Wong-Sandler Mixing Rules
These mixing rules use a relationship between the excess Helmholtz energy and equation-of-state.
They do not use a relationship between equation-of-state properties and excess Gibbs energy, as in
the Huron-Vidal mixing rules. The pressure-explicit expression for the equation-of-state is substituted
in the thermodynamic equation:
— (24 1
p= —(57)r ”
The Helmholtz energy is obtained by integration, Af is obtained by:
AE = Am — Yo aiAj — RT 2; nz (2)
Where both A and Aj, are calculated by using equation 1. A/* and A,, are written in terms of
equation-of-state parameters.
Like Huron and Vidal, the limiting case of infinite pressure is used. This simplifies the expressions for
Aj* and A,,. Equation 2 becomes:
$= Dadi — FAR (p= 00)
GB)
Where A depends on the equation-of-state (see Huron-Vidal Mixing Rules).
Equation 3 is completely analogous to the Huron-Vidal mixing rule for the excess Gibbs energy at
infinite pressure. (See equation 4, Huron-Vidal Mixing Rules.) The excess Helmholtz energy can be
approximated by the excess Gibbs energy at low pressure from any liquid activity coefficient model
Using the Helmholtz energy permits another mixing rule for b than the linear mixing rule, The mixing
rule for b is derived as follows. The second virial coefficient must depend quadratically on the mole
fraction:
BT) = OY o ain Bis o
oF
with
(Bat By) (5)
Bi z (1 — kis)
The relationship between the equation-of-state at low pressure and the virial coefficient is:
B=b- a (6)
By =b; — Sy ”
Wong and Sandler discovered the following mi:
ing rule to satisfy equation 4 (using equations 6 and
The excess Helmholtz energy is almost independent of pressure. It can be approximated by the Gibbs
energy at low pressure. The difference between the two functions is corrected by fitting k, until the
excess Gibbs energy from the equation-of-state (using the mixing rules 3 and 8) is equal to the excessGibbs energy computed by an activity coeffecient model. This is done at a specific mole fraction and
temperature.
This mixing rule accurately predicts the VLE of polar mixtures at high pressures. UNIFAC or other
activity coeffecient models and parameters from the literature are used. Gas solubilities are not
predicted. They must be regressed from experimental data.
Unlike other (modified) Huron-Vidal mixing rules, the Wong and Sandler mixing rule meets the
theoretical limit at low pressure. The use of ky does influence excess molar volume behavior. For
calculations where densities are important, check whether they are realistic.
Note: Under some conditions, particularly when large values for the ky are used, the Ap/RT term can become
negative, To avoid the impossible results (negative pressure or volume) this would cause, certain limits are applied.
This will lead to various kinds of anomalous results such as abrupt changes in enthalpy. To avoid this behavior,
review the parameters or use a different mixing rule.
References
D. S. Wong and S. I. Sandler, "A Theoretically Correct New Mixing Rule for Cubic Equations of State
for Both Highly and Slightly Non-ideal Mixtures," AIChE J., Vol. 38, (1992), pp. 671 - 680.
D. S. Wong, H. Orbey, and S. I. Sandler, "Equation-of-state Mixing Rule for Non-ideal Mixtures Using
Available Activity Coefficient Model Parameters and That Allows Extrapolation over Large Ranges of
‘Temperature and Pressure", Ind Eng Chem. Res., Vol. 31, (1992), pp. 2033 - 2039.
H. Orbey, S. I. Sandler and D. S. Wong, "Accurate Equation-of-state Predictions at High Temperatures
and Pressures Using the Existing UNIFAC Model," Fluid Phase Eq., Vol. 85, (1993), pp. 41 - 54.Predictive Soave-Redlich-Kwong-Gmehling Mixing Rules
These mixing rules by Holderbaum and Gmehling (1991) use a relationship between the excess
Helmholtz energy and equation-of-state. They do not use a relationship between equation-of-state
properties and excess Gibbs energy, as in the Huron-Vidal mixing rules. The pressure-explicit
expression for the equation-of-state is substituted in the thermodynamic equation:
q@)
The Helmholtz energy is calculated by integration. A® is obtained by
AR = Am — Sai Aj — RT 2; In; @
7 T
Where both A* and A, are calculated by using equation 1. A and A, are written in terms of
equation-of-state parameters.
The simplification of constant packing fraction (Vm/ b) is used:
(3)
(4)
Therefore:
VE (p = 00) =0 (5)
The mixing rule is:
(6)
a alae
$= doze — pAm(P)
, » & A
Where A’ is slightly different from A for the Huron-Vidal mixing rule:
%
Yar a”
v= szn( =)
Aa ay
Where A; and A», depend on the equation-of-state (see Huron-Vidal Mixing Rules). If equation 6 is
applied at infinite pressure, the packing fraction goes to 1. The excess Helmholtz energy is equal to
the excess Gibbs energy. The Huron-Vidal mixing rules are recovered.
The goal of these mixing rules is to be able to use binary interaction parameters for activity
coefficient models at any pressure. These parameters have been optimized at low pressures. UNIFAC
is chosen for its predictive character, Two issues exist: the packing fraction is not equal to one, and
the excess Gibbs and Helmholtz energy are not equal at the low pressure where the UNIFAC
parameters have been derived.
Fischer (1993) determined that boiling point, the average packing fraction for about 80 different
liquids with different chemical natures was 1.1. Adopting this value, the difference between liquid
excess Gibbs energy and liquid excess Helmholtz energy can be computed as:
(8)Vitab
paVy, — Say [ x
7
b
Vyt=Lib
u
The result is a predictive mixing rule for cubic equations of state. But the original UNTFAC formulation
gives the best performance for any binary pair with interactions available from UNIFAC. Gas-solvent
interactions are unavailable. However, it has poor accuracy for highly asymmetric such as CH4 — n-
C10H22. To address the issue, the Li correction (Li et al., 1998) is applied.
With the Li correction, the R and Q parameters for groups CH3, CH2, CH, and C in UNIFAC are
modified. R* and Q*, the effective values of R and Q, are calculated based on the original values and
ng, the number of alkyl carbons (single-bonded carbons in CH3, CH2, CH, and C groups), as follows,
for n.<45:
i = f(r) Re
Qi = Fre) Qe
1/2
(me) = 1.0 — 0.36983n¥/ + 1.0287n3/* — 1.0199n, — 0.41645n*/4 — 0.05556n°/
At the University of Oldenburg in Germany, the UNIFAC groups were extended with often-occurring
gases. New group interactions were determined from gas-solvent data, specific to the Redlich-Kwong-
Soave equation-of-state. The new built-in parameters to the Aspen Physical Property System are
activated when using the PSRK equation-of-state model (ESRKSV1 for mixtures, ESRKSV10 for pure
components).
‘The PSRK method has a lot in common with the Huron-Vidal mixing rules. The mole fraction is
dependent on the second virial coefficient and excess volume is predicted. These are minor
disadvantages.
References
K, Fischer, "Die PSRK-Methode: Eine Zustandsgleichung unter Verwendung des UNIFAC-
Gruppenbeitragsmodells,” (Diisseldorf: VDI Fortschrittberichte, Reihe 3: Verfahrenstechnik, Nr. 324,
VDI Verlag GmbH, 1993).
T. Holderbaum and J. Gmehling, "PSRK: A Group Contribution Equation-of-state based on UNIFAC,"
Fluid Phase Eq., Vol. 70, (1991), pp. 251-265.
J. Ui, K. Fischer, and J. Gmehling, "Prediction of vapor-liquid equilibria for asymmetric systems at low
and high pressures with the PSRK model," Fluid Phase Equilib., 1998, 143, 71-82
J. Gmehling, J. Li, and K. Fischer, "Further development of the PSRK model for the prediction of gas
solubilities and vapor-liquid-equilibria at low and high pressures 11," Fluid Phase Equilib., 1997, 141,
113-127.
S. Horstmann, K. Fischer, and J. Gmehling, "PSRK group contribution equation of state: revision and
extension III," Fluid Phase Equilib., 2000, 167, 173-186.
S. Horstmann, A. Jabloniec, J. Krafczyk, K. Fischer, and J. Gmehling, "PSRK group contribution
equation of state: comprehensive revision and extension IV, including critical constants and a-function
parameters for 1000 components," Fluid Phase Equilib., 2005, 227(2), 157-164.