9.

2 Solid-Liquid Equilibria
The prediction of solubility of a solid in a liquid is important for design of separation
processes that utilize either preferential dissolution or crystallization as a route to purification
of a species. The presence of a dissolved solid in a liquid also changes the freezing (as well as
boiling) point of the latter, and its estimation is important for design of heat exchange
equipments such as crystallizers (and evaporators). A wide variety of phase behaviour has
been observed in systems which comprise mixtures of solids, mixture of a solid and a liquid
and mixtures comprised of a liquid and more than a single solid (see R. T. DeHoff,
Thermodynamics in Materials Science, chaps. 9 and 10, McGraw-Hill, New York, 1993).
This section is devoted to consideration of thermodynamic relations that allow computation
of the solubility of solids in liquids as a function of temperature. Such relations may also be
extended to estimate changes in freezing or boiling point of a liquid in presence of dissolved
solids.
The phase equilibria relations developed here are limited to the simplest case of
systems comprising a solid (1) and a liquid (2) (solvent). As with all other phase equilibria,
one starts with the basis of equality of the species fugacity in the solid (s) and liquid (l)
phases respectively.

fˆi L  fˆi S ..(9.37)

xi iL fi L  zi iS fi S ..(9.38)

Where, xi and zi are, respectively, the mole fractions of species ‘i’ in the liquid and solid

solutions. Equivalently,

xi iL  zi iS i ..(9.39)

Where,  i  f i S (T , P) / f i L (T , P) ..(9.40)

The right side of this equation, defines  i as the ratio of fugacities at the temperature T and
pressure P of the system. Note that for each pure species at its (say, normal) melting point
Tm,i we have: fi S (Tm,i , P)  f i L (Tm ,i , P) ..(9.41)

However at T  Tm ,i ; fi S (T , P )  fi L (T , P )

Thus, in general one needs to derive a suitable expression for  i defined by eqn. 9.40.

1
This may be done by through estimation of the enthalpy and entropy of fusion at the system
temperature T, with values of the enthalpy and entropy at the melting point at the system
pressure as datum. However, since the properties of solids are not dependent on pressure in a
significant way, for convenience of computation one can assume the reference point as the
normal melting point Tm ,i . Accordingly, we may write for either the solid or the solvent

species:
Tm T
HTfus   CPL dT  H Tfus
m
  CPS dT ..(9.42)
T Tm

 C  CPL  dT
fus fus S
Or: H T  H Tm  P ..(9.43)
Tm

In the same manner one may write:

S fus
 S fus

T
C S
P - CPL 
dT ..(9.44)
T Tm
Tm
T

At temperature Tm , GTfus
m
 0 , thus:

HTfus
m
 Tm STfus
m
0 ..(9.45)

fus
HTfus
Or: S Tm  m
..(9.46)
Tm
Thus, using eqn. 9.46 in 9.44 we have:

S fus

HTfus
m

T
C S
P  CPL 
dT ..(9.47)
T
Tm Tm
T

Further, at temperature T, we have:

GTfus  H Tfus  T STfus ..(9.48)

Therefore, substituting eqns. 9.43 and 9.47 in 9.48 and re-arranging we have:

HTfus T

  C  C
T
C
S
P  CPL 
G T
fus
 H fus
Tm T m

Tm
S
P
L
P  dT  T  T
dT ..(9.49)
Tm Tm

If we further make the simplifying assumption that C S

P  CPL  is constant over the

temperature range Tm to T one obtains:

 T  T 
GTfus  H Tfus    C P  CP  (T  Tm )  (CP  CP )T ln  
S L S L
1  ..(9.50)
m
 Tm   Tm 

2
 S fS

Now,  dGT  RT  d ln f
L fL

Or:
fS
GTfus  RT ln ..(9.51)
fL
Thus, comparing eqn. 9.50 and 9.51:

fS  T  T 
 HTfus    CP  CP  (T  Tm )  (CP  CP )T ln  
S L S L
RT ln L 1  ..(9.52)
f m
 Tm   Tm 
Therefore using eqn. 9.52 in eqn. 9.40, we have:
 T  T 
RT ln  H Tfus    C P  C P  (T  Tm )  (CP  CP )T ln  
S L S L
m 
1 ..(9.53)
 Tm   Tm 
Upon suitable rearrangement:
H Tfus 1 1  (CPS  CPL )  Tm T 
ln  m
    1   ln  ..(9.54)
R  T Tm  R  T Tm 

Now by eqn. 9.39:

xi iL  zi iS i ..(9.39)

With  i given by the general expression as in eqn. 9.54, a suitable activity coefficient model

may be chosen to represent the activity coefficients  iS and  iL in the solid and liquid phases,

respectively and then eqn. 9.39 may be solved to yield the phase compositions. We illustrate
the typical methodology of such calculations for a binary system.
Let us assume a solid (1) is to be solubilized in a liquid (2). The starting point is eqn.
9.39. However, in applying it we make the simplifying assumption that the liquid does not
“dissolve” in the solid phase, and so the latter phase is a pure component solid. This is a
reasonable approximation for many system of practical interest in the chemical industry. The
liquid phase, however, will be a binary with the solid transforming to a “liquid” phase and
then mixing with the liquid solvent. Thus, one may rewrite the phase equilibrium as:
f1S (T , P)  fˆ1L (T , P, x1 ) ..(9.55)

Now fˆ1L (T , P, x1 )  x1 1 (T , P, x1 ) f1L ..(9.56)

Thus from the last two equations, the equilibrium solubility of the solid ( x1 ) in the liquid is
given by:

3
 f1S 1
x1  ( L
)( ) ..(9.57)
f1  1 (T , P, x1 )
Since activity coefficients are functions of temperature and composition only:
 f S  1 
x1   1L    ..(9.58)
 f1    1 (T , x1 ) 
It follows that if the temperature at which the solubility is desired is the melting point Tm of

the solid, then f1S (Tm )  f1L (Tm ).

Therefore, we have:
1
x1  ..(9.59)
 1 (Tm , x1 )
However, for any temperature T (other than Tm ) eqn. 9.58 needs to be employed. Thus, we

have:
 fS
ln x1   ln  1  ln  1L  ..(9.60)
 f1 
Or using eqns. 9.40 and 9.52 (or 9.53) we have:
 H Tfus  1 1  (CPS  CPL )  Tm T 
ln x1   ln  1   m
   1   ln   ..(9.61)
 R  T Tm  R  T Tm  

One may solve eqn. 9.61 (iteratively) for the solubility x1 , by assuming a suitable activity

coefficient model for the liquid phase (for ex: van Laar, Regular Solution, NRTL, etc). In the
special case where the temperature of interest T is close to the melting point Tm so that one

can make the approximation T Tm  1 eqn. 9.61 reduces to:

 H Tfus  1 1 
ln x1   ln  1   m
   ..(9.62)
 R  T Tm  
For the purpose of further illustration of the results that typically obtain from application of
the foregoing phase equilibria relations, we consider two limiting or special cases for a binary
system assuming that eqn. 9.62 applies. It may be noted, however, that this assumption is not
essential to the considerations below. Using eqn. 9.54 we write:
H Tfus 1 1 
ln  m
   ..(9.63)
R  T Tm 

Or, for any of the species:

4
  H fus  T  Tm ,i 
 i  exp  i   ..(9.64)
 RTm ,i  T  

The limiting cases are:

1. Both phases form ideal solutions, i.e.,  iL  1 and  iS  1 for all temperature and
compositions.
2. The liquid phase behaves as an ideal solution (  iL  1 ), and all species in the solid

state are completely immiscible, i.e., ( zi  iS  1 )

Case 1
Writing eqn. 9.39 for both species, we have:
x1  z1 1 ..(9.65)

x2  z2 2 ..(9.66)
Further:
x1  x2  1 ..(9.67)

z1  z2  1 ..(9.68)

Using eqns. 9.57 – 9.60, one obtains:

 1 1   2 
x1  ..(9.69)
 1  2
1  2
z1  ..(9.70)
 1  2
If one uses the simplified form for  , then by eqn. 9.64:

 H fus  T  Tm ,1 
 1  exp  1   ..(9.71)
 RTm,1  T 
 H 2fus  T  Tm ,2  
 2  exp    ..(9.72)
 RTm ,2  T  
On putting x1  z1  1 it follows that  1  1; thus, by eqn. 9.71, T  Tm ,1. In the same manner

if we put x2  z2  1 then 2  1; and T  Tm ,2 . On choosing a series of temperature T

between Tm ,1 and Tm ,2 and for each case solving for xi and zi one obtains a T  x  z plot that

is shown schematically in fig. 9.4. The upper curve in fig. 9.4 is the freezing curve and the

5
lower curve is the melting curve. Any T  x point within the lens-shaped region between the
two curves corresponds to a two phase situation, while points outside the region depict a
single phase. If a point is above the upper curve the state of the system is a homogeneous
binary liquid mixture, whereas if the point lies beneath the melting curve, the state of the
system is binary solid mixture. Any straight line parallel to the composition axis depicts a tie
line (for example XY), the compositions at the two ends signifying the liquid and solid phase
compositions. This behaviour is analogous to that of Raoult’s law T  x  y plot shown in
fig.7.7.

Figure 9.4 Case I, ideal liquid and solid solutions

An example of a system exhibiting a phase diagram of this type is a binary mixture of

nitrogen/carbon monoxide at low temperatures. However, as may be evident, this phase
behaviour is necessarily an idealized one. Its utility lies in that it can provide a reference
system for interpreting the behviour of more complex, real systems.

Case 2
For this case the liquid phase forms ideal solution (  il  1 ), and there is complete

immiscibility for all species in the solid state ( z i  is  1 ). Thus the governing phase equilibria

relations are:
x1   1 ..(9.73)

x2   2 ..(9.74)

6
Since,  1 and  2 are functions of temperature alone (by eqns.9.71 and 9.72), it follows that

x1 and x2 likewise are functions of temperature only. Thus eqns. 9.73 and 9.74 may be solved

independently upon which one obtains two distinct plots on the T  x1 diagram as shown in

fig. 9.5. However, if eqns. 9.64 and 9.65 apply simultaneously, then we have: 1   2  1 and

hence x1  x2  1. Thus, it follows that under this condition the following relation holds:

Fig. 9.5 T-x-z plot for Case II

 H 1fus  T  Tm ,1    H 2fus  T  Tm ,2  
exp      exp  RT  T    1 (9.75)
 RTm ,1  T    m ,2   

On solving the above equation one obtains the Eutectic temperature Te shown in fig. 9.5. This

particular phase diagram is to be interpreted as follows. When eqn. 9.73 applies one has:

 H1fus  T  Tm1 
x1  exp    ..(9.76)
 RTm1  T  

This equation is valid over the temperature range T  Tm1 , and the corresponding composition

range x1e  x  1, where x1  x1e the eutectic composition. Thus in this region (I) of the phase

diagram a liquid solution can be in equilibrium with only pure species 1 as a solid phase. This
is represented by region I on fig. 9.5 where liquid solutions with compositions x1 given by
line YE are in equilibrium with pure solid 1. Consider the system to be initially at a state

7
defined by point ‘M’. If one cools the system it eventually reaches the point ‘N’ on the T  x1

plot, at which condition only the component ‘1’ just begins to freeze. This happens till one
reaches the eutectic temperature Te at which point component ‘2’ begins to co-precipitate. In
the same manner region II obtains by solving eqn. 9.74, i.e.,

 H 2sl  T  Tm2  
x2  exp      1  x1 ..(9.77)
RT
 m2  T  

This equation is valid only from T  Tm 2 , where, x1  0 , to T  Te , where x1  x1e , the

eutectic composition. Equation 9.77 therefore applies where a liquid solution is in

equilibrium with pure species 2 as a solid phase. This is represented by region II on fig. 9.5
where liquid solutions with compositions x1 given by line XE are in equilibrium with pure

solid 2. Consider an initial system state defined by point ‘P’ which is a binary solution. On
cooling, the system eventually reaches the point ‘Q’ where only pure solid ‘2’ begins to
crystallize. On continued cooling, the eutectic temperature is reached, whereupon component
‘1’ also precipitates simultaneously.

Finally, if one commences cooling from the point R with a solution composition given
by x1  x1e , no precipitation of any solid occurs till the temperature Te (or point ‘E’) is

reached; at which state, as mentioned above, eqn. 9.75 holds. This is a state where two solids
are in equilibrium with the solution. Further cooling below Te results in co-precipitation of

both species, and the system enters a state where two immiscible solids are present.