Chapter 14 - Section B - Non-Numerical Solutions

14.2 Start with the equation immediately following Eq. (14.49), which can be modified slightly to read:

∂(nG R/RT ) ∂(n Z ) ∂ ln Z

ln φ̂i = − +n +1
∂n i ∂n i ∂n i
where the partial derivatives written here and in the following development without subscripts are
understood to be at constant T , n/ρ (or ρ/n), and n j . Equation (6.61) after multiplication by n can
be written:
nG R
ρ  3
ρ 2
= 2n(n B) + n 2 (nC) − n ln Z
RT n 2 n
Differentiate:
∂(nG R/RT )
ρ  3
ρ 2 ∂ ln Z
=2 (n B + n B̄i ) + (2n 2 C + n 2 C̄i ) − n − ln Z
∂n i n 2 n ∂n i

∂(nG R/RT ) 3 ∂ ln Z
or = 2ρ(B + B̄i ) + ρ 2 (2C + C̄i ) − n − ln Z
∂n i 2 ∂n i
   
∂(n B) ∂(nC)
By definition, B̄i ≡ and C̄i ≡
∂n i T,n j ∂n i T,n j

The equation of state, Eq. (3.40), can be written:

ρ 
ρ 2
Z = 1 + Bρ + Cρ 2 or n Z = n + n(n B) + n 2 (nC)
n n
∂(n Z )
ρ 
ρ 2
Differentiate: =1+ (n B + n B̄i ) + (2n 2 C + n 2 C̄i )
∂n i n n

∂(n Z )
or = 1 + ρ(B + B̄i ) + ρ 2 (2C + C̄i )
∂n i

When combined with the two underlined equations, the initial equation reduces to:

ln φ̂i = 1 + ρ(B + B̄i ) + 12 ρ 2 (2C + C̄i )

The two mixing rules are:

B = y12 B11 + 2y1 y2 B12 + y22 B22
C = y13 C111 + 3y12 y2 C112 + 3y1 y22 C122 + y23 C222
Application of the definitions of B̄i and C̄i to these mixing rules yields:

B̄1 = y1 (2 − y1 )B11 + 2y22 B12 − y22 B22

C̄1 = y12 (3 − 2y1 )C111 + 6y1 y22 C112 + 3y22 (1 − 2y1 )C122 − 2y23 C222
B̄2 = −y12 B11 + 2y12 B12 + y2 (2 − y2 )B22
C̄2 = −2y13 C111 + 3y12 (1 − 2y2 )C112 + 6y1 y22 C122 + 2y22 (3 − 2y2 )C222

709
 In combination with the mixing rules, these give:

B + B̄1 = 2(y1 B11 + y2 B12 )

2C + C̄1 = 3(y12 C111 + 2y1 y2 C112 + y22 C122 )
B + B̄2 = 2(y2 B22 + y1 B12 )
2C + C̄2 = 3(y22 C222 + 2y1 y2 C122 + y12 C112 )

In combination with the boxed equation these expressions along with Eq. (3.40) allow calculation of
ln φ̂1 and ln φ̂2 .
φisat
14.11 For the case described, Eqs. (14.1) and (14.2) combine to give: yi P = xi Pi sat
φ̂i

φisat
If the vapor phase is assumed an ideal solution, φ̂i = φi , and yi P = xi Pi sat
φi

When Eq. (3.38) is valid, the fugacity coefficient of pure species i is given by Eq. (11.36):
Bii P Bii Pi sat
ln φi = and φisat =
RT RT
φisat Bii Pi sat Bii P Bii (Pi sat − P)
Therefore, ln = ln φisat − ln φi = − =
φi RT RT RT

For small values of the final term, this becomes approximately:

φisat Bii (Pi sat − P)
=1+
φi RT
 
Bii (Pi sat − P)
Whence, yi P = xi Pi sat
1+
RT

xi Pi sat Bii (Pi sat − P)

or yi P − xi Pi sat =
RT
Write this equation for species 1 and 2 of a binary mixture, and sum. This yields on the left the
difference between the actual pressure and the pressure given by Raoult’s law:
x1 B11 P1sat (P1sat − P) + x2 B22 P2sat (P2sat − P)
P − P(RL) =
RT
Because deviations from Raoult’s law are presumably small, P on the right side may be replaced by
its Raoult’s-law value. For the two terms,

P1sat − P = P1sat − x1 P1sat − x2 P2sat = P1sat − (1 − x2 )P1sat − x2 P2sat = x2 (P1sat − P2sat )

P2sat − P = P2sat − x1 P1sat − x2 P2sat = P2sat − x1 P1sat − (1 − x1 )P2sat = x1 (P2sat − P1sat )

Combine the three preceding equations:
x1 x2 B11 (P1sat − P2sat )P1sat − x1 x2 B22 (P1sat − P2sat )P2sat
P − P(RL) =
RT
x1 x2 (P1sat − P2sat )
= (B11 P1sat − B22 P2sat )
RT

710
 Rearrangement yields the following:
 
x1 x2 (P1sat − P2sat )2 B11 P1sat − B22 P2sat
P − P(RL) =
RT P1sat − P2sat
 
x1 x2 (P1sat − P2sat )2 (B11 − B22 )P2sat
= B11 +
RT P1sat − P2sat
   
x1 x2 (P1sat − P2sat )2 B22 P2sat
= (B11 ) 1 + 1 −
RT B11 P1sat − P2sat
Clearly, when B22 = B11 , the term in square brackets equals 1, and the pressure deviation from the
Raoult’s-law value has the sign of B11 ; this is normally negative. When the virial coefficients are not
equal, a reasonable assumption is that species 2, taken here as the ”heavier” species (the one with
the smaller vapor pressure) has the more negative second virial coefficient. This has the effect of
making the quantity in parentheses negative and the quantity in square brackets < 1. However, if this
latter quantity remains positive (the most likely case), the sign of B11 still determines the sign of the
deviations.

14.13 By Eq. (11.90), the definition of γi , ln γi = ln fˆi − ln xi − ln f i

d ln γi d ln fˆi 1 1 d fˆi 1
Whence, = − = −
d xi d xi xi fˆi d xi xi

1 d fˆi
Combination of this expression with Eq. (14.71) yields: >0
fˆi d xi

d fˆi
Because fˆi ≥ 0, >0 (const T, P)
d xi

dµi d ln fˆi RT d fˆi

By Eq. (11.46), the definition of fˆi , = RT =
d xi d xi fˆi d xi

dµi
Combination with Eq. (14.72) yields: >0 (const T, P)
d xi

14.14 Stability requires that G < 0 (see Pg. 575). The limiting case obtains when G = 0, in which
event Eq. (12.30) becomes:
G E = −RT  xi ln xi
i
For an equimolar solution xi = 1/N where N is the number of species. Therefore,
1 1 1
G E (max) = −RT  ln = RT  ln N = RT ln N
i N N i N

For the special case of a binary solution, N = 2, and G E (max) = RT ln 2

711
 GE δ12 P
14.17 According to Pb. 11.35, G E = δ12 P y1 y2 or = y1 y2
RT RT

GE
This equation has the form: = Ax1 x2
RT
for which it is shown in Examples 14.5 and 14.6 that phase-splitting occurs for A > 2. Thus, the
formation of two immiscible vapor phases requires: δ12 P/RT > 2.
Suppose T = 300 K and P = 5 bar. The preceding condition then requires: δ12 > 9977 cm3 mol−1
for vapor-phase immiscibility. Such large positive values for δ12 are unknown for real mixtures.
(Examples of gas/gas equilibria are known, but at conditions outside the range of applicability of the
two-term virial EOS.)
GE
14.19 Consider a quadratic mixture, described by: = Ax1 x2
RT
It is shown in Example 14.5 that phase splitting occurs for such a mixture if A > 2; the value of
A = 2 corresponds to a consolute point, at x1 = x2 = 0.5. Thus, for a quadratic mixture,
phase-splitting obtains if:
1 1
G E > 2 · · · RT = 0.5RT
2 2
This is a model-dependent result. Many liquid mixtures are known which are stable as single phases,
even though G E > 0.5RT for equimolar composition.
14.21 Comparison of the Wilson equation, Eq. (12.18) with the modified Wilson equation shows that
(G E/RT )m = C(G E/RT ), where subscript m distinguishes the modified Wilson equation from
the original Wilson equation. To simplify, define g ≡ (G E/RT ); then
∂(ngm ) ∂(ng)
gm = Cg ngm = Cng =C ln(γ1 )m = C ln γ1
∂n 1 ∂n 1
where the final equality follows from Eq. (11.96). Addition and subtraction of ln x1 on the left side
of this equation and of C ln x1 on the right side yields:

ln(x1 γ1 )m − ln x1 = C ln(x1 γ1 ) − C ln x1

or ln(x1 γ1 )m = C ln(x1 γ1 ) − (C − 1) ln x1

d ln(x1 γ1 )m d ln(x1 γ1 ) C − 1
Differentiate: =C −
d x1 d x1 x1

As shown in Example 14.7, the derivative on the right side of this equation is always positive. How-
ever, for C sufficiently greater than unity, the contribution of the second term on the right can make
d ln(x1 γ1 )M
<0
d x1
over part of the composition range, thus violating the stability condition of Eq. (14.71) and implying
the formation of two liquid phases.
14.23 (a) Refer to the stability requirement of Eq. (14.70). For instability, i.e., for the formation of two
liquid phases,
d 2 (G E /RT ) 1
2
<−
d x1 x1 x2

712
 over part of the composition range. The second derivative of G E must be sufficiently negative so
as to satisfy this condition for some range of x1 . Negative curvature is the norm for mixtures for
which G E is positive; see, e.g., the sketches of G E vs. x1 for systems (a), (b), (d), (e), and (f ) in
Fig. 11.4. Such systems are candidates for liquid/liquid phase splitting, although it does not in
fact occur for the cases shown. Rather large values of G E are usually required.
(b) Nothing in principle precludes phase-splitting in mixtures for which G E < 0; one merely re-
quires that the curvature be sufficiently negative over part of the composition range. However,
positive curvature is the norm for such mixtures. We know of no examples of liquid/liquid phase-
splitting in systems exhibiting negative deviations from ideal-solution behavior.

14.29 The analogy is Raoult’s law, Eq. (10.1), applied at constant P (see Fig. 10.12): yi P = xi Pi sat
If the vapor phase in VLE is ideal and the liquid molar volumes are negligible (assumptions inherent
in Raoult’s law), then the Clausius/Clapeyron equation applies (see Ex. 6.5):

d ln Pi sat Hilv
=
dT RT 2
Integration from the boiling temperature Tbi at pressure P (where Pi sat = P) to the actual temperature
T (where Pi sat = Pi sat ) gives:
 T
P sat Hilv
ln i = 2
dT
P Tbi RT

Combination with Eq. (10.1) yields:

 T
Hilv
yi = xi exp dT
Tbi RT 2

which is an analog of the Case I SLE equations.

14.30 Consider binary (two-species) equilibrium between two phases of the same kind. Equation (14.74)
applies:
β β
xiα γiα = xi γi (i = 1, 2)
β β
If phase β is pure species 1 and phase α is pure species 2, then x1 = γ1 = 1 and x2α = γ2α = 1.
β β β β
Hence, x1α γ1α = x1 γ1 = 1 and x2α γ2α = x2 γ2 = 1
The reasoning applies generally to (degenerate) N -phase equilibrium involving N mutually immis-
cible species. Whence the cited result for solids.

14.31 The rules of thumb are based on Case II binary SLE behavior. For concreteness, let the solid be pure
species 1 and the solvent be liquid species 2. Then Eqs. (14.93) and (14.92a) apply:
 
H1sl T − Tm 1
x1 = ψ1 = exp
RTm 1 T

d x1 H1sl
(a) Differentiate: = ψ1 ·
dT RT 2
Thus d x1 /dT is necessarily positive: the solid solubility x1 increases with increasing T .
(b) Equation (14.92a) contains no information about species 2. Thus, to the extent that Eqs. (14.93)
and (14.92a) are valid, the solid solubility x1 is independent of the identity of species 2.

713
 (c) Denote the two solid phases by subscripts A and B. Then, by Eqs. (14.93) and (14.92a), the
solubilities x A and x B are related by:
 
xA H sl (Tm B − Tm A )
= exp
xB RTm A Tm B

where by assumption, H Asl = H Bsl ≡ H sl

Accordingly, x A /x B > 1 if and only if TA < TB , thus validating the rule of thumb.
(d) Identify the solid species as in Part (c). Then x A and x B are related by:
 
xA (H Bsl − H Asl )(Tm − T )
= exp
xB RTm T

where by assumption, Tm A = Tm B ≡ Tm
Notice that Tm > T (see Fig. 14.21b). Then x A /x B > 1 if and only if H Asl < H Bsl , in
accord with the rule of thumb.

14.34 The shape of the solubility curve is characterized in part by the behavior of the derivative dyi /d P
(constant T ). A general expression is found from Eq. (14.98), y1 = P1sat P/F1 , where the enhance-
ment factor F1 depends (at constant T ) on P and y1 . Thus,
   
dy1 P1sat P1sat ∂ F1 ∂ F1 dy1
= − 2 F1 + +
dP P P ∂ P y1 ∂ y1 P d P
   
y1 ∂ ln F1 ∂ ln F1 dy1
=− + y1 +
P ∂ P y1 ∂ y1 P d P
 
∂ ln F1 1
y1 −
dy1 ∂ P y1 P
Whence, =   (A)
dP ∂ ln F1
1 − y1
∂ y1 P
This is a general result. An expression for F1 is given by Eq. (14.99):

φ1sat V1s (P − P1sat )

F1 ≡ exp
φ̂1 RT

From this, after some reduction:

   
∂ ln F1 ∂ ln φ̂1 Vs ∂ ln F1 ∂ ln φ̂1
=− + 1 and =−
∂P y1 ∂P RT ∂ y1 P ∂ y1
y1 P
 
∂ ln φ̂1 V1s 1
y1 − + −
∂P RT P
dy1 y1
Whence, by Eq. (A), = (B)
dP ∂ ln φ̂1
1 + y1
∂ y1
P

714
 This too is a general result. If the two-term virial equation in pressure applies, then ln φ̂1 is given by
Eq. (11.63a), from which:

∂ ln φ̂1 1 ∂ ln φ̂1 2y2 δ12 P

= (B11 + y22 δ12 ) and =−
∂P RT ∂ y1 RT
y1 P

 
V1s − B11 − y22 δ12 1
y1 −
dy1 RT P
Whence, by Eq. (B), =
dP 2y1 y2 δ12 P
1−
RT
The denominator of this equation is positive at any pressure level for which Eq. (3.38) is likely to be
valid. Hence, the sign of dy1 /d P is determined by the sign of the group in parentheses. For very low
pressures the 1/P term dominates and dy1 /d P is negative. For very high pressures, 1/P is small,
and dy1 /d P can be positive. If this is the case, then dy1 /d P is zero for some intermediate pressure,
and the solubility y1 exhibits a minimum with respect to pressure. Qualitatively, these features are
consistent with the behavior illustrated by Fig. 14.23. However, the two-term virial equation is only
valid for low to moderate pressures, and is unable to mimic the change in curvature and “flattening”
of the y1 vs. P curve observed for high pressures for the naphthalene/CO2 system.

14.35 (a) Rewrite the UNILAN equation:

m  
ln(c + Pes ) − ln(c + Pe−s )
n= (A)
2s
As s → 0, this expression becomes indeterminate. Application of l’Hôpital’s rule gives:
 
m Pes Pe−s
lim n = lim +
s→0 s→0 2 c + Pes c + Pe−s
 
m P P
= +
2 c+P c+P

mP
or lims→0 n =
c+P
which is the Langmuir isotherm.

dn
(b) Henry’s constant, by definition: k ≡ lim
P→0 d P
 
dn m es e−s
Differentiate Eq. (A): = −
dP 2s c + Pes c + Pe−s
   
m es e−s m es − e−s m
Whence, k= − = or k= sinh s
2s c c cs 2 cs

(c) All derivatives of n with respect to P are well-behaved in the zero-pressure limit:

dn m
lim = sinh s
P→0 dP cs

715
 d 2n m
lim = − 2 sinh 2s
P→0 d P 2 c s
d 3n 2m
lim 3
= 3 sinh 3s
P→0 d P c s
Etc.
Numerical studies show that the UNILAN equation, although providing excellent overall corre-
lation of adsorption data at low-to-moderate surface coverage, tends to underestimate Henry’s
constant.
14.36 Start with Eq. (14.109), written as:
 n
dn
ln(P/n) = − ln k + (z − 1) +z−1
0 n
With z = 1 + Bn + Cn + · · ·, this becomes:
2

3
ln(P/n) = − ln k + 2Bn + Cn 2 + · · ·
2
Thus a plot of ln(P/n) vs. n produces − ln k as the intercept and 2B as the limiting slope (for
n → 0). Alternatively, a polynomial curve fit of ln(P/n) in n yields − ln k and 2B as the first
two coefficients.
14.37 For species i in a real-gas mixture, Eqs. (11.46) and (11.52) give:
g
µi = i (T ) + RT ln yi φ̂i P
g
At constant temperature, dµi = RT d ln yi φ̂i P
g
With dµi = dµi , Eq. (14.105) then becomes:
a
− d + d ln P +  xi d ln yi φ̂i = 0 (const T )
RT i
For pure-gas adsorption, this simplifies to:
a
d = d ln P + d ln φ (const T ) (A)
RT
which is the real-gas analog of Eq. (14.107). On the left side of Eq. (A), introduce the adsorbate
compressibility factor z through z ≡ a/RT = A/n RT :
a dn
d = dz + z (B)
RT n
where n is moles adsorbed. On the right side of Eq. (A), make the substitution:
dP
d ln φ = (Z − 1) (C)
P
which follows from Eq. (11.35). Combination of Eqs. (A), (B), and (C) gives on rearrangement (see
Sec. 14.8):
n dn dP
d ln = (1 − z) − dz + (Z − 1)
P n P
which yields on integration and rearrangement:
 P  n 
dP dn
n = k P · exp (Z − 1) · exp (1 − z) +1−z
0 P 0 n
This equation is the real-gas analog of Eq. (14.109).

716
14.39 & 14.40 Start with Eq. (14.109). With z = (1 − bm)−1 , one obtains the isotherm:
 
bn
n = k P(1 − bn) exp − (A)
1 − bn
 
bn bn
For bn sufficiently small, exp − ≈1−
1 − bn 1 − bn
kP
Whence, by Eq. (A), n ≈ k P(1 − 2bn) or n≈
1 + 2bk P
which is the Langmuir isotherm.
With z = 1 + βn, the adsorption isotherm is: n = k P exp(−2βn)
from which, for βn sufficiently small, the Langmuir isotherm is again recovered.
Ad dP
14.41 By Eq. (14.107) with a = A/n, =n
RT P
The definition of ψ and its derivative are:
A Ad
ψ≡ and dψ =
RT RT
dP
Whence, dψ = n (A)
P
By Eq. (14.128), the Raoult’s law analogy, xi = yi P/Pi ◦ . Summation for given P yields:
 yi
 xi = P Pi ◦ (B)
i i

By general differentiation,
 yi  yi
d  xi = P d ◦ + dP (C)
i i
Pi i
Pi ◦

The equation, i x i = 1, is an approximation that becomes increasingly accurate as the solution
procedure converges. Thus, by rearrangement of Eq. (B),

 yi i xi 1
= =
i
Pi ◦ P P

With P fixed, Eq. (C) can now be written in the simple but approximate form:
dP
d  xi =
i P

Equation (A) then becomes:

 
dψ = n d  xi or δψ = n δ  xi
i i

where we have replaced differentials by deviations. The deviation in i xi is known, since the true
value must be unity. Therefore,
 yi
δ  xi = P −1
i i
Pi ◦

717
 1
By Eq. (14.132), n=
i (xi /ni◦)
Combine the three preceding equations:
 yi
P −1
i
Pi ◦
δψ =
i (xi /ni◦)
When xi = yi P/Pi ◦ , the Raoult’s law analogy, is substituted the required equation is reproduced:
 yi
P −1
i
Pi ◦
δψ =  yi
P
i
Pi ◦ n i◦

14.42 Multiply the given equation for G E/RT by n and convert all mole fractions to mole numbers:
nG E n1n2 n1n3 n2n3
= A12 + A13 + A23
RT n n n
Apply Eq. (11.96) for i = 1:
   
1 n1 1 n1 n2n3
ln γ1 = A12 n 2 − 2 + A13 n 3 − 2 − A23 2
n n n n n
= A12 x2 (1 − x1 ) + A13 x3 (1 − x1 ) − A23 x2 x3

Introduce solute-free mole fractions:

x2 x2 x3
x2 ≡ = and x3 =
x2 + x3 1 − x1 1 − x1
Whence, ln γ1 = A12 x2 (1 − x1 )2 + A13 x3 (1 − x1 )2 − A23 x2 x3 (1 − x1 )2
For x1 → 0, ln γ1∞ = A12 x2 + A13 x3 − A23 x2 x3
Apply this equation to the special case of species 1 infinitely dilute in pure solvent 2. In this case,
x2 = 1, x3 = 0, and
∞
ln γ1,2 = A12 Also ∞
ln γ1,3 = A13

Whence, ln γ1∞ = x2 ln γ1,2

∞
+ x3 ln γ1,3
∞
− A23 x2 x3
In logarithmic form the equation immediately following Eq. (14.24) on page 552 may be applied to
the several infinite-dilution cases:

ln H1 = ln f 1 + ln γ1∞ ∞
ln H1,2 = ln f 1 + ln γ1,2 ∞
ln H1,3 = ln f 1 + ln γ1,3

Whence, ln H1 − ln f 1 = x2 (ln H1,2 − ln f 1 ) + x3 (ln H1,3 − ln f 1 ) − A23 x2 x3

or ln H1 = x2 ln H1,2 + x3 ln H1,3 − A23 x2 x3

718
14.43 For the situation described, Figure 14.12 would have two regions like the one shown from α to β,
probably one on either side of the minimum in curve II.

V2
14.44 By Eq. (14.136) with V̄2 = V2 : = − ln(x2 γ2 )
RT
Represent ln γ2 by a Taylor series:
 
d ln γ2  1 d 2 ln γ2 
ln γ2 = ln γ2 |x1 =0 + x1 + x2 + · · ·
d x1 x1 =0 2 d x12 x1 =0 1

But at x1 = 0 (x2 = 1), both ln γ2 and its first derivative are zero. Therefore,
 
1 d 2 ln γ2
ln γ2 = 2
x12 + · · ·
2 d x1 x1 =0

x12 x3 x4
Also, ln x2 = ln(1 − x1 ) = −x1 − − 1 − 1 − ···
2 3 4
  
1 d 2 ln γ2
Therefore, ln(x2 γ2 ) = + ln x2 + ln γ2 = −x1 − 1
2
1− x12 − · · ·
2 d x12 x1 =0

  
V2 1 1 d 2 ln γ2
and =1+ 1− x1 + · · ·
x1 RT 2 2 d x12 x1 =0

  
1 1 d 2 ln γ2
Comparison with the given equation shows that: B= 1−
2 2 d x12 x1 =0

 
∂(G E/RT ) HE
14.47 Equation (11.95) applies: =−
∂T P,x RT 2
For the partially miscible system G /RT is necessarily ”large,” and if it is to decrease with increasing
E

T , the derivative must be negative. This requires that H E be positive.

φ̂i yi γi Pi sat φisat

14.48 (a) In accord with Eqs. (14.1) and (14.2), yi P = xi γi Pi sat ⇒ Ki ≡ = ·
φisat xi P φ̂i

K1 γ1 P1sat φ1sat φ̂2

α12 ≡ = · ·
K2 γ2 P2sat φ̂1 φ2sat

γ1∞ P1sat φ1 (P1sat ) φ2 (P2sat ) γ1∞ P1sat φ1 (P1sat )

(b) α12 (x1 = 0) = · · = · ∞ sat
P2sat φ̂1∞ (P2sat ) φ2 (P2 )
sat
P2sat φ̂1 (P2 )

P1sat φ1 (P1sat ) φ̂2∞ (P1sat ) P1sat φ̂2∞ (P1sat )

α12 (x1 = 1) = · · = ·
γ2∞ P2sat φ1 (P1sat ) φ2 (P2sat ) γ2∞ P2sat φ2 (P2sat )
The final fractions represent corrections to modified Raoult’s law for vapor nonidealities.

719
 (c) If the vapor phase is an ideal solution of gases, then φ̂i = φi for all compositions.
 
∂ ln γi H̄ E
14.49 Equation (11.98) applies: =− i2
∂ T P,x RT
Assume that H E and H̄iE are functions of composition only. Then integration from Tk to T gives:
    
γi (x, T ) H̄ E T
dT H̄iE 1 1 H̄ E T
ln = i = − =− i −1
γi (x, Tk ) R Tk T2 R T Tk RT Tk

  
H̄iE T
γi (x, T ) = γi (x, Tk ) · exp − −1
RT Tk
 
∂G E
14.52 (a) From Table 11.1, p. 415, find: = −S E = 0 and G E is independent of T .
∂T P,x

GE FR (x)
Therefore =
RT RT
 
∂(G E /RT ) HE GE
(b) By Eq. (11.95), =− =0 ⇒ = FA (x)
∂T P,x RT 2 RT

(c) For solutions exhibiting LLE, G E /RT is generally positive and large. Thus α and β are positive
for LLE. For symmetrical behavior, the magic number is A = 2:
A<2 homogeneous; A=2 consolute point; A>2 LLE
With respect to Eq. (A), increasing T makes G E /RT smaller. thus, the consolute point is an up-
per consolute point. Its value follows from:
α α
=2 ⇒ TU =
RTU 2R
The shape of the solubility curve is as shown on Fig. 14.15.

14.53 Why? Because they are both nontoxic, relatively inexpensive, and readily available. For CO2 , its
Tc is near room temperature, making it a suitable solvent for temperature-sensitive materials. It is
considereably more expensive than water, which is probably the cheapest possible solvent. However,
both Tc and Pc for water are high, which increases heating and pumping costs.

720