.

, Dynamics
cycle-Part
( 1967).

of a Tubular Reactor with Re11. Nature of the Transient State, ibid., 13, 519

Robertson, D C., and A. J. Engel, Particle Separation by Controlled Cycling, Ind. Eng. Chem. Process Design Develop.,
6,2 (1967).

Robinson, R. C., and A. J. Engel, An Analysis of Controlled Cycling Mass Transfer Operations, lnd. Eng. Chem., 59, 22
(1967).

Root, R. B., An Experimental Study of Steady State Multiplicity in a Loop Reactor, Ph.D. thesis, Univ. Illinois, Urbana

., and R. A. Schmitz, An Experimental Study of Steady State Multiplicity in a Loop Reactor, AlChE J., 15, 607
(1969).

(1968).

Scale Study o Controlled CycIic Distillation, Chem. Eng. f Sci., 22,766 ( 1967). Schmeal, W. R., and N. R. Amundson, The Effect of Recycle on a Linear Reactor. AlChE J., 12, 1202 (1966). Silverman, R. P., Theoretical and Experimental Investigation of a Staged, Controlled-Cycling, Stirred Tank Reactor System, M.S. thesis, Penn. State Univ., University Park (1969). Vejtasa, S. A., Ah Experimental Study of Steady State Multiplicity in an Adiabatic Stirred Reactor, Ph.D. thesis, Univ. Illinois, Urbana ( 1969). -, . R. A. Schmitz, An Experimental Study o Steady f State Multiplicity in a Adiabatic Stirred Reactor, AIChE J., 16,410 (1970).
Manuscript received June 14, 1973; revision received and accepted November 5, 1973.

Schrodt, V. W., J. T. Sommerfeld, and 0. R. Martin, Plant

An Empirical Correlation of Second Vi riaI Coefficients

A new correlation of second virial coefficients of both polar and nonpolar systems is presented. It uses the Pitzer-Curl correlation for nonpolar compounds, but in a modified form. The second virial coefficient of nonhydrogen bonding compounds (ketones, acetaldehyde, acetonitrile, ethers) and weakly hydrogen bonding compounds (phenol) is fitted satisfactorily with only one additional parameter per compound, which is shown to be a strong function of the reduced dipole moment. Two parameters are needed for hydrogen bonding compounds (alcohols, water), but for alcohols, one parameter has been kept constant and the other expressed as a function of the reduced dipole moment. The extension of the correlation to mixtures is satisfactory, direct, and involves only one coefficient per binary.

CONSTANTINE TSONOPOULOS
Esso Reseach and Engineering Company
Florham Park, New Jersey 07932

Vapor-phase nonideality should properly be taken into account in the analysis and correlation of vapor-liquid equilibria, especially at superatmospheric pressures. For / reduced densities up to 14 (roughly speaking, up to pressures of 15 atmospheres), the effect of vapor-phase imperfection can be reliably calculated using only the second virial coefficient B of the pure components and crosscoefficients for each binary (for example, see Prausnitz, 1969). A very successful correlation of the B of nonpolar gases has been proposed by Pitzer and Curl (1957); see Equation (6). The soundness and reliability of the Pitzer-Curl correlation make its use as the basis of a correlation of the B of polar systems highly desirable. The objective of this study was to extend the Pitzer-Curl correlation to the following polar compounds of interest: ketones, aldehydes, acetonitrile, ethers, alcohols, phenol, and water. A similar approach has been taken by OConnell and Prausnitz (1967) and by Halm and Stiel (1971), but both

of these correlations are lacking in several respects, including reliability; they are fully discussed in the Appendix. The OConnell-Prausnitz correlation has been used widely (for example, in two important monographs on computer calculations for vapor-liquid equilibria; Prausnitz et al., 1967; Renon et al., 1971); it is therefore used as the basis of comparison with the new correlation. Another correlation of the B of polar system which, while not using the Pitzer-Curl correlation, is based on a similar idea to ours is that proposed by Black (1958); it is discussed in the Appendix. Other correlations that are not based on the Pitzer-Curl correlation and are not considered here are those of Kreglewski (1969), Nothnagel et al. (1973), OConnell(1971), and of Polak and Lu (1972).
The Appendix has been deposited as Document No. 02293 with the National Auxiliary Publications Service (NAPS), c/o Microfiche Publications, 305 E. 46 St., New York, N. Y. 10017 and may be obtained for $2.00 for microfiche or $5.00 for photocopies.

AlChE Journal (Vol. 20, No. 2)

March, 1974

Page 263

CONCLUSIONS A N D SIGNIFICANCE
A modified form of the Pitzer-Curl correlation is used for nonpolar compounds. The modification mainly improves the fit at low reduced temperatures (TR < 0.75) but is also shown to have some advantage at high reduced temperatures. The actual acentric factor is used for all polar compounds, thus obviating the need of determining and using homomorphs (for example, ethane is methanol's homomorph). The second virial coefficient of the nonhydrogen bonding compounds investigated (ketones, acetaldehyde, acetonitrile, ethers), and even of the weakly hydrogen bonding phenol, is satisfactorily fitted with only one additional parameter per compound. A strong correlation is shown to exist between this parameter and the compound's reduced dipole moment pR. Hydrogen bonding compounds (alcohols, water) require two parameters for the polar term, but all alcohol data are fitted with one parameter fixed; the other has a weak dependence on p R .
The polar contribution to B can be readily modified when new or better data become available-without invalidating the model. The form of the polar term is such that its value drops to zero at high reduced temperatures in accordance with theoretical and experimental evidence. The new correlation is especially superior to the other correlations in its extension to mixtures. A characteristic binary constant is introduced, as it has already been done for nonpolar/nonpolar mixtures (Chueh and Prausnitz, 1967; Hiza and Duncan, 1970), to account for the deviation of the characteristic critical temperature T,ij from the geometric mean; see Equation (18). The principal significance of this work is that the new B correlation makes possible the reliable calculation of fugacity coefficients and therefore improves the fit and extends the useful pressure range of vapor-liquid equilibrium data obtained at pressures of, roughly, 15 atmospheres or below. This application is the most sensitive as far as the second virial cross-coefficients are concerned.

THE VlRlAL EQUATION OF STATE

coefficient BM is given exactly by

n n

The virial equation is a power series in the reciprocal molar volume, l / u

(4)
Bij ( i # j ) is the second virial cross-coefficient; it is a function only of temperature like the pure-component coefficients (Bii and Bjj) . Equations (2) and ( 4 ) can be used to calculate vaporphase densities, fugacities, etc. if we know (1) the temperature dependence of the second virial coefficients and (2) how Bij is related to Bii and Bjj.
NONPOLAR GASES

z=-

Pu B C =1+-+-+ RT 2, 02

...

where z is the compressibility factor. B is the second virial coefficient, and C is the third virial coefficient. For pure components, these coefficients are functions only of temperature. Equation ( 1 ) predicts z accurately up to one half the critical density. The third virial coefficient of polar gases, however, is rarely known. When the equation is truncated after the second virial coefficient term, it is reliable up to one fourth of the critical density, or roughly up to 15 atm. For most chemicals applications, the pressure is under 15 atm., and hence the second virial coefficient alone can account for vapor phase nonidealities. Thus, the equation of state that will be used is

z=l+-

B
v

Nonpolar gases, for example, inert gases and kydrocarbons, are characterized by a zero dipole moment. In addition to critical temperature T, and critical pressure P c , Pitzer and Curl (1957) had to use a third parameter for the correlation of B for nonpolar gases. This parameter is the acentric factor o defined as (Pitzer et al., 1955)

For the evaluation of B from low-density P-V-Tdata, however, Equation ( 1 ) should be used, that is,

limit [ u ( z - l ) ]
l/V+O

(3)

(If the limit to zero density is not taken, then the apparent B calculated will have a density dependence, as it will incorporate the contribution of C, etc.) A great advantage of the virial equation is that it can be extended to mixtures without arbitrary assumptions. For a mixture of n components, the mixture second virial

Fluids with spherical molecules, for example, Ar, Kr, have a zero acentric factor; the quantum gases, for example, He, HB, have negative 0's; and everything else has a positive O . The Pitzer-Curl correlation is the best available for second virial coefficients of nonpolar gases. In reduced form, it is (Pitzer and Curl, 1957)

where

* This

form is superior to the pressure-explicit form

z= 1

+ BP/RT
>
1.

(Pa)

f :

(TR) = 0.1445 - 0 . 3 3 0 / T ~

for subcritical temperatures, but the reverse is true at TR

Page 264

March, 1974

AlChE Journal (Vol. 20, No. 2)

-100 /I

-A
T ,
= 150 65K)

in Figure 2 have also been reported; see Dymond and Smith ( 1969) .] At high reduced temperature ( T R > 1.5), f ( l ) becomes progressively more negative than f p C . Data are available for ethane ( o = 0.105) up to TR = 1.67 (Reamer et al., 1944). The results from the two correlations, which differ by 3 cc/gmol, bracket the experimental value. A better example is offered by carbon dioxide ( w = 0.225), whose quadrupole moment has an entirely negligible effect at supercritical temperatures. B data are available up to T R = 2.87 (MacCormack and Schneider, 1950; Vukalovich and Masalov, 1966) and the modified correlation is clearly superior to the Pitzer-Curl correlation, as shown in Figure 3.
(1)

J
75
100

12 5
T.
OK

150

175

Fig. 1. Second virial coefficient of gases with zero acentric factor.

POLAR GASES

f:

( T R ) = 0.073

+ 0.46/T~
(8)
(0)

- 0 . 5 0 / T ~ ~ 0 . 0 9 7 / T ~ ~ 0.0073/T~' -

Pitzer and Curl determined the coefficients of f p C by fitting the U data for Ar, Kr, and Xe, which have zero acentric factor. The tempei'ature range of the data used, however, was limited. Recent measurements on Ar and Kr (Brewer, 1967; Byrne et al., 1968; Crain and Sonntag, 1966; Weir et al., 1967) down to T R = 0.53 have shown that fpc is unsatisfactory for T R the error is 10 percent.
(0)

Polar compounds are characterized by a nonzero dipole moment p which expresses the effect of electrostatic forces between molecules. Polar compounds considered in this study are alcohols, phenols, water, ketones, aldehydes, ethers, and alkyl nitriles. The first three are hydrogen bonding compounds ( a bond forms between the hydrogen atom attached to oxygen in one molecule with the oxygen atom of another molecule); this makes their behavior in the vapor different and more complex than that of nonhydrogen bonding compounds (ketones, etc.) . These two broad classes of compounds will be treated separately.

< 0.75;

at T R = 0.53,
225
250 275

300

A slight modification to makes it possible to fit all these data to within 1% ; the modified f 0 ) is (9) The improvement in the fit is clearly shown in Figure 1, where the most recent B data for Ar and Kr have been
~ plotted along with the calculated values. f"' and f p are virtually identical for TR > 0.8 (120K for Ar; 168K for Kr)

f'''( T R ) = jpc' ( TR)- O.O00607/T~*

5
(D

-1000

(0)

/ijJ
9" .
/'

-- --

'Ullod f8ed P ~ l z e r - C t : r l Corlclal8 > PlLrei-Ctlrl curreia:,o

can be checked against reliable B data for compounds with large acentric factors. Two such compounds are n-butane ( O = 0.200) and n-octane (o = 0.398). Dymond and Smith (1969) have smoothed the data for n-butane, while for n-octane we used the data of McGlashan and Potter (1962) and Connolly and Kandalic
( 1960). The comparison showed that f: predicted results that were too positive at low temperatures (6% error at TR = 0.55), and too negative at 0.8 6 T R 4 1.0 ( 2 to 4% errors). These errors are rather small, but they indicate the need for some adjustment. The modified f"'

fp:

1500

25

50

75

100

125
C

150

175

200

2?

Fig. 2. Second virial coefficient of benzene.

+20

- --

has one term less than f p c :

(1)

f"' (TR) 0.0637 =

+ 0.331/T~~
- 0.423/TR3 - 0.008/T~'
(10)
0
Vukalavlch & Masalov

The differences between the Pitzer-Curl correlation and our modification are illustrated in Figure 2 for benzene (o = 0.211). The spread in the experimental data (Baxendale and Enustun, 1951; Bottomley and Spurling, 1966; Connolly and Kandalic, 1960; Hajjar et al., 1969; Knoebel and Edmister, 1968; Zaalishvili and Belousova, 1964) is larger than the difference between the two correlations. [More negative and more positive data than those plotted AlChE Journal (Vol. 20, No. 2)

-", ,

-'Yo0

v
I

D TR = 2 . 0
1 1

(1966) MacCormack . Schneider 8 (19501

I
t.

300

400

500

600

Fig. 3. Second virial coefficient of carbon dioxide a t high temperatures.

March, 1974

Page 265

Nonhydrogen Bonding Compounds

T h e second virial coefficients of ketones, acetaldehyde, acetonitrile, and ethers were fitted very satisfactorily by adding a one-parameter term to the modified Pitzer-Curl correlation:
-- - f c o ) ( T R ) BPc

RTc

ofc1(T R ) )

+ f"'

(T R )

(11)

where the polar term is given by

f'2'(TR)

are plotted in Figure 4. [The data of Anderson et al. ( 1 9 6 8 ) , Kappallo e t al. ( 1 9 6 3 ) , and Zaalishvili et al. (1960; 1964) differ from the other data in their temperature dependence and were not used in determining the value of a ] T h e new correlation gives a good fit. For comparison, we have also plotted the results from the O'Connell-Prausnitz correlation ( 1967), which is clearly poor at low and high temperatures. T h e discontinuity at 210C ( T R = 0.95) is inherent in the correlation: at T R

=TR'

The second virial coefficient of polar compounds is more

negative than that of nonpolar compounds (for the same TR, 0 ) ; hence, f'2' < 0, or a < 0. The contribution of fC2) to B diminishes rapidly with increasing temperature, in agreement with observed behavior. For a = - 0.01, f C 2 ) = - 0.64 a t TR = 0.50, but at T R = 1.5 it is only f'2' = - 0.0009. The results of fitting Equations (11) to (12) to B data for five ketones, acetaldehyde, acetonitrile, and three ethers are given in Table 1. T h e root mean square deviations of B are small, but the accuracy can be clearly illustrated only when B data for a given compound are available over a wide temperature range. A sensitive check of the new correlation is provided by the extensive data for acetone reported since 1960, which

C l m l q er ril , 1 9 6 5 )
~

-3000

A
b

i -3 6 0 0 0 -.-.__.__--___A -L - - i - ~ 50 100 150 200 2 50

', "C

Aiidersoi el a1 '19681 Z n a l ~ s l i v i l b BeIo.,sova (1964, i I K a w a I l o et al k1963: Z a a l # s , l i l l ! b K o I y S k 3 !1960,

Fig. 4. Second virial coefficient of acetone.

TABLE NOSHYDROGEK 1. BONDIXG POLAR CordPouPI'ns Root mean square deviation of B , cc/gmol (no. points) 32.3 (36)

Compound Acetone

PRa

Referencesb Abbott ( 1965); Bottomley and Spurling (1962; 1967); Chang et al. (1963); Hajjar et al. (1969); Hauthal and Sackmann (1969); Knoebe1 and Edmister ( 1968). Chang et al. (1965); Hauthal and Sackmann (1969); Hofmann and Pfestorf (1968); Nickerson et al. (1961). Nickerson et al. (1961). Chang et a1 (1965), Hales et al. (1967); Hauthal and Sackmann ( 1969). Hauthal and Sackinann (1969); Hofmann and Pfestorf ( 1968). Aleunder and Lambert ( 1941) ; Prausnitz and Carter ( 1960). Lainbert et a1 (1949); Prausnitz and Carter (1960). Cawood and Patterson (1933); Haworth and Sutton (1971); Kennedy et al. (1941); Osipiuk and Stryjek (1970); Tripp and Dnnlap (1962). Bottomley and Reeves ( 1958); Chevalier (1969); Knoebel and Edmister (1968); Lambert et al. (1949); Riitzsch and Bittrich (1965), Russell and Maass (1931); Stryiek and Kreglewski ( 1965); Zaalishvili and Kolysko (1960; 1962). Chevalier ( 1969).

0.31

147.0

-0.03090

Methyl ethyl ketone Methyl n-propyl ketone DiethyI ketone Methyl tert-butyl ketone Acetaldehyde Acetonitrile Methyl ether Ethyl ether

0.33 0.34 0.35 0.29 0.314 0.323 0.21 0.282

112.0 90.0 87.4 81.4 188 250.4 56.0 22.6

-0.02224 -0.02277 -0.01686 -0.01065

-0.0508 1 -0.1198f

70.4c (19) 8.4d ( 4 ) 66.3e (11) 25.5 (10)

33.8 ( 1 6 )

151f

(23)

-0.01513 -0.00202

19.2 ( 1 4 ) 50.39 ( 3 7 )

Isopropyl ether
#R

0.34

14.5

-0.00089

31.@

(4)

= 10VPc/Tc2;the units are Dehyes*-atm-'K-Z. Dipole moment data taken from McClellan (1963). Only those used in the regression analysis are listed. Most of the deviation is due to three widely scattered points at 41.5 to 60C. 'The 61.7"C point was discarded because i t is ahout 300 cc/gmol t o o negative. With it, the average deviation increases from 7.4 to 64.4 cc/ gmOl. e T h r e e of Chang's (1965)data are clearly inconsistent with all the rest; without them, the root mean square deviation drops to 29.7 cc/gmol for Q = -0.01873. f Including the 4 data points of Zaalishvili et al. (1971) makes a = -0.1211 and the root mean square deviation of B 156 cc/t?mol. PRernoving the questionable references identified in the text drops the root mean square deviation to 23.4 cc/gmol for a = -0.00246. &Root mean square deviation = 36.8 cc/gmol for o = 0.0 (i.e., no polar correction).

Page 266

March, 1974

AlChE Journal (Vol. 20, No. 2)

= 0.95 the polar corrections (analogous to f ( 2 ) ) are set equal to zero. More about this correlation can be found in the Appendix. Figure 5 is a plot of experimental and calculated B values for acetaldehyde and acetonitrile. The new correlation is again shown to be most satisfactory, while the O'Connell-Prausnitz correlation leads to serious errors. The new correlation is very satisfactory for methyl ether (maximum deviation is 42 cc/g-mol), but less so for ethyl ether. All available B data were used in determining the value of a; however, the results of Knoebel and Edmister (1968) and Ratzsch and Bittrich (1965) have too steep a temperature dependence, while those of Russell and Maass (1931) and Lambert e't al. (1949) are consistently too positive. There is good agreement (less than 30 cc/g-mol difference) between the new correlation and that of O'Connell and Prausnitz. Listed in Table 1 are values for the reduced dipole moment PR defined by

H y d r o g e n Banding Compounds

Molecules containing hydrogen linked to an electronegative atom (for example, 0, N ) tend to associate with each other by forming hydrogen bonds, that is, they dimerize. This is especially true of alcohols. Phenols associate in the liquid phase (Prausnitz, 1969), but their dimerization in the vapor ( u p to moderate densities) is much weaker than that of alcohols. Similarly, dimerization in water is limited (Tsonopoulos and Prausnitz, 1970; this reference treats the very strong dimerization of carboxylic acids). The effect of dimerization is to make more complex the temperature dependence of the polar contribution to the second virial coefficient of dimerizing compounds: it is steeper at very low T R (+ 0 5 ) , but flatter at TR > 0.8. The polar term of Equation (11) must now have two parameters :

O'Connell and Prausnitz (1967) introduced the factor lo5 to make the pR values greater than unity (when P is given in Debyes", P, in atmospheres, and T , in degrees Kelvin). Eubank and Smith (1962) have shown that p~ by itself is inadequate as a correlating parameter since the location and direction of the dipole moment are also important. It follows, therefore, that the polar contribution to B cannot be given for all polar compounds by a unique function of pR. Figure 6 clearly demonstrates the diminishing importance of a, and hence of f C 2 ) , with decreasing PR. However, the reduced dipole moment is a satisfactory correlating parameter only when each family of compounds is considered separately. In the case of ketones, a is best expressed as a linear function of p R :

Both a and b assume positive values. At low temperatures the last term predominates, while at high temperatures both terms become negligible. The B data for eight alcohols, phenol, and water were

t.

-2 7

10 -100 r

50

100
I

150
I

200 210 -100

, ,

a = - 0.00020483 ( p R );ketones (14) The predicted value of a for methyl isobutyl ketone is - 0.016, with which B at 120C is calculated to be - 1590 cc/g-mol; the experimental value is - 1580 cc/ g-mol (Hauthal and Sackmann, 1969). With a values from Equation (14), it was possible to predict satisfactorily the B data of Hales et al. (1967) for methyl isopropyl ketone (average error of 122 cc/g-mol or 6.6% ) and ethyl n-propyl ketone (average error of 19 cc/g-mol Rough but useful values for the a of CQf or 1.1%). aldehydes and alkyl nitriles can be estimated from Figure 6. There is an excellent linear relationship between In ( - a ) and In ( p R ) for ethers:
In (- a ) = - 12.63147

.; '

t.

"C

Fig. 5. Second viriol coefficient of acetaldehyde ond acetonitrile.

-0.14

a = - 2 . 1 4 0 ~ 0-4(~R)-4.308x10'21 1

-0.10

+ 2.09681 In ( p ~ )ethers ;
(15)
CO

0 Ketoiies
Acetaldehyde V Acetoliltrile D Ethers

The value of - a drops very rapidly with decreasing PR; the fit of the B data for isopropyl ether is nearly equally good when a = 0.0 (see Table 1).The predicted a values for methyl ethyl ether ( p R = 33.7) and ethyl n-propyl ether ( p R = 17.2) are, respectively, - 0.0052 and - 0.0013, but the limited available data (Osipiuk and Stryjek, 1970; Chevalier, 1969) lead to small positive values for a.

-0.05

100

2 00

300

REDUCED DIPOLE MOMENT, 1 Debye = 10-18 esu-cm. An extensive compilation of dipole moments is given by McClellan (1963).
Q

pR

Fig. 6. Dependence of o on reduced dipole moment for nonhydrogen bonding compounds.

AlChE Journal (Vol. 20, No. 2)

March, 1974

Page 267

TABLE HYDRWEN 2. BONDING POLAR COMPOUNDS

Compound Methanol

CLRa

( 1

Root mean square deviation of B, cc/gmol (no. points)

(22)

Reference8

0.572

87.9

0.0878

0.0560 28.2

Ethanol I-Propanol 2-Propanol (isopropanol) I-Butanol 2-Butanol (sec-Butanol) 2-Methyl-1-Propanol ( isobutanol) 2-Methyl-2-Propanol ( tert-Butanol) Phenol Water
1

0.635 0.625 0.666 0.590 0.578 0.586 0.614 0.443 0.344

68.3 51.2 52.6 39.7 41.6 40.8 44.2 24.6 172.6

0.0878 0.0878 0.0878 0.0878 0.0878 0.0878 0.0878 -0.0136 0.0279

0.0572 71.3c 0.0447 0.0537 0.0367 0.0487 0.0481 0.0508

15.1 (4) 19.2 (13) 61.2 14.6 23.2 33.1

Kell and McLaurin (1969); Knoebel and Edmister ( 1968); Kretschmer and Wiebe (1954); Kudchadker and Eubank ( 1970). ( 8 ) Knoebel and Edmister (1968); Kretschmer and Wiebe (1954).

0.0229

16.2

1.74

Cox (1961). Cox (1961); Kretschmer and Wiebe (1954); Moreland et al. (1967). ( 6 ) Cox (1961); Foz et al. (1954)d. ( 4 ) Cox ( 1961). ( 4 ) Cox (1961). ( 4 j Cox (1961). ( 8 ) Ope1 (1969). (22) Kell et a1 ( 1968); Vukalovich et a1 ( 1967).

See Table 1. All alcohols have p = 1.70 0.03 Debyes (McClellan, 1963). Only those used in the regression analysis are listed. c This large deviation is misleading; the experimental values at 60C are + 109 and value is removed, b = 0.056. d The 98'C point is inconsistent with the rest and has been discarded.
8

- 126.5 cc/gmol

off the calculated value. If the too negative

0.07 0.06-

I
Methanol

Ethanol
2-Propanol
tert-Butanol

P
1-Propanal

0 05-

0.04Dl

0.03-

b = 0.00908 +0.0006957

vR)

0.01
-2800

o'021
REDUCED DIPOLE MOMENT, /JR

-~
100
L

25

50

75

125

150

175

200

225

"C

Fig. 7. Second virial coefficient of methanol.

Fi.9. 8. Dependence of

b on reduced dipole moment for olcohols.

analyzed with Equations (11) and (16). The results are given in Table 2. The same value of a was used for all eight alcohols. A good example of the fit is illustrated in Figure 7 for methanol-for which we have the most extensive information. Only the most recent data (Kell and McLaurin, 1969; Knoebel and Edmister, 1968; Kudchadker and Eubank, 1970) have been plotted, together with the results of an earlier careful investigation (Kretschmer and Wiebe, 1954)'. The new correlation provides an excellent fit of the data over the whole temperature range. The O'Connell-Prausnitz correlation ( 1967) only fits the less precise data of Lambert et al. (1949)'. The discontinuity in the correlation at TR = 0.95 is 80 cdg-mol, or 40% of the B value. The B data for ethanol and 2-propanol are adequate,

but those for 1-propanol and all C4 alcohols are limited

in number and temperature range. In every case .the new

correlation has provided a satisfactory fit (superior to that obtained with the OConnell-Prausnitz correlation). Figure 8 shows thaet the alcohols' b has a weak dependence on reduced dipole moment. Reliable B data for C4+ alcohols are needed to clarify this, but it appears that the b of C2+ straight-chain 1-alkanols is given by

0.00908

* The data of Bottomley and Spurling (1967) have been discarded. Their values at 125" and 150'C are 150 and_ 100 cc/g-mol, respectively,
more negative than those plotted in Figure 7. 0 Other alcohol data discarded were those of Hanks and Lambert (1951)for ethanol, and of Foz et al. (1954) for CI-Ca alcohols.

(17) The branching in all other alcohols should make their B less negative, and hence their b less positive, but the reverse is shown in Figure 8. Opel's (1969) data for phenol were adequately fitted with only one parameter. The data of Andon et al. (1960) were inconsistent with Opel's and were discarded. (If they had also been considered, the value for a would have been -0.0101.) A two-parameter f"' may be needed to fit low-temperature ( T R < 0.6)data. The data for water are plotted in Figure 9. The

0.0006957 ( p R ) ; straight-chain 1-alkanols ( a = 0.0878)

Page 268

Moech, 1974

AlChE Journal (Vol. 20, No. 2)

of Keyes et al. (1936, 1938) are in serious disagreement with the recent measurements of Kell et al. (1968), especially for t < 250C. Only Kell's data (150 to 450C) were used in the regression analysis. At higher temperatures (500" to 900C) we used the values of Vukalovich et a]. (1967). As shown in Figure 9, the fit of Kell's data with the new correlation is excellent. (An almost equally good fit can be obtained using only one parameter.) The maximum deviation in B, up to 900C ( T R = 1.81) is less than 4 cc/g-mol.
CROSS-COEFFICIENTS FOR MIXTURES

50

100

150

200

1. g: 250

300

350

400

450

-200 -300 -

-400

The second virial cross-coefficient Bij has the same

temperature dependence that Bii and Bjj have, but the parameters to be used with Equations ( l l ) , (12), and (16) are Pcij, Tcij, uij, q,and bij. The mixing rules given below make it possible to relate these characteristic constanlts to pure-component parameters:
Tcij

- New Correlation - -- O'Connell-Pra"sn,n

0 D

Correiallo"

Kell e l a1 (19681 Kwes e l a1 (1936). Collins & Keves (1938)-

t,

g:

Fig. 9. Second virial coefficient of water.

(TciTcj) '

( 1 - kij)

(18)

oij

= 0.5 (mi + uj)

(20)

where kij is a characteristic constant for each binary. Equations ( 18) to (20) suffice for nonpolar/nonpolar binaries. An extensive tabulation of characteristic binary constants for nonpolar systems is given by Chueh and Prausnitz ( 1967). Hiza and Duncan (1970) have presented both data and a correla'tion for the kij of nonpolar binaries of He, Hz, Ne, etc, For polar/nonpolar binaries, Bij is assumed to have no polar term: Uj = 0 i (21)

bij = O (22) For polar/polar binaries, the polar contribution to Bij is calculated by assuming that
~ i j

(24) The suitability of the mixing rules, Equations (18) to ( 2 4 ) , will be indicated by the degree of success in fitting Bij, or BM data. [The relationship between Bii, Bjj, Bij, and BM is given by Equation ( 4 ) . ] The most sensitive
bij

= 0.5 (bi + bj)

= 0.5 (% + ~

j )

(23)

mixing rule is Equation (18). Tcij can be assumed to be the geometric mean of TCi and Tcj (kij = 0) only when i and i are very similar in size and chemical nature. For methane/n-heptane, kij = 0.10 (Chueh and Prausnitz, 1967). For systems with polar components, kij can be even larger. If there is strong solvation (hydrogen bonding between molecules of digerent species), kij can assume a negative value. There are relatively few mixture data for polarhonpolar or polar/polar sys'tems. The property usua!ly measured is BM, but it is also possible to determine Bij directly by measuring the solubility of liquid i. in gas j at elevated pressures. The average deviation of the calculated Bij values for methanol/nonpolar and water/nonpolar binaries is given in Table 3. The cross-coefficients have been directly determined from measurements of the solubility of liquid methanol or water in the compressed gas. The new correlation, with a temperature-independent k,i for each binary, provides an excellent fit of the data, even over a range of 213C (for H20/CH4). The O'Connell-Prausnitz correlation gives Bij's that are always much too negative (except for the CH30H/C02 binary). This correlation uses Equation (18) with kij = 0 (see the Appendix), which is clearly unsatisfactory. Most of the BM da'ta are of unsatisfactory quality. The uncertainty in Bij reflects the uncertainty in both the mixture and the pure-component data (which can be sub-

3. VIRIAL CROSS-COEFFICIENTS TABLE SECOND

Average deviation of
i/i

New
1.9 3.2 3.4 7.6 6.4 9.5 7.5 1.7 1.7 2.9 6.4 16.8 7.6

( kij )

Bij, cc/gmol t range, O'ConnellPrausnitz "C (no. points)

References Hemmaplardh and King ( 1972). Hemmaplardh and King ( 1972). Hemmaplardh and King ( 1972 ) . Hemmaplardh and King ( 1972 ) . Hemmaplardh and King ( 1972). Hemmaplardh and King ( 1972). Hemmaplardh and Kinq ( 1972). Rigby and Prausnitz ( 1968). Rigby and Prausnitz ( 1968). Olds et al. (1942); Rigby and Prausnitz
(1968). Coan nnd King (1971); Reamer et al. (1943). Coan and Kine, ( 1971). Coan and King ( 1971).

MethanoVnitrogen MethanoVargon MethanoVmethane MethanoVethylene MethanoVethane MethanoVcarbon dioxide Methanol/nitrous oxide Water/nitrogen Water/argon Watedmethane Water/ethane Watedcarbon dioxide Water/nitrous oxide

(0.05) (0.07) (0.13) (0.10) (0.12) ( 0.01) (0.13)

(~.30)
(0.35) (0.34) (0.37) (0.14) (0.16)

15 19 40 39 54 22 55 41 49 52 117 41 45

15-60 15-60 15-60 15-60 15-60 15-60 15-60 25-100 25-100 25-237.8

(5) (5) (5) (5) (5) (5) (5) (4) (4) (9)

25-171.1 ( 7 ) 25-100 25-100 (4) (4)

AlChE Journal (Vol. 20, No. 2)

March, 1974

Page 269

stantial, as already shown). It is therefore preferable to compare BM at y = 0.5, rather than Bij. The average deviation of the calculated BM ( y = 0.5) values for eleven polar/nonpolar and two polar/polar binaries is listed in Table 4. The new correlation is again satisfactory and superior to the O'Connell-Prausnitz correlation. The recent phenol/water data (Ope1 et al., 1969) have been fitted extremely well, but the fit for acetone/ ethyl ether (Zaalishvili and Kolysko, 1960) is less satisfactory. A temperature-dependent lc,j is needed to fit the BM ( y = 0.5) data for acetone/benzene and acetaldehyde/ acetonitrile, as shown in Table 5. In the case of acetone/ benzene, experimental error cannot be excluded [the data of Zaalishvili and Belousova (1964) should certainly be discarded; see Table 51, but the steep temperature dependence of B M has a physicochemical explanation: there is spectroscopic evidence ( Weimer and Prausnitz, 1966) that benzene (electron donor) forms a complex with acetone (electron acceptor). As temperature is raised,

the complex becomes weaker, and kij approaches its temperature-independent value of 0.12 (which it assumes when there is no significant complex formation). The negative ki;s for acetaldehyde/acetonitrile are a clear indication of the existence of a strong complex, probably formed by hydrogen bonding between nitrogen and the hydrogen attached to the carbonyl group. The existence of this complex has been recognized by Prausnitz and Carter (1960). Because of the strong chemical interaction between acetaldehyde and acetonitrile, which weakens as the temperature is raised, kij should be more negative at 40C than it is at a higher temperature (compare acetone/benzene). This is supported by the data: at 40C, the optimum Icij is - 0.32; at 100C, it is - 0.23. (When there is complex formation, O'Connell and Prausnitz add a specific association constant for the dimer. With such a constant, the O'Connell-Prausnitz correlation should provide a much better fit than that indicated in Table 5.) The paucity of experimental B A data will make it neces~ sary to predict rough kij values in most cases. The avail-

TABLE MIXTURESECOND 4. VIRIAL COEFFICIENTS y = 0.5 AT

Average deviation of B M ( y = 0.51, cc/gmol O'Connellt range, New (kij) Prausnitz "C (no. points) Acetone/n-butane Acetone/n-hexane Acetone/c yclohexane Acetone/carbon disulfide Acetone/ethyl ether Methyl ethyl ketonehenzene Diethyl ketonehenzene Acetonitrile/cyclohexane Ethyl ether/n-hexane Ethyl etherhenzene Methanol/benzene E thanol/benzene PhenoVwater
11 19 23 118 5.6 29 19 18 25 22 20 46 27 2.8

References Kappallo et al. (1963). Abbott (1965). Abbott ( 1985). Lambert et al. (1954).a Bottomley and Spurling (1967). Zaalishvili and Kolysko ( 1960). Chang et al. (1965): -950 & 100. Chang et al. (1965): -1120 & 100. Lambert et al. (1954). Fox and Lambert (1952). Knoebel and Edrnister ( 1968 ). Knoebel and Edmister ( 1968). Knoebel and Edmister ( 1968). Ope1 et al. (1969).

(0.07) (0.13) (0.20) (0.20) (0.10) (0.10) (0.12)b (0.12)b (0.40) (0.08) (0.10) (0.20). (0.20) (0.15)

47 143 178 61 42 58 122 263 216 88 82 160 158 164

9.15-47.85 45-90 75,90 53,76 51.7-158.9 50-80

80 80

(4) (4) (2) (2) (5)

(4) (1) (1)

53,76 53.2,78 60-100 40-1OOc 60-100 175-250

(2) (2) (3) (4) (3) (6)

=Their data are inconsistent with the more reliable results of Abbott (1965). b0.12 is the limiting k i j value for acetone/benzene (see Table 5); the agreement is even better with k,j = 0.14. The 40C is most probably too negative ( B of methanol is 500 cdgmol too negative!); without this point, the optimum

ktj

value is 0.25.

TABLE BM ( y = 0.5) 5.

OF

CROSS-DIMERIZING BINARIES

BM ( y = 0.5), cc/gmol
Calculated
i/i
t,

"C

Expenmentala

New ( k i j )
-1384 -1060 -858 -824 -824 -761 -761 -706 -706 -657 -5854 -4209 -2553 -1856 (0.00) (0.05) (0.12) (0.12) (0.12) (0.12) (0.12) (0.12) (0.12) (0.12) ( -0.32) (-0.32) (-0.25) (-0.23)

O'Connell-Prausnitz
-1258 - 1069 -955 -920 -920 -857 -857 -800 -800 -747 -2081 - 1757 - 1477 - 1232

Acetone/benzene

40 60 75 80 80 90 90 100
100 110

- 1060

- 1450b

-860 1100 -855 (-960). -750 & 100

(-85O)C

-710
( -765)'

Acetaldehyde/acetonitrile

40 60 80.1 100.4

-5900

(-710)'

-4180 -2540 - 1810

a Acetone/benzene: Abbott, 1965 (75, 90C); Chang et al. 1965 (80C); Knoehel and Edmister, 1968 (40, GO, SO, 100C); Zaalishvili and Belousova, 1964 (SO, 90. 100, 110C). Chang et al. are in agreement with Knoebel and Edmister at 80C. Acetaldehyde/acetonitrile: Prausnitz and Carter (1960). b Too negative, like the other mepsurements by Knoebel end Edmister (1968) at 40C. 0 The BY data of Zaalishvili and Belousova (1964) have the wrong composition dependence.

Page 270

March, 1974

AlChE Journal (Vol. 20, No. 2)

TABLE AVERAGE 6. VALUES THE BINARY FOR CONSTANT kij ( kn. - k t.3. ) . -

Hydrocarbons Ketones Ethers Alcohols Water

1 2 3 4 5

1 2 3 4 5

0.13

0.10 0.13

0.15 0.05 0.13

0.40 0.15 0.35 0.10

* kcj

= 0.15 for phenol/water.

able information can be used as the basis for all such predictions. Consideration of relative size and chemical nature should facilitate the prediction of t h e binary constants. As a rough guide, average values for kij are given in Table

6.
ACKNOWLEDGMENT

The author is grateful to Robert H. Johnston, for extensive guidance and support, and to Esso Research and Engineering Company for the permission to publish this paper.
NOTATION

a,

b = parameters of polar contribution term to B,

f2);

B
C
f(O),

kij

Po
R T
u
z

yi

see Equations ( 1 2 ) a n d ( 1 6 ) = second virial coefficient = third virial coefficient f ( l ) , fez) = dimensionless terms o Equation ( l l ) , f the new correlation of second virial coefficients = characteristic binary constant; see Equation (18) = pressure = vapor pressure of p u r e component = gas constant = absolute temperature = molar volume = vapor mole fraction of component i = compressibility factor

Greek Letters p = dipole moment o = acentric factor

Subscripts c = critical property

i, j ii

= property of = mixture

component i, j

= characteristic property used in the calculation of

t h e second virial cross-coefficient property = terms in Pitzer-Curl correlation = reduced property

M PC R

LITERATURE CITED

Abbott, M. M., Ph.D. thesis, Second Virial Coefficients of Some Organic Vapors and Vapor Mixtures, Rensselaer Polytechnic Inst. (1965); private communication (1970). Alexander, E. A,, and J. D. Lambert, The Second Virial Coefficient of Acetaldehyde, Trans. Faraday SOC., 37, 421 (1941); data reported by Dymond and Smith (1969). Anderson, L. N., A. P. Kudchadker, and P. T. Eubank, Volumetric Properties of Gaseous Acetone, J. Chem. Eng. Data, 13,321 (1968). Andon, R. J. L., D. P. Biddiscombe, J. D. Cox, R. Handley, D. Harrop, E. F. G. Herington, and J. F. Martin, Preparation and Physical Properties of Pure Phenol, Cresols, and Xylenols, 1. Chem. SOC., 5246 ( 1960). Baxendale, J. H., and B. V. Eniistdn, Deviations from Ideal

Gas Behavior of Benzene Vapor, Phil. Trans. Roy. SOC. London Ser. A, 243, 176 (1951); B recalculated by Dymond and Smith ( 1969). Black, C., Vapor Phase Imperfections in Vapor-Liquid Equilibria, Ind. Eng. Chem., 50,391 ( 1958). Bottomley, G. A., and C. G. Reeves, The Virial Coefficients of Carbon Disulfide, Diethyl Ether, and n-Hexane, J. Chem. SOC.,3794 ( 1958). Bottomley, G. A. and T. H. Spurling, Acetone Isotherms, Nature, 195, 900 (1962). ., The Second Virial Coefficients of Benzene, Austral. J. Chem., 19, 1331 (1966). ., The Second Virial Coefficient of Carbon Disulfide, Methyl Chloride, Acetone, and Methanol, and Certain of their Binary Mixtures, ibid., 20, 1789 ( 1967). Brewer, J., Determination o Mixed Virial Coefficients, f AFOSR No. 67-2795, (Dec., 1967). Byrne, M. A., M. R. Jones, and L. A. K. Staveley, Second Virial Coefficients of Argon, Krypton and Methane and their Binary Mixtures at Low Temperatures, Trans. Faraday SOC.,64, 1747 (1968). Cawood, W., and H. S. Patterson, The Coinpressibilities of Certain Gases at Low Pressures and Various Temperatures, J. Chem. SOC., 619 (1933). Chang, Y.-C., H.-T. Wang and T. Huang, Second Virial Coefficients of Vapors of Organic Compounds, KO Hsueh Tung Pao, 437 ( 1965); in Chinese. Chevalier, J.-L., Determination of the Second Virial Coefficients of Aliphatic Ethers, C . R. Acad. Sci. Paris Sec. C . , 268,747 ( 1969); in French. Chueh, P. L., and J. M. Prausnitz, Vapor-Phase Fugacity Coefficients in Nonpolar and Quantum-Gas Mixtures, Ind. Eng. Chem. Fundamentals, 6,492 ( 1967). Coan, C. R., and A. D. King, Jr., Solubility of Water in Compressed Carbon Dioxide, Nitrous Oxide, and Ethane. Evidence for Hydration of Carbon Dioxide and Nitrous Oxide in the Gas Phase, J. Am. Chem. SOC., 93, 1857 ( 1971). Connolly, J. F., and G. A. Kandalic, Virial Coefficients and Intermolecular Forces of Hydrocarbons, Phys. Fluids, 3, 463 ( 1960). Cox, J. D., Second Virial Coefficients of the Propanols and Butanols, Trans. Faraday SOC.,57, 1674 (1961). Crain, R. W., Jr., and R. E. Sonntag, The P-V-T Behavior of Nitrogen, Argon, and their Mixtures, Aduan. Cryog. Eng., 11,379 (1966). Dymond, J. H., and E. B. Smith, The Virial Coefficients of Gases, Clarendon Press, Oxford (1969). Eubank, P. T., and J. M. Smith, Thermodynamic Properties of Polar Gases in the Dilute Phase, AIChE J., 8, 117 (1962). Fox, J. H. P., and J. D. Lambert, The Second Virial Coefficients of Mixed Organic Vapors, Proc. Roy. SOC. Ser. A, 210, 557 ( 1952). Foz, 0. R., J. Morcillo, A. P. Mas% and A. MBndez, Compressibility of Alcohol Vapors, Anales Real SOC. Esp. Fis. Quim. (Madrid), Ser. B, 17, 23 ( 1954); in Spanish. Values presented graphically and reported by Dymond and Smith ( 1969 ). Hajjar, R. F., W. B. Kay, and G. F. Leverett, Second Virial Coefficients of Eight Compounds in the Range 40 to 200C, J. Chem. Eng. Data, 14, 377 (1969). Hales, J. L., E. B. Lees, and D. J. Ruxton, Vapor Heat Capacities and Heat of Vaporization of Ethyl Ketone, Ethyl Propyl Ketone, Methyl Isopropyl Ketone, and Methyl Phenyl Ether, Trans. Faraday SOC.,63, 1876 (1967). Halm, R. L., and L. I. Stiel, Second Virial Coefficients of Polar Fluids and Mixtures, AIChE J., 17, 259 ( 1971). Hanks, P. A., and J. D. Lambert, 1951; data given by Dymond and Smith (1969). Hauthal, W. H., and H. Sackmann, Second Virial Coefficients of Some Alkyl Ketones, Proc. 1st Intern. Conf. Calorimetry Thermodynamics ( W a r s a w ) ,625 ( 1969). Haworth, W. S., and L. E. Sutton, The Second Density Virial Coefficients of Some Polar Gases, Trans. Faraday SOC., 67, 2907 (1971). Hemmaplardh, B., and A. D. King, Jr., Solubility of Methanol in Compressed Nitrogen, Argon, Methane, Ethylene, Ethane, Carbon Dioxide, and Nitrous Oxide. Evidence for Association of Carbon Dioxide with Methanol in the Gas Phase,

AlChE Journal (Vol. 20, No. 2)

March, 1974

Page 271

j . Phys. Chem., 76,2170 ( 1972). Hiza, M. J., and A. G. Duncan, A Correlation for the Prediction of Interaction Energy Parameters for Mixtures of Small Molecules, AZChE J., 16, 733 ( 1970). Hofmann, H. P., and R. Pfestorf, 1968; data given by Hauthal and Sackmann ( 1969). Kappallo, W., N. Lund, and K. Schaffer, Intermolecular Forces Between Same and Different Molecules from Virial Coefficients, Z . Phys. Chem. (Frankfurt), 37, 196 (1963); in German. Kell, G. S., and G. E. McLaurin, Virial Coefficients of Methanol from 150 to 300C and Polymerization in the Vapor, J. Chem. Phys., 51,4345 ( 1969). ., and E. Whalley, PVT Properties of Water. 11. Virial Coefficients in the Range 150-450C without Independent Measurement of Vapor Volumes, J. Chem. Phys., 48,3805 ( 1968) . Kennedy, R. M., M. Sagenkahn, and J. G. Aston, . . . The Density of Gaseous Dimethyl Ether, J. Am. Chem. SOC., 63,2267 (1941). Keyes, F. G., L. B. Smith, and H. T. Gerry, 1936, and S. C. Collins, and F. G. Keyes, 1938; data given by Dymond and Smith ( 1969). Knoebel. D. H.. and W. C. Edmister. Second Virial Coefficients of Binary Mixtures of Benzene with Methanol, Ethanol, Acetone, and Diethyl Ether, J. Chem. Eng. Data, 13, 312 ( 1968). Kreglewski, A., On the Second Virial Coefficient of Real Gases, J. Phys. Chern., 73, 608 ( 1969). Kretschmer, C. B., and R. Wiebe, Pressure-Volume-Temperature Relationships of Alcohol Vapors, J. Am. Chem. SOC., 76,2579 ( 1954 ) Kudchadker, A. P., and P. T. Eubank, Second Virial Coefficient of Methanol, J. Chem. Eng. Data, 15, 7 ( 1970). Lambert, J. D., S. J. Murphy, and A. P. Sanday, The Second Virial Coefficients of Mixtures of Polar and Nonpolar Vapors, Proc. Roy. SOC. Ser. A, 226, 394 (1954); see also Dymond and Smith ( 1969). Lambert. 1. D.. G. A. H. Roberts, 1. S. Rowlinson, and V. I. Wilkinsin, The Second Virial coefficients of Organic Vapors, Proc. Roy. SOC. Ser A, 196, 113 (1949); unpublished data given by Dymond and Smith ( 1969). MacCormack, K. E., and W. G. Schneider, Carbon Dioxide in the Temperature Range 0-500 and Pressures up to Fifty Atmospheres, J. Chem. Phys., 18, 1269 ( 1950). McClellan, A. L., Tables of Experimental Dipole Moments, W. H. Freeman, San Francisco ( 1963). McGlashan, M. L., and D. J. B. Potter, An Apparatus for the Measurement of the Second Virial Coefficients of Vapors; the Second Virial Coefficients of Some n-Alkanes and of Some Mixtures of n-Alkanes, Proc. Roy. SOC. Ser. A, 267, 478 (1962). Moreland, M. P., J. J. McKetta, and I. H. Silberberg, Compressibility Factors and Virial Coefficients of 2-Propanol, J. Chem. Eng. Data, 12,329 (1967). Nickerson, J. K., K. A. Kobe, and J. J. McKetta, The Thermodynamic Properties of the Methyl Ketone Series, J. Phys. Chem., 65,1037 ( 1961). Nothnagel, K.-H., D. S. Abrams, and J. M. Prausnitz, Generalized Correlation for Fugacity Coefficients in Mixtures at Moderate Pressures, Ind. Eng. Chem. Process Design Develop., 12, 25 ( 1973). OConnell, J. P., A Method for Predicting Second Virial Coefficients for Vapor-Liquid Calculations, paper presented at 64th Annual AIChE Meeting, San Francisco (1971). ., and J. M. Prausnitz, Empirical Correlation of Second Virial Coefficients for Vapor-Liquid Equilibrium Calculations, Ind. Eng. Chem. Process Design Develop., 6, 245 (1967). Olds, R. H., B. H. Sage, and W. N. Lacey, Composition of the Dew-Point Gas of the Methane-Water System, Ind. Eng. Chern., 34, 1223 (1942); data reduced to Bijs and reported by Rigby et al. ( 1969). Opel, G., Second Virial Coefficients of Phenol Between 205 and 350C, Chem. Tech. (Leipzig), 21, 776 (1969); in German. ., V. Steinert, and U. von Weher, Second Virial Coefficients BIZ for the Water-Phenol Vapor Mixture, Proc. 1st Intern. Conf. Calorimetry Thermodynamics (Warm w , 635 (1969). a )
v
~~

Osipiuk, B., and R. Stryjek, Second Virial Coefficients of Binary Methyl n-Alkyl Ether-Sulfur Dioxide Mixtures, Bull. Acad. Polon. Sci. Ser. Sci. Chem., 18, 289 (1970). Pitzer, K. S., and R. F. Curl, Jr., Empirical Equation for the Second Virial Coefficient, J. Am. Chem. SOC., 79, 2369 (1957). Pitzer, K. S., D. Z. Lippmann, R. F. Curl, Jr., C. M. Huggins, and D. E. Petersen, Compressibility Factor, Vapor Pressure and Entropy of Vaporization, ibid., 77, 3433 (1955). Polak, J., and B. C.-Y. Lu, Second Virial Coefficients of Gases-A Correlation with Stockmayer Potential Function, Can. J . Chem. Eng., 50,553 (1972). Prausnitz, J. M., iMolecular Thermodynamics of Fluid-Phase Equilibria, Prentice-Hall, Englewood Cliffs, N. J. ( 1969). ., and W. B. Carter, Second Virial Coefficients of the Acetonitrile-Acetaldehyde System, AIChE J., 6, 611 ( 1960). Prausnitz, J. M., A. Eckert, R. V. Orye, and J. P. OConnell, C. Computer Calculations for Multicomponent Vapor-Liquid Equilibria, Prentice-Hall, Englewood Cliffs, N. J. ( 1967). Ratzsch, M., and H.-J. Bittrich, Second Virial Coefficients of Mixtures with Alkyl Halides, Z. Phys. Chem. (Leipzig ), 228, 81 ( 1965); in German. Reamer, H. H., R. H. Olds, B. H. Sage, and W. N. Lacey, Composition of Dew-Point Gas in Ethane-Water System, Ind. Eng. Chem., 35, 790 (1943); data reduced to Bijs and reported by Rigby et a1 ( 1969). ., 1914; B calculated by S. D. Hamann, and W. J. McManamey, 1953, and given by Dymond and Smith (1969). Renon, H., L. Asselineau, G. Cohen, and C. Raimbault, Computer Calculations for Vapor-Liquid and Liquid-Liquid Equilibria, Technip, Paris, ( 1971); in French. Rigby, M., J. P. OConnell, and J. M. Prausnitz, Intermolecular Forces in Aqueous Vapor Mixtures, Ind. Eng. Chem. Fundamentals, 8, 460 ( 1969) . Rigby, M.,and J. M. Prausnitz, Solubility of Water in Compressed Nitrogen, Argon, and Methane, J. Phys. Chem., 72, 330 (1968). Russell, J., and 0. Maass, Two-Component Systems Involving Compound Formation, Can. J. Research, 5,436 (1931). Stryjek, R., and A. Kreglewski, The Second Virial Coefficients of Gaseous Mixtures. I. The Diethyl Ether-Sulfur Dioxide Systems, Bull. Acad. Polon. Sci. Ser. Sci. Chem., 13, 201 (1965). Tripp, T. B., and R. D. Dunlap, Second Virial Coefficients for the Systems: n-Butane Perfluro-n-Butane and Dimethyl Ether + 1-Hydroperfluorpropane, J. Phys. Chem., 66, 635 (1962); B obtained from z with Equation ( 2 ) . Tsonopoulos, C., and J. Xf. Prausnitz, Fugacity Coefficients in Vapor-Phase Xfixtures of Water and Carboxylic Acids, Chem. Eng. J., 1, 273 (1970); erratum, 4, 96 (1972). Vukalovich, M. P., and Ya. F. Masalov, Equation of State for Gaseous Carbon Dioxide, Thermal Eng., 13, ( 5 ) , 73 (1966); the units in column 2 of Table 3 should be: cmZ/kgf x 105. Vukalovich, M. P., M. S. Trakhtengerts, and G. A. Spiridonov, An Equation of State for Steam at Temperatures above 500C, Thermal Eng., 14, ( 7 ) , 86 (1967); analysis of data obtained by Vukalovich .et al. ( 1962). Weimer, R. F., and J. M. Prausnitz, Ultraviolet Spectra and Complex Formation in Mixtures of Polar Organic Solvents and Aromatic Hydrocarbons, Spectrochim. Acta, 22, 77 (1966). Weir, R. D., I. Wynn Jones, J. S. Rowlinson, and G. Saville, Second Virial Coefficicnt of Argon and Krypton, Trans. Faraday SOC., 63, 1320 (1967). Zaalishvili, Sh. D., and Z. S. Belousova, The Second Virial Coefficient of Vapors and their Mixtures. IV. The AcetoneBenzene System, Russ. J. Phys. Chem., 38, 269 ( 1964). Zaalishvili, Sh. D., and L. E. Kolysko, . . . I. The Ethyl EtherAcetone System, ibid., 34, 1223 ( 1960). Zaalishvili, Sh. D., . . . 111. The Ethyl Ether-Methyl Iodide System, ibid., 36, 440 ( 1962). ., and E. Ya. Gorodinskaya, New Method for Determination of the Molar Volume of a Vapor. 111. Acetonitrile, ibid., 45, 1500 ( 1971).

Manuscript received June 12, 1973; revision receiued October 2 and accepted October 3, 1973.

Page 272

March, 1974

AlChE Journal (Vol. 20, NO. 2)