Intermediate Thermodynamics

Fugacity

Definitions

The specific Gibbs function for a simple compressible substance is:

dg = v dP − s dT (1)

As in a pure substance the specific Gibbs function equals the chemical potential, we can
write for a isothermal process:

dµ T = v dP (2)

and replacing by the ideal gas EOS we obtain:

RT dP
dµ T ,ideal = = R T d ln P (3)
P

From Eq. (3) we can calculate the chemical potential of a pure substance that
behaves as an ideal gas. For a real gas we can use an EOS and calculate the chemical
potential by integration. This approach is not followed. Instead, a new thermodynamic
property is defined such that the form of Eq. (3) still holds for a real gas. This new function
is the fugacity f, defined as:

dµ T ,real = R T d ln f (4)

In addition, as the real gas and the ideal gas behave the same at very low pressure, it is
obvious that:

f
lim =1 (5)
P →0 P

Therefore, with the definition of Eq. (4) and with the reference value of f =0 at zero
pressure the fugacity is completely defined.

Evaluating the fugacity

Using the definition of the isothermal chemical potential, Eq. (2), and the fugacity,
Eq. (4) we can write:

1
 v dP
d ln f = (6)
RT

Eq. (6) in conjunction with an EOS (explicit in the specific volume) can be used to
calculate the fugacity. Integrating between two pressures we get:

∫ v dP
f
RT ln = (7)
f0 P0

If the EOS is explicit in pressure, we can use the relation:

d (vP ) = vdP + Pdv (8)

Replacing Eq. (8) in Eq. (6) and integrating we get:

∫ P dv
f
RT ln = P v − P0 v0 − (9)
f0 v0

RT a
Using the Redlich-Kwong EOS, p = − , we can use Eq. (9) to
v−b T v (v + b )
obtain:

f v−b a (v + b )v0
RT ln = P v − P0 v0 − RT ln − ln (10)
f0 v0 − b b T (v0 + b )v

Taking P0 → 0 the gas behaves as ideal, and we can write:

ln f =
Pv
− 1 + ln
RT
−
a
ln
(v + b ) (11)
RT v − b bRT 3 / 2 v

and replacing by the definition of the pressure we get:

b RT a  1 1 (v + b ) 
ln f = + ln −  v + b + b ln v  (12)
v−b v − b RT 3 / 2

A similar procedure using Van der Waals EOS leads to:

b RT 2a
ln f = + ln − (13)
v−b v − b RTv

2
Example:
Determine the fugacity, in bars, for R134a for a Redlich-Kwong gas at 90 oC and 10 bar. Compare
against Van der Waals EOS.

The constants in RK EOS must be first evaluated. Using the standard conditions
(∂P ∂v )Tc ( )
= ∂ 2 P ∂v 2 T = 0 the constants are:
c

RT a R 2 Tc2.5 RTc
p= − , a = 0.42748 , b = 0.08664
v−b T v (v + b ) Pc Pc

In R134, (C 2 F4 H 2 ) , we have Tc = 374.3 K , Pc = 40.6 bar , M = 102.3 kg / kmol . From direct substitution
we get:

bar m 3 K 2.5 m3
a = 197.1 , b = 0.06634
kmol 2 kmol

( )
Calculating the specific volume from directly from the EOS v = 2.724 m 3 / kg and replacing in Eq.
(12) we get f = 9.09 bar . A similar procedure leads to a value of f for a Van der Waals gas f = 9.21 bar .

Evaluation of the fugacity from tables or EOS’s is usually done using the fugacity
coefficient φ, defined as:

f
φ≡ (14)
p

that can be differentiated to obtain:

d ln φ = d ln f − d ln p (15)

and combining Eq. (15) with Eq. (6) we get:

 v 1
d ln φ =  − dP (16)
 RT P 

Integration of Eq. (16) at constant temperature from zero pressure ( φ = 1 ) to a state

pressure P gives:

∫
f  v 1
ln =  − dP (17)
P 0  RT P 

3
which relates PvT data with the fugacity. Eq. (17) can be integrated numerically from data
or an EOS of state can be used to evaluate the integral analytically. If we replace the
definition of the Z factor in Eq. (17) we obtain:

P P

∫ ∫
f  v 1  Z −1
ln =  − dP =  dP (18)
P 0  RT P  0  P 

or, in terms of reduced properties,

Pt

∫
f  Z −1
ln =  dPr (19)
P 0  Pr 

Recall that the integral in Eq. (19) has been already solved if we evaluated residual
or departure functions for the entropy, Eq. (47) in notes “Thermodynamic Properties”.
Examination of the Z chart shows that for Pr smaller than 0.4 the Z-Pr curves are straight
lines with slope (Z − 1) / Pr . Then the integral in Eq. (19) can be readily evaluated:

Pt P

∫ ∫
f  Z −1  Z −1 t
ln =  dPr =   dPr = Z − 1 (20)
P 0  Pr   Pr  0

that can be written as:

f
= e Z −1 ≅ 1 + (Z − 1) +
(Z − 1)2 + (Z − 1)3 + " (21)
P 2! 3!

for Z > 0.9 the series on Eq. (21) can be approximated well as:

f
=Z (22)
P

so at low pressures and for Z close to 1 we can use Eq. (22) with fair accuracy.

To obtain f from tabular data we integrate Eq. (4) between a reference state and the
state of interest along a isotherm, to get:

f
µ T , P − µ T , Pref = R T ln (23)
f ref

As the chemical potential is the specific Gibbs function for a pure substance, we can
write:

4
 f g − g ref
ln = (24)
f ref RT

We can use Eq. (24) to evaluate the fugacity of a real gas if we use a reference state
with low enough pressure such that the reference fugacity is the pressure (ideal gas) and
recalling that g = h − Ts . Then Eq. (24) becomes:

1  h − href 
ln
f
Pref
= 
R T
(
− s − s ref  ) (25)


Eq. (25) is used in the following way: a reference state is chosen at the lowest
pressure available at the state temperature. If the pressure is not low enough to be in the
ideal gas region, extrapolation of the properties to a lower pressure region might be
necessary. Then the evaluation of h and s from tables allow the calculation of f.

f can be also estimated from a three-parameter principle of corresponding states

using Pitzer acentric factor ω. Due to the close relation between Z and f, the fugacity
coefficient is tabulated as:

0 1
f  f  f 
log10 =  log10  + ω log10  (26)
P  P  P

Eq. (23) for a real gas can be simplified for the case of an ideal gas:

(µ T ,P − µ T , Pref )
ig
= R T ln
P
Pref
(27)

where the reference state is usually chosen at unit pressure, usually 1 atm. This allows us to
write the following two relations:

(µT ,P )ig = (gT ,P )ig = gT0 + R T ln P (28)

µ T , P = g T0 + R T ln f (29)

Subtracting Eq. (28) from Eq. (29) we obtain:

µ T , P − (µ T , P )ig = R T ln
f
(30)
P

5
Liquids and solids

From Eq. (6) we can obtain the change in fugacity between the saturation liquid
state and a compressed state (of course at higher pressure):

∫ v dP
v dP f 1
d ln f = ⇒ ln = (31)
RT f sat liq
RT Psat

As no approximation has been made so far, Eq. (31) applies also to solids using a reference
pressure other than the saturation pressure, and the corresponding fugacity.

Most liquids and solids have fairly constant specific volume over wide ranges of
pressure variations, i.e. are essentially incompressible, thus we can approximate:

f v(P − Psat )
ln ≅ (32)
f sat liq
RT

In addition to assuming that the specific volume is constant, notice that specific
volumes of liquids and solids are much smaller than the corresponding specific volumes of
gases at the same pressure, for small pressures. Thus we can write for small pressure
changes:

liq
f liq ≅ f sat (33)

We want now to relate the fugacity of saturated vapor with the fugacity of saturated
liquid. To this end we will use Eq. (1) and the fact that phase changes are isobaric and
isothermal (see also properties of the saturation state in notes “Property Relations”):

dg = dµ = v dP − s dT = 0 (34)

and combining Eq. (34) with Eq. (4), dµ = R T d ln f we obtain:

vap liq
f sat = f sat (35)

which is a consequence of equilibrium considerations. Thus, Eq. (33) can be rewritten as:

liq
f liq ≅ f sat (36)

furthermore, if Psat is small enough so that the vapor behaves as an ideal gas we have:

f liq ≅ Psat (T ) (37)

6
Real gas mixtures

Recall that from Eq. (38) in Homogeneous Mixtures we have that the partial molar
volume is related to the Gibbs function by:

∂ µi ∂V
= = vi (38)
∂P T ,N
∂ Ni T ,P, N j

For an ideal gas, we can express the total volume as:

RT
V = ( N1 + N 2 + " + N i + ") (39)
P

and from Eq. (38) we have:

∂V RT
vi = = (40)
∂ Ni T ,P, N j
P

and using again Eq. (38) the relation between the chemical potential and the pressure is:

∂ µi RT
= (41)
∂P T ,N
P

and thus:

dµ i = RTd ln p (42)

Recall also that Dalton’s Law states that (“any gas is a vacuum to any other gas mixed with
it”):

p i = xi p (43)

as at constant xi we have dpi pi = ( xi dp ) xi p ⇒ d ln pi = d ln p we obtain

dµ i ,ig = RTd ln pi (44)

which is equivalent to Eq. (4) valid for a pure substance. Integration yields (compare to Eq.
(27) for a pure substance):

(µ i,T − µ ref )ig = R T ln p pi (45)

i ,ref

7
where the reference value is usually selected at the ideal gas value at unit pressure (usually
1 atm) and thus:

µ i ,T ,ig = g T0 + R T ln pi (46)

In a real gas mixture we use the same type of definition for the fugacity as was done
in a pure substance (recall Eq. (23)), and then we can write:

dµ i = dg i = RTd ln f i (47)

where pi is defined in Eq. (43). To evaluate fi we make use of Eq. (38), that allows us to
write:

dµ i ,T = vi dP (48)

that can be replaced into Eq. (47) yielding:

RTd ln f i = vi dP (49)

and integrating between a low pressure state to a given state we obtain:

∫ v dP
fi
RT ln = i (50)
f i* P*

A similar equation can be easily derived for an ideal gas, resulting in:

∫
P RT
RT ln = dP (51)
P* P* P

using Eqs. (50) and (51) yields:

P
f i*
∫
fi  RT 
RT ln = RT ln * +  vi − dP (52)
P P P*  P 

we want to find the limit for P*→0. Note that:

fi xi f i xi f i
lim = lim = lim = xi (53)
P * →0 P* P * →0 x i P * P * →0 pi*

and then we can write:

8
 P

∫
 vi 1
ln f i = ln xi P +  − dP (54)
0  RT P 

∫
f  v 1
notice the similarity with Eq. (17), ln =  − dP . Introducing the fugacity
P 0  RT P 
coefficient for mixtures we have:

∫
f  vi 1
ln i = ln φi =  − dP (55)
xi P 0  RT P 

Notice that we need the partial molar volumes as a function of pressure for each
temperature to calculate the fugacities of the components. If such data is not available we
will need to make suitable approximations.

Duhem-Margules relation

It is possible to obtain an important relation for the variation of the component

fugacities as a function of the composition for fixed pressure and temperature. In a binary
system with components 1 and 2, the Gibbs-Duhem relation (Eq. (36) in Homogeneous
Mixtures) states:

x1dµ1 + x 2 dµ 2 = 0 (56)

and from Eq. (47) we have:

x1d ln f1 + x 2 d ln f 2 = 0 (57)

As the fugacity is only a function of composition at constant pressure and temperature, we

can write:

∂ ln f i
d ln f i = dxi (constant T, P) (58)
∂xi

and replacing into Eq. (57) gives:

∂ ln f1 ∂ ln f 2
x1 dx1 + x 2 dx 2 = 0 (59)
∂x1 ∂x 2

and since dx1 = − dx 2 for a two-component system we have:

9
 ∂ ln f1 ∂ ln f 2
x1 = x2 (60)
∂x1 ∂x 2

or

∂ ln f1 ∂ ln f 2
= (61)
∂ ln x1 P ,T
∂ ln x 2 P ,T

that are the Duhem-Margules relations. For low pressure we can replace the component
fugacities by the partial pressure to get:

∂ ln p1 ∂ ln p 2
= (62)
∂ ln x1 P ,T
∂ ln x 2 P ,T

or, using Eq. (60),

∂ ln p1 ∂ ln p 2
x1 = − x2 (63)
∂x1 P ,T
∂x1 P ,T

As a consequence of the Duhem-Margules relation, partial pressure – composition plots

have a trend as shown in Fig. 1.

Figure 1: typical partial pressure plot for a two-component system.

10
Ideal Solutions

Going back to Eq. (46), the chemical potential in an ideal-gas mixture can be
written as:

µ i ,ig = g T0 + R T ln pi (64)

From the definition of partial pressure we have that pi = xi p , so Eq. (64) can be rewritten
as:

µ i ,ig = g T0 + R T ln P + R T ln xi = µ *i + R T ln xi (65)

or, in general,

µ i = µ *i + R T ln xi (66)

where µ *i is the Gibbs function of the pure substance at T and the total pressure P. Eq. (66)
represents the definition of an ideal solution: a solution is an ideal solution if the chemical
potential of every component can be expressed as a linear function of the logarithm of its
mole fraction. Of course, ideal-gas mixtures satisfies the definition. Many other non-ideal
liquid and gaseous mixtures behave like ideal mixtures at least in a range of compositions.

We will develop some relations for ideal solution, and then discuss the case of real
solutions.

From Eq. (30) of Homogeneous Mixtures, the enthalpy of formation was expressed
as:

∆H =
∑ N (h − h )
i
i i i (67)

In an ideal solution, this expression can be derived from manipulation of Eq. (37) of
Homogeneous Mixtures:

∂ µi T 1 ∂ µi µi si µ i h
= − =− − 2 = − i2 (67)
∂T P, N
T ∂T P, N T 2 T T T

and a similar derivation for a pure component results in:

∂ gi T h
= − i2 (68)
∂T T

11
Now, differentiating Eq. (66) respect to temperature at constant pressure after dividing by T
we obtain:

∂µ i T ∂µ *i T ∂µ i T ∂g i T
− = − =0 (69)
∂T ∂T ∂T ∂T

which implies in conjunction with Eqs. (67) and (68) that hi = hi in an ideal solution, and
therefore the heat or enthalpy of mixing will be:

∆H =
∑ N (h − h ) = 0
i
i i i (70)

In a similar manner, differentiation of Eq. (66) with respect to pressure at constant

temperature yields:

∂µ i ∂g i
− =0 (71)
∂P T ∂P T

and using ∂µ i ∂P = vi we obtain:

vi = vi (72)

so the partial molar volume is the pure specific volume.

Differentiation of Eq. (66) with respect to temperature leads to:

∂µ i ∂g i
= + R ln xi = 0 (73)
∂T ∂T

and using Eq. (37) in Homogeneous Mixtures we obtain:

si = si − R ln xi (74)

In consequence, according to Eqs. (70), (72), and (74) we can write for an ideal
solution:

hmixt =
∑x h = ∑x h
i
i i
i
i i , pure (∆hmix = 0) (75)

v mixt =
∑x v = ∑x v
i
i i
i
i i , pure (∆vmix = 0) (76)

12
 s mixt =
∑ i
xi s i =
∑
i
xi si , pure − R
∑
i 
xi ln xi (77)



∆smix

g mixt = hmix − (Ts )mix =

∑ i
xi hi , pure −
∑i
xi si , pure + RT
∑
i 
xi ln xi (78)


∆g mix

To calculate the fugacity of an ideal solution let’s recall the definitions of the
fugacity, Eq. (47) dµ i = RTd ln f i , and of the chemical potential for an ideal solution, Eq.
(66), µ i = µ *i + R T ln xi . Differentiating Eq. (66) at constant temperature and equating the
chemical potentials we obtain:

RTd ln f i = RTd ln xi (79)

and upon integration between states 1 and 2 with different composition we obtain:

f i,2 xi , 2
= (80)
f i ,1 xi ,1

where if we choose state 1 with a pure substance i, ie xi = 1, f i = f i , P (fugacity of pure

component i at the total pressure), then:

f i = xi f i , P (81)

which is known as Lewis-Randall rule, which

states that the partial molar fugacity in an ideal
solution is the pure substance fugacity multiplied
by the molar fraction.

In a binary solution with components A

and B we have x A + x B = 1 , and thus Eq. (75)
reduces to:

hmixt =
∑ i
xi hi , pure = x A h A + (1 − x A )hB

(82)

that can be easily interpreted in a x-h diagram, as

shown in Fig. 2.
Figure 2: molal enthalpy-composition

13
The entropy of the ideal mixture is shown in Fig. 3. Notice that the mixture entropy is
always larger than the individual entropies. A similar statement is true for the Gibbs
function but with lower Gibbs function after the mixture.

2.5

1.5
smix

0.5

0
0 0.2 0.4 0.6 0.8 1
xA

Figure 3: molal entropy vs. composition diagram

Heat/work interactions in ideal solutions

Let’s analyze the case of an ideal solution undergoing a process in which heat and
work interactions occur. Consider an internally reversible process in which the volume is
changed isothermally in a flow device (compressor or turbine). In this case the heat transfer
can be calculated from:

Q = T ∆S = T
∑
i
N i (si, 2 − si ,1 ) (82)

and the work is calculated from:

W = m (∆h + ∆ke + ∆pe ) − Q (83)

As shown by Eqs. (82) and (83), evaluation of mixture properties, namely h and s, is
necessary. These property data might not be readily available and some modeling will be
necessary to estimate the changes in enthalpy and entropy. In the case of an ideal solution
we have:

14
 ∆si = (si − R ln xi )2 − (si − R ln xi )1 = ∆si (84)

thus Eq. (82) reduces to:

Q = T ∆S = T
∑ i
N i (si , 2 − si ,1 ) = N T
∑
i
xi (si , 2 − si ,1 ) (85)

The change in enthalpy is calculated from Eq. (75) as:

∆hm =
∑ i
xi (hi, 2 − hi,1 ) (86)

Example: a gas mixture consisting of 70 % molar methane and 30 % nitrogen is compressed reversibly and
isothermally from 10 bar to 100 bar at 250 K. Flow rate is 0.2 m3/min. Determine a) the heat transfer, b) the
power requirement assuming ideal solution and ideal gas.

The pure component properties are taken from tables at 10 bar and at 100 bar, 250 K, We have: Nitrogen at 10
bar: v =73.6 cm3/g, h = 356.3 kJ/kg, s = 5.962 kJ/kg K. Nitrogen at 100 bar: v =7.12 cm3/g, h = 330.4 kJ/kg, s
= 5.190 kJ/kg K. Methane at 10 bar: v =125.4 cm3/g, h = 1077.9 kJ/kg, s = 10.002 kJ/kg K. Methane at 100
bar: v =8.88 cm3/g, h = 928.5 kJ/kg, s = 8.364 kJ/kg K.

The molar flow rate can be calculated from N = Vm v m , with Vm = 0.2 m 3 min and vm the specific molar
volume. For an ideal solution, see Eq. (76),

1 m3
v m = x1 v1 + x 2 v 2 = (0.7 125.4 16.04 + 0.3 73.6 28.01) = 2.026
10 3 kmol

thus the molar flow rate is:

0.2 kmol
N = = 0.0987
2.026 min

a) The heat transfer can now be evaluated from Eq. (85):

kJ
Q = 0.0987 250[0.7(8.364 − 10.002)16.04 + 0.3(5.190 − 5.962)28.01] = −624
min

b) The work is:

kJ
W = N ∆h − Q = N (x1 ∆h1 + x 2 ∆h2 ) − Q = 427 = 7.12 kW
kmol

For the ideal gas model we have that the volume is:

RT 0.08314 250 m3
vm = = = 2.079
P 10 kmol

and the molar flow rate is:

15
 0.2 kmol
N = = 0.0962
2.079 min

Using Eq. (85) for the heat transfer we get:

 P  kJ
Q = T∆s = 0.0987 250 − R ln 2  = −460
 P1  min

And the work is:

kJ
W = −Q = 460 = 7.67 kW
kmol

since ∆h = 0 in an isothermal process for an ideal gas.

Non-ideal solutions

Enthalpy-composition diagram

In non-ideal solutions the enthalpy of mixing is non-zero, and therefore the molal
enthalpy vs. composition diagram will no longer be a straight line. Fig. 4 shows a typical
molal enthalpy – composition diagram for a solution with positive enthalpy of formation.

Figure 4: molal enthalpy vs. composition diagrams for an ideal solution (left) and a real solution (right).

16
 Notice in Fig. 4 that the enthalpy of mixing can be defined as the difference
between the molal enthalpy for a real solution minus the molal enthalpy for an ideal
solution. Using Eq. (70) on a molar basis:

∆hmix =
∑ i
xi (hi − hi ) =
∑
i
xi hi −
∑
i
xi hi = hmix − [(1 − x )h A + x hB ]

(87)

The enthalpy of mixing is the amount of heat that needs to be added or extracted
from a closed system to keep the temperature of the mixture constant during the formation
of a solution. A negative enthalpy of mixing implies that heat must be removed from the
system to keep it at constant temperature or, under adiabatic conditions, the temperature of
the system will increase. Conversely, heat must be added to undergo an isothermal mixing
if the enthalpy of mixing is positive.

Figure 5: enthalpy of mixing vs. composition diagram for a water/ethanol solution.

Fig. 5 shows the enthalpy of mixing as a function of composition for a binary

solution of ethanol and water for three different temperatures. Notice that at 0 oC the

17
enthalpy of mixing is negative for all compositions, thus heat has to be removed from the
system to keep it at constant temperature. This type of enthalpy of mixing/composition
diagram can be used to solve problems involving real mixtures. Note that the composition
is given in mass fraction and not in molar fraction.

A second (and preferred) format in which the data may be available is in mixture
(molal) enthalpy vs. composition diagram for several temperatures, at a given pressure.
Temperatures are interpolated for use in actual problems from the temperatures of the
diagram. This type of diagram is shown in Fig. 6 for a ethanol/water binary system.

Figure 6: molal enthalpy vs. composition diagram for a water/ethanol solution.

18
Example: the mixing of two fluid streams.

Two streams of water and ethanol are mixed in an adiabatic chamber. The flow rates are 0.7 kg/s of
water and 0.3 kg/s of ethanol. Inlet temperatures are 70 °C for ethanol and 10 °C for water. Find the
temperature of the exit mixture assuming a) ideal solution and b) real solution. c) What is the amount of heat
that should be removed from the mixer to overcome the heat of mixing?

The temperature at the exit can be found using the mass and energy conservation equations:

m w + m e = m mixt

m w hw + m e he = m mixt hmixt

mass conservation implies that m mixt = 1 kg s . From the diagram for ethanol and water, we have,
hw = 45 kJ kg , he = 180 kJ kg . Thus, the mixture enthalpy will be:

m w hw + m e he 0.7 45 + 0.3 180

hmixt = = = 85.5 kJ kg
m mixt 1

a) assuming ideal solution, for a mass fraction of ethanol of 0.3, we can interpolate the temperature
values of the diagram of Fig. 5, using straight lines. For 20 °C, xm=0.3 we have an enthalpy of 75
kJ/kg. For 30 °C we have an enthalpy of 110 kJ/kg. Interpolation gives a temperature of 23 °C.
b) for a real solution, we have 80 kJ/kg for 30 °C and 120 kJ/kg for 40 °C. Interpolation results in a
temperature of 31.3 °C.

For question c), the energy equation writes:

m w hw + m e he + Q = m mixt hmixt

where the mixture enthalpy must be such that the temperature is that of the ideal solution (in which case
the enthalpy of mixing is zero and there is no overheating in the mixer). For Tmixt = 20 °C we get from Fig. 5
that hmixt = 35 kJ/kg, and for Tmixt = 30 °C we get hmixt = 65 kJ/kg. Thus, for Tmixt = 23 °C we obtain by
interpolation hmixt = 44 kJ/kg. Replacing into the energy equation results in:

Q = m mixt hmixt − m w hw − m e he = 1 44 − 0.7 45 − 0.3 180 = −41.5 W

19