This paper describes a new statistical approach to the theory of multicomponent systems. A ‘conformal solution’ is defined as one satisfying the following conditions: (i) The mutual potential energy of a molecule of species
L
r
and one of species
L
s
at a distance
ρ
is given by the expression
u
rs
(ρ)
=
f
rs
u
00
(
g
rs
ρ
), where
u
00
is the mutual potential energy of two molecules of some reference species
L
0
at a distance
ρ
, and
f
rs
and
g
rs
are constants depending only on the chemical nature of
L
r
and
L
s
. (ii) If
L
0
is taken to be one of the components of the solution, then
f
rs
and
g
rs
are close to unity for every pair of components. (iii) The constant
g
rs
equals ½(
g
rr
+
g
ss
). From these assumptions it is possible to calculate
rigorously
the thermodynamic properties of a conformal solution in terms of those of the components and their interaction constants. The non-ideal free energy of mixing is given by the equation ∆*
G
=
E
0
ƩƩ
rs
x
r
x
s
d
rs
, where
E
0
equals
RT
minus the latent heat of vaporization of
L
0
,
x
r
is the mole fraction of
L
r
and
d
rs
denotes 2
f
rs
—
f
rr
—
f
ss
. This equation resembles that defining a regular solution, with the important difference that
E
0
is a measurable function of
T
and
p
, which makes it possible to relate the free energy, entropy, heat and volume of mixing to the thermodynamic properties of the reference species; and the predicted relationships between these quantities agree well with available data on non-polar solutions. The theory makes no appeal to a lattice model or any other model of the liquid state, and can therefore be applied both to liquids and to imperfect gases, and to two-phase two-component systems near the critical point.
