ON THE DIETERICI EQUATION OF STATE.

1229

In addition to quartz and the compounds KzSiz06and KHSizO6,the

following new compounds occur as solid phases : Potassium disilicate
monohydrate, K2Si20s.H20; potassium metasilicate, K2Si03; potassium
metasilicate hemihydrate, KzSi03.0.5Hz0; and potassium metasilicate mono-
hydrate, K&i03.H20.
The data obtained are presented by means of tables, and graphically
by means of curves and solid models. Curves are given showing the solu-
bility relations in the binary systems Hz0-KzSi03and H20-KzSiz05 (Fig. 7),
the isothermal polybaric saturation curves (Fig. 4)) the variation of pressure
with SiO2/K20ratio along the isotherms (Fig. 5), the isobaric polythermal
saturation curves (Pig. 13), the P-T curves of the various monovariant
systems (Fig. g), and the boundary curves of the different solid phases in
the ternary system (Pig. 8). In addition, photographs are given of the
solid models showing the variation in the composition of the saturated
solutions with temperature, under the corresponding 3-phase pressure
(Fig. 11), and the variation of the composition of the saturated solutions
with pressure, at the corresponding 3-phase temperatures (Fig. 12).
Brief mention is made of some of the theoretical relations governing the
equilibrium in binary and ternary systems containing a volatile component,
and a short discussion of the proper application of the term “solubility”
is given.
WASHINGTON, D. e.

[CONTRIBUTION FROM THE SCHOOL OF CHEMISTRY, UNIVERSITY OF MINNESOTA. ]

ON THE DIETERICI EQUATION OF STATE.
BY F. H. MACDOUGALL.
Received March 29, 1917.

In a recent article’ I discussed at some length the Dieterici equation

of state, p = -
RT and applied it to a number of substances,
v-b
among which may be mentioned isopentane, carbon dioxide and hydro-
gen. Various methods were developed of calculating the values of a
and b (two of the “constants” of the equation) below, at, and above
the critical temperature ; and formulas were deduced connecting the pres-
sure of saturated vapor and the latent heat of vaporization with the densi-
ties of liquid and saturated vapor.
In the present paper I desire to bring forward an additional method of
calculating a and further to urge strongly the advantages, both theoretical
and practical, which the Dieterici equation has over any equation of the
van der Waals type.
1 THISJOURNAL, 38,528-555 (1916).
1230 F. 11. MAcDOUGALL.

General Considerations.
It is a simple matter to modify the equation for an ideal gas so as to
indicate the effect produced by the intermolecular forces and by the
+
volume of the molecules. One can write ( p K) (v - b) = RT,where
K is the “cohesive” pressure and b is the volume correction, and this equa-
tion is doubtless in a general way perfectly correct, but the exact form
of the functions which express the dependence of K and b on the tempera-
ture and volume may still be unknown. As is well known, van der Waals
assumed that b is an absolute constant and that K is of the form a/v2,
where a is assumed to be a constant. The resulting equation known by
the name of van der Waals gives good results only for small pressures
and entirely false results when applied to the critical state. For instance,
RT
the value of -‘ deduced from van der Waals’ equation is 8/3, while the
PCVC
values experimentally found’ in a large number of cases are around 3.7.
It may well be that the chief cause of the failure of the van der Waals
equation, to be much more than qualitative in its application, is to be
sought in the form of the function assumed for the “cohesive” or “internal”
pressure. This view will be confirmed if we consider carefully some of
the results obtained by Boltzmann in his studies of the kinetic theory of
gases.
Let us consider, first of all, a mass of gas, made up of infinitesimal
molecules which, however, attract each other, the attraction being ap-
preciable only over extremely small distances. In the interior of the gas,
the forces of attraction will, on the average, balance, but in a thin layer
a t the boundary there will be a resultant force directed towards the in-
terior. Boltzmann has shown that the existence of such a field of force
will not affect the distribution of velocities or of kinetic energies, but i t
will affect the spatial distribution of the molecules, i. e., the density of the
gas2 A certain amount of work will be done by us if we drag a molecule
from the interior of the gas to the wall of the containing vessel. If A
represents the work done against the field of force in dragging N mole-
cules to the boundary, where N is the number of molecules in one gram-
molecular weight, then the pressure exerted on the wall is given by the
following expression :2
p = -RT e -A/RT
2,

If we introduce now a correction for the volume of the molecules, we

obtain p = - RT For low densities2 a t least we may set A pro-
v-b
portional to the density or A = a/v, finally obtaining in this way the
‘Young’s “Stoichiometry,” 228 (1908).
8Jeans, “Dynamical Theory of Gases,” 126 (1916).
 ON THE DIETERICI EQUATION OF STATE. 1231

Dieterici equation of state, p = -

RT This then is the correct
v-b
form of the equation of state; the “correction” on account of the “cohe-
sive” pressure is given by an exponential function. As is well known,
for low pressures and, therefore, large volumes, the Dieterici equation
degenerates into the van der Waals form; for when V is large compared

RT
U RT
with band - , we have p = - ( I
v-b
-
a
vRT
-) and this is approximately

the same as p = --
RT the van der Waals formula. The van der
v -b v2’
Waals equation cannot, therefore, be expected to hold at the critical point
where v is of the same order of magnitude as b and as a matter of fact,
it does not. The Dieterici equation in view of the considerations advanced
above, might well be expected to have a much wider range of validity.
And if, as I think I have shown to be the case, the Dieterici form is the
correct one, the Dieterici equation should hold over any range if we de-
termine a and b as volume and temperature functions. The kinetic theory
leads to the conclusion that a and b may be both volume and tempera-
ture functions and it will be worth while to determine the nature of these
functions, using the Dieterici equation as a basis. To attempt to express
the a and b of van der Waals’ equation as functions of volume or tempera-
ture seems to me clearly a waste of time and effort, since its deduction
from the Dieterici equation indicates that its validity is restricted to
cases where the volume of a gas is very large compared with the volume
of the molecules.
Variations of a and b with the Volume and Temperature.
Assuming a and b in Dieterici’s equation to be independent of the vol-
ume, we obtain the following relations a t the critical point:
ac ac RT, e*
V , = 2bc; T, = -* P C = -. __ - - = 3.695
4Rbc’ 4e2bc2’pcvc 2
These results are independent of whether a and b are functions of the
temperature or not.
RTc
Now the value of the ratio - thus obtained is very close to the one
PcVc
that actually holds for a large number of substances. We therefore in-
fer that in the case of these substances, a and b are actually independent
of the volume even when the volume is as small as corresponds to the
critical point. More accurately stated, the inference is that a t the critical
point, (e)
dvT
and T
)
@ are very small, if not zero, or they happen to be of
such a magnitude that they actually neutralize each other’s effect. There
is the possibility that at the critical point, a and b may have minimum or
1232 P. H . MAcDOUGALL.

maximum values in which case g)T and &)T would be equal to zero.
As regards the variation of a and b with the temperature, the general
conclusion reached in my previous paper was that b increases as we
approach the critical temperature, reaches a maximum and then decreases ;
a decreases always with rise in temperature.
Work of Bakker on Surface Films.
Any theoretical isotherm on a p - v diagram as given by the equation
of state is cut by a line of constant pressure in three points below the
critical temperature. If this pressure is that at which the liquid and vapor
are in equilibrium, two of the points, corresponding to the vdumes V I
and VZ, represent the volumes of liquid and of saturated vapor, respec-
tively, while the third volume, 213, is not realizable by the substance in a
homogeneous state. In my previous article on this subject, reasons
were given for the assumption that this volume v3 is equal to 2b (not
2 b J . An equivalent assumption is that the density of the substance
corresponding to v3 on the theoretical isotherm is the arithmetic mean
of the densities of liquid and saturated vapor. What seems to me to be
a confirmation of this view is found in the conclusions reached by Bakker'
in regard to the average density of the surface layer between liquid and
vapor. I became acquainted with the work of Bakker after my paper
was published. According to Bakker, the average density of the surface
layer is the mean of the densities of liquid and vapor. This result de-
pends, it should be noted, on certain assumptions which may not be strictly
correct. Now Bakker points out the analogy between the surface layer
and the homogeneous substance in a state corresponding to a point on
the theoretical isothermal close to the point we have indicated by v3. In
other words, when liquid and vapor are in equilibrium, we have three
specific volumes, namely, those of homogeneous liquid, homogeneous
vapor and of transition layer corresponding to the three values on the
theoretical isothermal, which we have denoted by vl, v2 and v3, If, then,
the average density of the surface layer is equal to that of the homo-
geneous substance in the unrealizable state corresponding to v3, Bakker's
conclusion is equivalent to the equation:
2
- = -1 + -,
1
and this, on the
vs v1 v2
basis of the Dieterici equation, as shown in my previous paper means
that v3 is equal to 2 b .
Reference List of Equations.
It will be convenient to set down for reference a number of mathematical
results taken from my previous paper on this subject.2
1 Z.phy's. Chem., 73, 664 (1910).

* THISJOURNAL, 38,529, 531,536, 539, 540 (1916).

 ON THE DlETERfCI EQUATION Ol? STilTG. 1233

The Dieterici equation, p = - R T e-a/vRT . . . . . . . . . . . . . . . . . . . .

(1)
v-b
V , = ab,; Tc = -;
a, pc = -
Relations at critical point, 4Rbc .. (2)
= 5' = 3.695
pcvc 2

Co-existing phases, ........................ (3)

Pressure of saturated vapor, p s = __

R T e-a/PbRT . . . . . . . . . . . . . . . . . (4)
b
Total heat of vaporization,

al = -
w J 2
--
R T log, v2--b 2v1v2 R T log, 3..
-- ... (6)
v2 -Vl VI-b VZ-V~ v1
RT
UZ = 2bRT loge- ............................ (7)
Pb
RT
a3 = vlRT log, .........................
Pb1- b)
RT
a4 = vzRT log, .......................
P(VZ b) -
Heat of Vaporization.
a a
It is usually stated that - - - measures the internal latent heat of
Vl v2
vaporization and this is correct if one uses the original van der Waals
equation. For, according to this equation, the total work done in evap-
orating at constant temperature would be
jwla RT dv = f ( p v); +=~ ( V-
Z vl> a -a- = L.
- +
Vl v2
a a
Hence, - - - is the internal heat of vaporization. But if we assume the
v1 vz
Dieterici equation, we obtain the following results.
Since
RT -a/wlRT = _ R e-a/vrRT 5 -E
r
Ps = - e , v1 vz =
V I -b V Z -b
V Z -b -
R T log, v1--b -
"
v - b dv = L,
~ jw,
a a
and hence, - -- measures the total latent heat of vaporization. In
v1 v2
I234 F. H. MAcDOUGALL.

applying the van der Waals equation in the above calculations, a is as-
sumed to be independent of the volume and in applying the Dieteric
equation, the same assumption is made in the case of b.
As a matter of fact, the values of L, obtained from the equation'
I4 = a/v1 - a/vz,
differ somewhat from those calculated by means of the thermodynamic
formula

The inference to be drawn from this is that b does vary slightly with the
volume when the latter is that of the substance in the liquid state. A
study of these deviations might lead to a better knowledge of b as a func-
tion of the volume.
New Method of Calculating a.
Assuming, however, that the variation of b with changes in volume is
negligible, we can deduce an additional method of calculating a and it
may be of interest to compare these new values which we shall represent
by as with those obtained in my previous paper and represented by U I ,
a2,us and a4 (see equations numbered 6 , 7, 8 and g of the present article).
Since L = (VZ - vJT !& from thermodynamics, and since we have
dT
L=u---
( il iz),we easily deduce
4%
vlvzT -
dT
as =
I have-carried out the calculations for isopentane, using the data as
given in my previous paper. The numerical results refer to I g. of iso-

values of *
pentane, pressures are given in mm. of mercury and volumes in cc. The

dT
were obtained by taking three corresponding values of ps

and T,and using the interpolation formula, log p = a log T - ;+c.

-
From this, the value of d9S was calculated for the intermediate pressure
dT
and temperature. By repeating this process on different sets of corre-
sponding values of p and d-p'
T,the values of dT from o o up to the critical
temperature were obtained. The values of a6 found from the equation
US = vlvzT 4%
- are given in the seventh column of Table I and agree very
dT
well with those of al, a,a8 and u4. At the critical temperature, @e value
of U S is much lower than that of any of the others. The explanation of
this divergence may be due to the fact, pointed out by Mills,2 that inter-
1 THISJOURNAL, 38, 542, 544 (1916).
2 J. Phys. Chenz., 9, 402 (1905).
 DIFFERENTIAL IODIMETRY. IV. I235

polation formulas give results for dP

-' at the critical point which are con-
dT
--
siderably too low. In fact, he finds that at the critical point dPs - 2R
-. dT vc
This result is in perfect agreement with the Dieterici equation, for if we
2R
substitute the value - for & ! in our formula a = vlvZT&' we obtain
v, dT dT
for the critical point, a, = 2vcRTc, a relation easily deducible from the
Dieterici equation.
TABLEI.-ISOPENTANE
UNIT MASS = IG.
f 0. d$s/dT. (11 X 104. MX 104. arX lo*. 0tX 104. M X 104.
0 11.17 4.713 4.704 4.708 2.223 4.292
20 21.02 4.583 4.582 4.582 4.450 4.217
40 35.65 4.461 4.466 4.467 4.434 4.175
60 55.67 4.354 4.353 4.354 4,310 4.115
80 81.90 4.248 4.236 4.242 3.987 4.071
Io0 114.5 4.150 4.120 4.135 3.747 4 . 6
I IO 133.9 4.093 4.059 4.075 3.727 3.970
I20 155.4 4.036 3.991 4.012 3.650 3.954
130 178.7 3-977 3.922 3.948 3.604 3.894
140 204.2 3.916 3.860 3.881 3.564 3.832
150 232.9 3.850 3.771 3.806 3.512 3,801
I 60 263.8 3.778 3.690 3.726 3.480 3.725
170 298.4 3.687 3.599 3.633 3.459 3.624
180 335.7 3.579 3.503 3.528 3.436 3.459
187.8 367.7 3.404 3.423 3.423 3.423 3.087
(Critical temperature)
Summary.
In the present paper, it is shown that the Dieterici equation of state is
of the correct form and that it degenerates into the equation of van der
Waals for a gas under low pressure. The equation of Dieterici might
therefore be expected to have a much wider range of validity than one
of the van der Waals type.
A new method of calculating a is described and the values so obtained
are compared with those calculated by other methods.
MINNEAPOLIS,
MI".

[CONTRIBUTION
FROM LABORATORY
THE CHEMICAL OF THE UNIVERSITY ]
OF WISCONSIN.

DIFFERENTIAL IODIMETRY 1V.-THE ANALYSIS OF PYRO-

LUSITE AND OTHER OXIDIZED MANGANESE ORES.
BY 0. .I BARNEBBY AND GBO.M. BISHOP.
Received April 2 5 . 1917.

In a previous paper1 a method has been described for the determina-

tion of the available oxygen in precipitated higher oxides of manganese
THISJOURNAL, 39, 607 (1917).