The Framework of Solution Thermodynamics
The Framework of Solution Thermodynamics
The Framework of Solution Thermodynamics
Our purpose in this chapter is to lay the theoretical foundation for applications of
thermodynamics to gas mixtures and liquid solutions. Throughout the chemical, energy,
microelectronics, personal care, and pharmaceutical industries, multicomponent fluid mixtures
undergo composition changes brought about by mixing and separation processes, the transfer
of species from one phase to another, and chemical reaction. Thus, measures of composition
become essential variables, along with temperature and pressure, which we already considered
in detail in Chap. 6. This adds substantially to the complexity of tabulating and correlating
thermodynamic properties, and leads to the introduction of a menagerie of new variables and
relationships among them. Applying these relationships to practical problems, such as phase
equilibrium calculations, requires that we first map out this “thermodynamic zoo.” Thus, in
the present chapter, we:
348
10.1. Fundamental Property Relation 349
Measures of Composition
The three most common measures of composition in thermodynamics are mass fraction, mole
fraction, and molar concentration. Mass or mole fraction is defined as the ratio of the mass or
number of moles of a particular chemical species in a mixture to the total mass or number of
moles of mixture:
m i m i n i n i
∙ ∙
x i ≡ ___
= ___
∙ or x i ≡ __
= __∙
m m n n
Molar concentration is defined as the ratio of the mole fraction of a particular chemical
species in a mixture or solution to the molar volume of the mixture or solution:
x i
C i ≡ __
V
This quantity has units of moles of i per unit volume. For flow processes, convenience sug-
gests its expression as a ratio of rates. Multiplying and dividing by molar flow rate n gives:
∙
n i
∙
C i ≡ __
q
where n iis molar flow rate of species i, and q is volumetric flow rate.
∙
ℳ ≡ ∑ x i ℳ i
i
In the present chapter, we develop the framework of solution thermodynamics using
mole fractions as composition variables. For nonreacting systems, virtually all of the same
development can be done using mass fractions, yielding identical definitions and equations.
Thus, we may take xi to represent either a mole fraction or mass fraction in nonreacting
systems. In reacting systems, it is nearly always best to work in terms of mole fractions.
Equation (6.7) relates the total Gibbs energy of any closed system to its canonical variables,
temperature and pressure:
d(nG)= (nV )dP − (nS)dT (6.7)
where n is the total number of moles of the system. It applies to a single-phase fluid in a closed
system wherein no chemical reactions occur. For such a system the composition is necessarily
constant, and therefore:
[ ∂ P ] [ ∂ T ]
∂ (nG) ∂ (nG)
_____
= nV and _____
= − nS
T, n P, n
The subscript n indicates that the numbers of moles of all chemical species are held constant.
350 CHAPTER 10 . The Framework of Solution Thermodynamics
For the more general case of a single-phase, open system, material may pass into and
out of the system, and nG becomes a function of the numbers of moles of the chemical species
present. It remains a function of T and P, and we can therefore write the functional relation:
nG = g(P, T, n 1, n 2, . . ., n i, . . .)
The summation is over all species present, and subscript nj indicates that all mole numbers
except the ith are held constant. The derivative in the final term is given its own symbol and
name. Thus, by definition the chemical potential of species i in the mixture is:
[ ∂ n i ]
∂ (nG)
μ i≡ _____
(10.1)
P, T, n j
With this definition and with the first two partial derivatives replaced by (nV) and −(nS), the
preceding equation becomes:
d(nG) = (nV)dP − (nS)dT + ∑ μ i dn i
(10.2)
i
Equation (10.2) is the fundamental property relation for single-phase fluid systems
of variable mass and composition. It is the foundation upon which the structure of solution
thermodynamics is built. For the special case of one mole of solution, n = 1 and ni = xi:
dG = VdP − SdT + ∑ μ i dx i
(10.3)
i
Implicit in this equation is the functional relationship of the molar Gibbs energy to its canonical
variables, here: T, P, and {xi}:
G = G(T, P, x 1, x 2, . . ., x i, . . .)
Equation (6.11) for a constant-composition solution is a special case of Eq. (10.3).
Although the mole numbers ni of Eq. (10.2) are independent variables, the mole fractions xi in
Eq. (10.3) are not, because ∑ i x i = 1.This precludes certain mathematical operations which
depend upon independence of the variables. Nevertheless, Eq. (10.3) does imply:
Other solution properties come from definitions; e.g., the enthalpy, from H = G + TS. Thus,
by Eq. (10.5),
( ∂ T )P, x
∂ G
H = G − T ___
10.2. The Chemical Potential and Equilibrium 351
Practical applications of the chemical potential will become clearer in later chapters that treat
chemical and phase equilibria. However, at this point one can already appreciate its role in
these analyses. For a closed, single-phase PVT system containing chemically reactive species,
Eqs. (6.7) and (10.2) must both be valid, the former simply because the system is closed and
the second because of its generality. In addition, for a closed system, all differentials dni in
Eq. (10.2) must result from chemical reaction. Comparison of these two equations shows that
they can both be valid only if:
∑ μ i dn i= 0
i
This equation therefore represents a general criterion for chemical-reaction equilibrium
in a single-phase closed PVT system, and is the basis for the development of working equa-
tions for the solution of reaction-equilibrium problems.
With respect to phase equilibrium, we note that for a closed nonreacting system consist-
ing of two phases in equilibrium, each individual phase is open to the other, and mass transfer
between phases may occur. Equation (10.2) applies separately to each phase:
d(nG) α = (nV) α dP − (nS) α dT + ∑ μ iα dn iα
i
β β
d(nG) β = (nV) β dP − (nS) β dT + ∑ μ i dn i
i
where superscripts α and β identify the phases. For the system to be in thermal and mechanical
equilibrium, T and P must be uniform.
The change in the total Gibbs energy of the two-phase system is the sum of the equations
for the separate phases. When each total-system property is expressed by an equation of the form,
nM = (nM) α+ (nM) β
β β
the sum is: d(nG)= (nV)dP − (nS)dT + ∑ μ iα dn iα + ∑ μ i dn i
i i
Because the two-phase system is closed, Eq. (6.7) is also valid. Comparison of the two equa-
tions shows that at equilibrium:
β β
∑ μ
iα dn iα + ∑ μ
i dn i = 0
i i
352 CHAPTER 10 . The Framework of Solution Thermodynamics
β
The changes d n iα and d n i result from mass transfer between the phases; mass c onservation
therefore requires:
β β
dn iα = − dn i and ∑ (μ iα − μ i )dn iα = 0
i
A similar but more comprehensive derivation shows (as we have supposed) that for equilib-
rium T and P must be the same in all phases.
Thus, multiple phases at the same T and P are in equilibrium when the
chemical potential of each species is the same in all phases.
The application of Eq. (10.6) in later chapters to specific phase-equilibrium problems
requires models of solution behavior, which provide expressions for G and μi as functions of
temperature, pressure, and composition. The simplest of these, the ideal-gas state mixture and
the ideal solution, are treated in Secs. 10.4 and 10.8, respectively.
The definition of the chemical potential by Eq. (10.1) as the mole-number derivative of nG
suggests that other derivatives of this kind may prove useful in solution thermodynamics.
¯ i of species i in solution as:
Thus, we define the partial molar property M
[ ∂ n i ] P, T, n
∂ ( nM)
¯ i ≡ ______
M
(10.7)
j
subscript, and the symbol is Mi. In summary, the three kinds of properties used in solution
thermodynamics are distinguished using the following notation:
Solution properties M, for example : V, U, H, S, G
Partial properties M¯ i , for example : V¯ i , U¯ i , H¯ i , S¯ i, G¯ i
Pure-species properties M i, for example : V i, U i, H i, Si , G i
Comparison of Eq. (10.1) with Eq. (10.7) written for the Gibbs energy shows that the
chemical potential and the partial molar Gibbs energy are identical; i.e.,
μ i≡ G¯ i
(10.8)
Example 10.1
The partial molar volume is defined as:
[ ∂ n i
] P, T, n j
∂ (nV )
¯ i ≡ _____
V
(A)
Solution 10.1
Suppose an open beaker containing an equimolar mixture of ethanol and water
occupies a total volume nV at room temperature T and atmospheric pressure P.
Add to this solution a drop of pure water, also at T and P, containing Δnw moles,
and mix it thoroughly into the solution, allowing sufficient time for heat exchange
to return the contents of the beaker to the initial temperature. One might expect
that the volume of solution increases by an amount equal to the volume of the
water added, i.e., by VwΔnw, where Vw is the molar volume of pure water at T and
P. If this were true, the total volume change would be:
Δ(nV)= Vw Δ n w
However, experimental observations show that the actual volume change is somewhat
less. Evidently, the effective molar volume of water in the final solution is less
than the molar volume of pure water at the same T and P. We may therefore write:
Δ(nV)= V˜ w Δ n w
(B)
˜ wrepresents the effective molar volume of water in the final solution. Its
where V
experimental value is given by:
Δ( nV)
˜ w = ______
V (C)
Δ n w
In the process described, a drop of water is mixed with a substantial amount of solu-
tion, and the result is a small but measurable change in composition of the solution.
354 CHAPTER 10 . The Framework of Solution Thermodynamics
For the effective molar volume of the water to be considered a property of the original
equimolar solution, the process must be taken to the limit of an infinitesimal drop.
Whence, Δ n w→ 0,and Eq. (C) becomes:
Δ( nV) _____
d(nV)
˜ w = lim ______
V
=
Δn w→0 Δ n w dn w
Because T, P, and na (the number of moles of alcohol) are constant, this equation
is more appropriately written:
[ ∂ n w ]
∂ ( nV)
˜ w = _____
V
P, T, n a
¯ w
˜ wis the partial molar volume V
Comparison with Eq. (A) shows that in this limit V
of the water in the equimolar solution, i.e., the rate of change of the total solution
volume with nw at constant T, P, and na for a specific composition. Written for the
addition of dnw moles of water to the solution, Eq. (B) is then:
d(nV)= V¯ w dn w
(D)
When V¯ w is considered the molar property of water as it exists in solution, the total
volume change d(nV) is merely this molar property multiplied by the number of
moles dnw of water added.
If dnw moles of water is added to a volume of pure water, then we have every
reason to expect the volume change of the system to be:
d(nV)= Vw dn w (E)
where Vw is the molar volume of pure water at T and P. Comparison of Eqs. (D)
¯ w = Vw when the “solution” is pure water.
and (E) indicates that V
1Mere functionality does not make a set of variables into canonical variables. These are the canonical variables
only for M ≡ G.
10.3. Partial Properties 355
where subscript n indicates that all mole numbers are held constant, and subscript nj that all
mole numbers except ni are held constant. Because the first two partial derivatives on the
right are evaluated at constant n and because the partial derivative of the last term is given by
Eq. (10.7), this equation has the simpler form:
dn i= x i dn + n dx i
Moreover,
d(nM)= n dM + M dn
The terms containing n are collected and separated from those containing dn to yield:
Equation (10.10) is in fact just a special case of Eq. (10.9), obtained by setting n = 1,
which also makes ni = xi. Equations (10.11) and (10.12) on the other hand are new and vital.
Known as summability relations, they allow the calculation of mixture properties from partial
properties, playing a role opposite to that of Eq. (10.7), which provides for the calculation of
partial properties from mixture properties.
356 CHAPTER 10 . The Framework of Solution Thermodynamics
One further important equation follows directly from Eqs. (10.10) and (10.11).
Differentiation of Eq. (10.11), a general expression for M, yields a general expression for dM:
This equation must be satisfied for all changes occurring in a homogeneous phase. For the
important special case of changes in composition at constant T and P, it simplifies to:
Eq. 10.14 shows that the partial molar properties cannot all vary independently. This
constraint is analogous to the constraint on mole fractions, which are not all independent
because they must sum to one. Similarly, the mole-fraction-weighted sum of the partial molar
properties must yield the overall solution property (Eq. 10.11), and this constrains the varia-
tion in partial molar properties with composition.
For a partial property of a species that approaches its infinite-dilution limit, i.e., a par-
tial property value of a species as its mole fraction approaches zero, we can make no general
statements. Values come from experiment or from models of solution behavior. Because it is
an important quantity, we do give it a symbol, and by definition we write:
¯ i∞
M ≡ lim M¯ i
x i→0
[ ∂ n i ]
∂ ( nM)
¯ i ≡ ______
Definition: M
(10.7)
P, T, n j
( ) ( ∂ T )P, x
∂ M ∂ M
∑ x i dM¯ i = ___
Gibbs/Duhem: dP + ___
dT (10.13)
i ∂ P T, x
which shows that the partial properties of species making up a solution are not independent of
one another.
When M is known as a function of x1 at constant T and P, the appropriate form of the Gibbs/
Duhem equation is Eq. (10.14), expressed here as:
x 1 dM¯ 1 + x 2 dM¯ 2 = 0
(C)
Because x1 + x2 = 1, it follows that dx1 = −dx2. Eliminating dx2 in favor of dx1 in Eq. (B) and
combining the result with Eq. (C) gives:
dM
= M¯ 1 − M¯ 2
____
(D)
dx 1
Two equivalent forms of Eq. (A) result from the elimination separately of x1 and x2:
M = M¯ 1 − x 2(M¯ 1 − M¯ 2 ) and M = x 1(M¯ 1 − M¯ 2 ) + M¯ 2
358 CHAPTER 10 . The Framework of Solution Thermodynamics
dM dM
M¯ 1 = M + x 2 ____
(10.15) M¯ 2 = M − x 1 ____
(10.16)
dx 1 dx 1
Thus for binary systems, the partial properties are readily calculated directly from an expres-
sion for the solution property as a function of composition at constant T and P. The corre-
sponding equations for multicomponent systems are much more complex. They are given in
detail by Van Ness and Abbott.4
Equation (C), the Gibbs/Duhem equation, may be written in derivative forms:
dM¯ 1 dM¯ 2 ¯
dM 1 ¯
x 2dM 2
x 1 ____ ____
dx + x 2 dx = 0 (E)
____
dx = − ___ ____
x dx (F)
1 1 1 1 1
¯ 1 and M¯ 2 are plotted vs. x1, the slopes must be of opposite sign. Moreover,
Clearly, when M
dM¯ 1 dM¯ 2
lim ____
= 0 (Provided lim ____
is finite)
x 1→1 dx 1 x 1→1 dx 1
Similarly,
dM¯ 2 dM¯ 1
lim ____
= 0 (Provided lim ____
is finite )
x 2→1 dx 1 x 2→1 dx 1
Thus, plots of M ¯ 1 and M ¯ 2 vs. x1 become horizontal as each species approaches purity.
Finally, given an expression for M ¯ 1( x 1), integration of Eq. (E) or Eq. (F) yields an
expression for M ¯ 2 (x 1) that satisfies the Gibbs/Duhem equation. This means that expressions
cannot be specified independently for both M ¯ 1(x 1) and M¯ 2 (x 1).
Example 10.2
Describe a graphical interpretation of Eqs. (10.15) and (10.16).
Solution 10.2
Figure 10.1(a) shows a representative plot of M vs. x1 for a binary system. The
tangent line shown extends across the figure, intersecting the edges (at x1 = 1 and
x1 = 0) at points labeled I1 and I2. As is evident from the figure, two equivalent
expressions can be written for the slope of this tangent line:
dM M − I 2 dM
____ = _____
and ____
= I 1− I 2
dx 1 x 1 dx 1
4H. C. Van Ness and M. M. Abbott, Classical Thermodynamics of Nonelectrolyte Solutions: With Applications to
Constant T, P Constant T, P
M1
I1 M1
M
M
M2
I2
M2
0 1 0 1
x1 x1
(a) (b)
Figure 10.1: (a) Graphical construction of Example 10.2. (b) Infinite-dilution values of partial
properties.
The first equation is solved for I2; it combines with the second to give I1:
dM dM
I 2= M − x 1 ____
and I 1= M + (1 − x 1)____
dx 1 dx 1
Comparisons of these expressions with Eqs. (10.16) and (10.15) show that:
I 1= M¯ 1 and
I 2= M¯ 2
Thus, the tangent intercepts give directly the values of the two partial properties.
These intercepts of course shift as the point of tangency moves along the curve,
and the limiting values are indicated by the constructions shown in Fig. 10.1(b).
For the tangent line drawn at x1 = 0 (pure species 2), M ¯ 2 = M 2, and at the opposite
¯ 1 = M
intercept, M ¯ 1 ∞ . Similar comments apply to the tangent drawn at x1 = 1 (pure
species 1). In this case M ¯ 1 = M 1 and M¯ 2 = M¯ ∞ .
2
Example 10.3
The need arises in a laboratory for 2000 cm3 of an antifreeze solution consisting of
30 mol-% methanol in water. What volumes of pure methanol and of pure water at
25°C must be mixed to form the 2000 cm3 of antifreeze, also at 25°C? Partial molar vol-
umes for methanol and water in a 30 mol-% methanol solution and their pure-species
molar volumes, both at 25°C, are:
360 CHAPTER 10 . The Framework of Solution Thermodynamics
Solution 10.3
The summability relation, Eq. (10.11), is written for the molar volume of the
binary antifreeze solution, and known values are substituted for the mole fractions
and partial molar volumes:
V = x 1 V¯ 1 + x 2 V¯ 2 = (0.3)(38.632)+ (0.7)(17.765)= 24.025 cm 3⋅mol −1
Because the required total volume of solution is Vt = 2000 cm3, the total num-
ber of moles required is:
V t 2000
n = ___
= ______
= 83.246 mol
V 24.025
Of this, 30% is methanol, and 70% is water:
n 1= (0.3)(83.246)= 24.974 n 2= (0.7)(83.246)= 58.272 mol
1t = n i V i; thus,
The volume of each pure species is V
V 1t = (24.974)(40.727)= 1017 cm 3 V 2t = (58.272)(18.068)= 1053 cm 3
Example 10.4
The enthalpy of a binary liquid system of species 1 and 2 at fixed T and P is repre-
sented by the equation:
where H is in J·mol–1. Determine expressions for H ¯ 1 and H¯ 2 as functions of x1, numer-
ical values for the pure-species enthalpies H1 and H2, and numerical values for the
¯ 1∞ and H¯ 2 ∞ .
partial enthalpies at infinite dilution H
Solution 10.4
Replacing x2 by 1 – x1 in the given equation for H and simplifying gives:
H = 600 − 180 x 1− 20 x 13 (A)
and
dH
____
= − 180 − 60 x 12
dx 1
By Eq. (10.15),
dH
¯ 1 = H + x 2 ____
H
dx 1
10.3. Partial Properties 361
Then,
¯ 1 = 600 − 180x 1− 20x 3 − 180x 2− 60x 2 x 2
H 1 1
Application of the criterion of exactness, Eq. (6.13), yields the Maxwell relation,
where subscript n indicates constancy of all ni, and therefore of composition, and subscript nj
indicates that all mole numbers except the ith are held constant. We recognize the terms on the
right-hand side of these equations as the partial volume and partial entropy, and thus we can
rewrite them more simply as:
These equations allow us to calculate the effects of P and T on the partial Gibbs energy (or
chemical potential). They are the partial-property analogs of Eqs. (10.4) and (10.5). Many
additional relationships among partial properties can be derived in the same ways that rela-
tionships among pure species properties were derived in earlier chapters. More generally, one
can prove the following:
Every equation that provides a linear relation among thermodynamic
properties of a constant-composition solution has as its counterpart
an equation connecting the corresponding partial properties of each
species in the solution.
An example is based on the equation that defines enthalpy: H = U + PV. For n moles,
nH = nU + P(nV)
( ∂ T ) ( ∂ P )
∂ G¯ i ∂ G¯ i
dG¯ i = ___
dT + ___
dP
P, x T, x
This may be compared with Eq. (6.11). These examples illustrate the parallelism that exists
between equations for a constant-composition solution and the corresponding equations for
the partial properties of the species in solution. We can therefore write simply by analogy
many equations that relate partial properties.
10.4. The Ideal-Gas-State Mixture Model 363
Despite its limited ability to describe actual mixture behavior, the ideal-gas-state mixture
model provides a conceptual basis upon which to build the structure of solution thermody-
namics. It is a useful property model because it:
At the molecular level, the ideal-gas state represents a collection of molecules that do not
interact and occupy no volume. This idealization is approached for real molecules in the limit
of zero pressure (which implies zero density) because both the energies of intermolecular
interactions and the volume fraction occupied by the molecules go to zero with increasing
separation of the molecules. Although they do not interact with one another, molecules in
the ideal-gas state do have internal structure; it is differences in molecular structure that give
rise to differences in ideal-gas-state heat capacities (Sec. 4.1), enthalpies, entropies, and other
properties.
Molar volumes in the ideal-gas state are Vig = RT/P [Eq. (3.7)] regardless of the
nature of the gas. Thus for the ideal-gas state, whether of pure or mixed gases, the molar v olume
is the same for given T and P. The partial molar volume of species i in the ideal-gas-state
mixture is found from Eq. (10.7) applied to the volume; superscript ig denotes the ideal-gas
state:
[ ∂ n i ]
∂ (n V ig)
[ ∂ n i ] ( )
¯ iig = _______ ∂ (nRT / P) RT ∂ n RT
V
= ________
= ___ = ___
___
T, P, n P
∂
n
i n j P
T, P, n j j
where the final equality depends on the equation n = n i+ ∑ j n j. This result means that for the
ideal-gas state at given T and P the partial molar volume, the pure-species molar volume, and
the mixture molar volume are identical:
RT
¯ iig = V iig = V ig= ___
V (10.20)
P
We define the partial pressure of species i in the ideal-gas-state mixture (pi) as the
pressure that species i would exert if it alone occupied the molar volume of the mixture. Thus,5
y i RT
p i≡ _____
ig = y i P (i = 1, 2, . . ., N)
V
where yi is the mole fraction of species i. The partial pressures obviously sum to the total
pressure.
Because the ideal-gas-state mixture model presumes molecules of zero volume that
do not interact, the thermodynamic properties (other than molar volume) of the constituent
5Note that this definition does not make the partial pressure a partial molar property.
364 CHAPTER 10 . The Framework of Solution Thermodynamics
species are independent of one another, and each species has its own set of private properties.
This is the basis for the following statement of Gibbs’s theorem:
A partial molar property (other than volume) of a constituent species in
an ideal-gas-state mixture is equal to the corresponding molar property
of the species in the pure ideal-gas state at the mixture temperature but
at a pressure equal to its partial pressure in the mixture.
¯ i ≠ V¯ i by the equation:
This is expressed mathematically for generic partial property M
ig ig
Comparing this with Eq. (10.21), written for the entropy, yields:
¯ ig (T, P)= S ig (T, P)− R ln y i
S i i
or
¯ ig = S ig − R ln y i
S (10.23)
i i
ig
where S i is the pure-species value at the mixture T and P.
For the Gibbs energy in the ideal-gas-state mixture, Gig = Hig − TSig; the parallel rela-
tion for partial properties is:
¯ ig = H¯ ig − T S¯ ig
G i i i
or
Differentiation of this equation in accord with Eqs. (10.18) and (10.19) confirms the results
expressed by Eqs. (10.20) and (10.23).
The summability relation, Eq. (10.11), with Eqs. (10.22), (10.23), and (10.24) yields:
ig
H ig= ∑ y i H i (10.25)
i
ig
S ig= ∑ y i S i − R∑ y i ln y i (10.26)
i i
ig
G ig= ∑ y i G i + RT ∑ y i ln y i (10.27)
i i
ig
Equations analogous to Eq. (10.25) may be written for both C P and V ig. The former
appears as Eq. (4.7), but the latter reduces to an identity because of Eq. (10.20).
When Eq. (10.25) is written,
ig
H ig− ∑ y i H i = 0
i
the difference on the left is the enthalpy change associated with a process in which appropriate
amounts of the pure species at T and P are mixed to form one mole of mixture at the same T
and P. For the ideal-gas state, this enthalpy change of mixing is zero.
When Eq. (10.26) is rearranged as:
1
S ig− ∑ y iS i = R∑ y iln __
ig
i i
y i
the left side is the entropy change of mixing for the ideal-gas state. Because 1/yi > 1, this
quantity is always positive, in agreement with the second law. The mixing process is inher-
ently irreversible, so the mixing process must increase the total entropy of the system and
surroundings together. For ideal-gas-state mixing at constant T and P, using Eq. (10.25) with
an energy balance shows that no heat transfer will occur between the system and surroundings.
Therefore, the total entropy change of system plus surroundings is only the entropy change of
mixing.
ig ig
An alternative expression for the chemical potential μ
i results when G
i in Eq. (10.24)
is replaced by an expression giving its T and P dependence. This comes from Eq. (6.11)
written for the ideal-gas state at constant T:
RT
dG i = V i dP = ___
ig ig
dP = RT d ln P (const T)
P
Integration gives:
ig
G i = Γ i (T)+ RT ln P (10.28)
366 CHAPTER 10 . The Framework of Solution Thermodynamics
where the argument of the logarithm is the partial pressure. Application of the summabil-
ity relation, Eq. (10.11), produces an expression for the Gibbs energy for the ideal-gas-state
mixture:
As is evident from Eq. (10.6), the chemical potential μi provides the fundamental criterion
for phase equilibrium. This is true as well for chemical-reaction equilibria. However, it exhib-
its characteristics that discourage its direct use. The Gibbs energy, and hence μi, is defined
in relation to internal energy and entropy. Because absolute values of internal energy are
ig
unknown, the same is true for μi. Moreover, Eq. (10.29) shows that μ i approaches negative
infinity when either P or yi approaches zero. This is true not only for the ideal-gas state, but
for any gas. Although these characteristics do not preclude the use of chemical potentials, the
application of equilibrium criteria is facilitated by the introduction of the fugacity,7 a property
that takes the place of μi but does not exhibit its less desirable characteristics.
The origin of the fugacity concept resides in Eq. (10.28), valid only for pure species i in
the ideal-gas state. For a real fluid, we write an analogous equation that defines fi, the fugacity
of pure species i:
G i≡ Γ i (T)+ RT ln fi (10.31)
This new property fi , with units of pressure, replaces P in Eq. (10.28). Clearly, if Eq. (10.28)
is viewed as a special case of Eq. (10.31), then:
ig
f i = P (10.32)
6A dimensional ambiguity is evident with Eq. (10.28) and with analogous equations to follow in that P has units,
whereas ln P must be dimensionless. This difficulty is more apparent than real, because the Gibbs energy is always
expressed on a relative scale, absolute values being unknown. Thus in application only differences in Gibbs energy
appear, leading to ratios of quantities with units of pressure in the argument of the logarithm. The only requirement is
that consistency of pressure units be maintained.
7This quantity originated with Gilbert Newton Lewis (1875–1946), American physical chemist, who also
developed the concepts of the partial property and the ideal solution. See http://en.wikipedia.org/wiki/Gilbert N.
Lewis.
10.5. Fugacity and Fugacity Coefficient: Pure Species 367
and the fugacity of pure species i in the ideal-gas state is necessarily equal to its pressure.
Subtraction of Eq. (10.28) from Eq. (10.31), both written for the same T and P, gives:
fi
G i− G i = RT ln __
ig
P
ig ig
By the definition of Eq. (6.41), G i− G i is the residual Gibbs energy, G i ; thus,
fi
G iR = RT ln _
= RT ln ϕi (10.33)
P
where the dimensionless ratio fi/P has been defined as another new property, the fugacity
coefficient, given by symbol ϕi:
fi
ϕi ≡ _
(10.34)
P
These equations apply to pure species i in any phase at any condition. However, as a special
iR = 0, ϕi = 1, and Eq. (10.28) is
case they must be valid for the ideal-gas state, for which G
recovered from Eq. (10.31). Moreover, we may write Eq. (10.33) for P = 0 and combine it
with Eq. (6.45):
P→0 ( RT )
G R
lim ___
i = lim ln ϕi = J
P→0
As explained in connection with Eq. (6.48), the value of J is immaterial and is set equal to
zero. Whence,
(P)
fi
lim ln ϕi = lim ln _
= 0
P→0 P→0
and
fi
lim ϕi = lim __
= 1
P→0 P→0 P
The identification of ln ϕi with G iR / RTby Eq. (10.33) permits its evaluation by the inte-
gral of Eq. (6.49):
∫0
dP
(const T) (10.35)
P
Fugacity coefficients (and therefore fugacities) for pure gases are evaluated by this equation
from PVT data or from a volume-explicit equation of state.
For example, when the compressibility factor is given by Eq. (3.36), written here with
subscripts to indicate that it is applied to a pure substance:
B ii P
Z i− 1 = ____
RT
368 CHAPTER 10 . The Framework of Solution Thermodynamics
Because the second virial coefficient Bii is a function of temperature only for a pure
species, substitution into Eq. (10.35) gives:
and
B ii P
ln ϕi = ____
(10.36)
RT
By difference,
f v
G iv − G il = RT ln __il
f i
This equation applies to the change of state from saturated liquid to saturated vapor, at temper-
isat
ature T and at the vapor pressure P iv − G il = 0; therefore:
. According to Eq. (6.83), G
f i v = f i l = f isat
(10.39)
This equation, expressing the equality of fugacity coefficients, is an equally valid crite-
rion of vapor/liquid equilibrium for pure species.
8The word fugacity is based on a Latin root meaning to flee or escape, also the basis for the word fugitive. Thus
fugacity has been interpreted to mean “escaping tendency.” When the escaping tendency is the same for the two
phases, they are in equilibrium. When the escaping tendency of a species is higher in one phase than another, that
species will tend to transfer to the phase where its fugacity is lower.
10.5. Fugacity and Fugacity Coefficient: Pure Species 369
All terms are at the temperature of interest. Inspection reveals that cancellation of
numerators and denominators produces a mathematical identity.
Ratio (A) is the vapor-phase fugacity coefficient of pure vapor i at its vapor/liquid satu-
ration pressure, designated ϕ isat
. It is given by Eq. (10.35), written,
∫0
dP
sat
P
( ) (const T) (10.42)
As shown by Eq. (10.39), expressing the equality of liquid and vapor fugacities at
e quilibrium, ratio (B) is unity. Ratio (C) reflects the effect of pressure on the fugacity of pure
liquid i. This effect is generally small. The basis for its calculation is Eq. (6.11), integrated at
constant T to give:
∫ P i
G i− G isat
Another expression for this difference results when Eq. (10.31) is written for both Gi and G isat
;
subtraction then yields:
fi
= RT ln ____
G i− G isat
f isat
The two expressions for G i− G isat
are set equal:
Because V il , the liquid-phase molar volume, is a very weak function of P at temperatures well
below Tc, an excellent approximation is often obtained by taking V il to be constant at the value
for saturated liquid. In this case,
V il (P − P isat
)
fi = ϕ isat
exp ___________
P isat
(10.44)
RT
370 CHAPTER 10 . The Framework of Solution Thermodynamics
If Z iv is given by Eq. (3.36), the simplest form of the virial equation, then:
B ii P B ii P isat
Z iv − 1 = ____
= exp _______
and ϕ isat
RT RT
In the following example, data from the steam tables form the basis for calculation of
the fugacity and fugacity coefficient of both vapor and liquid water as a function of pressure.
Example 10.5
For H2O at a temperature of 300°C and for pressures up to 10,000 kPa (100 bar)
calculate values of fi and ϕi from data in the steam tables and plot them vs. P.
Solution 10.5
Equation (10.31) is written twice: first, for a state at pressure P; second, for a
low-pressure reference state, denoted by *, both for temperature T:
G i= Γ i (T ) + RT ln fi and G i* = Γ i (T ) + RT ln f i*
fi 1
ln __* = ___
( G − G i* )
f i RT i
By definition Gi = Hi − TSi and G i* = H i* − T S i* ; substitution gives:
f i R [ T ]
fi 1 _______ H i− H i*
ln __* = __
− (Si − S i* ) (A)
9John Henry Poynting (1852–1914), British physicist. See http://en.wikipedia.org/wiki/John Henry Poynting.
10.5. Fugacity and Fugacity Coefficient: Pure Species 371
The lowest pressure for which data at 300°C are given in the steam tables is
1 kPa. Steam at these conditions is for practical purposes in its ideal-gas state, for
which f i* = P *= 1 kPa. Data for this state provide the following reference values:
H i* = 3076.8 J⋅g −1 S i* =10.3450 J⋅g −1⋅K −1
Equation (A) may now be applied to states of superheated steam at 300°C for various
values of P from 1 kPa to the saturation pressure of 8592.7 kPa. For example, at
P = 4000 kPa and 300°C:
H i= 2962.0 J⋅g −1 Si = 6.3642 J⋅g −1⋅K −1
Values of H and S must be multiplied by the molar mass of water (18.015 g⋅mol−1)
to put them on a molar basis for substitution into Eq. (A):
[ ]
fi 18.015 ______________
2962.0 − 3076.8
ln ___
* = ______
− (6.3642 − 10.3450) = 8.1917
f 8.314 573.15
and
fi / f *= 3611.0
fi = (3611.0)( f *)= (3611.0)(1 kPa)= 3611.0 kPa
7 fisat 1.0
6 fi 0.9
i
5 0.8 sat
i
fi P
/kPa
4 0.7
3
3 0.6
Pisat
1
0 2 4 6 8 10
P 10 3/kPa
discontinuity in slope, and the curve then rises very slowly with increasing pres-
sure, indicating that the fugacity of liquid water at 300°C is a weak function of
pressure. This behavior is characteristic of a liquid at a temperature well below
its critical temperature. The fugacity coefficient ϕi decreases steadily from its
zero-pressure value of unity as the pressure rises. Its rapid decrease in the liquid
region is a consequence of the near-constancy of the fugacity itself.
The definition of the fugacity of a species in solution is parallel to the definition of the
pure-species fugacity. For species i in a mixture of real gases or in a solution of liquids, the
equation analogous to Eq. (10.29), the ideal-gas-state expression, is:
μ i≡ Γ i (T)+ RT ln fˆ i
(10.46)
374 CHAPTER 10 . The Framework of Solution Thermodynamics
This result combined with Eq. (10.50) and the identity μ i≡ G¯ i gives:
where by definition,
( RT ) RT
nG 1 nG
d _
≡ ___
d(nG ) − ____
2 dT
R T
In this equation d(nG) is eliminated by Eq. (10.2) and G is replaced by its definition,
H − TS. The result, after algebraic reduction, is:
G¯ i
( RT ) RT
nG nV nH
d _
= _
dP − _
2 dT + ∑ ___
d n i (10.54)
R T i
RT
All terms in Eq. (10.54) have the units of moles; moreover, in contrast to Eq. (10.2), the
enthalpy rather than the entropy appears on the right side. Equation (10.54) is a general
relation expressing nG/RT as a function of all of its canonical variables, T, P, and the mole
numbers. It reduces to Eq. (6.37) for the special case of 1 mol of a constant-composition
phase. Equations (6.38) and (6.39) follow from either equation, and equations for the other
thermodynamic properties then come from appropriate defining equations. Knowledge of
G/RT as a function of its canonical variables allows evaluation of all other thermodynamic
properties, and therefore implicitly contains complete property information. Unfortunately,
we do not have the experimental means by which to exploit this characteristic. That is, we
cannot directly measure G/RT as a function of T, P, and composition. However, we can obtain
complete thermodynamic information by combining calorimetric and volumetric data. In this
regard, an analogous equation relating residual properties proves useful.
Because Eq. (10.54) is general, it may be written for the special case of the ideal-gas state:
( RT )
G¯
ig
n G ig n V ig n H ig
d _
= ____
dP − ____
2 dT + ∑ ___
i d n i
RT R T i
RT
10.6. Fugacity and Fugacity Coefficient: Species in Solution 375
In view of the definitions of residual properties [Eqs. (6.41) and (10.50)], subtracting this
equation from Eq. (10.54) gives:
( RT ) RT
n G R n V R nH R G¯ R
d _
= _
dP − _
2 dT + ∑ ___
i d n i (10.55)
RT i
RT
Equation (10.55) is the fundamental residual-property relation. Its derivation from Eq. (10.2)
parallels the derivation in Chap. 6 that led from Eq. (6.11) to Eq. (6.42). Indeed, Eqs. (6.11)
and (6.42) are special cases of Eqs. (10.2) and (10.55), valid for 1 mol of a constant-composition
fluid. An alternative form of Eq. (10.55) follows by introduction of the fugacity coefficient as
given by Eq. (10.51):
( RT ) RT
n G R n V R nH R
d _
= _
dP − _
2 dT + ∑ ln ϕˆ i dn i (10.56)
RT i
Equations so general as Eqs. (10.55) and (10.56) are most useful for practical applica-
tion in restricted forms. Division of Eqs. (10.55) and (10.56), first, by dP with restriction to
constant T and composition, and second, by dT and restriction to constant P and composition
leads to:
[ ] [ ]
V R ∂ (G R / RT ) H R ∂ (G R / RT )
___ = __________
(10.57) ___ = − T __________
(10.58)
RT ∂ P RT ∂ T
T, x P, x
These equations are restatements of Eqs. (6.43) and (6.44) wherein the restriction of the deriv-
atives to constant composition is shown explicitly. They lead to Eqs. (6.46), (6.48), and (6.49)
for the calculation of residual properties from volumetric data. Moreover, Eq. (10.57) is the
basis for the direct derivation of Eq. (10.35), which yields fugacity coefficients from volumet-
ric data. It is through the residual properties that this kind of experimental information enters
into the practical application of thermodynamics.
In addition, from Eq. (10.56),
[ ]
∂ (n G R / RT )
ln ϕˆ i = __________
(10.59)
∂ n i P, T, n j
Example 10.6
Develop a general equation to calculate ln ϕˆ ivalues from compressibility-factor data.
376 CHAPTER 10 . The Framework of Solution Thermodynamics
Solution 10.6
For n mol of a constant-composition mixture, Eq. (6.49) becomes:
= (nZ − n) ___
n G
R P
∫0
____ dP
RT P
In accord with Eq. (10.59) this equation may be differentiated with respect to ni at
constant T, P, and nj to yield:
ln ϕˆ i = _
∫ 0 [ ∂ n i ] P, T, n P
P ∂ (nZ − n ) dP
___
j
∫0
dP
(10.60)
P
where yi and yj represent mole fractions in a gas mixture. The indices i and j identify species,
and both run over all species present in the mixture. The virial coefficient Bij characterizes bimo-
lecular interactions between molecules of species i and species j, and therefore B ij = B ji. The
double summation accounts for all possible bimolecular interactions.
For a binary mixture i = 1, 2 and j = 1, 2; the expansion of Eq. (10.61) then gives:
B = y 1 y 1B 11+ y 1y 2B 12+ y 2y 1B 21+ y 2y 2B 22
or
Two types of virial coefficients have appeared: B11 and B22, for which the successive sub-
scripts are the same, and B12, for which the two subscripts are different. The first type is a
pure-species virial coefficient; the second is a mixture property, known as a cross coefficient.
Both are functions of temperature only. Expressions such as Eqs. (10.61) and (10.62) relate
mixture coefficients to pure-species and cross coefficients. They are called mixing rules.
Equation (10.62) allows the derivation of expressions for l n ϕˆ 1 and ln ϕˆ 2for a binary gas
mixture that obeys Eq. (3.36). Written for n mol of gas mixture, it becomes:
nBP
nZ = n + ____
RT
Differentiation with respect to n1 gives:
[ ∂ n 1 ] [ ∂ n 1 ]
∂ (nZ )
¯ 1 ≡ _____ P ∂ (nB )
Z
= 1 + ___
______
P, T, n
RT T, n
2 2
_
RT ∫ 0 [ ∂ n 1 ] RT [ ∂ n 1 ] T, n
1 P ∂ (nB) P ∂ (nB )
ln ϕˆ 1 = ___
dP = ___
______
T, n 2 2
where the integration is elementary because B is not a function of pressure. All that remains is
evaluation of the derivative.
Equation (10.62) for the second virial coefficient may be written:
B = y 1 (1 − y 2 ) B 11 + 2 y 1 y 2 B 12 + y 2 (1 − y 1 ) B 22
= y 1 B 11 − y 1 y 2 B 11 + 2 y 1 y 2 B 12 + y 2 B 22 − y 1 y 2 B 22
or
B = y 1 B 11+ y 2 B 22+ y 1 y 2 δ 12 with δ 12≡ 2 B 12− B 11− B 22
Multiplying by n and substituting yi = ni/n gives,
n 1 n 2
nB = n 1 B 11+ n 2 B 22+ ____
δ
n 12
By differentiation:
[ ] ( n n 2) 2 12
∂ ( nB) 1 n 1
_____
= B 11+ __
− ___
n δ
∂ n 1 T, n 2
= B 11+ (1 − y 1) y 2 δ 12= B 11+ y 22 δ 12
Therefore,
P
ln ϕˆ 1 = ___
( B 11+ y 22 δ 12) (10.63a)
RT
Similarly,
P
ln ϕˆ 2 = ___
( B 22+ y 12 δ 12) (10.63b)
RT
378 CHAPTER 10 . The Framework of Solution Thermodynamics
Equations (10.63) are readily extended for application to multicomponent gas mixtures; the
general equation is:10
RT [ ]
P 1
ln ϕˆ k = ___
B kk+ _
∑ ∑ y i y j (2δ ik− δi j ) (10.64)
2 i j
where the dummy indices i and j run over all species, and
δ ik≡ 2 B ik− B ii− B kk δi j≡ 2B ij− B ii− B jj
with
δi i= 0, δ kk= 0, etc., and δ ki= δ ik, etc.
Example 10.7
Determine the fugacity coefficients as given by Eqs. (10.63) for nitrogen and methane
in a N2(1)/CH4(2) mixture at 200 K and 30 bar if the mixture contains 40 mol-% N2.
Experimental virial-coefficient data are as follows:
Solution 10.7
By definition, δ 12= 2 B 12− B 11− B 22. Whence,
30
(83.14 ) (200 ) [
ln ϕˆ 2 = ___________
− 105.0 + (0.4 ) 2(20.6 )]= − 0.1835
Whence,
Note that the second virial coefficient of the mixture as given by Eq. (10.62) is
B = −72.14 cm3⋅mol−1, and that substitution in Eq. (3.36) yields a mixture
compressibility factor, Z = 0.870.
10H. C. Van Ness and M. M. Abbott, Classical Thermodynamics of Nonelectrolyte Solutions: With Applications to
Example 10.8
Estimate from Eq. (10.67) a value for the fugacity of 1-butene vapor at 200°C and
70 bar.
380 CHAPTER 10 . The Framework of Solution Thermodynamics
Solution 10.8
At these conditions, with Tc = 420.0 K, Pc = 40.43 bar from Table B.1, we have:
Tr = 1.127 Pr = 1.731 ω = 0.191
By interpolation in Tables D.15 and D.16 at these conditions,
ϕ 0= 0.627 and ϕ 1= 1.096
Equation (10.67) then gives:
ϕ = (0.627) (1.096) 0.191= 0.638
and
f = ϕP = (0.638)(70)= 44.7 bar
A useful generalized correlation for ln ϕ results when the simplest form of the virial equation
is valid. Equations (3.57) and (3.59) combine to give:
Pr
Z − 1 = ___
( B 0+ ω B 1)
Tr
Substitution into Eq. (10.65) and integration yield:
Pr
ln ϕ = ___
( B 0+ ωB 1)
Tr
or
[ Tr ]
Pr
ϕ = exp _
( B 0+ ω B 1 ) (10.68)
This equation, used in conjunction with Eqs. (3.61) and (3.62), provides reliable values of ϕ
for any nonpolar or slightly polar gas when applied at conditions where Z is approximately
linear in pressure. Figure 3.13 again serves as a guide to its applicability.
Named functions HRB(TR,PR,OMEGA) and SRB(TR,PR,OMEGA) for the evaluation
of HR/RTc and SR/R by the generalized virial-coefficient correlation are described in
Sec. 6.4. Similarly, we introduce here a function named PHIB(TR,PR,OMEGA) for the
evaluation of ϕ11:
ϕ = PHIB(TR,PR,OMEGA)
It combines Eq. (10.68) with Eqs. (3.61) and (3.62) to evaluate the fugacity coefficient for
given reduced temperature, reduced pressure, and acentric factor. For example, the value of ϕ
for carbon dioxide at the conditions of Example 6.4, Step 3, is denoted as:
PHIB(0.963,0.203,0.224) = 0.923
11Sample programs and spreadsheets for the evaluation of these functions are available in the Online Learning
Center at http://highered.mheducation.com:80/sites/1259696529.
10.7. Generalized Correlations for the Fugacity Coefficient 381
Extension to Mixtures
The generalized correlation just described is for pure gases only. The remainder of this section
shows how the virial equation may be generalized to allow the calculation of fugacity coeffi-
cients ϕˆ i for species in gas mixtures.
The general expression for calculation of ln ϕˆ kfrom second-virial-coefficient data is
given by Eq. (10.64). Values of the pure-species virial coefficients Bkk, Bii, etc., are found
from the generalized correlation represented by Eqs. (3.58), (3.59), (3.61), and (3.62). The
cross coefficients Bik, Bij, etc., are found from an extension of the same correlation. For this
purpose, Eq. (3.59) is rewritten in the more general form:12
ˆ ij = B 0+ ω ij B 1
B (10.69a)
where
B ij P cij
ˆ ij ≡ ______
B
(10.69b)
RT cij
and B0 and B1 are the same functions of Tr as given by Eqs. (3.61) and (3.62). The combining
rules proposed by Prausnitz et al. for the calculation of ωij, Tcij, and Pcij are:
ω i+ ω j
ω ij= ______
(10.70) T cij= (T ci T cj ) 1/2(1 − ki j ) (10.71)
2
( )
1∕3 1∕3 3
V ci + V cj (10.74)
V cij= __________
2
In Eq. (10.71), kij is an empirical interaction parameter specific to a particular i-j molecular
pair. When i = j and for chemically similar species, kij = 0. Otherwise, it is a small positive
number evaluated from minimal PVT data or, in the absence of data, set equal to zero. When
i = j, all equations reduce to the appropriate values for a pure species. When i ≠ j, these equations
define a set of interaction parameters that, while they do not have any fundamental physical
significance, do provide useful estimates of fugacity coefficients in moderately non-ideal gas
mixtures. Reduced temperature is given for each ij pair by T rij≡ T / T cij. For a mixture, values of
Bij from Eq. (10.69b) substituted into Eq. (10.61) yield the mixture second virial coefficient B,
and substituted into Eq. (10.64) [Eqs. (10.63) for a binary] they yield values of ln ϕˆ i.
The primary virtue of the generalized correlation for second virial coefficients presented
here is simplicity; more accurate, but more complex, correlations appear in the literature.13
12J. M. Prausnitz, R. N. Lichtenthaler, and E. G. de Azevedo, Molecular Thermodynamics of Fluid-Phase
Equilibria, 3rd ed., pp. 133 and 160, Prentice-Hall, Englewood Cliffs, NJ, 1998.
13C. Tsonopoulos, AIChE J., vol. 20, pp. 263–272, 1974, vol. 21, pp. 827–829, 1975, vol. 24, pp. 1112–1115, 1978;
C. Tsonopoulos, Adv. in Chemistry Series 182, pp. 143–162, 1979; J. G. Hayden and J. P. O’Connell, Ind. Eng. Chem.
Proc. Des. Dev., vol. 14, pp. 209–216, 1975; D. W. McCann and R. P. Danner, Ibid., vol. 23, pp. 529–533, 1984; J. A.
Abusleme and J. H. Vera, AIChE J., vol. 35, pp. 481–489, 1989; L. Meng, Y. Y. Duan, and X. D. Wang, Fluid Phase
Equilib., vol. 260, pp. 354–358, 2007.
382 CHAPTER 10 . The Framework of Solution Thermodynamics
Example 10.9
Estimate ϕˆ 1 and ϕˆ 2by Eqs. (10.63) for an equimolar mixture of methyl ethyl ketone(1)/
toluene(2) at 50°C and 25 kPa. Set all kij = 0.
Solution 10.9
The required data are as follows:
contains a final term that gives it the simplest possible composition dependence. Indeed, this
functional form could reasonably serve as the basis for the composition dependence of chem-
ical potential in dense gases and liquids. In it, the composition dependence arises only from
the entropy increase due to random intermixing of molecules of different species. This entropy
increase due to mixing is the same in any random mixture, and thus can be expected to be
present in liquids and dense gases as well. However, the pure-species behavior implied by the
ig
term G i (T, P)is unrealistic except for the ideal-gas state. A natural extension of Eq. (10.24)
ig
therefore replaces G i (T, P)with Gi(T, P), the Gibbs energy of pure i in its real physical state
of gas, liquid, or solid. Thus, we define an ideal solution as one for which:
( )T, x
∂ G¯ id
( ∂ P )T
¯ iid = ____ ∂ G i
V i = ____
∂ P
( ∂ T )P, x
∂ G¯ id
( ∂ T )P
¯ id = − ____ ∂ G i
S i i = − ____
− R ln x i
By Eq. (10.5),
¯ id = Si − R ln x i
S (10.77)
i
¯ i id = G¯ i id + T S¯ iid , substitutions by Eqs. (10.75) and (10.77) yield:
Because H
¯ iid = G i+ RT ln x i+ T Si − RT ln x i
H
or
¯ iid = H i
H (10.78)
The summability relation, Eq. (10.11), applied to the special case of an ideal solution,
is written:
G id= ∑ x iG i+ RT∑ x iln x i (10.79) S id= ∑ x i Si − R∑ x i ln x i (10.80)
i i i i
Note the similarity of these equations to Eq. (10.25), (10.26), and (10.27) for
ideal-gas-state mixtures. For all properties of an ideal solution, the composition dependence
is, by definition, the same as that of ideal-gas-state mixtures, for which this composition
dependence is well defined and arises entirely from the entropy increase of random mixing.
However, the temperature and pressure dependence are not those of ideal gases but are given by
mole-fraction-weighted averages of the pure species properties.
If in Example 10.3 the solution formed by mixing methanol(1) and water(2) were
assumed ideal, the final volume would be given by Eq. (10.81), and the V-vs.-x1 relation
would be a straight line connecting the pure-species volumes, V2 at x1 = 0 with V1 at x1 = 1.
For the specific calculation at x1 = 0.3, the use of V1 and V2 in place of partial volumes yields:
V 1t = 983 V 2t = 1017 cm 3
Both values are about 3.4% low.
f ˆ i id
= x i fi (10.83)
This equation, known as the Lewis/Randall rule, applies to each species in an ideal solution
at all conditions of temperature, pressure, and composition. It shows that the fugacity of each
species in an ideal solution is proportional to its mole fraction; the proportionality constant is
the fugacity of pure species i in the same physical state as the solution and at the same T and P.
10.9. Excess Properties 385
Thus the fugacity coefficient of species i in an ideal solution is equal to the fugacity coef-
ficient of pure species i in the same physical state as the solution and at the same T and P.
Phases comprised of liquids whose molecules are of similar size and similar chemical nature
approximate ideal solutions. Mixtures of isomers closely conform to these conditions. Mix-
tures of adjacent members of homologous series are also examples.
The residual Gibbs energy and the fugacity coefficient are directly related to experimental
PVT data by Eqs. (6.49), (10.35), and (10.60). Where such data can be adequately correlated
by equations of state, thermodynamic-property information is readily provided by residual
properties. However, liquid solutions are often more easily dealt with through properties that
measure their departures, not from ideal-gas-state behavior, but from ideal-solution behavior.
Thus the mathematical formalism of excess properties is analogous to that of the residual
properties, but with ideal-solution behavior rather than ideal-gas-state behavior as a basis.
If M represents the molar (or unit-mass) value of any extensive thermodynamic property
(e.g., V, U, H, S, G, etc.), then an excess property ME is defined as the difference between the
actual property value of a solution and the value it would have as an ideal solution at the same
temperature, pressure, and composition. Thus,
M E≡ M − M id (10.85)
For example,
G E≡ G − G id H E≡ H − H id S E≡ S − S id
Moreover,
G E= H E− T S E (10.86)
which follows from Eq. (10.85) and Eq. (6.4), the definition of G.
The definition of ME is analogous to the definition of a residual property as given by
Eq. (6.41). Indeed, excess properties have a simple relation to residual properties, found by
subtracting Eq. (6.41) from Eq. (10.85):
M E− M R= − (M id− Mig )
As already noted, ideal-gas-state mixtures are ideal solutions of pure gases in the ideal-gas
state. Equations (10.79) through (10.82) therefore become expressions for Mig when Mi is
i𝗀
i . Equation (10.82) becomes Eq. (10.25), Eq. (10.80) becomes Eq. (10.26), and
replaced by M
Eq. (10.79) becomes Eq. (10.27). The two sets of equations, for Mid and Mig, therefore provide
a general relation for the difference:
ig
M id− M ig= ∑ x i M i− ∑ x i M i = ∑ x i M iR
i i i
386 CHAPTER 10 . The Framework of Solution Thermodynamics
( RT ) RT
n G E n V E n H E G¯ E
d _
= _
dP − _
2 dT + ∑ ___
i dn i (10.89)
R T i
RT
This is the fundamental excess-property relation, analogous to Eq. (10.55), the fundamental
residual-property relation.
The exact analogy that exists between properties M, residual properties MR, and excess
properties ME is indicated by Table 10.1. All of the equations that appear are basic property
relations, although only Eqs. (10.4) and (10.5) have been shown explicitly before.
Example 10.10
PE is a constant, independent of T, find expressions for GE, SE, and HE as
(a) If C
functions of T.
(b) From the equations developed in part (a), find values for GE, SE, and HE for an
equimolar solution of benzene(1)/n-hexane(2) at 323.15 K, given the following
excess-property values for an equimolar solution at 298.15 K:
Solution 10.10
(a) Let C PE = a, where a is a constant. From the last column of Table 10.1:
Integration yields:
( ∂ T )P, x
∂ G E
____
= − a ln T + b
Table 10.1: Summary of Equations for the Gibbs Energy and Related Properties
CP = (∂ H / ∂ T ) P, x C
RP = (∂ H R / ∂ T ) P, x C
EP = (∂ H E / ∂ T ) P, x
= − T (∂ 2 G / ∂ T 2) P, x = − T (∂ 2 G R / ∂ T 2) P, x = − T (∂ 2 G E / ∂ T 2) P, x
387
388 CHAPTER 10 . The Framework of Solution Thermodynamics
By Eq. (A),
c = H 0E − aT 0= 1750.6
By Eq. (C),
G E + a(T 0 ln T 0− T 0 ) − c
0
_____________________
b = = − 18.0171
T 0
Substitution of known values into Eqs. (A), (B), and (C) for T = 323.15 yields:
G E= 344.4 J·mol −1 S E= 1.492 J·mol −1 ·K −1 H E= 826.4 J·mol −1
HE HE GE
1
J mol GE
500 1000
J mol
500
J mol
TS E
E TS E
TS
GE
HE
0 1 0 1 0 1
x1 x1 x1
GE
GE
1000 500
E
H
GE
500 0
1
1
H
HE
J mol
J mol
J mol
0 0
TS E
500 TS E
500 TS E 1000
0 1 0 1 0 1
x1 x1 x1
(d ) (e) (f )
Figure 10.3: Excess properties at 50ºC for six binary liquid systems: (a) chloroform(1)/n-heptane(2);
(b) acetone(1)/methanol(2); (c) acetone(1)/chloroform(2); (d) ethanol(1)/n-heptane(2); (e) ethanol(1)/
chloroform(2); (f) ethanol(1)/water(2).
10.10 SYNOPSIS
After studying this chapter, including the end-of-chapter problems, one should be able to:
∙ Recognize that partial properties cannot vary independently, but are constrained by the
Gibbs/Duhem equation
∙ Understand that all linear relationships among thermodynamic properties of pure spe-
cies also apply to partial properties of species in a mixture
∙ Compute partial properties of a mixture in the ideal-gas state
∙ Define and use fugacity and fugacity coefficients for pure species and for species
in a mixture, and recognize that these are defined for convenience in solving phase-
equilibrium problems
∙ Compute fugacities and fugacity coefficients of pure species from PVT data or general-
ized correlations
∙ Estimate fugacity coefficients of species in gas mixtures using second virial coefficient
correlations
∙ Recognize that the ideal-solution model provides a reference for description of liquid mixtures
∙ Relate partial properties in an ideal solution to corresponding pure species properties
∙ Define and use excess properties of species in a mixture
10.11 PROBLEMS
10.1. What is the change in entropy when 0.7 m3 of CO2 and 0.3 m3 of N2, each at 1 bar and
25°C, blend to form a gas mixture at the same conditions? Assume ideal gases.
10.2. A vessel, divided into two parts by a partition, contains 4 mol of nitrogen gas at 75°C
and 30 bar on one side and 2.5 mol of argon gas at 130°C and 20 bar on the other. If
the partition is removed and the gases mix adiabatically and completely, what is the
change in entropy? Assume nitrogen to be an ideal gas with CV = (5/2)R and argon to
be an ideal gas with CV = (3/2)R.
10.3. A stream of nitrogen flowing at a rate of 2 kg·s−1 and a stream of hydrogen flowing
at a rate of 0.5 kg·s−1 mix adiabatically in a steady-flow process. If the gases are
assumed ideal, what is the rate of entropy increase as a result of the process?
10.4. What is the ideal work for the separation of an equimolar mixture of methane and
ethane at 175°C and 3 bar in a steady-flow process into product streams of the pure
gases at 35°C and 1 bar if Tσ = 300 K?
10.5. What is the work required for the separation of air (21-mol-% oxygen and 79-mol-%
nitrogen) at 25°C and 1 bar in a steady-flow process into product streams of pure
oxygen and nitrogen, also at 25°C and 1 bar, if the thermodynamic efficiency of the
process is 5% and if Tσ = 300 K?
10.11. Problems 391
10.6. What is the partial molar temperature? What is the partial molar pressure? Express
results in relation to the T and P of the mixture.
(a) The “partial molar mass” of a species in solution is equal to its molar mass.
(b) A partial specific property of a species in solution is obtained by division of the
partial molar property by the molar mass of the species.
10.8. If the molar density of a binary mixture is given by the empirical expression:
10.9. For a ternary solution at constant T and P, the composition dependence of molar prop-
erty M is given by:
M = x 1 M 1+ x 2 M 2+ x 3 M 3+ x 1 x 2 x 3 C
where M1, M2, and M3 are the values of M for pure species 1, 2, and 3, and C is a
parameter independent of composition. Determine expressions for M ¯ 1, M¯ 2 , and M¯ 3 by
application of Eq. (10.7). As a partial check on your results, verify that they satisfy
the summability relation, Eq. (10.11). For this correlating equation, what are the M ¯ i at
infinite dilution?
10.10. A pure-component pressure pi for species i in a gas mixture may be defined as the
pressure that species i would exert if it alone occupied the mixture volume. Thus,
y i Z i RT
p i≡ ______
V
where yi is the mole fraction of species i in the gas mixture, Zi is evaluated at pi and
T, and V is the molar volume of the gas mixture. Note that pi as defined here is not
a partial pressure yiP, except for an ideal gas. Dalton’s “law” of additive pressures
states that the total pressure exerted by a gas mixture is equal to the sum of the pure-
component pressures of its constituent species: P = ∑ i p i. Show that Dalton’s “law”
implies that Z = ∑ i y i Z i, where Zi is the compressibility factor of pure species i
evaluated at the mixture temperature but at its pure-component pressure.
10.11. If for a binary solution one starts with an expression for M (or MR or ME) as a
function of x1 and applies Eqs. (10.15) and (10.16) to find M ¯ 1 and M
¯ 2 (or M
¯ 1 R
and M ¯ R ¯ E ¯
2 or M 1 and M 2 )and then combines these expressions by Eq. (10.11), the
E
initial expression for M is regenerated. On the other hand, if one starts with expres-
sions for M¯ 1 and M¯ 2 , combines them in accord with Eq. (10.11), and then applies
392 CHAPTER 10 . The Framework of Solution Thermodynamics
(a) Apply Eq. (10.7) to Eq. (A) to verify Eqs. (B) and (C).
(b) Show that Eqs. (B) and (C) combine in accord with Eq. (10.11) to regenerate Eq. (A).
(c) Show that Eqs. (B) and (C) satisfy Eq. (10.14), the Gibbs/Duhem equation.
(d) Show that at constant T and P,
10.13. The molar volume (cm3·mol−1) of a binary liquid mixture at T and P is given by:
V = 120 x 1+ 70 x 2+ (15 x 1+ 8 x 2 ) x 1 x 2
(a) Find expressions for the partial molar volumes of species 1 and 2 at T and P.
(b) Show that when these expressions are combined in accord with Eq. (10.11) the
given equation for V is recovered.
(c) Show that these expressions satisfy Eq. (10.14), the Gibbs/Duhem equation.
(d) Show that (dV¯ 1 / dx 1 ) x 1=1 = (dV¯ 2 / dx 1 ) x 1=0 = 0.
(e) Plot values of V, V¯ 1 , and V ¯ 2 calculated by the given equation for V and by the
equations developed in (a) vs. x1. Label points V1, V2, V¯ 1 ∞ , and V¯ 2 ∞ and show their
values.
10.14. For a particular binary liquid solution at constant T and P, the molar enthalpies of
mixtures are represented by the equation
H = x 1(a 1+ b1 x 1 ) +x 2(a 2+ b 2 x 2 )
where the ai and bi are constants. Because the equation has the form of Eq. (10.11), it
¯ i = a i+ bi x i. Show whether this is true.
might be that H
[ ∂ n i ]
∂ ( nM)
˜ i ≡ ______
M
T, V, n j
( ∂ V )T, x
∂ M
˜ i = M¯ i + (V − V¯ i ) ___
M
10.16 From the following compressibility-factor data for CO2 at 150°C, prepare plots of
the fugacity and fugacity coefficient of CO2 vs. P for pressures up to 500 bar.
Compare results with those found from the generalized correlation represented by
Eq. (10.68).
P/bar Z
10 0.985
20 0.970
40 0.942
60 0.913
80 0.885
100 0.869
200 0.765
300 0.762
400 0.824
500 0.910
10.17 For SO2 at 600 K and 300 bar, determine good estimates of the fugacity and of GR/RT.
G R
= ∑ x i ln ϕˆ i ∑ x i d ln ϕˆ i = 0 (const T, P)
___
RT i i
10.21 From data in the steam tables, determine a good estimate for f/fsat for liquid water at
150°C and 150 bar, where fsat is the fugacity of saturated liquid at 150°C.
10.22 For one of the following, determine the ratio of the fugacity in the final state to that in
the initial state for steam undergoing the isothermal change of state:
10.23 Estimate the fugacity of one of the following liquids at its normal-boiling-point tem-
perature and 200 bar:
(a) n-Pentane;
(b) Isobutylene;
(c) 1-Butene.
10.24 Assuming that Eq. (10.68) is valid for the vapor phase and that the molar volume of
saturated liquid is given by Eq. (3.68), prepare plots of f vs. P and of ϕ vs. P for one of
the following:
(a) Chloroform at 200°C for the pressure range from 0 to 40 bar. At 200°C the vapor
pressure of chloroform is 22.27 bar.
(b) Isobutane at 40°C for the pressure range from 0 to 10 bar. At 40°C the vapor pres-
sure of isobutane is 5.28 bar.
10.26 Rationalize the following expression, valid at sufficiently low pressures, for estimat-
ing the fugacity coefficient: ln ϕ ≈ Z − 1.
10.27 For the system methane(1)/ethane(2)/propane(3) as a gas, estimate f ˆ 1 , fˆ 2 , fˆ 3 , ϕˆ 1 , ϕˆ 2 ,
ˆ
and ϕ 3 at t = 100°C, P = 35 bar, y1 = 0.21, and y2 = 0.43:
10.28 Given below are values of GE /J·mol−1, HE /J·mol−1, and C PE / J·mol −1 ·K −1 for some
equimolar binary liquid mixtures at 298.15 K. Estimate values of GE, HE, and SE at
328.15 K for one of the equimolar mixtures by two procedures: (I) Use all the data;
PE = 0.Compare and discuss your results for the two procedures.
(II) Assume C
10.29 The data in Table 10.2 are experimental values of VE for binary liquid mixtures of
1,3-dioxolane(1) and isooctane(2) at 298.15 K and 1(atm).
10.11. Problems 395
(a) Determine from the data numerical values of parameters a, b, and c in the corre-
lating equation:
x1 VE /10−3 cm3·mol−1
0.02715 87.5
0.09329 265.6
0.17490 417.4
0.32760 534.5
0.40244 531.7
0.56689 421.1
0.63128 347.1
0.66233 321.7
0.69984 276.4
0.72792 252.9
0.77514 190.7
0.79243 178.1
0.82954 138.4
0.86835 98.4
0.93287 37.6
0.98233 10.0
R. Francesconi et al., Int. DATA Ser., Ser. A, Vol. 25, No. 3,
p. 229, 1997.
10.30 For an equimolar vapor mixture of propane(1) and n-pentane(2) at 75°C and 2 bar,
estimate Z, HR, and SR. Second virial coefficients, in cm3·mol−1 are:
10.31 Use the data of Prob. 10.30 to determine ϕ ˆ 1 and ϕˆ 2 as functions of composition for
binary vapor mixtures of propane(1) and n-pentane(2) at 75°C and 2 bar. Plot the
results on a single graph. Discuss the features of this plot.
396 CHAPTER 10 . The Framework of Solution Thermodynamics
10.32 For a binary gas mixture described by Eqs. (3.36) and (10.62), prove that:
dδ 12
E= δ 12 P y 1 y 2
G S E = − ____
P y 1 y 2
dT
( dT )
dδ 12 d 2 δ 12
H E= δ 12− T _ P y 1 y 2 C PE = − T _____
2 P y 1 y 2
dT
See also Eq. (10.87), and note that δ 12= 2 B 12− B 11− B 22.
10.33 The data in Table 10.3 are experimental values of HE for binary liquid mixtures of
1,2-dichloroethane(1) and dimethyl carbonate(2) at 313.15 K and 1(atm).
(a) Determine from the data numerical values of parameters a, b, and c in the corre-
lating equation:
H E= x 1 x 2(a + b x 1+ c x 12 )
(b) Determine from the results of part (a) the minimum value of HE. At what value of
x1 does this occur?
¯ 1E and H
(c) Determine from the results of part (a) expressions for H ¯ 2 E . Prepare a plot
of these quantities vs. x1, and discuss its features.
x1 HE /J·mol−1
0.0426 −23.3
0.0817 −45.7
0.1177 −66.5
0.1510 −86.6
0.2107 −118.2
0.2624 −144.6
0.3472 −176.6
0.4158 −195.7
0.5163 −204.2
0.6156 −191.7
0.6810 −174.1
0.7621 −141.0
0.8181 −116.8
0.8650 −85.6
0.9276 −43.5
0.9624 −22.6
10.34 Make use of Eqs. (3.36), (3.61), (3.62), (6.54), (6.55), (6.56), (6.70), (6.71), (10.62),
and (10.69)–(10.74), to estimate V, HR, SR, and GR for one of the following binary
vapor mixtures:
10.35 Laboratory A reports the following results for equimolar values of GE for liquid mix-
tures of benzene(1) with 1-hexanol(2):
G E= 805 J·mol −1 at T = 298 K G E= 785 J·mol −1 at T = 323 K
Laboratory B reports the following result for the equimolar value of HE for the same system:
Are the results from the two laboratories thermodynamically consistent with one
another? Explain.
10.36 The following expressions have been proposed for the partial molar properties of a
particular binary mixture:
10.37 Two (2) kmol·hr−1 of liquid n-octane (species 1) are continuously mixed with
4 kmol·hr−1 of liquid iso-octane (species 2). The mixing process occurs at constant T
and P; mechanical power requirements are negligible.
10.38 Fifty (50) mol·s−1 of enriched air (50 mol-% N2, 50 mol-% O2) are produced by con-
tinuously combining air (79 mol-% N2, 21 mol-% O2) with a stream of pure oxygen.
398 CHAPTER 10 . The Framework of Solution Thermodynamics
All streams are at the constant conditions T = 25°C and P = 1.2(atm). There are no
moving parts.
10.40 For a multicomponent mixture containing any number of species, prove that
( ∂ x i) ( ∂ x k)
¯ i = M + ___∂ M ∂ M
M − ∑ x k ___
T, P k T, P
where the summation is over all species. Show that for a binary mixture this result
reduces to Eqs. (10.15) and (10.16).
10.41 The following empirical two-parameter expression has been proposed for correlation
of excess properties of symmetrical liquid mixtures:
(a) Determine from the given equation the implied expressions for M ¯ 1E and M¯ 2 E .
(b) Show that the results of part (a) satisfy all necessary constraints for partial excess
properties.
(c) Determine from the results of part (a) expressions for (M¯ 1 E ) ∞ and (M¯ 2 E ) ∞.
10.42 Commonly, if ME for a binary system has a single sign, then the partial properties M ¯ 1E
and M¯ 2 have the same sign as ME over the entire composition range. There are occa-
E
sions, however, where the M ¯ iE may change sign even though ME has a single sign. In
fact, it is the shape of the ME vs. x1 curve that determines whether the M ¯ iE change sign.
Show that a sufficient condition for M ¯ 1 and M¯ 2 to have single signs is that the curva-
E E
ture of ME vs. x1 have a single sign over the entire composition range.
10.11. Problems 399
10.43 An engineer claims that the volume expansity of an ideal solution is given by
Is this claim valid? If so, show why. If not, find a correct expression for βid.
10.44 Following are data for GE and HE (both in J·mol−1) for equimolar mixtures of the same
organic liquids. Use all of the data to estimate values of GE, HE, and TSE for the equi-
molar mixture at 25°C.
Suggestion: Assume C PE is constant and use material developed in Example 10.10.