Fluid Phase Equilibria 281 (2009) 163–171

Contents lists available at ScienceDirect

Fluid Phase Equilibria

journal homepage: www.elsevier.com/locate/fluid

Electrolyte-UNIQUAC-NRF model for the correlation of the mean activity

coefficient of electrolyte solutions
Ali Haghtalab ∗ , Kiana Peyvandi
Chemical Engineering Department, Tarbiat Modares University, P.O. Box 14115-143, Tehran, Iran

a r t i c l e i n f o a b s t r a c t

Article history: The new electrolyte-UNIQUAC-NRF excess Gibbs function is obtained for calculation of the activity
Received 6 December 2008 coefficient of the binary electrolyte solutions. The excess Gibbs energy of the model consists of the
Received in revised form 15 April 2009 Pitzer–Debye–Hückel equation, describing the long-range electrostatic contribution and the electrolyte-
Accepted 19 April 2009
UNIQUAC-NRF model to account for the short-range contributions. With two adjustable parameters per
Available online 3 May 2009
electrolyte, the new model is applied to correlation of the mean activity coefficients of more than 130
binary aqueous electrolyte solutions at 25 ◦ C. Also the binary parameters, obtaining from regression of
Keywords:
mean activity data, are used for prediction of osmotic coefficient data for the same electrolytes. The
Local composition
Activity coefficient
results are compared with various local composition models such as the electrolyte-NRTL, electrolyte
Excess Gibbs energy NRF-Wilson, electrolyte-NRTL-NRF, N-Wilson-NRF models. The comparison of the results with exper-
Electrolyte-UNIQUAC-NRF iment demonstrates that the new model can correlate the experimental activity coefficient data and
Binary electrolyte solution predict the osmotic coefficient data of binary electrolytes accurately.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction
Electrolyte systems represent an important part in chemical, gas
and petroleum industries such as natural gas treatment, extractive
distillation, crystallization, desalination of water, hydrometal-
lurgy, etc. [1,2]. The thermodynamics of electrolyte are used
in the environmental protection industries and computation of
liquid–liquid equilibria in aqueous two-phase systems for extrac-
tion of biomolecules in downstream processing in biotechnology
[1]. Thus, the accurate representation of thermodynamic proper-
ties of electrolyte solutions is vital in design of several industrial
processes. Calculation activity coefficient of salts, bases and acids
in aqueous and nonaqueous solutions are based on Gibbs energy
functions that are usually obtained from two-liquid theory using
local composition concept. In the last decades many correlations
have been developed to represent the nonideality of electrolyte
solutions and for accurate calculations of activity coefficient of
ionic species [1]. One class of models, which widely used in chem-
ical industries, is the local composition models. These types of
models are based on two terms of the long-range, electrostatic
interactions, and the short-range interactions. The long-range con-
tribution of activity coefficient is based on the electrostatic theory
of Poisson which results the brilliant equation of Debye–Hückel
or Pitzer–Debye–Hückel equation [2,3]. During recent decays, for
the short-range interaction variety of the local composition excess
∗ Corresponding author. Tel.: +98 21 82883313; fax: +98 21 82883381.
E-mail address: haghtala@modares.ac.ir (A. Haghtalab).
Gibbs functions have been proposed for electrolyte and nonelec-
trolyte systems. The different excess Gibbs functions have been
developed for molecular systems which have been successfully
used in calculation of activity coefficient of nonelectrolyte systems
[1]. These functions are the various versions of two-liquid theo-
ries such as Wilson, NRTL, UNIQUAC and UNIQUAC-NRF that are
based on the Wilson local composition expression [4–7]. The two-
liquid theory was used to develop the excess Gibbs function for
binary single-electrolyte systems such as Cruz and Renon [8], Chen
et al. [9,10], Haghtalab and Vera [11], Jaretun and Aly [12], Sadeghi
[13], Haghtalab and Mazloumi [14]. Dissolution of electrolytes in
a solvent dissociates into ionic or true species which with their
physico-chemical behavior in the solution demonstrate different
characteristic from original electrolyte. So, in developing the excess
Gibbs functions, those models that used the pure ionic species as a
reference state are not realistic such as electrolyte-NRTL and Cruz
and Renon models [8–10]. Moreover, the pure ion reference state
does not satisfy the electroneutrality principle. In the models such
as the electrolyte-NRTL-NRF, the electrolyte-Wilson-NRF and the
N-Wilson-NRF models [11,13,14], using the random state as a refer-
ence state, show the realistic behavior of ions in vanishing of local
composition. In construction of an excess Gibbs function for elec-
trolyte systems, two general approaches were applied. In the first
method, the molecular local composition models were directly used
for the electrolyte mixtures without changing the form of the excess
Gibbs functions. At the other method, the configuration of the local
cells was modified so that the new excess Gibbs functions were
obtained for electrolyte solutions. Sander et al. [15], Macedo et al.
[16] and Rasmussen [17] applied the original UNIQUAC equation to

0378-3812/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.fluid.2009.04.013
164 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171

  
correlation of the mean activity coefficient of electrolyte solutions. UW = CW UCW + AW UAW + WW UWW (3c)
Using ion pair assumption, Haghtalab and Mazloumi [14] devel-
Following the approach of Haghtalab and Vera [11], the energies of
oped a nonelectrolyte-local composition model for correlation of
the reference cells in the random state are expressed as:
the activity coefficient data of electrolyte solutions that results
accurate correlation of the mean activity coefficient of many binary UA0 = C UCA + W

UWA (4a)
electrolyte systems from dilute solution up to the saturation point
UC0 = A UAC 
+ W UWC (4b)
of the solution. As mentioned, in the second class of models, based
on two assumptions of the “like-ion repulsion” and “electroneu- 0
UW = C UCW + A UAW 
+ W UWW (4c)
trality of solvent cell”, the Wilson local composition expression
was applied for ionic and molecular species, so that the electrolyte By combining Eqs. (2)–(4), one can obtain the configurational excess
excess Gibbs functions were constructed by using the different ver- Gibbs energy of the cells as:

sions of two-liquid theory [9–13]. UAE = (CA − C )(UCA − UWA ) (5a)
In most electrolyte-local composition models such as

electrolyte-NRTL [9], Electrolyte-NRTL-NRF [11] and the sim- UCE = (AC − A )(UAC − UWC ) (5b)
ilar models the two adjustable energy parameters were used, 
E
UW = (AW − A )UAW 
+ (CW − C )UCW 
+ (WW 
− W )UWW (5c)
however, in extended UNIQUAC [17,18] models the structural
parameters of ionic species such as area and volume parameters After dissolution of m moles of a single electrolyte in a solvent with
of ions are used with binary interaction parameters as adjustable complete dissociation into NA of anions and NC of cations, the total
variables. So in UNIQUAC model, the four adjustable parameters excess internal energy of the mixture will be written as:
per binary electrolyte system were used. As a summary, one may 1 1 1
conclude, so far the developed local composition models used both UE = zqA NA UAE + zqC NC UCE + zqW NW UW
E
(6)
2 2 2
molecular and electrolyte versions of Wilson [13,14], NRTL and
where “z” is the coordination number and usually set to 10.
UNIQUAC [9,17,18], nevertheless, the UNIQUAC-NRF [7] model has
By substitution of Eq. (5) into Eq. (6), the molar excess internal
not been developed to correlation of binary electrolyte systems
energy of the binary system is obtained as:
yet.

At this work, following approach of Haghtalab and Asadol- uE = qA xA uCA (CA − C ) + qC xC uAC (AC

− A )
lahi [7], the electrolyte-UNIQUAC-NRF model is developed for the 
binary electrolyte solutions. In applying the model, the random + qW xW uEw [(AW − A ) + (CW

− C )] (7)
state of ions is assumed as the reference state and the structural where
parameters of ions, volume and area, are estimated using van der 1
Waals area and volume of ions. In contrast to electrolyte-UNIQUAC uCA = z(UCA − UWA )NA (8a)
2
model [18] which used four adjustable parameters, only two energy
1
parameters per electrolyte were used as the adjustable variables. uAC = z(UAC − UWC )NA (8b)
2
2. Electrolyte-UNIQUAC-NRF model 1 1
uEW = z(UAW − UWW )NA = z(UCW − UWW )NA (8c)
2 2
Following the general approach of the electrolyte-local composi- Following the UNIQUACL-NRF model [7], the Wilson local area com-
tion models, the excess Gibbs energy of a binary electrolyte solution position expression is written as:
includes the contribution of two terms as: ij i
 = ij,j (9)
g ex ex ex
= (g )PDH + (g )E-UNIQUAC-NRF (1) j 

where the first term denotes the long-range contribution which where
is presented by the Pitzer–Debye–Hückel [3] and the second term    a 
ij −(1/2)z(Uij − Uj ) ij
denotes the contribution of the short-range interaction that is ij,j = = exp = exp − (10)
j KT T
expressed by electrolyte-UNIQUAC-NRF model.
The short-range effects are originated by molecule–molecule, where the effective local area fraction is defined as:
molecule–ion and ion–ion interactions. So at this work, following
Haghtalab and Vera [11], for configuration of the species we assume ij = i ij (11)
the existence of three types of central cells. Two types of cells are The parameter “ ”, the nonrandom factor, is used for deviating
those with a cation or anion as central species. For cation and anion of local composition from bulk composition. To show the effect of
cells the assumption of like ion repulsion leads to the local mole charge of ions on nonideality of electrolyte solutions, one can define
fraction of cation around cation and anion around anion vanishes. the effective area fraction as:
The third type of cell is the solvent molecules in the center of the
X qi
cell and all the ionic and molecular species in surrounding. Follow- i = i (12)
Xq
j j j
ing approach of the UNIQUAC-NRF model [7], one can express the
excess internal energy for the each central cell as: where qj is surface parameter of species j and Xi is the effective
UiE = Ui − Ui0 i = C, A, W (2) mole fraction that is presented by Xi = Ci xi . Ci = Zi for ion “i” and
Ci = 1 for water. The mole fractions of cation and anion are defined,
Based on the like ion repulsion assumption, the effective local area respectively, as
 ,  =
fractions of like ions in ionic central cells are vanished, i.e. CC AA xA = A xS and xC = C xS (13)
0. So, the internal energy of the cells are expressed in terms of the
potential energies and the local area fractions as: Following Stockes and Robinson [19], the mole fraction of the elec-
  trolyte is defined as:
UA = CA UCA + WA UWA (3a)
m
  xS = (14)
UC = AC UAC + WC UWC (3b) m + 1000/18.02
 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171 165

where v = vA + vC and m is the molality of the electrolyte. Table 1

The structural parameters r and q for various anions [24].
Using the relation (11), Eq. (7) can be written a function of non-
random factor as: Anion R (Pm) r q Anion R (Pm) r q
1− 2−
uE = qA xA uCA C (CA − 1) + qC xC uAC A (AC − 1) Cl 181 1.020 1.025 Te 221 1.856 1.529
Br1− 196 1.295 1.202 SeO4 2− 243 2.467 1.848
+ qW xW [uAW A (AW − 1) + uCW C (CW − 1)] (15) I1− 220 1.831 1.515 MoO4 2− 254 2.818 2.019
F1− 133 0.405 0.554 N2 O2 2− 170 0.845 0.905
OH1− 133 0.405 0.554 SO4 2− 230 2.092 1.656
Using the local electroneutrality assumption around the solvent
NO3 1− 179 0.986 1.003 SO3 2− 200 1.376 1.252
central cell, one can obtain as: ClO2 1− 250 2.687 1.956 CO3 2− 178 0.970 0.992
ClO3 1− 200 1.376 1.252 S2 O3 2− 250 2.687 1.956
AW = CW = EW (16) ClO4 1− 240 2.377 1.803 S2 O4 2− 260 3.022 2.116
H2 PO4 1− 200 1.376 1.252 CrO4 2− 240 2.377 1.803
also by assuming UCA = UAC and UWA = UWC , CNS1− 213 1.662 1.420 HPO4 2− 200 1.376 1.252
BrO3 1− 191 1.198 1.142 S2 O6 2− 280 3.774 2.454
CA = AC = EE (17) BrO4 1− 250 2.687 1.956 S2 O8 2− 290 4.193 2.632
PF6 1− 245 2.529 1.879 HCO2 2− 204 1.460 1.303
On the other hand, the local area balances for the three central cells NO2 1− 192 1.217 1.154 HasO4 2− 210 1.592 1.380
are written as: MnO4 1− 240 2.377 1.803 WO4 2− 270 3.384 2.282
HSO4 1− 190 1.179 1.130 Cr2 O7 2− 320 5.634 3.205
 
CA + WA =1 (18a) HSO3 1− 170 0.845 0.905 C2 O4 2− 210 1.592 1.380
HseO4 1− 210 1.592 1.380 SiO3 2− 214 1.685 1.433
 
AC + WC =1 (18b) CH3 CO3 1− 232 2.147 1.685 SeO3 2− 239 2.347 1.788
AlO2 1− 260 3.022 2.116 TeO3 2− 250 2.687 1.956
  
AW + CW + WW =1 (18c) BO2 1− 240 2.377 1.803 N3− 171 0.860 0.915
CN1− 191 1.198 1.142 PO4 3− 238 2.318 1.773
Using Eqs. (11) and (18) and applying the relations (16) and (17), SCN1− 213 1.662 1.420 VO4 3− 260 3.022 2.116
the electrolyte and the solvent nonrandom factors, respectively, are SeCN1− 225 1.958 1.585 S2− 184 1.071 1.060
O2− 140 0.472 0.613 Se2− 198 1.335 1.227
derived as:
EE
EE =   (19)
E EE + W
The combinatorial term is identical to the term used in the tra-
EW ditional UNIQUAC equation, but we use the effective mole fraction
EW =   ) (20)
W + (1 − W EW instead of mole fraction. So the Guggenheim combinatorial term is
written [21] as:
where E = (A + C )/2  = 1 − 2  ;
and W E   
 a  gE
˚i z
˚i
EE = Xi ln − qi Xi ln (26)
EE = exp − (21a) RT Xi 2 i
T c
 a 
EW where ˚i , the effective volume fraction, is defined as:
EW = exp − (21b)
T
X ri
The parameters aEE and aEW are the electrolyte and solvent ˚i = i (27)
Xr
j j j
adjustable parameters for a binary electrolyte solution. Finally,
using the overall area balance and the relations (16) and (17), Eq.
(15) for the molar excess internal energy function is presented as:
Table 2
The structural parameters r and q for various cations [24].
E
u = (qA xA + qC xC )E uEE (EE − 1)

Cation R (Pm) r q Cation R (Pm) r q
+ qW xW uEW [(1 − W )(EW − 1)] (22)
1+ 2+
H 30 0.005 0.028 Pb 118 0.282 0.436
where Na1+ 102 0.182 0.326 Fe2+ 78 0.082 0.190
K1+ 138 0.452 0.596 EU2+ 117 0.275 0.428
uAC = uCA = uEE (23a) Li1+ 69 0.056 0.149 Yb2+ 113 0.248 0.400
SC1+ 170 0.845 0.905 Pt2+ 80 0.088 0.200
uCW = uAW = uEW (23b) Ag1+ 115 0.261 0.414 Hg2+ (I) 390 10.199 4.761
Tl1+ 150 0.580 0.704 Hg2+ (II) 102 0.182 0.326
Rb1+ 149 0.569 0.695 VO2 2+ 220 1.831 1.515
Substitution Eq. (22) into the Gibbs–Helmohltz equation, the molar PH4 1+ 180 1.003 1.014 Tl3+ 88 0.117 0.242
excess Helmohltz energy of the system is obtained as follows: NH4 1+ 148 0.557 0.686 PU3+ 101 0.177 0.319
 1/T 1 Cs1+ 170 0.845 0.905 Co3+ 61 0.039 0.116
aE Ni2+ 69 0.056 0.149 N2 H6 2+ 190 1.179 1.130
= uE d + integration constant (24) Pd2+ 86 0.109 0.231 Ti3+ 67 0.052 0.141
T 1/T0
T
Sr2+ 113 0.248 0.400 V3+ 64 0.045 0.128
CO2+ 75 0.073 0.176 Mn3+ 65 0.047 0.132
It should be noted that u are assumed to be independent of Mg2+ 72 0.064 0.162 Fe3+ 65 0.047 0.132
temperature and as shown by Hildebrand and Scott [20], as T0 Ca2+ 99 0.167 0.307 Sb3+ 77 0.078 0.186
approaches infinity the lower limit of molar excess Gibbs energy Ba2+ 136 0.433 0.579 In3+ 79 0.085 0.195
will be contributed from combinatorial part due to entropy so, Mn2+ 83.3 0.099 0.217 AL3+ 53 0.026 0.088
Zn2+ Cr3+
    Cd2+
75
95
0.073
0.147
0.176
0.282 Y3+
62
90
0.041
0.125
0.120
0.254
aE gE gE gE
≈ = + (25) UO2+ 280 3.774 2.454 Sm3+ 96 0.152 0.288
RT RT RT RT Cu2+ 73 0.067 0.167 SC3+ 75 0.073 0.176
T,V T,P c r
Ni2+ 69 0.056 0.149 Pr3+ 100 0.172 0.313
where the subscripts c and r denote the combinatorial and residual Be2+ 35 0.007 0.038 Nd3+ 99 0.167 0.307
UO2 2+ 280 3.774 2.454 La3+ 105 0.199 0.345
terms, respectively.
166 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171

Table 3
The binary adjustable parameters of the present model for 1:1, 1:2, 2:1, 2:2, 3:1 and 3:2 aqueous electrolytes at 298.15 K.

Electrolyte Max. m aEE /T aEW /T Electrolyte Max. m aEE /T aEW /T

Uni-univalent Uni-divalent
AgNO3 15 −3.059 0.006 Rb2 SO4 1.707 −2.136 2.971
CsBr 5 −2.278 4.867 Rb2 S2 O8 0.075 −3.998 20.050
CsCl 11 −2.151 4.557 Cs2 SO4 1.631 −1.961 2.937
CsI 3 −2.088 4.096 Cs2 S2 O8 0.109 −3.833 15.306
CsNO3 1.5 −2.977 4.317 Di-univalent
HBr 11 −3.301 36.941 PbCl2 0.039 −6.964 −3.088
HCl 16 2.247 30.725 Pb(ClO4 )2 10.830 −2.493 9.660
HClO4 10 −3.020 20.576 Pb(NO3 )2 1.830 −3.890 5.497
HI 10 −2.870 27.262 CuCl2 5.750 −1.089 −1.683
HNO3 28 −4.315 −2.397 Cu(ClO4 )2 3.557 −2.594 15.667
KAc 3.5 −0.548 6.717 CuBr2 3.606 2.355 −1.848
KBr 5.5 −1.516 4.784 Cu(NO3 )2 7.840 −1.864 14.471
KCl 5 −1.888 5.635 MnCl2 7.699 −4.452 −2.279
KCNS 5 −0.528 2.583 Mn(ClO4 )2 3.450 −2.380 16.879
KF 6 −2.707 16.017 MnBr2 5.640 −0.538 16.180
KH2 PO4 1.8 −3.004 4.336 UO2 Cl2 3.174 −0.478 4.259
KI 4.5 −0.521 3.853 UO2 (ClO4 )2 5.458 −2.624 8.333
KNO3 3.5 −2.786 1.528 UO2 (NO3 )2 5.511 −3.421 −1.772
KOH 20 −3.025 26.801 ZnF2 0.142 −6.201 −2.654
LiAc 4 −0.556 7.008 ZnCl2 23.193 −3.080 10.661
LiBr 6 −2.947 26.214 Zn(ClO4 )2 4.300 −3.074 19.190
LiCl 19.219 −2.832 28.137 ZnBr2 20.100 −1.744 9.951
LiClO4 4.5 −0.748 12.076 ZnI2 11.892 −0.568 8.532
LiI 3 −0.569 15.494 Zn(NO3 )2 6.750 −1.686 19.789
LiNO3 20 −0.488 14.587 Cd(ClO4 )2 1.928 −0.618 10.579
LiOH 5 −4.310 27.657 Cd(NO2 )2 7.840 −4.491 −1.488
NaBr 9 −2.264 12.054 Cd(NO3 )2 2.638 6.075 −1.852
NaBrO3 2.617 −2.464 4.330 FeCl2 2.050 −0.632 17.145
NaCl 6.144 −2.299 11.891 NiCl2 5.060 −0.681 20.182
NaClO3 3 −0.607 2.700 Ni(ClO4 )2 3.501 −2.801 18.810
NaClO4 6 −0.563 4.035 NiBr2 4.693 −2.720 24.519
NaCNS 18 −1.202 8.148 Ni(NO3 )2 4.623 −1.794 20.718
NaF 1 −3.000 12.673 CoCl2 4.118 −0.564 18.194
NaH2 PO4 6.5 −3.131 −1.429 Co(ClO4 )2 3.514 −2.709 18.196
NaI 12 −2.346 11.974 CoBr2 5.445 −1.999 21.074
NaNO3 10 −2.939 7.556 CoI2 6.500 −2.434 19.996
NaOH 29 −2.849 31.532 Co(NO3 )2 5.620 −2.271 19.833
NH4 Cl 7.405 −0.811 3.199 Tri-univalent
NH4 NO3 25 −1.878 1.054 AlCl3 1.8 −0.680 30.593
RbAc 3.5 −0.552 6.493 CeCl3 2 −0.475 16.166
RbBr 5 −1.974 4.674 CrCl3 1.2 0.379 −2.152
RbCl 7.8 −2.000 5.138 Cr(NO3 )3 1.4 0.459 −2.124
RbI 5 −1.894 4.421 EuCl3 2 −0.604 17.967
RbNO3 4.5 −2.787 1.597 LaCl3 2 −0.530 15.599
TiAc 6 −2.215 4.435 NdCl3 2 −0.610 16.600
Uni-divalent PrCl3 2 −0.612 16.232
(NH4 )2 HPO4 3.107 −3.380 2.531 ScCl3 1.8 −0.567 23.347
Li2 SO4 3.14 −0.528 3.658 SmCl3 2 −0.481 17.466
Na2 SO3 2.058 −0.526 1.654 YCl3 2 −0.581 19.302
Na2 SO4 1.957 −2.090 2.394 Di-divalent
Na2 S2 O3 4.052 −2.050 3.610 BeSO4 4 −3.177 7.624
Na2 S2 O6 0.852 −0.492 2.513 UO2 SO4 6 −1.884 2.585
Na2 S2 O8 0.098 −0.735 3.908 MgSO4 3.6 −2.841 5.975
Na2 HPO4 2.121 −1.924 0.013 MnSO4 4.9664 −3.034 5.537
Na2 fumarate 2.077 −2.178 7.101 NiSO4 2.5 −2.858 5.045
Na2 CO3 2.767 −2.033 2.881 CuSO4 1.4 −3.454 6.773
Na2 HAsO4 1.029 −0.552 1.964 ZnSO4 3.5 −3.146 5.917
Na2 CrO4 4.363 −2.180 4.426 CdSO4 3.5 −2.927 4.710
Na2 WO4 2.523 −1.609 3.848 Tri-bivalent
K2 SO4 0.692 −3.253 7.524 Al2 (SO4 )3 1 0.227 −1.412
K2 HPO4 0.873 −0.717 0.568 Cr2 (SO4 )3 1.2 4.187 −1.538
K2 HAsO4 0.886 −0.523 3.863
K2 CrO4 3.372 −1.711 2.680
K2 Cr2 O7 0.507 −1.884 −1.453

where rj is the volume parameter for species j. Finally, by integration where Eq. (28) presents the new electrolyte-UNIQUAC-NRF excess
of Eq. (24) yields: Gibbs model.


gE   3. The new activity coefficient equation
= (xA qA + xC qC ) ln EE − (1 − E ) ln EE
RT
r
 
 By differentiation of the excess Gibbs functions for the long-
+ xW qW ln EW − W ln EW (28) rang and the short-range terms, the activity coefficient of ionic or
 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171 167

Table 4
The deviation comparison of mean activity and osmotic coefficients using the present model with N-Wilson-NRF, E-Wilson-NRF, E-NRTL, E-NRTL-NRF and experimental data
for 1:1 aqueous electrolyte [25] solutions at 298.15 K.

Electrolyte Max. m This work ( ) N-Wilson-RF ( ) E-Wilson-NRF ( ) E-NRTL ( ) E-NRTL-NRF ( ) This work (
) E-NRTL (
)

Uni-univalent
AgNO3 15 0.036 0.0266 0.0215 0.0136 0.014 0.074 0.010
CsBr 5 0.007 0.0066 0.0093 0.0048 0.005 0.053 0.004
CsCl 11 0.017 0.0129 0.0236 0.0058 0.013 0.054 0.004
CsI 3 0.007 0.0074 0.0081 0.0069 0.005 0.042 0.006
CsNO3 1.5 0.002 0.0021 0.0024 0.0019 0 0.046 0.002
HBr 11 0.034 0.0523 0.0257 0.1829 0.04 0.071 0.163
HCl 16 0.057 0.0415 0.0238 0.1984 0.024 0.034 0.182
HClO4 10 0.068 0.0732 0.0375 0.1823 0.06 0.110 0.165
HI 10 0.055 0.0709 0.0431 0.1795 0.062 0.086 0.160
HNO3 28 0.066 0.0183 0.0264 0.0243 0.018 0.101 0.023
KAc 3.5 0.011 0.0065 0.0063 0.0063 0.008 0.019 0.006
KBr 5.5 0.013 0.0136 0.0135 0.0137 0.004 0.025 0.022
KCl 5 0.001 0.0012 0.0011 0.0035 0.021 0.028 0.003
KCNS 5 0.003 0.0019 0.0018 0.0019 0.003 0.014 0.002
KF 6 0.003 0.0056 0.0021 0.0124 0.006 0.034 0.011
KH2 PO4 1.8 0.004 0.0043 0.0047 0.0041 0.002 0.048 0.004
KI 4.5 0.008 0.0031 0.0027 0.0032 0.005 0.018 0.003
KNO3 3.5 0.006 0.0066 0.0070 0.0062 0.004 0.041 0.005
KOH 20 0.022 0.0645 0.0370 0.2736 0.039 0.050 0.256
LiAc 4 0.006 0.0033 0.0032 0.0033 0.005 0.012 0.003
LiBr 6 0.014 0.0175 0.0105 0.0410 0.045 0.053 0.037
LiCl 19.219 0.070 0.0486 0.0852 0.2662 0.052 0.079 0.245
LiClO4 4.5 0.015 0.0150 0.0145 0.0220 0.017 0.020 0.019
LiI 3 0.022 0.0189 0.0185 0.0192 0.02 0.020 0.018
LiNO3 20 0.082 0.0144 0.0284 0.0387 0.016 0.083 0.033
LiOH 5 0.020 0.0217 0.0255 0.0197 0.021 0.038 0.019
NaBr 9 0.007 0.0079 0.0062 0.0325 0.061 0.044 0.028
NaBrO3 2.617 0.001 0.0007 0.0006 0.0008 0.002 0.026 0.001
NaCl 6.144 0.008 0.0066 0.0068 0.0163 0.011 0.035 0.014
NaClO3 3 0.005 0.0041 0.0047 0.0035 0.005 0.008 0.003
NaClO4 6 0.007 0.0068 0.0065 0.0067 0.009 0.021 0.005
NaCNS 18 0.044 0.0368 0.0512 0.0652 0.039 0.057 0.056
NaF 1 0.001 0.0009 0.0009 0.0009 0.002 0.011 0.001
NaH2 PO4 6.5 0.004 0.0031 0.0074 0.0021 0.003 0.067 0.002
NaI 12 0.023 0.0336 0.0135 0.0909 0.028 0.069 0.081
NaNO3 10 0.072 0.0662 0.0614 0.0866 0.072 0.094 0.084
NaOH 29 0.155 0.0788 0.1670 0.3288 0.057 0.146 0.306
NH4 Cl 7.405 0.002 0.0023 0.0034 0.0015 0.002 0.016 0.001
NH4 NO3 25 0.017 0.0230 0.0191 0.0125 0.01 0.050 0.008
RbAc 3.5 0.010 0.0059 0.0057 0.0057 0.008 0.019 0.005
RbBr 5 0.004 0.0038 0.0049 0.0030 0.001 0.036 0.002
RbCl 7.8 0.007 0.0045 0.0085 0.0025 0.003 0.041 0.002
RbI 5 0.005 0.0043 0.0057 0.0033 0.002 0.043 0.003
RbNO3 4.5 0.007 0.0094 0.0103 0.0082 0.007 0.051 0.007
TiAc 6 0.012 0.0117 0.0139 0.0099 0.01 0.046 0.008

Overall 0.023 0.0193 0.0195 0.0492 0.0186 0.047 0.045

molecular species in a binary electrolyte system can be written as: respectively, that is defined as:
1
2
ln i = (ln i )PDH + (ln i )∗c + (ln i )∗r i = A, C, W (29) Ix = xi zi (32)
2
i

where the Pitzer–Debye–Hückel equation [9] for ions is written as: On the other hand, the electrolyte-UNIQUAC-NRF term is divided
to two parts of the combinatorial or entropic term, and the residual

 1000 0.5 2zi2
1/2
zi2 Ix
3/2
− 2Ix or enthalpic term. Thus, by proper differentiation of Eqs. (26) and
1/2
(ln i )PDH = −A
ln(1 + Ix ) + (28), the combinatorial activity coefficient is obtained as:
MW  1 + Ix
1/2
     
˚i ˚i z ˚i ˚i
(ln i )c = ln +1− − q ln +1−
(30) Xi Xi 2 i i 
i

and using the Gibbs–Duhem equation, the long-range activity coef- i = C, A, W (33)
ficient of water is obtained as: and the residual terms for anion, cation and water, respectively, are
   derived as:
xS
 1000 1/2 ZC ZA  Ix1/2 − 2Ix3/2 1 ZA 
(ln W )PDH = A
(31) (ln A )r = ln EE + E (1 − EE ) + 
 (1 − EE ) + ZA W (1−EW )
xW MW 1 + Ix
1/2 qA ZC E
  ZA 

−(1 − E ) ln EE 1 − E + + ZA   W ln EW
2

where A
and  are the Debye–Hückel parameters. Mw and Ix are the ZC E
molecular weight of solvent and the mole fraction ionic strength, (34)
168 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171

Table 5
The deviation comparison of mean activity and osmotic coefficients using the present model with N-Wilson-NRF, E-Wilson-NRF, E-NRTL, E-NRTL-NRF and experimental data
for 1:2 and 2:1 [26–29] aqueous electrolyte solutions at 298.15 K.

Electrolyte Max. m This work ( ) N-Wilson-RF ( ) E-Wilson-NRF ( ) E-NRTL ( ) E-NRTL-NRF ( ) This work (
) E-NRTL (
)

Uni-divalent
(NH4 )2 HPO4 3.107 0.050 0.0404 0.0704 0.0548 0.0640 0.097 0.041
Li2 SO4 3.14 0.041 0.0412 0.0412 0.0432 0.0415 0.026 0.016
Na2 SO3 2.058 0.006 0.0059 0.0059 0.0060 0.0087 0.019 0.005
Na2 SO4 1.957 0.005 0.0050 0.0051 0.0048 0.0011 0.019 0.003
Na2 S2 O3 4.052 0.020 0.0177 0.0145 0.0294 0.0178 0.026 0.027
Na2 S2 O6 0.852 0.022 0.0145 0.0145 0.0198 0.0252 0.027 0.011
Na2 S2 O8 0.098 0.002 0.0011 0.0012 0.0011 0.0003 0.009 0.001
Na2 HPO4 2.121 0.011 0.0102 0.0101 0.0104 0.0140 0.022 0.009
Na2 fumarate 2.077 0.010 0.0095 0.0101 0.0082 0.0059 0.022 0.049
Na2 CO3 2.767 0.010 0.0101 0.0102 0.0105 0.0128 0.024 0.010
Na2 HAsO4 1.029 0.035 0.0185 0.0186 0.0332 0.0230 0.023 0.025
Na2 CrO4 4.363 0.039 0.0379 0.0337 0.0542 0.0376 0.050 0.048
Na2 WO4 2.523 0.011 0.0114 0.0121 0.0097 0.0077 0.032 0.006
K2 SO4 0.692 0.013 0.0134 0.0142 0.0130 0.0098 0.062 0.009
K2 HPO4 0.873 0.003 0.0014 0.0014 0.0022 0.0045 0.002 0.002
K2 HAsO4 0.886 0.038 0.0199 0.0200 0.0350 0.0243 0.025 0.024
K2 CrO4 3.372 0.008 0.0069 0.0073 0.0131 0.0093 0.016 0.012
K2 Cr2 O7 0.507 0.121 0.1273 0.1277 0.2071 0.1316 0.072 0.110
Rb2 SO4 1.707 0.009 0.0095 0.0098 0.0087 0.0049 0.030 0.005
Rb2 S2 O8 0.075 0.002 0.0027 0.0027 0.0026 0.0021 0.060 0.002
Cs2 SO4 1.631 0.013 0.0131 0.0135 0.0166 0.0087 0.027 0.010
Cs2 S2 O8 0.109 0.004 0.0043 0.0044 0.0043 0.0035 0.078 0.003

Overall 0.021 0.019 0.020 0.027 0.021 0.034 0.019

Di-univalent
PbCl2 0.039 0.018 0.0211 0.0216 0.0217 0.0230 0.092 0.017
Pb(ClO4 )2 10.83 0.061 0.0471 0.0359 0.2989 0.0368 0.160 0.277
Pb(NO3 )2 1.83 0.042 0.0483 0.0514 0.0458 0.0840 0.093 0.003
CuCl2 5.75 0.016 0.0161 0.0162 0.0188 0.0486 0.019 0.016
Cu(ClO4 )2 3.557 0.067 0.0627 0.0582 0.1222 0.0611 0.120 0.110
CuBr2 3.606 0.023 0.0241 0.0245 0.0242 0.0332 0.026 0.016
Cu(NO3 )2 7.84 0.034 0.0321 0.0316 0.1252 0.0348 0.055 0.122
MnCl2 7.699 0.081 0.0179 0.0162 0.0159 0.0624 0.167 0.011
Mn(ClO4 )2 3.45 0.082 0.0868 0.0772 0.1311 0.0271 0.116 0.115
MnBr2 5.64 0.063 0.0301 0.0299 0.0692 0.0522 0.055 0.066
UO2 Cl2 3.174 0.037 0.0223 0.0225 0.0228 0.0249 0.068 0.018
UO2 (ClO4 )2 5.458 0.192 0.0771 0.0656 0.3499 0.0295 0.239 0.321
UO2 (NO3 )2 5.511 0.069 0.0338 0.0344 0.0298 0.0783 0.215 0.029
ZnF2 0.142 0.005 0.0058 0.0060 0.0058 0.0108 0.039 0.005
ZnCl2 23.193 0.094 0.1039 0.0954 0.2293 0.0605 0.143 0.225
Zn(ClO4 )2 4.3 0.082 0.0689 0.0579 0.1926 0.0159 0.172 0.177
ZnBr2 20.100 0.156 0.1510 0.1463 0.1838 0.1157 0.169 0.172
ZnI2 11.892 0.283 0.2241 0.2218 0.2227 0.1387 0.250 0.193
Zn(NO3 )2 6.75 0.047 0.0476 0.0453 0.1450 0.0176 0.050 0.141
Cd(ClO4 )2 1.928 0.044 0.0426 0.0419 0.0423 0.0045 0.034 0.034
Cd(NO2 )2 7.84 0.033 0.1118 0.1532 0.0681 0.1626 0.157 0.047
Cd(NO3 )2 2.638 0.023 0.0233 0.0236 0.0235 0.0229 0.092 0.018
FeCl2 2.05 0.020 0.0147 0.0142 0.0145 0.0259 0.015 0.012
NiCl2 5.06 0.028 0.0284 0.0283 0.0768 0.0115 0.023 0.066
Ni(ClO4 )2 3.501 0.078 0.0735 0.0683 0.1414 0.0219 0.141 0.128
NiBr2 4.693 0.044 0.0481 0.0414 0.1277 0.0164 0.087 0.114
Ni(NO3 )2 4.623 0.047 0.0480 0.0465 0.0910 0.0226 0.049 0.090
CoCl2 4.118 0.047 0.0363 0.0357 0.0470 0.0231 0.030 0.038
Co(ClO4 )2 3.514 0.076 0.0727 0.0679 0.1381 0.0742 0.135 0.124
CoBr2 5.445 0.043 0.0461 0.0428 0.1293 0.0332 0.059 0.112
CoI2 6.5 0.094 0.0890 0.1041 0.1863 0.0740 0.044 0.166
Co(NO3 )2 5.62 0.039 0.0438 0.0372 0.1111 0.0263 0.045 0.100

Overall 0.064 0.056 0.055 0.108 0.046 0.098 0.099

1 ZC  Since the infinite dilution standard state is used for the

(ln C )r = ln EE + E (1 − EE ) + 
 (1 −EE )+ZC W (1 −EW )
qC ZA E Pitzer–Debye–Hückel term activity coefficient, so the short-range
  ZC
 activity coefficient of ionic species should be normalized on the
− (1 − E ) ln EE 1 − E + + ZC   W ln EW
2
E (35) same basis. Thus, the infinite dilution activity coefficients are pre-
ZA
sented as:

1
1 1
  EE

(ln W )r = + E 1− − E ln EE
qW ZC ZA EE (ln i∗ )c = (ln i )c − (ln i∞ )c i = A, C (37)
 EW

 
+ ln EW − W ln EW + W 1− − 2E ln EW
EW
(36) (ln i∗ )r = (ln i )r − (ln i∞ )r i = A, C (38)
 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171 169

Table 6
The deviation comparison of mean activity and osmotic coefficients using the present model with N-Wilson-NRF, E-Wilson-NRF, E-NRTL, E-NRTL-NRF and experimental data
for 3:1 [19], 2:2 and 3:2 [30] aqueous electrolyte solutions at 298.15 K.

Electrolyte Max. m This work ( ) N-Wilson-RF ( ) E-Wilson-NRF ( ) E-NRTL ( ) E-NRTL-NRF ( ) This work (
) E-NRTL (
)

Tri-univalent
AlCl3 1.8 0.084 0.0802 0.0766 0.115 0.080 0.075 0.104
CeCl3 2 0.083 0.0647 0.0613 0.084 0.063 0.069 0.076
CrCl3 1.2 0.048 0.0494 0.0474 0.069 0.073 0.046 0.039
Cr(NO3 )3 1.4 0.047 0.0488 0.0469 0.054 0.07 0.046 0.047
EuCl3 2 0.088 0.0708 0.0666 0.091 0.068 0.074 0.082
LaCl3 2 0.087 0.0647 0.062 0.082 0.063 0.073 0.074
NdCl3 2 0.078 0.0636 0.0595 0.083 0.061 0.067 0.075
PrCl3 2 0.081 0.0638 0.0603 0.082 0.062 0.069 0.075
ScCl3 1.8 0.072 0.0604 0.0565 0.078 0.059 0.069 0.071
SmCl3 2 0.084 0.0667 0.0627 0.087 0.064 0.070 0.078
YCl3 2 0.074 0.0660 0.0623 0.093 0.064 0.065 0.085

Overall 0.074 0.064 0.060 0.083 0.066 0.065 0.073

Di-divalent
MgSO4 3.6 0.041 0.035 0.034 0.031 0.118 – –
MnSO4 4.9664 0.047 0.078 0.080 0.073 0.165 – –
NiSO4 2.5 0.074 0.072 0.071 0.081 0.093 – –
CuSO4 1.4 0.105 0.045 0.048 0.042 0.213 – –
ZnSO4 3.5 0.078 0.107 0.109 0.099 0.127 – –
CdSO4 3.5 0.081 0.082 0.084 0.078 0.132 – –
BeSO4 [19] 4 0.052 0.054 0.062 0.039 0.078 0.105 0.062
UO2 SO4 [19] 6 0.053 0.075 0.095 0.050 0.054 0.088 0.050

Overall 0.066 0.069 0.073 0.062 0.123 0.096 0.056

Tri-divalent
Al2 (SO4 )3 1 0.080 0.143 0.143 0.122 0.138 – –
Cr2 (SO4 )3 1.2 0.133 0.113 0.080 0.062 0.080 – –

Overall 0.107 0.128 0.111 0.092 0.109

Using Eqs. (33), the combinatorial infinite dilute activity coefficient The volume and surface parameters of the ions are calculated by
is written as: using the radius of the species and the following relations [23]:
r  r  z
 r qW r qW

(ln i∞ )c = ln i
+1− i
− q ln i +1− i VWK 4 3
rW rW 2 i rW qi rW qi ri = ; VWK = R (43)
15.17 3 i
i = A, C, W (39) AWK
qi = ; AWK = 4Ri2 (44)
and using (34) and (35) the residual infinite activity coefficient is 2.5 × 109
obtained as: where the ionic radii, Ri , are obtained from Yizhak Marcus [24].
The structural parameters, r and q, of ions that are calculated by
(ln i∞ )r = qi Zi [1 − EW + ln EW ] i = A, C (40) Eqs. (43) and (44) are shown in Tables 1 and 2.
Finally the mean ionic activity coefficient of a single electrolyte in Using an optimization program based on Nelder–Mead algo-
a binary mixture is calculated as: rithm, the binary interaction parameters for binary electrolyte
systems are obtained. The new model has been used for many
A ln A∗ + C ln C∗ uni-univalent and nonuni-valent electrolyte systems at 25 ◦ C. The
ln ±∗ = (41)
A + C binary interaction parameters resulting from regression of the
mean ionic activity coefficient for all type of electrolytes are shown
4. Results and discussion in Table 3. Table 4 shows the accuracy of the present model with
the other models such as E-NRTL, E-NRTL-NRF, Wilson-NRF and
The new present model is applied to correlation of the N-Wilson-NRF for uni-univalent electrolytes. As one can see accu-
experimental data of the mean ionic activity coefficient for the racy of the mean activity coefficient of the present model is better
single-electrolyte solutions that were collected from the litera- than the Electrolyte-NRTL, but comparable in precision with the
ture [19,25–30]. To obtain the electrolyte binary parameters the other models. Also Table 4 shows the accuracy of the osmotic coef-
mean square standard deviation of mean activity coefficient data, ficient using the present model and E-NRTL. The osmotic coefficient
is employed as objective function: is calculated by the binary parameters resulting from fitting of the
 2
1/2 activity coefficient data for the same electrolyte and the deviation
Exp
(ln ± − ln ±Calc ) of prediction is calculated as:
 = (42)
N  2
1/2
(
Exp −
Calc )

= (45)
where the superscripts show the calculated and experiment, N
respectively. At this work, the energy parameters: aEE and aEW are
only adjusted and the surface and volume parameters of cation, As Table 4 shows the prediction of osmotic coefficient is
anion and water are calculated by van der Waals area and volume very good and very close to E-NRTL prediction. Tables 5 and 6
of the species. The volume and area parameters for water are set show the results of fitting for molality mean ionic activity coef-
r = 0.920 and q = 1.4, respectively [22]. ficients for nonuni-univalent electrolytes at 298.15 K. As shown in
170
Fig. 1. Experimental and calculated molality mean activity coefficient of various
uni-univalent electrolytes at 298.15 K.
Fig. 2. Experimental and calculated molality mean activity coefficient of various
nonuni-univalent electrolytes at 298.15 K.
Debye–Hückel constant
Fig. 3. Experimental and calculated molality mean activity coefficient of various
acids and bases at 298.15 K.
A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171
Fig. 4. Experimental and predicted values of the osmotic coefficient of various elec-
trolytes at 298.15 K.
Tables 5 and 6, the results of the present work are compared with
the N-Wilson-NRF, E-Wilson-NRF, E-NRTL and E-NRTL-NRF models.
The results of the present work for uni-divalent and di-univalent
electrolytes show better accuracy respect to Electrolyte-NRTL
model. Fig. 1 shows the variation of mean activity coefficient of
highly soluble electrolyte versus the molality of salts for different
electrolytes up to saturation point. A one can see the Electrolyte-
UIQUAC-NRF model correlates the experimental data very good.
Also Fig. 2 shows the results of the model for nonuni-univalent elec-
trolytes with very good agreement of the model with experiment.
Moreover, Fig. 3 shows the mean activity coefficient of several acids
against the molality of acid. Finally, Fig. 4 shows the results of pre-
diction of osmotic coefficient for some electrolyte solutions. As one
can see the prediction results are in very good agreement with the
experiment.
5. Conclusion
A new electrolyte-local composition model, E-UNIQUAC-NRF,
based on an ionic framework, was developed. The model was
adopted for representation of the short-range contribution of
excess Gibbs energy of electrolyte solutions so that the struc-
tural parameters of the ions such as volume and area of ions
were taken into account. For the long-range contribution, the
Pitzer–Debye–Hückel model has been used. The present model has
two adjustable parameters per electrolyte and by correlating of
mean ionic activity coefficients, the binary parameters were deter-
mined for more than 131 aqueous electrolyte solutions at 25 ◦ C. The
comparison of the results with experiment showed that the new
model is applicable over the whole range of electrolyte molality,
from dilute solution up to saturation. Moreover, it is shown that the
model is an accurate model for prediction of the osmotic coefficient
of the electrolyte solutions.
List of symbols
A
g Gibbs energy function
u internal energy function
uE molar excess internal energy
Uij energy interaction parameter
Ix ionic strength on mole fraction basis
m molality
MW molecular weight of water
R universal gas constant
T absolute temperature
 A. Haghtalab, K. Peyvandi / Fluid Phase Equilibria 281 (2009) 163–171 171

x mole fraction E ion

X effective mole fraction W water
z coordination number i,j,k any compound
Z charge number of ionic species ± mean
Ni moles of species i
ri volume parameter of species i
References
qi surface parameter of species i
[1] J.M. Prausnitz, R.N. Lichtenthaler, E.G. de Azevedo, Molecular Thermodynamics
Greek symbols of Fluid Phase Equilibrium, 3rd ed., Prentice Hall PTR, NJ, 1999.
[2] J.F. Zemaitis, D.M. Clark, M. Rafal, N.C. Scrivner, Handbook of Aqueous Elec-
activity coefficient trolyte Thermodynamics, American Institute of Chemical Engineers, NY, 1986.

osmotic coefficient [3] K.S. Pitzer, J. Am. Chem. Soc. 102 (1980) 2902–2906.
v stoichiometric number [4] G.M. Wilson, J. Am. Chem. Soc. 86 (1964) 127–130.
[5] H. Renon, J.M. Prausnitz, AIChE J. 14 (1968) 135–144.
 closest approach parameter [6] G. Maurer, J.M. Prausnitz, Fluid Phase Equilib. 2 (1978) 91–99.
␴ standard deviation for mean activity coefficient [7] A. Haghtalab, M.A. Asadollahi, Fluid Phase Equilib. 171 (2000) 77–90.
␴
standard deviation for osmotic coefficient [8] J.L. Cruz, H. Renon, AICHE J. 24 (1978) 817–829.
[9] C.C. Chen, H.I. Britt, J.F. Boston, L.B. Evans, AICHE J. 28 (1982) 588–596.
 ij nonrandom factor
[10] C.C. Chen, L.B. Evans, AICHE J. 32 (1986) 1655–1664.
␶EE interaction energy parameter between two ions [11] A. Haghtalab, J.H. Vera, AICHE J. 34 (1988) 803–813.
␶EW interaction energy parameter between ion and solvent [12] A. Jaretun, G. Aly, Fluid Phase Equilib. 163 (1999) 175–193.
i effective area fractions of i [13] R. Sadeghi, Fluid Phase Equilib. 231 (2005) 53–60.
[14] A. Haghtalab, S.H. Mazloumi, Fluid Phase Equilib. 275 (2008) 70–77.
ij effective local area fractions of j around i [15] B. Sander, A. Fredenslund, P. Rasmussen, Chem. Eng. Sci. 41 (1986)
˚i effective volume fraction 1171–1183.
[16] A. Macedo, P. Skovborg, P. Rasmussen, Chem. Eng. Sci. 45 (1990) 875–882.
[17] K. Thomsen, P. Rasmussen, R. Gani, Chem. Eng. Sci. 51 (1996) 3675–3683.
Superscripts [18] B. Messnaoui, S. Ouiazzane, A. Bouhaouss, T. Bounahmidi, CALPHAD 32 (2008)
Calc calculated 566–576.
[19] R.A. Robinson, R.H. Stokes, Electrolyte Solutions, 2nd ed., Butterworth, London,
Exp experimental 1965.
E excess [20] J.H. Hildebrand, R.L. Scott, The Solubility of Non-electrolytes, 3rd ed., Reinhold
LR long-range Publishing Corporation, New York, 1950.
[21] E.A. Guggenheim, Mixtures, Oxford Univ. Press, Oxford, 1952.
SR short-range [22] Y.L. Gao, Q.H. Peng, Z.C. Li, Y.G. Li, Fluid Phase Equilib. 63 (1991) 157–171.
* unsymmetrical convention [23] A. Bondi, Physical Properties of Molecular Crystals, Liquids, and Glasses, John
∞ infinite dilution Wiley & Sons, New York, 1968.
[24] Y. Marcus, Ion Properties, 3rd ed., Marcel Dekker, Inc., New York, 1997.
[25] W.J. Hamer, Y.C. Wu, J. Phys. Chem. Ref. Data 1 (1972) 1074–1099.
Subscripts [26] R.N. Goldberg, J. Phys. Chem. Ref. Data 10 (1981) 671–764.
o reference state [27] R.N. Goldberg, J. Phys. Chem. Ref. Data 8 (1979) 1005–1050.
[28] R.N. Goldberg, J. Phys. Chem. Ref. Data 10 (1981) 1–55.
c combinatorial [29] R.N. Goldberg, R.L. Nuttall, B.R. Staples, J. Phys. Chem. Ref. Data 8 (1979)
r residual 923–1003.
A anion [30] L.D. Zaytsev, G.G. Aseyev, Properties of Aqueous Solutions of Electrolytes, CRC
Press, Boca Raton, FL, 1992.
C cation