
4 Latent enthalpy: vaporization, fusion, and sublimation
5 Heat capacity: ideal gas, gas, liquid, and solid
6 Density: gas, liquid, and solid
7 Viscosity: gas and liquid
8 Thermal conductivity: gas and liquid
9 Surface tension
10 Flammability properties: flash point, flammability limits, and
autoignition temperature
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 8 / 45
 Local Composition Models

Local Composition Models

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 9 / 45
 Local Composition Models Limitations

Limitations

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 10 / 45
 Local Composition Models Limitations

Limitations I

No theoretical foundation
Cannot be extended to mixtures with three or more components
No temperature dependence

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 11 / 45
 Local Composition Models Local Composition Models

Local Composition Models

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 12 / 45
 Local Composition Models Local Composition Models

Local Composition Models I

composition in a local volume around a molecule

may be different from overall mixture composition
Non-ideality in solution behavior from:
Molecular size and shape difference
Molecular interactions

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 13 / 45
 Wilson Equation

Wilson Equation

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 14 / 45
 Wilson Equation

Wilson Equation I

1 For a multicomponent system:

 
GE X X
=− xi ln  xj Λij 
RT
i j

 
X X xk Λki
ln γi = 1 − ln  xj Λij  − P
j j xj Λkj k

Vj −aij
Λij = 1 for i = j Λij 6= Λji Λij = exp i 6= j
Vi RT

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 15 / 45
 Wilson Equation

Wilson Equation II

2 For a binary system:

GE
= −x1 ln (x1 + x2 Λ12 ) − x2 ln (x2 + x1 Λ21 )
RT
 
Λ12 Λ21
ln γ1 = − ln (x1 + x2 Λ12 ) + x2 −
x1 + x2 Λ12 x2 + x1 Λ21
 
Λ12 Λ21
ln γ2 = − ln (x2 + x1 Λ21 ) − x1 −
x1 + x2 Λ12 x2 + x1 Λ21

ln γ1∞ = − ln Λ12 + 1 − Λ21 ln γ2∞ = − ln Λ21 + 1 − Λ12

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 16 / 45
 Wilson Equation

Wilson Equation III

Works well for mixtures of polar and nonpolar species, e.g. alcohols
and alkanes.
Works well for hydrocarbon mixtures and is readily extended to
multicomponent mixtures.
Wilson parameters, Λij and Λji however, must be positive to be valid
for infinite dilution cases.
Unable to describe systems exhibiting partial miscibility

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 17 / 45
 Wilson Equation

Wilson Equation IV

Derived from a molecular basis

aij and aji represent energetic parameter that describe how the i-i and
j-j interaction, respectively, varies from i-j interaction.
The size of the molecules are taken into account thrugh the ratio of
their molar volumes.
aij and aji are relatively insensitive to temperature.

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 18 / 45
 NRTL Equation

NRTL Equation

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 19 / 45
 NRTL Equation

NRTL Equation I
1 Three-parameter equation, for a multicomponent system
P
j xj τji Gji
X xi Gij  P 
m xm τmj Gmj
ln γi = P + P τij − P
k xk Gki k xk Gkj j k xk Gkj

2 Three-parameter equation, for a binary system

GE G21 τ21 G12 τ12

= +
x1 x2 RT x1 + x2 G21 x2 + x1 G12
"  2 #
G21 G12 τ12
ln γ1 = x22 τ21 +
x1 + x2 G21 (x2 + x1 G12 )2
"  2 #
G12 G21 τ21
ln γ2 = x12 τ12 +
x2 + x1 G12 (x1 + x2 G21 )2
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 20 / 45
 NRTL Equation

NRTL Equation II

ln γ1∞ = τ21 + τ12 exp (−ατ12 )

ln γ2∞ = τ12 + τ21 exp (−ατ21 )

G21 = exp (−ατ21 ) G12 = exp (−ατ12 )

b12 b21
τ12 = τ21 =
RT RT
Developed by Renon and Prausnitz
Extension of Wilson’s concept to multicomponent LL and VLL
systems
Widely used for liquid-liquid extraction
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 21 / 45
 NRTL Equation

NRTL Equation III

The parameter αji characterizes the tendency of species j and I to be

distributed nonrandomly. When αji = 0, local mole fractions equal
solution mole fractions.
αji is independent of temperature and depends on molecule properties
Introduces a third binary parameter that generates an extremely
flexible functional form for fitting activity coefficients.
Not very appealing from a theoretical perspective, but its flexibility
has led to a broad range of applications including combinations with
electrolyte models.
Default is αij = 0.3 but is adjusted for additional flexibility when
necessary (when modeling LLE, it is commonly increased).

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 22 / 45
 NRTL Equation

NRTL Equation IV
1 αji = 0.20 for hydrocarbons and polar, nonassociated species (e.g.,
n-heptane/acetone).
2 αji = 0.30 for nonpolar compounds (e.g., benzenel n-heptane), except
fluorocarbons and paraffins; nonpolar and polar, nonassociated species
(e.g., benzenel acetone); polar species that exhibit negative deviations
from Raoult’s law (e.g., acetone/chloroform) and moderate positive
deviations (e.g., ethanol/water); mixtures of water and polar
nonassociated species (e.g., water/ acetone).
3 αji = 0.40 for saturated hydrocarbons and homolog perfluorocarbons
(e.g., n-hexane/perfluoro- n-hexane).
4 αji = 0.47 for alcohols or other strongly self-associated species with
nonpolar species (e.g., ethanol/benzene); carbon tetrachloride with
either acetonitrile or nitromethane; water with either butyl glycol or
pyridine.

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 23 / 45
 UNIQUAC Model

UNIQUAC Model

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 24 / 45
 UNIQUAC Model

UNIQUAC Model I

Universal Quasichemical Model

Abrams and Prausnitz used statistical mechanics in an attempt to
place calculations of activity coefficients on a more theoretical basis.
Rather than local volume fractions or local mole fractions, UNIQUAC
uses local area fraction θij as the primary concentration variable.
Local area fraction, θij
Determined by representing a molecule by a set of bonded segments.
Each molecule is characterized by two structural parameters
determined relative to a standard segment.
Equivalent sphere of a unit of a linear, infinite-length, polymethylene
molecule.

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 25 / 45
 UNIQUAC Model

UNIQUAC Model II

Structural parameters
Relative number of segments per molecule
r - volume parameters
Relative surface area
q-surface parameters
Parameters are computed from bond angles and bond distances, given
for many species by Abrams and Prausnitz.

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 26 / 45
 UNIQUAC Model

UNIQUAC Model III

For a multicomponent liquid mixture:

C C
gE
   
X Ψi Z̄ X θi
= xi ln + qi xi ln
RT xi 2 Ψi
i=1 i=1
 
X C XC
− qi xi ln  θi Tji 
i=1 j=1

Combinatorial Effects - due to differences in size and shape

Residual Effects - due to differences in intermolecular forces

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 27 / 45
 UNIQUAC Model

UNIQUAC Model IV
For a multicomponent liquid mixture:
C C
gE
   
X Ψi Z̄ X θi
= xi ln + qi xi ln
RT xi 2 Ψi
i=1 i=1
 
XC XC
− qi xi ln  θi Tji 
i=1 j=1

xi ri
Ψi = PC = segment fraction
i=1 xi ri
x i qi
θ = PC = area fraction
i=1 xi qi
 
uji − uii
Tji = exp
RT
z = 10
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 28 / 45
 UNIQUAC Model

UNIQUAC Model V

For a multicomponent liquid mixture:

ln γi = ln γiC + ln γiR
C
X
= ln (Ψi /xi ) + (Z̄ /2)qi ln (θi /Ψi ) + li − (Ψi /xi ) xj lj
j=1
| {z }
C , combinatorial
   !
C C
X X θj Tij
+ 1 − ln  θj Tji  − PC 
j=1 j=1 k=1 θk Tkj
| {z }
R, residual
 
Z̄
z = 10 lj = (rj − aj ) − (rj − 1)
2

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 29 / 45
 UNIQUAC Model

UNIQUAC Model VI

Often gives good representation of VLE and LLE for binary and
multicomponent mixtures containing nonelectrolytes
For hydrocarbons, ketones, esters, water, amines, alcohols, nitriles,
etc.

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 30 / 45
 UNIQUAC Model

Sample Reference Values I

Poling, Prausnitz, & O’Connell. The Properties of Liquids and
Gases

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 31 / 45
 UNIQUAC Model

Sample Reference Values II

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 32 / 45
 UNIQUAC Model

Sample Reference Values III

 
uji − Uii
Tji = exp
RT
 
Umn − Unn
Ψmn = exp −
RT
 a 
mn
= exp −
T

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 33 / 45
 Local Composition Models

Local Composition Models

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 34 / 45
 Local Composition Models

Local Composition Models: Advantages I

Foundation from statistical mechanics

Can be extended to multi-component mixtures requiring only binary
interaction parameters
With temperature dependence

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 35 / 45
 Local Composition Models

Local Composition Models: Applications of rational

functions and local composition models I
Simple mixtures (similar in chemical nature and molecular size)
One-parameter Margules
Moderately non-ideal binary mixtures
Margules
van Laar
Wilson
NRTL
UNIQUAC
Strongly non-ideal binary mixtures (ex. alcohols + hydrocarbons)
Not Two-parameter Margules and van Laar
Wilson; particularly suitable in the region dilute with respect to alcohol
Solutions with miscibility gap (LLE)
Not Wilson
NRTL and UNIQUAC

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 36 / 45
 Local Composition Models

Local Composition Models: Applications of rational

functions and local composition models II

UNIQUAC vs. NRTL

UNIQUAC:
Has only two vs. three adjustable parameters
Parameters have smaller dependence on temperature
Applicable to solutions containing small or large molecules (ex.
polymers) due to consideration of surface fraction over mole fraction

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 37 / 45
 UNIFAC Model

UNIFAC Model

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 38 / 45
 UNIFAC Model

UNIFAC Model I

UNIQUAC Functional-group Activity Coefficients group-contribution

method
Theoretically-based
Parameters are essentially independent of temperature
Size and binary interaction parameters are available for a range of
functional groups
Predictions can be made over a temperature range of 275-425 K and
for pressures to a few atmospheres
Extensive comparisons with experimental data are available
All components must be condensable at nearambient conditions

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 39 / 45
 UNIFAC Model

UNIFAC Model II

Activity Coefficients from Contribution Method

Molecule - aggregate of functional groups
Fluid mixture property is the sum of contributions from the
molecules’ functional groups
Activity coefficients related to the interaction of the functional groups
Data from experimentally-obtained activity coefficients is used to
obtain parameters characterizing interactions between pairs of
structural groups

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 40 / 45
 UNIFAC Model

UNIFAC Model III

The UNIQUAC MODEL

ln γi = ln γiC + ln γiR
C
X
= ln (Ψi /xi ) + (Z̄ /2)qi ln (θi /Ψi ) + li − (Ψi /xi ) xj lj
j=1
| {z }
C , combinatorial
   !
C C
X X θj Tij
+ 1 − ln
  θj Tji −
 PC 
j=1 j=1 k=1 θk Tkj
| {z }
R, residual

Combinatorial part - accounts for molecular size and shape difference

Residual part - accounts for molecular interactions
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 41 / 45
 UNIFAC Model

UNIFAC Model IV

Combinatorial Part
(i) (i)
X X
ri = vk Rk qi = vk Qk
k k

Rk = group volume Qk = area parameter

Vwk Awk
Rk = Qk =
15.17 2.5 × 109

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 42 / 45
 UNIFAC Model

UNIFAC Model V

Combinatorial Part

Φi z θi Φi X
ln γic = ln + qi ln + li − xj lj
xi 2 Φi xi
j

ri x i qi x i
Φi = P θi = P
j rj xj j qj xj

(i) (i)
X X
ri = νk Rk qi = νk Qk
k k

z = 10
z
li = (ri − qi ) − (ri − 1)
2

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 43 / 45
 UNIFAC Model

UNIFAC Model VI
Residual Part
 
(i) (i)
X
ln γiR = vk ln Γk − ln Γk
k

Γk = group residual activity coefficient

(i)
Γk = residual activity coefficient of a group k in a reference solution
containing only molecules of type i.
" ! #
X X θm Ψkm
ln Γk = Qk 1 − ln θm Ψmk − P
m m n θn Ψnm
" ! #
X θm (i)
(i)
X (i) Ψkm
ln Γk = Qk 1 − ln θm Ψmk − P (i)
m m n θn Ψnm
 
Umn − Unn  amn 
Ψmn = exp − = exp −
RT T
Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 44 / 45
 UNIFAC Model

The End

Engr. Paul Gabriel E. Lerona Local Composition Models June 18, 2022 45 / 45