Equation of State for Saturated Liquids

HAROLD G. RACKETT
Oklahoma State Unversity, Stillwater, Okla. 74074

7
An empirical equation, log {Vf/Vc) = (1 TV)2 log Zc, relates the reduced volume
-

of saturated liquids to the reduced temperature and the critical compressibility factor.
This equation holds within experimental uncertainty for 16 diverse substances, including
the highly polar ammonia, acetone, and hydrogen cyanide. The equation and the
similar equations of Riedel, Yen and Woods, and Bradford and Thodos are compared
See [Link] for options on how to legitimately share published articles.

using 32 sets of liquid-volume data for these 16 substances; this equation is the
most precise. Polar substances for which the equation fails—the alcohols, carboxylic
acids, and nitriles—are those whose molecules combine into cyclic groups in the
liquid state. For conforming substances, with the exception of water, the applicable
range is from the triple point to the critical point.
Downloaded via DURHAM UNIV on July 13, 2018 at [Link] (UTC).

General equations for reduced saturated-liquid density

~
r''":
Vfivr =
z,+l(1
orvolume as a function of reduced temperature and critical
compressibility factor have been developed (3, 23, 31). Z,)(l 7
= 1 + (In -

T,)17 + (1 -

TrY (4)
Herein is presented another such equation (21),
7
log (Vf/V,) =
(1 -

71)2 log 2, (1)

simpler than the other similar equations in that it contains

but one right-hand term. The exponent of (1 TV) is -

or
its only arbitrary component.
T'f~]
Equation 1 becomes an identity at the critical point,
“

pi/Pc =
zrl(1 (6)
TV = 1. As TV approaches 0, Equation 1 approaches

log (Vo/ log Zc (2) (_1" Z,)

V7) =
= 1 + (-In Z,.)(l -

T, 7
+ (1 -

TrY
1
+ ...
(7)
or
which then more resembles the other correlation equations.
po/p, =
1/Z, (3)
The equation of Riedel (23) contains two (1 TV) terms: -

Timmermans (27) demonstrated that the actual behavior

Pr= 1 + [(1/Z, -

1.211)/1.3)(1 -

Tr)13 + 0.85 (1 -

Tr) (8)
of most liquids is in accord with this relation. He measured
the densities of 12 liquids at temperatures between +15° C The equation of Yen and Woods (31) contains three (1
and their freezing points and fitted cubic equations in tem- -

Tr) terms, the exponents of which are A, %, and 4A,

perature to these densities. For all but one of these liquids and the coefficients of which are cubic functions of Z(.
he found that his equations gave hypothetical liquid The equation of Bradford and Thodos (3) also contains
densities at absolute zero that satisfy Equation 3. Earlier, three (1 TV) terms. The three coefficients and the smallest
-

van der Waals (29) set forth this relation as a general of the three exponents are linear functions of Z<.
rule, but without verification. A well-known text (11) states that the reduced saturated-
The reduced volumes given by Equation 1 at the two liquid density table of Lydersen, Greenkorn, and Hougen
end points, V, = 1 at T, 1 and V, =
=
Zc at TV = 0, (15) was derived from the data of Riedel (23). Riedel’s
are independent of the value of the exponent of (1 Tr). -

equation replicates the Z, = 0.25, 0.27, and 0.29 columns

There is no known theoretical basis for Equation 1 or of this table with minimum standard errors of 0.07, 0.05,
for the specific value selected for this exponent. The value and 0.11%, respectively, when Zc 0.2508, 0.2710, and =

that most closely reproduces a given set of liquid-volume 0.2955, respectively, are used in this equation.
data generally lies between about 0.275 and 0.295 (21).
But the uncertainty of these best exponent values is at
TESTING AND COMPARISON WITH OTHER EQUATIONS
least ±0.01, owing to the uncertainty of the Zc values
that must be used in their determination. The exponent Equation 1 contains the three parameters TV, Z„, and
cannot be made a function of Zc, because a relation between Vc Substituting ZcRTjPc for Vc merges the uncertainty
the exponent and Zr is not even detectable. The particular of Zc and W,
value, , that was selected as a universal exponent for log Z, log (P./RT,)
log V/ =
[1 + (1 Tr)1 (9)
- -

]
all conforming substances lies near the middle of the possible
range of a universal value. In testing the proposed equation and comparing it and
The exponential form of Equation 1 can be expanded the three similar equations cited, each equation was put
into its Maclaurin’s series, into a Vf Vf(T,TcPcZ<:) form. For each of 16 test sub-
=

514 Journal of Chemical and Engineering Data, Vol. 15, No. 4, 1970
stances (Table I or II), a single value of Tc and of Pc Rather than the mean of the Zc values given by Equation
(Table II) was used in this testing, but for each substance 10, the Zc value given by the linear-regression equation,
several values of Z, were used, as described below.
In the initial test Z„ values taken from five literature log ( . ,) [1 + (1 -
TT :] + Ell + (1 -

Tr)‘log V,
(3, ¡0, 15. 22. 31) were used. These Z, values log Z, =
sources
and the standard per cent errors of estimate when these
Ell + (1 -

, 7
(11)
Z, values are used with each equation are listed in Table
I. The uncertainty of recorded Zr values and the sensitivity was used in the determination of the precision of the pro-
of calculated Vt to variations of Z, are here exemplified. posed equation. This Zc minimizes the standard error of
An error listed in Table I is not really a measure of the estimate of log V¡. The Zt value given by Equation 11,
accuracy of that equation for that substance, but rather however, is generally almost identical to the mean of those
is primarily a measure of how close the Z, value used given by Equation 10. Their difference is less than 0.0001
is to the best-fitting Z, for that equation and that substance. for each of the data sets of Table II.
Because of the uncertainty of Z, values (or equivalently, The standard errors of estimate of In V¡ (equivalent
of V. values), the accuracy of the equations cannot be to the standard relative errors of estimate of Vf) when
determined. Equation 11 Zr values are used in Equation 9 are listed
The precision of any of the equations, however, can be in Table II. These standard errors of Vf are generally
determined by making Z, a free parameter—that is, using less than 0.5%. A Zc value listed in Table II is not implied
that Z value which brings about the best agreement of to be correct, but only to be most consistent with Equation
the equation and a given set of density data. Equation 9, the particular density-data set, and the values of T<
9 can be rearranged into a form explicit in Zr, and Pc that were used.
Best Zc values to be used in the other liquid-density
log V, + log (P./RT,) equations were obtained by using equations or methods
log Z, (10)
F+ (1 -

Tr)-' analogous to Equation 10. The standard errors that were

found when these Zc values are used in these other equations
For the same 16 substances of the initial test, the mean are not listed here, but the largest of the individual V¡
of the Z, values given by Equation 10 for each data set errors number as follows. The Bradford-Thodos equation
and the standard deviation of Z, from this mean are listed gives 13 errors greater than 2%, but all but one are for
in Table II. These standard deviations are generally less substances for which Z< < 0.255. Bradford and Thodos
than 0.3%. For each of 12 of the substances two or three formulated the expressions for their equation coefficients
sets of data were examined and the results listed separately, and variable exponent to fit best the observed densities
so that the agreement or lack of agreement of the mean of 12 hydrocarbons, for each of which Z, > 0.255; their
calculated Z, values would be shown. equation cannot be extrapolated to substances with lower

Table I. Testing of Saturated-Liquid Volume Equations Using Zt Values Taken from Literature Sources
Standard % Error of Estimate of V,-
Yen & Bradford of Liquid
Source (s) Riedel Woods & Thodos Volume
Substance Z, of Z< Value Eq. 9 eq. eq. eq. Data
Neon 0.296 (22) 2.81 2.69 2.90 3.34 (18)
0.304 (15) 1.58 2.16 1.84 1.09
Nitrogen 0.288 (10) 0.87 0.98 1.30 1.19 (20)
0.291 (15. 22, 31) 0.95 1.21 2.35 0.58
Hydrogen cyanide 0.197 (15, 22) 4.99 3.69 24.59 15.02 (4)
Water 0.230 (15, 22, 31) 0.41 2.77 2.98 1.59 (1)
Ammonia 0.242 (22, 31) 3.25 4.39 1.39 1.81 (7
0.243 (15) 2.57 3.69 1.02 1.46
n-Pentane 0.268 (10) 0.40 0.61 0.81 0.71 (32)
0.269 (3, 15, 22, 31) 0.51 0.51 1.31 0.80
2-Methvlbutane 0.268 (15, 22) 0.93 1.21 0.51 0.98 (32)
0.269 (10) 0.44 0.67 0.67 0.61
0.272 (31) 1.44 1.33 2.29 1.49
Ethylene 0.270 (15, 22) 7.02 7.34 6.27 6.97 (19)
0.283 (10) 0.48 0.59 1.58 0.29
0.284 (3) 0.98 1.09 2.08 0.72
0.285 (31) 1.52 1.65 2.60 1.24
Acetylene 0.274 (15, 22. 31) 0.51 0.53 1.52 0.31 (17)
Benzene 0.270 (3) 0.62 0.37 1.34 0.93 (32)
0.274 (15, 22, 31) 2.98 2.74 3.74 3.07
Tetrachloromethane 0.272 (15) 0.22 0.53 0.59 0.34 (32)
Diethyl ether 0.255 (22) 5.22 6.02 4.79 4.70 (32)
0.261 (15) 1.59 2.16 1.30 1.42
Acetone 0.237 (22) 5.48 7.19 3.11 3.94 (12)
0.238 (15, 31) 4.82 6.47 2.65 3.36
Ethyl acetate 0.252 (15, 22) 1.39 2.19 0.68 1.06 (32)
Diethvlamine 0.270 (22) 5.78 5.51 6.34 5.92 (12)
0.286 (15) 14.88 15.17 15.36 14.66
Ethyl mercaptan 0.274 (15, 22) 2.28 2.08 3.04 2.30 (2)
Data used presented in (12).
"
as

Journal of Chemical and Engineering Data, Vol. 15, No. 4, 1970 515
 Table II. Indirect Testing of Proposed Saturated-Liquid Volume Equation
Std. Source
Std. Error of No. of of Liquid
Mean Devn., Estimate. Data Volume
Tr Range of Data Used
Substance Z<, Eq. 10 Z<,xlO' In V, xlO1 Points Data T. K
°
Pr, Atm
Neon 0.3011 4.4 21.8 10 0.567 0.9689 (18) 44.4 26.2

Nitrogen 0.2895 8.4 40.9 11 0.513 0.9906 (20) 126.2 33.5

0.2886 4.2 25.4 9 0.496 0.718 (30)
0.2901 1.5 8.6 15 0.615 0.830 (26)

Hydrogen cyanide 0.1915 0.3 2.7 11 0.569 0.653 (4) 456.7 53.2
Water 0.2297 6.1 38.1 21 0.824 0.9910 (1) 647.3 218.2
Ammonia 0.2468 5.1 30.7 43 0.479 0.9944 (7Y‘ 405.4 111.4
0.2465 4.0 25.4 10 0.550 0.9698 (13)
rc-Pentane 0.2684 7.6 35.6 29 0.581 0.9999 (32) 470.3 33.3
0.2685 3.1 21.9 19 0.291 0.613 (27Y
0.2688 1.8 11.7 14 0.478 0.708 (14)

2-Methylbutane 0.2695 6.1 33.0 25 0.593 0.9993 (32) 460.9 32.9

0.2697 5.6 39.0 20 0.250 0.625 (27Y
0.2688 2.4 16.9 20 0.250 0.334 (8)

Ethylene 0.2824 5.5 30.0 15 0.453 0.9941 (19) 282.7 50.0

0.2818 1.1 6.6 8 0.582 0.721 (16)

Acetylene 0.2731 3.4 17.2 12 0.807 0.9885 (17) 308.7 61.2

0.2728 1.0 6.4 9 0.620 0.706 (16)
Benzene 0.2691 4.6 24.1 31 0.486 0.9949 (32) 562.1 48.2
0.2692 1.2 8.2 11 0.522 0.593 (5)
T etrachloromethane 0.2722 3.2 17.8 28 0.491 0.9762 (32) 556.4 45.0
0.2723 1.9 12.1 7 0.518 0.626 (9)

Diethyl ether 0.2636 6.0 33.1 23 0.585 0.9980 (32) 467.1 35.7
0.2639 2.5 17.6 19 0.321 0.618 (27)"
0.2639 2.1 14.3 39 0.328 0.735 (25)
Acetone 0.2452 9.8 61.2 19 0.648 0.9895 (12) 508.5 46.8
0.2463 2.4 18.3 12 0.351 0.537 (27)"

Ethyl acetate 0.2540 6.7 40.6 28 0.522 0.9980 (32) 523.2 37.9
0.2544 2.4 17.4 13 0.363 0.553 (27)"
Diethylamine 0.2604 6.5 35.2 18 0.662 0.9939 (12) 496.2 36.2
0.2602 7.3 46.9 10 0.470 0.833 (6)

Ethyl mercaptan 0.2701 4.9 26.5 19 0.547 0.9845 (2) 499.2 54.2
°
presented in (12).
“
Data used as Data used as presented in (28).

Z< values. The Yen-Woods equation gives 13 errors greater with temperature decrease in the critical region does not
than 1.5%, three of which are greater than 2%. Riedel’s belie the very low critical compressibility factors of these
equation gives 13 errors greater than 1%, two of which three compounds. But then this rapid rate diminishes faster
are greater than 1.5%. Equation 9 gives but three errors than these low Z, values and Equation 1 predict.
greater than 1%, and none is greater than 1.5%. Thus, Saum (24) has demonstrated, mainly on the basis of
for this group of saturated-liquid-density data sets, the observed viscosities, that molecules of carboxylic acids and
equation here proposed is more precise than any of the of nitriles tend to dimerize in their liquid phases, and that
others. alcohol molecules combine to form larger groups, possibly
tetramers. The attraction between the molecules of a cyclic
CONFORMING LIQUIDS AND NONCONFORMING LIQUIDS group will be much stronger than the attraction between
molecules in different groups. The rapid volume decrease
Equation 9 was tested indirectly with the liquid-density in the critical region for these compounds is brought about
data of 106 substances and found to hold within by the strong intragroup forces, while the less-than-
experimental uncertainty for 93. The results of this testing predicted rate of decrease at lower temperatures is brought
of 16 of the conforming liquids are presented in Table about by the weaker intergroup forces. Nitrogen dioxide,
II. The In (Vf/Vc) vs. Tr curves of several of these 16 whose molecules dimerize, also does not conform. At least
liquids are shown in Figure 1. Several of the conforming one additional parameter would be needed to correlate these
liquids are polar—e.g., hydrogen cyanide, ammonia, and departures from predicted behavior.
acetone. This correlation fails, however, with certain classes The correlation also fails for those two liquids that exhibit
of polar compounds—viz., the alcohols, the carboxylic acids, the largest quantum effects, helium and hydrogen. Water
and the nitriles. (Acetonitrile is that one of Timmermans’ conforms at reduced temperatures above ~0.82 (Table II);
12 liquids that he found did not comply with the Equation it does not conform at lower temperatures. For 11 of the
3 relation.) The In (V¡/Vc) points of a representative of test substances one of the data sets of Table II contains
each of these classes are plotted in Figure 2. The curves at least one point within 2° of the triple point. Except
on which these points would lie intersect the curves pre- for water (and deuterium oxide), no conforming substance
dicted by Equation 1. The rapid rate of volume decrease has been found that does not continue to conform down

516 Journal of Chemical and Engineering Data, Vol. 15, No. 4, 1970
Figure 1. Reduced volume as function of reduced temperature Figure 2. Reduced volume as function of reduced temperature
Liquids that conform Polar liquids that do not conform
• Neon (18) AHydrogen cyanide (4) • Methanol (32) Acetonitrile (12)
M Tetrachloromethane (32) —Calculated Acetic add (32) —Calculated
Ammonia (13)

to the lowest temperatures at which the liquid density (11) Hougen, O.A., Watson, K.M., Ragatz, R.A., “Chemical Pro-
has been measured. At the other temperature extreme, cess Principles,” Part II, “Thermodynamics,” 2nd ed., p. 577,

many of the data sets of Table II demonstrate that the Wiley, New: York, 1959.
(12) “International Critical Tables,” Vol. Ill, McGraw-Hill, New
proposed equation is applicable in the critical region.
York, 1928.
(13) Keyes, F.G., Brownlee, R.B., “Thermodynamic Properties of
NOMENCLATURE Ammonia,” Wiley, New York, 1916.
(14) Legatski, T.W., Nelson, W.R., Dean, M.R., Fruit, L.R., Ind.
In = natural logarithm of Eng. Chem. 34, 1240 (1942).
log =
logarithm of: base is optional (15) Lydersen, A.L., Greenkorn, R.A., Hougen, O.A., “Generalized
P, =
critical pressure Thermodynamic Properties of Pure Fluids,” Coll. Eng., Univ.
R =
gas-law constant Wisconsin, Eng. Expt. Sta., Rept. 4, (October 1955).
T =
temperature: absolute (16) Maass, 0., Wright, C.H., J. Amer. Chem. Soc. 43, 1098 (1921).
T, = critical temperature (17) Mathias, E., Compt. Rend. 148, 1102 (1909).
T, = reduced temperature = T T, (18) Mathias, E., Crommelin, C.A., Kamerlingh Onnes, H., Ibid.,
V,· = critical volume 175, 933 (1922).
V, =
saturated-liquid volume (19) Mathias, E., Crommelin, C.A.. Watts, H.G., Ibid., 185, 1240
V, = reduced saturated-liquid volume V,· /' V.
=
(1927).
V, =
hypothetical saturated-liquid volume at T. 0 =
(20) Mathias, E., Kamerlingh Onnes. H., Crommelin, C.A., Ibid.,
Z, = critical compressibility factor V,P,lRTr. Z, is dimension-
=
160, 237 (1915).
less; the units of R must be consistent with those of (21) Rackett, H.G., S. M. thesis, Massachusetts Institute of Tech-
P,, T, and V,.
nology, Cambridge, Mass., 1960.
µ = critical density (22) Reid, R.C., Sherwood, T.K., “Properties of Gases and
pi =
saturated-liquid density Liquids,” 2nd ed., Appendix A, McGraw-Hill, New York,
µ. = reduced saturated-liquid density pi!µ, =
1966.
µ,, =
hypothetical saturated-liquid density at T, = 0
(23) Riedel, L., Chem.-Ing.-Tech. 26, 259, 679 (1954).
(24) Saum, A.M., J. Polym. Sci. 42, 57 (1960).
(25) Taylor, R.S., Smith, L.B., J. Amer. Chem. Soc. 44, 2450 (1922).
LITERATURE CITED (26) Terry, M.J., Lynch, J.T., Bunclark, M., Mansell, K.R., Stave-
(1) ASME Research Committee on Properties of Steam, “Thermo- ly, L.A.K., J. Chem. Thermodyn. 1, 413 (1969).
(27) Timmermans, J., Bull. Soc. Chim. Beiges 26, 205 (1912); CA
dynamic and Transport Properties of Steam,” p. 83, New
7, 924 (1913).
York, 1967.
(28) Timmermans, J., “Physico-Chemical Constants of Pure
(2) Berthoud, A., Brum, R., J. Chim. Phys. 21, 143 (1924).
Organic Compounds,” Elsevier, Amsterdam, 1950.
(3) Bradford, M.L., Thodos, George, Can. J. Chem. Eng. 46, 277 van der Waals, J.D., Proc. Kon. Ned. Akad. Wetenschap. 13,
(29)
(1968).
107 (1910).
(4) Coates, J.E., Davies, R.H., J. Chem. Soc. 1950, 1194.
(30) Van Itterbeek, A., Verbeke, O., Physica 26, 931 (1960).
(5) Cohen, Ernst, Buij, J.S., Z. Physik. Chem. 35B, 270 (1937).
(31) Yen, L.C., Woods, S.S., [Link].E. J. 12, 95 (1966).
(6) Costello, J.M., Bowden, S.T., Rec. Trav. Chim. 78, 391 (1959).
(32) Young, S., Sci. Proc. Roy. Dublin Soc. 12, 374 (1910).
(7) Cragoe, C.S., Harper, D.R., Nat. Bur. Stand. Sci. Paper 420
(1921).
(8) Davenport, A.J.. Rowiinson, J.S.. Saville. G., Trans. Faraday
Soc. 62, 322 (1966).
(9) Gibson, R.E., Loeffler, O.H., J. Amer. Chem. Soc. 63, 898
(1941). Received for review July 14, 1969. Resubmitted January 19, 1970.
(10) Harmens, A., Chem. Eng. Sci. 20, 813 (1965). Accepted July 10, 1970.

Journal of Chemical and Engineering Data, Val. 15, No. 4, 1970 517