Journal of Solution Chemistry, Vol. 23, No.

3, 1994

Solubility of Carbon Dioxide in Aqueous

Solutions of Sodium Chloride: Experimental
Results and Correlation
B. R u m p f , 1 H. N i c o l a i s e n , z C. t3cal, 1 and
G. M a u r e r 1'3
Received July 30, 1993; Revised November 30, 1993

The solubility of carbon dioxide in aqueous solutions of sodium chloride

was measured in the temperature range from 40 to 160~ up to 6 mol-kg
salt solutions and total pressures up to 10 MPa. Pitzer' s(1) equations as
well as the Chen and Evans(z) model were used to correlate the new data.
Results are reported and compared to literature data and correlations.

KEY WORDS: Gas solubility; weak electrolytes; strong electrolytes;

carbon dioxide; sodium chloride; water; correlation.

1. Introduction
The solubility of weak electrolyte gases like ammonia, carbon
dioxide, sulfur dioxide or hydrogen sulfide must be known for designing
separation equipment in many technical applications, for example in the
chemical or oil-related industries or for the production of fertilizers.
Correlating and predicting phase equilibria in such systems still is a dif-
ficult task, partially because of chemical reactions in the liquid phase
but also because of the possible formation of one or more solid phases.
The situation becomes even more complicated when strong electrolytes
like chlorides, sulfates or nitrates are also present. Physico-chemical
models to correlate and predict the solubility of weak electrolyte gases

1Lehrstuhl ftir Technische Thermodynamik, Universitat Kaiserslautem, D-67653

Kaiserslauterrh Federal Republic of Germany.
2On leave from the Institutet for Kemiteknik, Danmarks Tekniske Hr DK-2800
Lyngby, Denmark.
3To whom correspondence should be addressed.
431
0095-9782/94/0100-0431507.00/0 9 1994PleonmPublishingCorporation
432 Rumpf, Nicolaisen, Ocal, and Maurer

cannot be tested or improved as long as published data on such systems

is as scarce as it is at present. This is not only the case for the simul-
taneous solubility of ammonia and sour gases, but also for the solubility
of the single gases in aqueous phases. As a continuation of earlier work
on the solubility of carbon dioxide in aqueous electrolyte solutions, ~
this publication reports on experimental and theoretical work on the
solubility of carbon dioxide in aqueous solutions of sodium chloride in
the temperature range from 313 to 433 K, up to 6 mol-kg salt solutions
and total pressures up to 10 MPa.

2. E x p e r i m e n t a l
The experimental equipment for measuring the solubility of a
single gas in aqueous solutions was described in a previous
publication, ~3)therefore only some essentials are given here.
In an experiment a high pressure optical cell is filled with carbon
dioxide. Pressure and temperature of the gas are measured after equi-
libration. The aqueous solvent is then added by a calibrated high pres-
sure displacer until the gas is completely dissolved. The amount of sol-
vent added into the ceil is slightly above the minimum needed to dis-
solve the gas completely. The pressure is then reduced stepwise by
withdrawing small amounts of the liquid mixture until the first stable
bubble appears.
The mass of carbon dioxide filled into the ceil is calculated from
the Bender equation of state ~4) using the experimental results for tem-
perature and pressure and the known volume of the cell. The volume of
solvent needed to dissolve the gas is determined by measuring the posi-
tion of the high pressure displacer piston before and after each experi-
ment. The density of the aqueous salt solution was taken from
Washbumfl ~ The temperature is measured with two calibrated resis-
tance thermometers with an accuracy of 0.1 K. The pressure of pure
carbon dioxide is measured with two pressure transducers (ranges: 0 -
0.6 MPa and 0 - 2.5 MPa).
Three transducers with ranges of 0 - 6 MPa, 0 - 10 MPa and 0 - 20
MPa were used to determine the solubility pressure. Before and after
each series of measurements the transducers were calibrated against a
high precision pressure gauge. A detailed calculation of the propagation
of errors shows that the uncertainty of the experimental results for the
solubility pressure (resulting from uncertainties in determining the con-
centration of the gas, concentration of the salt, temperature and total
pressure) are below 3% at total pressures up to 6 MPa and below 5% at
higher pressures.
Carbon Dioxide in Aqueous Solutions 433

Table I. Experimental Results for the Solubility of Carbon Dioxide

in Pure Water and Comparison with Literature Data

7"(K) ~o2 ~ pb pC ~/p ~ pd Ap/p %

323.17 0.196 1.059 1.076 1.6

323.16 0.230 1.247 1.285 3.1
323.17 0.259 1.432 1.464 1.7
323.17 0.299 1.660 1.719 3.5
323.17 0.323 1.809 1.870 3.3
323.17 0.581 3.480 3.384 -2.7
323.18 0.871 5.798 5.762 -0.6

~ Units: mol-kg, b Units: MPa. ~ Units: MPa; Ref. 6. '/Units: MPa; Ref. 7.

3. Substances
Carbon dioxide (>99.995 mole percent) was purchased from
Messer-Griesheim, Ludwigshafen and used without further purification.
Sodium chloride (>99.7 mass percent) was purchased from Merck
GmbH, Darmstadt. The salt was degassed and dried under vacuum.
Water was deionized and degassed by vacuum distillation.

4. Test of P r o c e d u r e
4.1. Results for COz-H20
As before,c3) the experimental equipment and procedure was
checked by measuring the solubility of carbon dioxide in pure water at
323.15 K. The experimental results which cover carbon dioxide
molalities up to about 0.9 mol-kg (corresponding to total pressures up to
6 MPa) are summarized and compared to literature data in Table
I. Good agreement between those results and literature data is observed.
For each of the new data points, the absolute and relative deviations be-
tween the new results for the total pressure and those interpolated from
Zawisza and Malesinska~~ and Zelvenski~ are given. As can be seen
from that comparison, the results of Zawisza and Malesinska~~ are
slightly above the new results for the total pressure. Relative deviations
do not exceed 3.5% corresponding to absolute deviations of 0.06 MPa.
In contrast to this, Zelvenskio) reports total pressures which are
predominantly below our results. The maximum relative deviation
amounts to 2.7% corresponding to an absolute deviation of about 0.1
MPa.
434 Rumpf, Nicolaisen, t)cal, and Maurer

T a b l e H . E x p e r i m e n t a l Results for the S o l u b i l i t y o f C a r b o n D i o x i d e

in A q u e o u s S o l u t i o n s o f S o d i u m C h l o r i d e
m a -- a -- a

T (K) ~/C02a mNac1 pb T (K) mco 2 mNac1 pb

313.16 0.0497 3.997 0.467 432.97 0.0000 4.003 0.515

313.14 0.0995 3.997 0.911 432.86 0.0506 4.003 1.662
313.14 0.1996 3.997 1.900 432.96 0.1013 4.003 2.858
313.19 0.3098 3.997 3.147 432.95 0.2188 4.003 5.912
313.19 0.4529 3.997 5.123 432.98 0.3254 4.003 9.048
313.19 0.5497 3.997 6.917 313.31 0.0517 5.999 0.602
333.13 0.0509 3.997 0.625 313.22 0.1035 5.999 1.232
333.16 0.0909 3.997 1.141 313.28 0.2074 5.999 2.640
333.16 O.1073 3,997 1.328 313.27 0.3133 5.999 4.430
333.15 0.1922 3.997 2.514 313.31 0.4177 5.999 6.799
333.16 0.2184 3.997 2.871 313.19 0.4693 5.999 8.427
333.14 0.2570 3.997 3.521 333.12 0.0526 5.999 0.820
333.17 0.3305 3.997 4.737 333.12 0.1047 5.999 1.678
333.15 0.4472 3,997 7.303 333.10 0.2106 5.999 3.657
333.16 0.5272 3.997 9.642 333.00 0.3388 5.999 6.844
353.12 0.0505 4.001 0.817 333.05 0.3965 5.999 8.670
353.12 0.1014 4.001 1.642 353.12 0.0509 5.999 0.997
353.10 0.2032 4.001 3.387 353.11 0.1021 5.999 2.032
353,10 0.3084 4,001 5.592 353.08 0.2043 5.999 4.394
353.12 0.3594 4,001 6.940 353.10 0.3085 5.999 7.610
353.11 0.4116 4,001 8.337 353.08 0.3553 5.999 9.044
353,11 0.4613 4,001 9.637 393.09 0.0000 5.999 0.151
393.06 0.0000 4.003 0.169 393.17 0.0518 5.999 1.421
393,07 0.0518 4.001 1.204 393.14 0.1031 5.999 2.778
393.19 0.1035 4,003 2.331 393.13 0.2067 5.999 5.937
393.10 0.2084 4.001 4.742 393.12 0.3049 5.999 9.135
393.19 0.3135 4.003 7.650 413.10 0.0000 5.999 0.275
393.12 0.3657 4.003 9.328 413.09 0.0511 5.999 1.661
413.06 0.0000 4.003 0,301 413,09 0.1024 5.999 3.195
413.07 0.0505 4.003 1.393 413.09 0.2051 5.999 6.291
413.08 O.1011 4.003 2.550 413.12 0.2878 5.999 9.201
413.09 0.2027 4,003 5.042 433.12 0.0000 5.999 0.471
413.07 0.3272 4.003 8.671 433.07 0.0512 5.999 1.898
433.05 0.1025 5.999 3.406
433,08 0.2052 5.999 6.578
433.05 0.2743 5.999 8.981

a Units: mol-kg, b Units: MPa.

Carbon Dioxide in Aqueous Solutions 435

10-

r~nect= 6 real tkg

/
/
/
/

02 OL, 06
rflc02 l(mo[/kg)
Fig. 1. Solubilityof carbon dioxide in aqueous soludortsof sodium chloride at 40~ o,
experimentalresults (this work); and -- ., calculatedresults for the systemCO2-H20.
4.2. Results for CO2-NaCI-H20
The experimental results for the solubility of carbon dioxide in
aqueous solutions of sodium chloride at salt molalities of 4 and 6 mol-
kg and temperatures between 313.15 K and 433.15 K are given in Table
II. The carbon dioxide molality ranged up to about 0.55 mol-kg, the
total pressure up to 10 MPa.
Figure 1 demonstrates the influence of sodium chloride on the
solubility of carbon dioxide. Carbon dioxide is salted out by sodium
chloride. For example at 313.15 K, the total pressure above a solution
containing 0.2 mol-kg of carbon dioxide and 4 or 6 mol-kg of sodium
chloride is 1.9 or 2.6 MPa whereas it is 0.9 MPa above pure water.
Conversely at a constant pressure of 2 MPa, the carbon dioxide molality
is 0.42 mol-kg in pure water whereas it is 0.2 or 0.15 mol-kg in a 4 or 6
mol-kg solution of sodium chloride.
Figure 2 shows the influence of different salts (ammonium sul-
fate, sodium chloride and sodium sulfate) on the solubility of carbon
dioxide. At constant ionic strength (about 6 mol-kg) and constant tem-
perature, ammonium sulfate causes a much smaller salting out effect
than both sodium sulfate and sodium chloride. For example at a carbon
dioxide molality of about 0.2 mol-kg, the total pressure above a solution
436 Rumpf, Nicolaisen, Ocal, and Maurer
10

?
A
/ / o/

/o
/
/ /

02 0.~ 06 08
rnco l ( m 0 l / k g )

Fig. 2. Solubility of carbon dioxide in aqueous electrolyte solutions at 40~ and con-
stant ionic strength: o, 2 mol-kg solution of ammonium sulfate (Ref. 3); n, 2 mol-kg
solution of sodium sulfate (Ref. 3); A, 6 mol-kg solution of sodium chloride (this work);
and , calculated results using Pitzer equations.
containing 2 moles of ammonium sulfate is about 1.6 MPa whereas it is
about 2.6 MPa above a 6 mol-kg sodium chloride and 3.1 MPa above a
2 mol-kg sodium sulfate solution.

5. Fitting Equation
The model to correlate and predict gas solubilities in aqueous
electrolyte solutions was described before in detail, ~ therefore only the
essentials are repeated here. The phase equilibrium condition for water
and dissolved carbon dioxide yields

PYw*w = pw~pwexp~, R7 )aw (1)

oo $

pyco2~)CO 2 =
m, , x (p
H~%w(T,pw)e vco2w(p
-if7:-pw) ~o2 ~nco2 (2)

m) denotes Henry's constant for the solubility of carbon

where H~o2.w
dioxide in pure water on the molality scale. We assume the salts to be
Carbon Dioxide in Aqueous Solutions 437

Table HI. Interaction Parameters for Pitzer's Equations

for the System CO2-NaC1-H20 a

Parameter a b c d

B(0)
CO2,NaC1 !0.254x -76.82 -10656 6312x103

FCO2,NaC1,NaC1 -0.0028

af ( T ) = a + b / T + c f f 2 + d]T 3.

fully ionized. At the concentrations of the dissolved gas considered

here, the dissociation of carbon dioxide in the liquid phase can be
neglected. For the calculations described below, all model parameters
,w, Pw, Vw, Vcoz,w, Oi) were calculated as described by Rumpf and
Maurerfl )
5.1. Pitzer Equations
The activity coefficient of a nonreacting gas G dissolved in an
aqueous solution of a fully dissociated strong electrolyte Cv+Av- (salt
molality ~ ) according to Pitzer's ~ equations is (a)

ln ,; = = o) +

q- 3 ~ F G , c A , C A -t- 6~/SfflGFG,G,CA (3)

where y~(m)(~ = 0) denotes the activity coefficient of the gas dissolved

in pure water
In 7d
* (m)(ms=)
- 0 = 2~;13{~?a+ 3~'I~'~G,G,G (4)
The parameters B~)cA, FG,CA,CA and Fa,G,CA are combinations of the
second and third virial coefficients I]f~ "q,i~ as used in Pitzer's original
equations

FG,CA,CA ----" V2"~G,C,C + 2V§ + V 2- + "I~G,A,A (6)

FG,G,CA = V+'I;G,6,C + V-~CG,G,A (7)

As can be seen from these equations, only certain combinations of inter-

action parameters can be determined from the experimental results.
To correlate the new results for the solubility of carbon dioxide in
438 Rumpf, Nicolaisen, I)cal, and Maurer

Table IV. Interaction Parameters for the Chen and Evans

Model for the Systems NaC1-H20 and CO2-NaCI-H20"

Parameter Aid Bid Cij

~W,NaQ 7.549 402.2044 -2.954

XNaCI,w -4.371 -53.667 3.042
~CO~NaO 5.563 174.333 22.308

ZNaCLCO2 -9.944 3009.45 27.445

a ~ij = Aij + BiJT + CIj(TR/T- 1 + InTffR).

aqueous solutions of sodium chloride, interaction parameters for the bi-

nary subsystem sodium chloride-water were calulated from the equa-
tions given by Silvester and Pitzer ts) (see also Rumpf and Maurert3)).
For CO2-H20, interaction parameters [3~!0 and xo,6,0 were set to
zerofl ) As the solubility of carbon dioxide is rather small even at high
pressures, the ternary parameter for interactions including two dissolved
carbon dioxide molecules, F6,0,CA was also set to zero. Correlating the
new results therefore only requires one binary (B~!cA) and one ternary
(FG,cA,CA) parameter. Isothermal fits showed that the temperature
dependence of the binary parameter had to be taken into account. It was
approximated by
bco2~cl CCOzN~Cl dco2~cl
B~2,Naca = aco2~cl + ~ + T-------T - -~ T3 (8)

The influence of temperature on the temary parameter could be

neglected. Parameters aco~yaca to dco2~acl as well as Fco2~,ca~raca were
fitted simultaneously to the new results for the total pressure. The
results are given in Table III.
The correlation describes the new measurements (total pressures)
with an average relative deviation of 1.9%, the maximum relative devia-
tion is 6.5% at 393.12 K and hlcoz = 0.305 mol-kg atp = 9.14 MPa.
5.2. Chen and Evans Model
Another semiempirical model to describe thermodynamic
properties of electrolyte systems is that developed by Chen and Evans (2)
which is widely used in commercially available process simulators. By
applying the local composition concept(9) together with several further
Carbon Dioxide in Aqueous Solutions 439

Table V. Comparison of Experimental Mean Ionic Activity

Coefficients for NaC1 at 25~ with Calculated Results a

7+ ~/• A~,+ 7• AT+

mNaCl (exp.) (calc. Pitzer) (Pitzer) (calc. Chen) (Chen)

0.001 0.965 0.965 0.0 0.965 0.0

0.005 0.928 0.927 0.1 0.927 0.1
0.010 0.903 0.902 0.1 0.902 0.1
0.050 0.822 0.820 0.2 0,819 0.4
0.100 0.779 0.777 0.3 0.773 0.8
0.500 0,681 0.680 0.2 0.668 1.9
1.000 0.657 0.656 0.2 0.646 1.7
3.500 0.746 0.745 0.1 0.760 -1.9
6.000 0.986 0.988 -0,2 0.943 4.3
6.144 1.004 1.007 -0,3 0.954 5.0

a Ref. 11; on molality scale.

Table VI. Comparison of Literature Data with Correlation

(Pitzer Equations) of Data from the Present Work

Ref. N Tmin Tmax mCO2.rnax mNaCa,rnax IA/Sco21rel IA/~lrel

(K) (K) (mol-kg) (mol-kg) (%) (%)

12 18 273.35 313.15 0.08 4.0 2.44

13 30 288.15 308.15 0.04 5.46 1.46
14 16 298.15 348.15 1.20 4.46 1.70
15 12 445.15 472.15 0.50 2.0 5.67
16 34 353,15 473.15 0.87 0,17 11.61

assumptions, Chen and Evans (2) derived an expression for the excess
Gibbs energy of a multicomponent electrolyte system. The resulting
equations for the excess Gibbs energy and for the activity coefficients
are summarized in the Appendix. For the present case of a system con-
taining a nonreacting gas (G) and a strong electrolyte (CA) dissolved in
water (w), the model requires interaction parameters Zw,CA,ZCn,w,~G.w(=
"Cw,a), ZG,CAand XCA,~. Parameters Xw,CAand "CCA,wCan be obtained from
the binary subsystem water-salt CA whereas XG,w can be obtained by
correlating gas solubility data in the binary subsystem gas-water. Thus,
there are two remaining parameters "%,CAand "CcA,~tO describe the effect
of the salt on the solubility of the gas.
To correlate the new data with the Chen and Evans model,
440 Rumpf, Nicolaisen, 0cal, and Maurer

parameters "Cw~acland XNaCl,wwere obtained from Chen(m (cf. Table IV).

In Table V, mean ionic activity coefficients of NaCI at 298.15 K cal-
culated with the Chen model and with the Pitzer equations are compared
to experimental data as taken from Hamer and Wu. m) As can be seen
from that comparison, the Pitzer equations with parameters as reported
by Silvester and Pitzer (s) result in numbers for 7• which are only
slightly, but systematically smaller than the numbers reported by Hamer
and Wu. (11) The relatwe deviations do not exceed 0.3% corresponding
to an absolute deviation in 7• of 3• 3. The Chen and Evans (2) model
with parameters as taken from Chen (~~ also results in numbers which
are systematically smaller than the experimental data, but the deviations
increase with increasing molality of sodium chloride. The maximum
relative deviation is 5% corresponding to an absolute deviation of
5• "2.
Following Chen and Evans, a) parameters Xco~w and "Cw,c02were
set to zero. The remaining parameters "Cc02~cn and XN~Cl,CO2were ob-
tained by fitting to the new data for the solubility of carbon dioxide in
aqueous solutions of sodium chloride. The influence of temperature on
interaction parameters was approximated by

"C = A +-~
B + C (~ -l+ln T
TRR) (9)

where TR = 298.15 K (cf. Chen and Evansa)). The resulting coefficients

are given in Table IV. Note that the fitted parameters are strongly inter-
correlated which is typical for local composition models. This set of
parameters correlates the new experimental results for the total pressure
with an average relative deviation of 1.6%, the maximum relative devia-
tion is 4.8% at 433.05 K and Ynco2= 0.274 mol-kg a t p = 8.98 MPa; i.e.,
the agreement is slightly better than for Pitzer's model. However, the
number of adjusted parameters for the model of Chen and Evans is
higher than that for Pitzer's model (6 vs. 5).

6. C o m p a r i s o n w i t h t h e L i t e r a t u r e D a t a
Literature data on the solubility of carbon dioxide in aqueous
solutions of sodium chloride is scarce. Most of the available data were
measured at temperatures below or above those investigated here. Table
VI gives an overview over the available data and the relative deviations
to the (Pitzer) correlation of the present work.
Markham and Kobe (m reported 18 data points in the temperature
range from 273.35 to 313.15 K at a partial pressure of carbon dioxide of
Carbon Dioxide in Aqueous Solutions 441

I~ 0.0~' o

\o

0.02" "0 '

OO : T='~05"15K

~nac~l(m0tlkg)

Fig. 3. Solubility of carbon dioxide in aqueous solutions of sodium c~odde: o, e•

pm-Jm6mm]results (Ref. 13) (p = 0.1013 MPa); and ~ , calculated results using Pitzer
equations(ffdswork).
0.1 MPa. These data points agree well with the correlation of the
present work. The average relative deviation in the partial pressure of
carbon dioxide is 2.4%, the maximum relative deviation is 8%.
u and Yoshida(~) reported 30 data points at temperatures
from 288.15 K to 308.15 K at a total pressure of 0.1 Ms Their data
are compared to the correlation of the present work in Fig. 3. The
average relative deviation in the total pressure is 1.5%, the maximum
relative deviation is 5.2%,
Malinin and Savelyevac~4)reported 16 data points at a partial pres-
sure of carbon dioxide of 4.8 MPa. An excellent agreement is observed
between these data and the correlation. The average relative deviation is
1,7% (IApl~x = 3%) (cf. Fig. 4).
The results for the partial pressure of carbon dioxide in the tem-
perature range from 445.15 to 472.15 K as reported by Ellis and
Golding c~s) show larger deviations from the correlation of the present
work. The average relative deviation is 5.7%, corresponding to absolute
deviations up to 0.6 MPa. However, it should be noted that model
parameters had to be extrapolated in temperature.
In a recent publication, Nighswander et al. (xs) reported 34 data
442 Rumpf, Nicolaisen, 0cal, and M a u r e r

~ 1.2

~2
lie 1.0
\ "0

0.6"

0.~

I
0.2

3
@~[tl[mot/kg)

Fig. 4. Solubility of carbon dioxide in aqueous solutions of sodium chloride: o, ex-

perimental results (Ref. 14) (Pco2 = 4.797 MPa); and - - , calculated results using Pit-
zer equations (this work).
points for the total pressure in the temperature range from 353.15 K to
473.15 K. Their data points agree only fairly with the correlation of the
present work. The average relative deviation is 11.6% corresponding to
absolute deviations up to 2.68 MPa. Neglecting data at temperatures
above 433.15 K, the mean average and absolute deviation is 12.3% and
2.68 MPa, respectively.

7. C o n c l u s i o n s
The solubility of carbon dioxide in aqueous solutions of sodium
chloride was determined by a synthetic method in a wide range of tem-
perature and composition. Interaction parameters for Pitzer's (~) equa-
tions and the Ctten and Evans (2) model were determined from the new
measurements. Both theories are able to correlate the new data rel~ably.
A comparison with lilerature data shows a good to fair agreement in
most cases.
Carbon Dioxide in Aqueous Solutions 443

Acknowledgment
Financial support of this investigation by the Bundesminister ftir
Forschung und Technologie of the Federal Republic of Germany
(BMFT-Contract No. 0312-4003-0326558C), BASF AG, Ludwig-
shafen, Bayer AG, Leverkusen, Linde KCA, Dresden, Lurgi AG,
Frankfurt, Hoechst AG, Frankfurt and Degussa AG, Hanau is gratefully
acknowledged.

References
1. K. S. Pitzer, J. Phys. Chem. 77, 268 (1973).
2. C. C. Chen and L. B. Evans, AIChE J. 32, 444 (1986).
3. B. Rttrnpf and G. Maurer, Ber. Bunsenges. Phys. Chem. 97, 85 (1993).
4. E. Bender, 5th Proc. Symp. Thermophys. Prop. (ASME, New York, (1970) p. 227.
5. E. W. Washburn, International Critical Tables of Numerical Data, Physics,
Chemistry and Technology, Vol. HI (McGraw Hill, New York, 1928).
6. A. Zawisza andB. Malesinska, J. Chem. Eng. Data 26, 388 (1981).
7. Y. D. Zelvenski, Zhur. Chem. Prom. 14, 1250 (1973).
8. L. F. Silvester and K. S. Pitzer, J. Phys. Chem. 81, 1822 (1977).
9. H. Renon and J. M. Prausnitz, AIChE J. 14, 135 (1968).
10. C. C. Chert, FluidPhase Equilibria 27, 457 (1986).
11. W. J. Hamer and Y.-C. Wu, J. Phys. Chem. Ref. Data 1, 1047 (1972).
12. A. E. Markham and K. A. Kobe, J. Am. Chem. Soc. 63,449 (1941).
13. Y. Yasunishi and F. Yoshida, J. Chem. Eng. Data 24, 11 (1979).
14. S. D. Malinin and N. I. Savelyeva, Geokhimiya 6, 643 (1972).
15. A. J. Ellis and R. M. Goldlng, Am. J. Sci. 261, 47 (1963).
16. J. A. Nighswander, N. Kalogerakis, and A. K. Mehrotra, J. Chem. Eng. Data 34,
355 (1989).
17. K. S. Pitzer, Activity Coefficients in Electrolyte Solutions, 2nd edn., (CRC Press,
Boca Raton, 1991).
18. C. C. Chen, H. I. Britt, J. F. Boston and L. B. Evans, AIChEJ. 28, 588 (1982).

Appendix
Brief Outline of the Chen and Evans Model
By applying the so-called local composition concept to a mtfl-
ticomponent electrolyte solution containing cations c, anions a and
neutral species m, Chen and Evans (z) derived the following expression
for the excess Gibbs energy per mole of mixture

gE,Ic---- Xm ZjXjGj'm'l~J'm
4-c~ Xca
~ Xa' Y'jXjGjc'a'c'l~jc'a'c
RT Z XkXkGk,ra ]~a,,Xa,----'~ ZkXkGkc a,c
444 Rumpf, Nicolaisen,Ocal, arid Maurer

Xc' ~jXjGja'c'a27ja'c'a (lO)

+~ Xa~ X~"X~-----=ZkXk~,ea
where Xj = xjlzjl (11)
for ions and Xm = Xm (12)
for neutral species. By differentiating this equation with respect to the
number of moles of component i (i = c or a or m), Chen and Evans (2)
derived the following expressions for the local composition contribution
to the activity coefficient of molecule m, cation c and anion a

ln@mC,(x)-- Zj~jGjmTjm Xm'G~' / ,_ ~:kXk~'~km?~

EkXkGkm +E J
Xa, XcGmc,a'c / 2k~,a'c'~
* ~c a~ ~a'--~a"~.kXkakc,a,ct"Cmc'a'r ~kXkGkc,ac J
Xc' XaGma,c'a / ZkSkaka,c'a'l;ka,c'a.'~
+ a~ c~~ ~C"~c,,,~,kXka..ka,c,,at'l~ma,c'a-- Y,kXk~rka,c,a j (13)
1 lnT~,(x)=~ Xa' ZkXkGkc,a'cqikc,a'c_l_~m XmGcm
] Zc ] Za"Xa" ZkXkGkc,a'c EkXkfrkm
~,kXkGkmTkm~ Xc' XaGca,e'a
(T, crn ~_,kXkGkra ] + a~ c~ ~c,--~c,,~kXkGka, a
EkXkGka,c'a'Cka,c'a"~
'l;ca,e'a - ~ ,] (14)

1 ln~ac'(x) = x--~c2
~ Xe" ~kXkakc,a'cTkc,c'a m~ XmGam
[ Za [ ~-'c"Xc" 2kXkakc,a"c + ~kXkGkm
EkXkGkm'~km'~ Xa" XcGac,a'c

EkXkGkc,a'cTkc,a'e'~
( %c,a'c - ~ ,] (15)

The following equations were given for calculation of the model

parameters Gji,ki
Carbon Dioxide in Aqueous Solutions 445

aji.ki = exp (-~ji&(qi&-0 (16)

where gj,i -- gk,i (17)

'l~ji,ki -- RT

and cqi,ki is the so-called nonrandomness factor. Further equations were

given to calculate the model parameters Tma,ca and '~mc,ac
Xrna,ca = Tam --Tca,m+ Tin,ca (18)

%c,ac = z ~ - z o . ~ + % , c . (19)
where %m and "Camcan be computed from
XaXaGca,m
Gc~. = exp(-0~m%m) - (20)
:Ca'Xa'
~,~X~G~a,m
and Gash = exp (--team%n) -- (21)
XcXr
These equations yield some inconsistency between Eq. (10) and Eqs.
(13 - 15) as it was assumed that model parameters Gcm and Gain are in-
dependent of concentration when differentiating Eq. (10). According to
Chen and Evans, all nonrandomness factors are fixed to 0.2. Adopting
the above rules for calculating the model parameters "c~ and Tam, the
Chen and Evans model then requires two parameters (Tm,~aand "c~.0n) for
a system containing a single salt dissolved in a single solvent.
To account for long range interactions, Chen and Evans adopted a
modified Debye-Htickel contribution
,,1000-,1/2 r2~ .2rlt2 -- ,~
~ilX ~ lI3t2
x ]

where Ix is the ionic strength on mole fraction scale

1
Ix = g 2 xiz] (23)
l

and 13 = 14.9. 4 A 0 is the D e b y - H t i c k e l parameter for the o s m o t i c c o e f -

ficient. Following Chen et al.,(18)the temperature dependence of A, was
calculated from the equation

4Applying Eq. (22) to a dissolved molecular species (zi = 0) (as proposed by Chen and
Evans) is questionable. For a detailed derivation of Eq. (22) see for example Ref. 16.
446 R u m p f , Nicolaisen, 0cal, and M a u r e r

Ar -61.44534exp(~o - 1 ) + 2.864468 [ e x p ( ~ o - 1 ) ] 2

+ 183.5379 In T - 0.6820223(T- To)

To

+ 7.875695x10_4(T2 Tg) + 58.95788 To

- (24)
T
where To = 273.15 K. Equations (14, 15) give the local composition
contribution to the activity coefficient of a salt ca on a symmetric scale
whereas Eq. (22) is normalized to infinite dilution. To change Eq. (14)
to the unsymmetric scale, the following equation has to be applied

v; = 1,1 x - in (25)

where lny~ ,(x)is lny~ '(x) = -1 zr [(Xwc,ac+ Gcw%w) (26)

Therefore, the final expression for calculating the activity coefficient of
cation c is
ln@ Cx) = lnyeJ)'u'(x) + ln~lee'(x)- I zc [(%c,ac + GewZcw) (27)
A similar equation holds for anions a. For neutral species (m) dissolved
in water (w), the unsymmetric activity coefficient is found to be
l n ~ (x) = lny*m
~'rI'(x) + lnTtx~'(x) - Xwm+ Grnw'Crnw) (28)
Finally, the activity coefficients have to be transformed to the molality
scale. For a dissolved species i, the resulting expression is
l n ' ~ "(m) = l n ~ ; '(x) + lnxw (29)

Symbols
A,B,C parameters in the model of Ref. 2
Ar Debye-Htickel parameter for the osmotic coefficient
aco2,NaC1 9 . .
dco2N~Cl coefficients for the temperature dependence of inter-
action parameter B~2,NaC 1
B(O) binary parameter in Pitzer's equations
ai activity of component i
g~, Gji,ki parameters in the model of Ref. 2
excess Gibbs energy per mole of mixture
Carbon Dioxide in Aqueous Solutions 447

interaction energy in the model of Ref. 2

Henry's constant for the solubility of carbon dioxide in
water (on molality scale)
/x ionic strength on mole fraction scale
Mw molar mass of water
mi true molality of component i
overall molality of component i
N number of data points
P total pressure
Pi partial pressure of component i
R universal gas constant
T temperature
reference temperature in the model of Ref. 2 (= 298.15 K)
To reference temperature in the expression for the
Debye-Htickel parameter (= 273.15 K)
V partial molar volume
xj modified mole fraction of component j, Xj =
xjl zj I for ions, Xj = xj for neutral species
Yi mole fraction of component i in vapor phase
zi number of charges of component i
Greek Letters
0~ji,ld nonrandomness factor in the model of Ref. 2
binary interaction parameters in Pitzer's equations
Yt activity coefficient of species i
Y~ mean ionic activity coefficient on molality scale
P parameter in modified Debye-Htickel expression
temary interaction parameter in Pitzer's equations
xji,va, xji interaction parameters of the model of Ref. 2
V+, V_ number of cations and anions in salt CA
fugacity coefficient of component i
Subscripts
a, a', a" anion
abs absolute
C, C', C" cation
cal calculated
C, a, m cation c, anion a, molecule m
exp experimental
G gas
448 Rumpf, Nicolaisen, 0cal, and M a u r e r

i,j,k component i, j, k
max maximum
min minimum
rel relative
S salt
w water
X on mole fraction scale
m on molality scale
Superscripts
D-H Debye-Htickel
lc local composition
(m) on molality scale
S saturation
(x) on mole fraction scale
normalized to infinite dilution
infinite dilution
I
liquid phase
tt
vapor phase
Definition of relative deviation of property x
Xk,cal -- Xk,exp
ZXX-
Xk,exp

Definition of average absolute deviation of property x

1 N
Iz~Clabs = ~ r ~ 1 IXk,cal--Xk,expl

Definition of average relative deviation of property x

1 N
I lre = IXk"
--Xk'explzk,exp