Professional Documents
Culture Documents
Thermodynamics Lecture Notes
Thermodynamics Lecture Notes
1 of 5
At 27C the reading on a manometer filled with mercury is 60.5 cm. The local acceleration of
gravity is 9.784 m s
-2
. To what pressure does this height of mercury correspond?
Solution: The pressure exerted by the column of mercury is equal to its height times its density
times the local acceleration of gravity, or in symbols P = hg. We have h = 60.5 cm = 0.605 m.
The density of mercury at 27C is 13.53 g cm
-3
= 13530 kg m
-3
(where would I find this?). So,
we have
P = 0.605 m 13530 kg m
-3
9.784 m s
-2
= 80088 kg m
-1
s
-2
= 8.0110
4
Pa = 0.801 bar
Work:
Usually, mechanical work can be defined as the product of a force times a distance or the integral
of a force over some distance through which it acts. If F is the component of force along the
direction of motion and dl is a differential displacement in that direction, then this can be
expressed as
dW = F dl
LECTURE NOTES
CHEMICAL ENGINEERING THERMODYNAMICS
What is thermodynamics?
One answer: a field of science and engineering that describes, macroscopically, the driving
forces for the flow of energy (as heat and/or work) and the conditions for equilibrium within
a system (where equilibrium is the state where these driving forces are absent and the system
has no tendency to change). Thermodynamics does not deal with rates of change, but only
with the driving forces that cause change. To do thermodynamics, we must understand basic
physics (like that learned in the mechanics portion of a first-year college physics course) and
be adept at using and converting units of measurement. In engineering, most numbers are
meaningless without units. We will not review units and unit conversions here, but you should
read the review of dimensions and units in chapter 1 of Smith, Van Ness, and Abbott (SVA).
Pay particular attention to the section on pressure. In general, while I will post detailed lecture
notes like the ones you are reading, and these will follow the textbook fairly closely, these are
not an adequate substitute for also reading the textbook chapters as we cover them.
Example 1.3 from SVA
where dW is the differential amount of work performed when the force F acts through the
displacement dl.
In thermodynamics, we often deal with the work done by an expanding fluid (or work done on a
fluid to compress it). In this case, the force F is the pressure multiplied by the area over which
the pressure is applied (F = PA). If the compression is done by a cylinder of constant area A then
the change in total volume of the fluid is dV
t
= A dl. Using these expressions for dl and F gives
dW = -P dV
t
The cross-sectional area cancels out (so the work is independent of the shape of the piston and
cylinder). The negative sign arises if we use the sign convention that work done on the gas by
the piston is positive. When the piston does work on the gas, the change in volume is negative
and the pressure is positive, so the negative sign is required to make the work come out to be
positive. In integral form, then, the work on a gas in compressing it from pressure P
1
to pressure
P
2
is:
2
1
P
P
W PdV =
How would we compute this, though? The limits of the integral are in terms of pressure, while
the variable of integration is dV. We need an equation of state that relates the pressure and
volume (and temperature, if temperature changes) in order to evaluate this. We will spend
substantial time this semester considering equations of state.
Forms of Mechanical Energy (Kinetic Energy and Gravitational Potential Energy)
Kinetic energy (E
K
= mu
2
where m is the mass of an object and u is its speed) and gravitational
potential energy (E
P
= mgz where m is the mass of an object, z is its elevation relative to a
reference elevation, and g is the local acceleration of gravity) are forms of mechanical energy
that can be converted (fully) to work. Note that only changes in or relative amounts of kinetic
and potential energy are meaningful, since the objects speed and elevation must be defined
relative to some stationary reference frame and reference position, respectively.
Mechanical Energy Conservation: Example 1.4 in SVA.
An elevator with a mass of 2,500 kg rests at a level 10 m above the base of an elevator shaft. It
is raised 100m above the base of the shaft, where the cable holding it breaks. The elevator falls
freely to the base of the shaft and strikes a strong spring. The spring is designed to bring the
elevator to rest and, by means of a catch arrangement, to hold the elevator at the position of
maximum spring compression. Assuming the entire process to be frictionless, and taking g = 9.8
m s
-2
, calculate:
(a) The potential energy of the elevator in its initial position, relative to the base of the shaft.
(b) The work done in raising the elevator
Lecture 1
(c) The potential energy of the elevator in its highest position relative to the base of the shaft.
(d) The velocity and kinetic energy of the elevator just before it strikes the shaft.
(e) The potential energy of the compressed spring.
(f) The energy of the system consisting of the elevator and spring (1) at the start of the
process, (2) when the elevator reaches its maximum height, (3) just before the elevator
strikes the spring, and (4) after the elevator has come to rest.
Solution:
The whole situation will probably be clearer if we draw a sketch of it:
10 m
100 m
(a) If our reference elevation is the bottom of the shaft, and the initial elevation of the elevator is
10 m above the bottom of the shaft, then the potential energy of the elevator (relative to the
bottom of the shaft) is
E
p
= mgz = 2500 kg 9.8 m s
-2
10 m = 245,000 kg m
2
s
-2
= 245 kJ
(b) The work done to raise the elevator is equal to the integral of the force applied to the elevator
over the distance it is raised. In this case, the force is constant F = mg (or more formally, F
= mg/g
c
to make it work in English units). So, the work done is
( )
2 2
1 1
2 1
z z
z z
W Fdz mgdz mg z z = = =
W = 2500 kg 9.8 m s
-2
90 m = 2,205,000 kg m
2
s
-2
= 2,205 kJ
Lecture 1
(c) Just as in part (a), we have
E
p
= mgz = 2500 kg 9.8 m s
-2
100 m = 2,450,000 kg m
2
s
-2
= 2,450 kJ
(d) We use the principle of conservation of mechanical energy (in this frictionless system) to
equate the change in gravitational potential energy to the change in kinetic energy.
( ) ( )
,
2 -2
,
0
0 2, 450 kJ 0 0
2, 450 kJ = 2,450,000 kg m s
P K
K after
K after
E E
E
E
+ =
+ =
=
That is, the 2,450 kJ of potential energy that the elevator has when it is at a height of 100 m
and its kinetic energy is zero (because its velocity is zero) is converted completely to kinetic
energy when it falls to a height of zero (where its potential energy is zero). The velocity
corresponding to this kinetic energy is obtained from
( )
2 -2 2
1
, 2
2 -2
1
2
2,450,000 kg m s
2,450,000 kg m s
44.3 m/s
2500
K after
E mu
u
kg
= =
= =
Note that we could get this same result by solving Newton's laws of motion for the elevator.
At t = 0, we have z = 100 m and u = 0. The acceleration of the elevator (by gravity) is a = g
= 9.8 m s
-2
. Integrating this to get the velocity gives:
-2
9.8 m s
, where C is a constant of integration
dv
a
dt
v at C
= =
= +
At t = 0, v = 0, so C = 0, and v = -9.8 t m/s
Integrating again,
-2
2 2
1
2
9.8 m s
4.9 , where C is a constant of integration
dz
v t
dt
z vt C C t
= =
= + =
At t = 0, z = 100 m, so C = 100 m, and z = (100 - 4.9t
2
) m
To find the velocity when the elevator reaches z = 0, we first must find the time when z = 0
by solving 100 - 4.9t
2
= 0, which gives t = 4.518 s. At this time, the velocity is then v = -9.8
m s
-2
4.517 s = 44.3 m/s. We could show (but won't) that the law of conservation of
mechanical and potential energy can be derived from Newton's laws of motion.
This is one of many examples in the course in which there is more than one logical and
appropriate way to reach the correct answer. On occasion, I will show methods that differ from
those in the textbook. In such cases, you are encouraged to try applying both methods, which
will hopefully deepen your understanding of the problem and its solution.
(e) The change in potential energy of the spring plus the change of kinetic energy of the elevator
must sum to zero (again, for a frictionless system), so E
P,spring
= 2,450 kJ after the elevator
comes to rest (converting all of its kinetic energy to potential energy of the spring). Note that
since the spring is described as 'very stiff', we are neglecting any change in the elevation of
the elevator during the compression of the spring. If the spring compresses a significant
distance, then the gravitational potential energy would become negative (since the elevation
of the elevator would be less than the reference elevation) and the potential energy of the
spring would be slightly more than 2,450 kJ.
(f) If the elevator and spring are taken together as the system, then the energy of the system only
changes when something outside the system does work on it, which only happens when the
elevator is initially raised from 10 m to 100 m. So, initially, the total energy is 245 kJ (as in
part (a)). After the elevator is raised to 100 m, the total energy of the spring plus the elevator
is 2,450 kJ. When the elevator falls and compresses the spring, energy is converted from one
form to another within the system, but no work is done on or by objects outside the system,
so the total energy remains 2,450 kJ.
Heat: See discussion in SVA section 1.9.
Heat is what flows when energy is transferred from a warmer object to a cooler one, raising the
internal energy of the cooler object and lowering the internal energy of the warmer one, until
their temperatures are equal. A more precise definition will require improved definitions of
temperature and internal energy. The first law of thermodynamics (up next) will show us that
heat and work are equivalent in some senses but not in others. As a result, heat is measured in
the same units (joules or btu) as work.
When we add heat to or do work on a substance, we may increase its internal energy. From a
macroscopic point of view, this is energy stored within the material that can later be recovered as
heat transferred to a cooler object. Microscopically, this energy is stored in the motions of the
molecules that make up the material and in the potential energy of the interactions between the
molecules. In classical thermodynamics, absolute values of internal energy aren't known, just
relative values, which are all that are required for thermodynamic analysis anyway.
The First Law: Energy is Conserved
Although energy assumes many forms, the total quantity of energy is constant, and when energy
disappears in one form it appears simultaneously in other forms.
We often analyze problems by dividing the universe into the system (the part we are explicitly
studying) and the surroundings (everything else).
Any changes in the energy of the system must be accompanied by changes in the energy of the
surroundings. That is
(Energy of the System) + (Energy of the Surroundings) = 0
Energy Balances for Closed Systems:
If our system is closed (no mass flows across its boundaries) then energy can only enter or leave
the system as heat or as work, and we have
(Energy of the System) = Q + W
The System
(the part of the uni verse
were looking at)
The Surroundi ngs
(Everythi ng Else)
Q
W
FIRST LAW OF THERMODYNAMICS (chapter 2 in SVA)
Internal energy:
Where Q is energy transferred from the surroundings to the system by heat, and W is energy
transferred from the surroundings to the system as work (work done on the system by the
surroundings). This implies that
(Energy of the Surroundings) = -Q - W
The sign convention for Q and W used here is the one used in SVA, but is not universal. You
should always be sure you understand which way energy is flowing and not just rely on the sign
in the equations to get it right.
If all energy transferred into the system goes to increase the internal energy of the system (as
opposed to increasing its kinetic or potential energy by changing its overall position or velocity),
then we have
U
t
= Q + W
where U
t
is the total internal energy of the system.
For infinitesimally small (differential) changes, this is
dU
t
= dQ + dW
The quantity U
t
is an extensive quantity - it is the total internal energy of the system, which if the
system is homogeneous, is proportional to the total amount of matter in the system.
Homogeneous means "the same throughout" - at every point within the system the properties
(temperature, pressure, density, composition, etc.) are the same as at every other point.
If the system is homogeneous, then we can define intensive properties like the internal energy
per mole or per mass
U = U
t
/n = internal energy per mole = total internal energy divided by the number of moles
U = U
t
/m = internal energy per mass = total internal energy divided by the total mass
Then, for a homogeneous closed system containing n moles of material, we can write
(nU) = nU = Q + W
d(nU) = ndU = dQ + dW
Often, we simply write thermodynamic equations for a representative amount of a homogeneous
material, in which case, we would just write (for one mole of material, for example)
U = Q + W
dU = dQ + dW
This last expression, in particular, allows us to keep track of the molar internal energy of a
substance and compute changes in it by computing the amount of work done on it by its
surroundings and the amount of heat that flows to it from its surroundings. By observing how its
temperature and other properties change as its internal energy changes we can also relate changes
in internal energy (a property of the substance) to changes in its other properties (like density,
etc.) that are observable.
Thermodynamic State and State Properties
At a given thermodynamic state, a homogenous system always has the same thermodynamic
properties. For a homogeneous pure substance, fixing two properties (i.e. temperature and
pressure or temperature and density or density and pressure) fixes all of its other thermodynamic
properties (temperature, pressure, density, internal energy, etc.). For more complex systems,
more than 2 thermodynamic properties may have to be specified to specify the state of the
system, but once the required number is specified, the thermodynamic state of the system is fixed
and all other thermodynamic properties are also fixed.
State functions depend only on the thermodynamic state of the system and do not depend on the
path by which the system arrived at that state. Internal energy is a state function.
Q and W are not state functions. They do not represent properties of the system, but represent
interactions of the system with its surroundings. As such, they depend on the path. The system
can get from one state to another through different paths, where by different paths we mean
through different interactions with its surroundings. If two paths move the system between the
same two states (they cause the same change in the state of the system) then the sum Q + W will
be the same for both paths, since the change in internal energy must be the same. However, Q
and W, individually, will be different for the two paths.
For example, suppose I have some water at a temperature of 280 K and a pressure of 1
atmosphere. Fixing the temperature and pressure specifies the state of the system. I could stir
the water violently until its temperature increased to 300 K (still at 1 atmosphere) while keeping
it thermally insulated. In this case, Q would be zero, and we would have U = W. Alternatively,
I could heat the water from 280 K to 300K by sitting the (non-insulated) beaker on a warm
surface. In this case, no work would be done on the water (W = 0), and we would have U = Q.
Both methods of heating the water result in the same change in the state of the water (they go
from the same initial thermodynamic state to the same final thermodynamic state), so U is the
same in both cases. However, Q and W are different for the two cases.
The Phase Rule:
For any non-reacting system at thermodynamic equilibrium, the number of independent variables
that must be fixed to establish its intensive state is given by
F = 2 - + N
where is the number of phases present, N is the number of chemical species present, and F is
the number of degrees of freedom that the system has. The number of degrees of freedom of the
system is the number of phase rule variables (temperature, pressure, composition of each phase,
etc.) that must be fixed to completely specify the state of the system, including all other phase-
rule variables.
See example 2.5 in SVA.
Reversible Processes
In thermodynamic analyses, we often make use of the idealized concept of a reversible process,
much like in first-year physics we make use of a lot of frictionless processes, even though such a
process can only be approximately realized in a real system.
A reversible process is one whose direction can be changed at any point by infinitesimal changes
in external conditions
(which means there must be no friction, no inertia (so it must happen infinitely slowly), no heat
transfer through finite temperature differences, and so on).
A reversible process
Is frictionless
Is never more than infinitesimally away from equilibrium
Goes through a set of equilibrium states
Is driven by infinitesimally small force or temperature imbalances
Can be reversed at any point by an infinitesimal change in external conditions
When reversed, retraces its original path back to the original state of the system
Constant Volume processes, Constant Pressure processes, and Enthalpy
For a homogeneous, closed system containing n moles of a substance, the energy balance was (in
differential form)
d(nU) = dQ + dW
For a mechanically-reversible closed-system process, the work term is
dW = -PdV
t
= -Pd(nV) = -nPdV
So, for a constant volume process, the work is zero and we have
dQ = n dU (at constant V)
or, in integral form
Q = nU (at constant V)
At constant pressure, the energy balance becomes
d(nU) = dQ - nPdV (at constant pressure)
or
dQ = n(dU + PdV) = nd(U + PV) (at constant pressure)
This motivates the definition of enthalpy as H = U + PV, so that at constant pressure dQ = n(dH),
just as at constant volume dQ = n(dU).
Note that even though Q is not a state function, H is defined in terms of U, P, and V, which are
state functions, so H is a state function as well.
See example 2.8 in SVA.
Heat Capacity
The constant volume heat capacity is defined as
v
V
dU
C
dT
For a constant volume process, we can then write
dU = C
v
dT (at constant V)
Or, integrating over a finite change in temperature
2
1
T
v
T
U C dT =
(at constant V)
For a mechanically reversible constant-volume process, we then have
2
1
T
v
T
Q n U n C dT = =
Then, for a constant pressure process on a fixed amount of a homogeneous substance, we have
dH = C
p
dT (at constant P)
and for finite temperature changes
2
1
T
p
T
H C dT =
(at constant P)
And finally, for a mechanically reversible, constant pressure process, this tells us that
2
1
T
p
T
Q n H n C dT = =
Isothermal compressibility () is defined as
1
T
V
V P
So, we can write
dV
dT dP
V
=
If the isothermal compressibility and the volume expansivity were both zero, then the fluid
would be incompressible, which is a common approximation for liquids. Using constant values
for and is the next-better approximation, which accounts, on average, for the pressure and
temperature dependence of the specific volume of liquids. An exact treatment would require
pressure and temperature dependent values of and (but this approach is rarely required).
See example 3.1 in SVA
For gases, we are familiar with the ideal gas equation of state, which can be written as
RT
V
P
=
or
RT
P
V
=
or
PV
T
R
=
Next, we'll consider more realistic (and therefore mathematically more complex) equations of
state for gases.
At sufficiently low pressure, all gases approach ideal gas behavior, which means that they obey
the ideal gas equation of state:
PV = RT
where V is the molar volume (volume per mole). The ideal gas equation of state provides the
basis for virtually all equations of state for gases. Equations of state that are more accurate at
higher pressures usually add corrections to the ideal gas equation of state. Since it is the baseline
against which other equations are derived, we will start by going through important properties of
ideal gases and and how to do process calculations on them.
An important property of an ideal gas is that its internal energy depends only on temperature
(and not on pressure). This, in turn, implies that the constant volume heat capacity is a function
only of temperature (and not of pressure or specific volume). That is
( )
( )
v v
V
U U T
U dU
C C T
T dT
=
= =
From the definition of enthalpy and the constant pressure heat capacity, we can show that H is
also a function only of temperature, and we can derive the relationship C
p
= C
v
+ R as follows:
( ) ( )
( )
p v
P
H U PV U T RT H T
H dH dU
C R C T R
T dT dT
+ = + =
= = + = +
Because enthalpy and internal energy are functions only of temperature, we have
dU = C
v
dT
and dH = C
p
dT
for all processes (not just the usual constant volume and constant pressure processes,
respectively).
So, for an ideal gas undergoing any process we have
2
1
2
1
T
v
T
T
p
T
U C dT
H C dT
=
=
Ideal Gas Equation,
regardless of the details of the process.
Process calculations with ideal gases
Calculating the heat and work required to affect changes in an ideal gas is particularly simple (at
least compared to similar calculations with more complex equations of state). For an ideal gas in
a closed system undergoing a mechanically reversible process, we have (on a per mole or per
mass basis)
dU = C
v
dT = dQ + dW
We also have (again for a mechanically reversible, closed-system process)
dW = -PdV
So we can write
dQ = C
v
dT + PdV
We can rearrange these relationships and use the ideal gas law to eliminate one of the 3 variables
(P, V, or T) to solve particular problems. If we substitute P = RT/V, we get
dW = -(RT/V) dV
dQ = C
v
dT + (RT/V) dV
If we substitute V = RT/P, from which
2
R RT
dV dT dP
P
P
= , we get
RT
dW RdT dP
P
= +
v p
RT RT
dQ C dT RdT dP C dT dP
P P
= + =
Finally, if we substitute T = PV/R, from which
V P
dT dP dV
R R
= + , we get
dW = -PdV
p
v
v
C P
C V V P
dQ C dP dV PdV dP dV
R R R R
= + + = +
We can use these relationships to solve for the heat and work requirements for all sorts of
processes.
Isothermal Processes:
Since the internal energy and enthalpy of an ideal gas are functions only of temperature, they do
not change in an isothermal process
U = H = 0 for an isothermal process
Using the equations from above with temperature and volume as the independent variables:
dW = -(RT/V) dV
dQ = C
v
dT + (RT/V) dV = (RT/V) dV
We can integrate to get the following expressions for the heat and work
W = -RT ln(V
2
/V
1
) = RT ln(P
2
/P
1
)
Q = RT ln(V
2
/V
1
) = -RT ln(P
2
/P
1
)
Isobaric Processes:
For a constant pressure (isobaric) process, we can start from the equations written above in terms
of pressure and temperature
RT
dW RdT dP RdT
P
= + =
p p
RT
dQ C dT dP C dT
P
= =
Integrating these, we get
( ) ( )
2 1 2 1
W R T T P V V = =
2
1
T
p
T
Q C dT =
We also have, as always,
2
1
2
1
T
v
T
T
p
T
U C dT
H C dT
=
=
Isochoric Processes:
For constant volume (isochoric) processes, we can use the equations in terms of T and V to get
dW = -(RT/V) dV = 0
dQ = C
v
dT + (RT/V) dV = C
v
dT
From which
W = 0
2
1
T
v
T
Q C dT =
We also have, as always,
2
1
2
1
T
v
T
T
p
T
U C dT
H C dT
=
=
Adiabatic Processes: Ideal gas with constant heat capacity
For an adiabatic process, by definition no heat enters or leaves the system, so dQ = 0.
We can therefore take any of the expressions we had above for dQ and set it equal to zero.
Doing so gives
dQ = C
v
dT + (RT/V) dV = 0, or C
v
dT = - (RT/V) dV
and 0, or
p p
RT RT
dQ C dT dP C dT dP
P P
= = =
and 0, or
p p
v v
C P C P
C V C V
dQ dP dV dP dV
R R R R
= + = =
If the heat capacities are constant, we can integrate each of these to get
2 2
1 1
2 1
1 2
ln ln
v
v
v
R
C
dT R dV
T C V
T V R
T C V
T V
T V
=
=
=
and
2 2
1 1
2 2
1 1
ln ln
p
p
p
R
C
dT R dP
T C P
T P R
T C P
T P
T P
=
=
=
and
2 2
1 1
2 1
1 2
ln ln
p
v
p
v
p
v
C
C
C
dP dV
P C V
C
P V
P C V
P V
P V
=
=
=
Often, we use the relationship C
p
= C
v
+ R and we define the heat capacity ratio as
p p
v
v v p
C C
C R
C C C R
+
= = =
To get
1
1 1
1
v
p v
p p
R
C
C C
R
C C
= = =
Which then gives
( )
( )
1
2 1
1 2
1
2 2
1 1
2 1
1 2
T V
T V
T P
T P
P V
P V
=
=
=
for an ideal gas with constant heat capacity in an adiabatic process
which we can also write as
( )
( )
1
1
constant
constant
constant
TV
TP
PV
=
=
=
for an ideal gas with constant heat capacity in an adiabatic process
Polytropic Processes:
We define a polytropic process as one for which
PV
= constant
For some value of . For this model, SVA give equations for relating T and V, for relating T and
P, and for computing heat and work requirements. Rather than derive them, we will now work
through examples 3.2 through 3.6 (or as many as possible before the end of class).
For gases, we know that at constant temperature, the product of pressure and specific volume is
approximately constant. Therefore, we might try to write an equation of state for gases that
looks like:
2 3
... PV a bP cP dP = + + + +
where the coefficients a, b, c, etc. are functions only of temperature (and not of pressure). This
is an infinite series, but if the PV product is a weak function of pressure, then it can be truncated
after a few terms. We could also write it as
( )
2 3
1 ' ' ' ... PV a B P C P D P = + + + +
where we've simply re-defined the constants as b = aB', etc.
It turns out that a in the above equation is the same temperature-dependent function for all gases,
though the other constants are, in general, different for each particular gas. Experimentally, as
the pressure becomes very small (mathematically, in the limit that P goes to zero) the pressure-
volume product (PV) goes to the same value for all gases. This is used to establish the ideal gas
temperature scale and, in turn, the value of the Universal Gas Constant. If we denote the gas-
independent value of PV as the pressure goes to zero as (PV)*, then we can define the ideal gas
temperature scale by assigning the quantity (PV)* to be proportional to temperature
( )* PV a RT = =
where the proportionality constant is the Universal Gas Constant. R is then defined by setting
the triple-point of water to be at T = 273.16 K and using the experimental value of (PV)* at the
triple-point of water, which is (PV)* = 22711.8 cm
3
bar mol
-1
. This gives
3 -1 -1 3 -1 -1 -1 -1
22711.8
83.1447 cm bar mol K 8.31447 m Pa mol K 8.31447 J mol K
273.16
R = = = =
So, this gives us the first term in our virial equation of state - the pressure (and specific volume)
independent term that is the same for all gases. Since it will always be the same, we could move
it to the other side of the equation and write
2 3
1 ' ' ' ...
PV PV
Z B P C P D P
RT a
= = + + + +
We define Z as the compressibility, which is equal to 1 for an ideal gas. Variations in Z away
from 1 then indicate non-ideal behavior. Sometimes the virial equation of state is written as an
infinite series expansion in inverse volume rather than in pressure, in which case the virial
coefficients are different.
VIRIAL EQUATIONS OF STATE:
2 3
1 ...
PV B C D
Z
RT V
V V
= + + + +
where, with some mathematical manipulation, we could relate the two forms of the virial
coefficients as
( )
2
2
'
'
B
B
RT
C B
C
RT
=
=
and so forth.
In either form (volume-expansion or pressure-expansion), the virial equations will only be useful
for practical applications when they converge within a few terms. This is true both because the
computations would become inconvenient and because only the first few coefficients are likely
to be available from experiments. So, for practical applications, we will truncate these to a one-,
two-, or three-term form. The one-term form, in which we keep only the first term is just the
ideal gas law:
1
PV
Z
RT
=
Truncating the pressure-expansion version after two terms gives
1 ' 1
PV BP
Z B P
RT RT
= + = +
where, in the second version, we have substituted '
B
B
RT
= . Alternatively, we could truncate
volume expansion as
1
PV B
Z
RT V
= +
but the other expression in terms of P is more convenient and more commonly used.
We could also truncate the volume expansion after 3 terms to get a 3-term virial equation
2
1
PV B C
Z
RT V
V
= + +
This equation gives P in terms of V directly, but is cubic in V and therefore if one wants to solve
for V, one generally does so iteratively. Remember that the coefficients B and C are temperature
dependent, and therefore the solution for T in terms of P and V (which we usually don't want to
do anyway) may be simple or complex depending on the temperature dependence.
Virial coefficients beyond the 3rd virial coefficient are rarely known. If we look at a plot of the
compressibility of a typical substance like this one for CO
2
, which I've prepared from data
downloaded from the NIST WebBook, we can get an idea of the range over which a 1, 2, or 3-
term virial expansion may be appropriate. A similar plot for methane is given on p. 87 of SVA.
Here is the same plot with the different truncations of the virial equation superimposed on it.
Compressibility of CO
2
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
0 10 20 30 40 50 60 70 80 90 100
Pressure (MPa)
Z
=
(
P
V
)
/
(
R
T
)600 K
400 K
Compressibility of CO
2
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
0 10 20 30 40 50 60 70 80 90 100
Pressure (MPa)
Z
=
(
P
V
)
/
(
R
T
)
400 K, actual
600 K, actual
600 K, 3-term
600 K, 2-term
1-term
400 K, 2-term and 3-term
The values of the virial coefficients are obtained from the slope and curvature of the curves as P
goes to zero:
0
2
2
0
'
1
'
2
P
P
dZ
B
dP
d Z
C
dP
=
=
For this particular case, we see that at 400 K, the 3-term expansion doesn't work any better than
the 2-term expansion, and both are good to about 10 or 15 MPa (100 to 150 atm). At 600K, the
2-term expansion is only good to about 10 MPa, but the 3-term expansion is good out to more
than 50 Mpa. Clearly both the values of the virial coefficients and the range over which a given
expansion is useful depend on temperature.
For precise fitting of measured PVT data, one can use an extended virial equation in which
additional terms are added to the 3-term virial expansion to allow more flexibility in data fitting.
An example is the Benedict/Webb/Rubin (BWR) equation of state:
2
0 0 0
2 3 6 3 2 2 2
1 exp
B RT A C T RT bRT a a c
P
V V V V V T V V
= + + + + +
This has 8 parameters, and sometimes some of these are taken to have additional temperature
dependence. We will not make much use of the BWR equation of state in this course, but for
substances for which the parameters are available, it can provide high accuracy.
Example 3.7 from SVA
Cubic Equations of State:
So far, we have talked about equations of state that are basically good for gases - ideal or non-
ideal, but still gas-like (that is, with compressibility reasonably near 1). There are many
situations where we would like to have an equation of state that describes both the liquid and the
vapor phase of a substance. The simplest and most widely used of such equations are cubic
equations of state. By cubic, we just mean that they are cubic polynomials in the specific
volume. The first (historically) and simplest of these is the van der Waals equation of state.
The van der Waals Equation of State:
This was the first practical cubic equation of state, and it is:
2
RT a
P
V b
V
=
where a and b are positive. If a and b were both zero, this would become the ideal gas equation
of state.
Some general features of cubic equations of state are illustrated for the van der Waals EOS as
plotted for CO
2
below.
For this, and all cubic equations of state, there is always a single pressure corresponding to a
given volume (at any temperature), but there may be either one or three volumes corresponding
to a given pressure. For temperatures above the critical temperature, there is only a single
volume for a given pressure (or at least only a single physically meaningful volume, which must
be greater than b). For temperatures below the critical temperature, there are 3. The smallest
corresponds to the liquid phase, while the largest corresponds to the vapor phase. The middle
one doesn't have a direct physical meaning. We will consider this all in more detail next time.
van der Waals EOS for CO2
0
50
100
150
200
250
0 200 400 600 800 1000 1200 1400
Volume (cm3/mol)
P
r
e
s
s
u
r
e
(
b
a
r
)
600 K
400 K
300 K
250 K
200 K
r
c
r
c
T
T
T
P
P
P
That is, if we scale the temperature and pressure by the critical temperature and pressure, all
fluids are about the same. This is nearly true for simple fluids (monatomic gases except helium),
but less nearly true for other substances.
The description of other fluids can be improved by introducing a third parameter. The most
popular 3rd parameter is the acentric factor, . The acentric factor of a pure substance is
defined in terms of its vapor pressure. If we plot the log of the vapor pressure vs. inverse
temperature, we get approximately a straight line. If we plot this as reduced pressure vs. inverse
reduced temperature, we get a plot like the one shown below.
of the word theorem. It might better be termed a 'coarse generalization of experimental
observations' or something like that. It is not a statement that is mathematically provable or that is
always true.
The reduced temperature and reduced pressure are defined as
The Theorem of Corresponding States
The two-parameter theorem of corresponding states, which
is based on experimental observations, says that all fluids, when compared at
the same reduced pressure and temperature, have about the same compressibility. Note that
this is an unusual use
Vapor Pressure Curves
-4
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
1/Tr (1/K)
l
o
g
1
0
(
P
r
,
s
a
t
)
Argon (blue)
Methane (magenta)
Hexane
Ethanol
Tr=0.7
The acentric factor is defined as the difference in the value of this curve from the value of -1.0
that is observed at T
r
= 0.7 for simple fluids (argon, neon, krypton). Thus, to find it one needs
one vapor pressure measurement (at T
r
= 0.7) as well as the critical properties. This is the basis
of the three-parameter theorem of corresponding states: all fluids that have the same
acentric factor, when compared at the same reduced pressure and temperature, have
about the same compressibility.
This is more generally true than is the 2-parameter theorem of corresponding states. That is, it
agrees with reality for more cases, because it contains one more parameter and is therefore more
flexible.
The acentric factor is used in some equations of state, like the Soave-Redlich-Kwong equation of
state and the Peng-Robinson equation of state, where it shows up in the function a(T). It is also
used in the generalized correlations for the compressibility of gases that we will discuss shortly.
Z = Z
0
+ Z
1
where Z is the compressibility of the gas (or supercritical fluid), is the acentric factor that we
just discussed, and Z
0
and Z
1
are both functions of the reduced temperature and reduced pressure.
The most popular of these correlations is the Lee/Kesler correlation. For it, values of Z
0
and Z
1
are found in appendix E of the textbook. To use these, you first look up the critical temperature,
critical pressure, and acentric factor for the substance of interest, then compute the reduced
pressure and temperature, then look up Z
0
and Z
1
in the tables (interpolating between entries) and
then finally put them in the above equation for Z. For nonpolar or slightly polar gases, this
correlation should give the compressibility to with 2 or 3 percent.
Generalized correlations for gases:
There are a variety of generalized correlations that can be used to estimate the PVT behavior of
real (non-ideal) gases. These are often of the form
Unsolved Problems
prob. 3.6, 3.8, 3.14, 3.27, 3.46
Over a limited range of reduced temperatures and pressures (shown on p. 103 of SVA), this
correlation can be represented by a two-term truncated virial equation of state as follows. The
compressibility is written as
1 1
c r
c r
BP P BP
Z
RT RT T
= + = +
with the correlation
0 1 c
c
BP
B B
RT
= +
which gives
0 1
1
r r
r r
P P
Z B B
T T
= + +
or, in the form of the original Lee/Kesler correlation:
0 1
0 0
1 1
1
r
r
r
r
Z Z Z
P
Z B
T
P
Z B
T
= +
= +
=
The coefficients B
0
and B
1
are functions only of temperature (remember that the virial
coefficients themselves are functions only of temperature). They are well represented by:
0
1.6
1
4.2
0.422
0.083
0.172
0.139
r
r
B
T
B
T
=
=
Generalized correlations for liquids
The Lee/Kesler correlation discussed above for gases also includes data for subcooled liquids.
Figure 3.14 on p. 101 of SVA shows its predictions for both liquids and gases.
For saturated liquids (those in equilibrium with the vapor) some simpler correlations work. The
Rackett equation gives good results and only requires the critical parameters as inputs:
( )
0.2857
1
r
T sat
c c
V V Z
=
Another graphical correlation for liquid densities is shown in figure 3.17 on p. 109 of SVA.
So, now we have a variety of equations of state and correlations to compute the PVT behavior of
fluids. In the homework and in-class examples, we will practice applying these for such direct
calculations (i.e. compute volume given temperature and pressure), and then we will go on to use
them throughout the rest of the semester to solve more complicated and more interesting
problems.
and
v
V T T
U U U
dU dT dV C dT dV
T V V
| | | | | |
= + = +
| | |
\ . \ . \ .
where we have used the definition of C
v
that we have already discussed in previous lectures.
The last term (the dV term) is zero for
Any constant-volume process, with any fluid
Any fluid where internal energy does not depend on molar volume, including
Ideal gases
Incompressible liquids
In those cases, we have simply
2
1
v
T
v
T
dU C dT
U C dT
=
=
Similarly for enthalpy, we write
H = H (T,P)
and
p
P T T
H H H
dH dT dP C dT dP
T P P
| | | | | |
= + = +
| | |
\ . \ . \ .
where we have used the definition of C
p
that we have already discussed.
The last term (the dP term) is zero for
Any constant-pressure process, with any fluid
When we put heat into (or remove heat from) a fluid, this heat might be used (in
whole or in part) to do work on the surroundings, to change the temperature of the fluid, to
change the phase of the fluid, or to provide the energy required to carry out a chemical reaction.
In the next couple of lectures, we will talk about these heat effects. The first ones that we will
consider are sensible heat effects, which involve heating that changes the temperature of the
system.
Sensible Heat Effects
If we consider a closed system where there are no phase transitions, no chemical reactions, and
no changes in the composition of the system, then adding heat to or removing heat
from the system will change its temperature and/or cause it to do work on the surroundings.
Our goal here is to relate the temperature change and work done to the
amount of heat added. For a homogeneous substance of constant composition,
the phase rule tells us that two intensive properties must be fixed to specify the
state of the fluid completely. Thus the internal energy or enthalpy can be written as a function
of two intensive variables, like temperature, pressure, and molar volume. Thus, using T and V as
the independent variables for U, we write
U = U (T,V)
SENSIBLE HEAT AND LATENT HEAT
Any fluid where the enthalpy does not depend on pressure, particularly ideal gases
In those cases, we have simply
2
1
p
T
p
T
dH C dT
H C dT
=
=
Furthermore, we know that for constant-volume processes, Q = U, and for constant-pressure
processes, Q = H. This lets us calculate things like the amount of heat required to change the
temperature of a substance by a certain amount.
Heat capacities are often expressed in the form of polynomials in temperature, or as other simple
functions. The one used in SVA most commonly is
2 2 p
C
A BT CT DT
R
= + + +
Note that for ideal gases, the constant volume and constant pressure heat capacities are related by
C
p
= C
v
+ R
So for an ideal gas, if C
p
is given by the form shown above, then C
v
is given by
2 2
1
v
C
A BT CT DT
R
= + + +
For incompressible liquids, C
v
and C
p
are the same (a constant pressure process on an
incompressible substance is also a constant-volume process, since the volume is always constant
for an incompressible substance).
The above form of the heat capacity is easily integrated - for example
( )
( )
( ) ( )
2 2
1 1
2 2
2 2 3 3
2 1 2 1 2 1
2 1
1 1
2 3
T T
p
T T
H C dT R A BT CT DT dT
B C
H R A T T T T T T D
T T
= = + + +
( | |
= + +
( |
( \ .
,
all
p i p i
i
C y C =
SVA gives a longer description of evaluation of this, but I assume you can all do integrals,
so let's not spend time on it. They also give a representative computer program for doing
this in appendix D.
We might similarly have a handy function in an Excel spreadsheet
For mixtures of gases at low pressure, one can simply take the heat capacity of the mixture as the
average of the heat capacity of the components, weighted by their mole fractions in the mixture.
That is, if y
i
is the mole fraction of species i in the mixture, then the molar heat capacity of the
mixture is simply
Latent Heats of Pure Substances
When a pure substance is converted from one phase to another (solid to liquid, liquid to vapor,
etc.) heat flows to or from the substance, but there is no change of temperature. The energy per
mole required to vaporize a liquid is called the latent heat of vaporization, and the energy per
mole required to melt a solid is called the latent heat of fusion. We know from the phase rule
that for a 1-component, 2-phase system, specifying one intensive variable specifies the state of
the system. Thus, the latent heat of a phase change is a function only of one intensive variable,
and can therefore be written as a function of temperature alone. This function can be written as
sat
dP
H T V
dT
=
where H is the latent heat for the phase change (the difference between the molar enthalpies of
the two phases), V is the volume change for the phase change (big for vaporization, small for
melting), and P
sat
is the pressure of the two-phase system (which is a function only of
temperature, since specifying the temperature specifies all other intensive properties of the 2-
phase system). For a vaporization phase change, P
sat
is the vapor pressure of the liquid. This
equation, which we will derive in a few weeks, is called the Clapeyron equation. So, to get the
heat of vaporization, we could either make a direct measurement (using a calorimeter), or we
could compute it from the vapor pressure curve and the volumetric properties of the liquid and
gas. Some correlations for estimating heats of vaporization are given on p. 131 of SVA. I won't
bother to repeat them here, but clearly you should be able to apply them.
The process is carried out at constant pressure, and no non-flow work is done, so
Q = H
That is, the heat added or removed is equal to the change in enthalpy of the stream flowing
through the calorimeter. This, in turn, is equal to the difference in enthalpy between the products
and the reactants, at constant temperature. The heat of reaction for the reaction is defined as this
change in enthalpy at constant temperature.
Now, consider the generic balanced chemical reaction
aA + bB lL + mM
The standard heat of reaction for this reaction at temperature T is defined as the enthalpy change
when a moles of A plus b moles of B in their standard states at temperature T react to form l
moles of L and m moles of M in their standard states at the same temperature T. A standard state
is a particular state of a species at temperature T and at specified conditions of pressure,
composition, and physical condition (e.g. gas, liquid, or solid). In SVA, the standard state
conditions for computing heats of reaction are:
Gases: the pure substance in the ideal-gas state at 1 bar.
Liquids and solids: the pure substance at 1 bar.
Older data tabulations often use 1 atm rather than 1 bar as the standard pressure, but this makes a
negligible difference for most applications. It is particularly important to note and remember
Reactants in at T
o
Products out at T
o
Q (i n or out)
Isothermal
Flow
Calorimeter
Heats of Reaction
So far, the heat effects we have discussed (sensible heat and latent heat
of pure substances) have been for physical processes (processes where no chemical reactions
occur). In fact, pretty much everything that we have done so far this semester has
been for physical processes. However, as you might guess, chemically reacting systems are
an important part of chemical engineering. Heat effects due to chemical reaction can be
quite large compared to those associated with physical processes. During a chemical
process, energy is stored in and released from the chemical bonds that form between atoms in
molecules. Since these bonds are, usually, much stronger than the physical interactions
between molecules, they can store and release much greater amounts of energy.
Consider an isothermal flow calorimeter for measuring the heat released or consumed by
reaction. In such a device, a stoichiometric mixture of reactants flows in and reacts, and
we measure the heat (addition or removal) required to return the reaction products to
the inlet temperature of the reactants.
that the heat of reaction is defined for a particular set of stoichiometric coefficients of the
reaction. For example, we could write the reaction for the partial oxidation of methane to CO
and H
2
(synthesis gas) as
CH
4
+ O
2
CO + 2 H
2
or as 2 CH
4
+ O
2
2 CO + 4 H
2
Either of these is a valid way to write the reaction, but the heat of reaction for the second version
is twice that of the first version, so it is important that one know which version is being used. A
heat of reaction can only be interpreted with respect to a particular writing of the reaction.
Standard Heats of Formation:
It would be impossible (or at least impractical) to tabulate the heats of reaction for all possible
reactions of interest. Therefore, they are calculated from a standard set of heats of reaction. This
set is the heats of formation of the chemical species. The heat of formation of a substance is
defined as the heat of reaction when the substance is formed, in its standard state at temperature
T, from the elements in their standard states at temperature T. For example, the heat of formation
of ethanol (CH
3
CH
2
OH) is equal to the heat of reaction for the reaction
2 C(s) + 3 H
2
(g) + O
2
(g) CH
3
CH
2
OH(l)
This is the formation reaction for ethanol. In this hypothetical reaction, two moles of graphite
(the standard state of carbon) plus 3 moles of hydrogen gas plus one-half of a mole of oxygen
gas produce one mole of liquid ethanol. Heats of formation are defined as the heat of reaction
per mole of the species formed, and therefore these reactions are written so that the
stoichiometric coefficient of the species formed is one.
The value of heats of formation is that we can write any reaction as the sum of formation
reactions. Consider again the partial oxidation of methane to give synthesis gas
CH
4
+ O
2
CO + 2 H
2
This can be written as the sum of the following formation reactions:
C + O
2
CO
CH
4
C + 2 H
2
CH
4
+ O
2
CO + 2 H
2
Note that we don't have to include formation reactions for H
2
or O
2
, since they are elements. The
heat of reaction of the first reaction (C + O
2
CO) is the heat of formation of CO (H
f
(CO)).
The heat of reaction of the second reaction is minus the heat of formation of CH
4
(-H
f
(CH
4
))
(reversing the direction of the reaction reverses the sign of the heat of reaction). So, the heat of
reaction for the partial oxidation is H
rxn
= H
f
(CO) - H
f
(CH
4
). In general, any heat of reaction
can be computed as the heat of formation of the products minus the heat of formation of the
reactants. That is, for the generic reaction
aA + bB lL + mM
the heat of reaction is
H
rxn
= lH
f
(L) + mH
f
(M) - aH
f
(A) - bH
f
(B)
Thus, if we have a table of the enthalpies of formation of species (at some standard temperature),
then we can compute the heat of reaction for any reaction involving those species (at the same
standard temperature). The heats of formation of elements are zero by definition. They form the
reference point for defining the heats of formation of everything else. Usually, heats of
formation are tabulated at a standard temperature of 298 or 298.15 K. To get heats of formation
or heats of reaction at other temperatures, we must integrate the heat capacities of the reactants
and products to see how the enthalpies of the reactions and products change with temperature.
Heats of Combustion:
Most formation reactions cannot actually be carried out in practice. For example, we probably
could not devise a calorimeter where we form ethanol (in 100% yield) from a stoichiometric
mixture of graphite, hydrogen, and oxygen. So, heats of formation usually have to be
determined from calorimetric measurements that can easily be carried out. The most common
class of reactions that goes nicely to completion is combustion reactions. Imagine that we have a
flow calorimeter in which we can burn things. Then we could first burn some hydrogen
H
2
(g) + O
2
(g) H
2
O(l)
From which we could get the heat of formation of water (this is a formation reaction that we can
measure directly). Next, we might be able to burn some graphite directly, as
C(s) + O
2
(g) CO
2
(g)
From which we would get the heat of formation of CO
2
(this is another formation reaction that
we can measure directly). Knowing these, we could boldly start burning all sorts of compounds
made up of only C, H, and O, and could determine their heats of formation from the heats of
reaction for the combustion reaction
C
x
H
y
O
z
+ (x + y - z) O
2
(g) xCO
2
(g) + y H
2
O(l)
and the known heats of formation for carbon dioxide and water. This particular heat of reaction
(to convert 1 mole of the substance burned plus a stoichiometric amount of oxygen to carbon
dioxide and water) is known as the heat of combustion of the material. For example, we could
burn ethanol as
CH
3
CH
2
OH(l) + 3 O
2
(g) 2 CO
2
(g) + 2 H
2
O(l)
The standard heat of reaction for this reaction at 298 K is
( ) ( ) ( )
,298 ,298 2 ,298 2 ,298 3 2
2 ( ) 2 ( ) ( )
o o o o
rxn f f f
H H CO g H H O l H CH CH OH l = +
This is the heat of combustion for ethanol. So, we might also write it as
( ) ( ) ( ) ( )
,298 3 2 ,298 2 ,298 2 ,298 3 2
( ) 2 ( ) 2 ( ) ( )
o o o o
comb f f f
H CH CH OH l H CO g H H O l H CH CH OH l = +
Note that in these reactions we have explicitly indicated the phase (s for solid, l for liquid, or g
for gas) that is being taken as the standard state for each species. From the above
From the above relationship between the heat of formation and the heat of combustion, we have
( ) ( ) ( ) ( )
,298 3 2 ,298 2 ,298 2 ,298 3 2
( ) 2 ( ) 2 ( ) ( )
o o o o
f f f comb
H CH CH OH l H CO g H H O l H CH CH OH l = +
So, let's imagine for a moment that we don't know the heats of formation of anything, but we
have a flow calorimeter that we operate at a constant pressure of 1 bar and an inlet and outlet
temperature of 25 C. We first burn H
2
in it, and find a heat of combustion of -68.3174 kcal/mol.
Next, we burn some graphite and find a heat of combustion of -94.0518 kcal/mol. Finally, we
burn some ethanol and find a heat of combustion of -258.3884 kcal/mol. The formation
reactions for H
2
O and CO
2
are the same as the combustion reactions of H
2
and C, respectively,
so the heats of formation for H
2
O and CO
2
are equal to the heats of reaction that we measure for
the H
2
and C combustion reactions (both are at the same standard state conditions of 25 C and 1
bar). So, we get
( )
( )
,298 2
,298 2
( ) 95.0518 kcal/mol
( ) 68.3174 kcal/mol
o
f
o
f
H CO g
H H O l
=
=
Then, we can compute the heat of formation of ethanol from these, and from the heat of
combustion that we measured for it:
( ) ( ) ( ) ( )
( )
,298 3 2
,298 3 2
( ) 2 94.0518 2 68.3174 267.6988 kcal/mol
( ) 66.35 kcal/mol = -277.61 kJ/mol
o
f
o
f
H CH CH OH l
H CH CH OH l
= +
=
Reassuringly, this agrees with the value in the table in the back of SVA to within 0.1 kJ/mol. I
took the data for the example from Perry's Handbook.
The Temperature Dependence of Heats of Reaction:
So, now we know how to use heats of formation to compute heats of reaction at a given
temperature, but what if we want to know a heat of reaction at some temperature other than the
standard temperature of 25 C at which standard heats of formation are generally tabulated? One
way that we could compute this would be to compute the enthalpy change for a hypothetical
process in which we heat or cool the reactants to the standard state temperature, then carry out
the reaction isothermally, then cool or heat the products back to the temperature of interest. The
total enthalpy change for this 3-step process would be equal to the entropy change for the 1-step
process of carrying out the reaction at the initial temperature. For the first step of heating or
cooling the reactants to 298.15 K, we have
298.15 K
total
1 ,reactants
o
p
T
H C dT =
where
total
,reactants p
C is the total heat capacity of the reactants for the reaction as written. For
example, if we had the generic reaction
aA + bB lL + mM
it would be
total
,reactants
( ) ( )
p p p
C aC A bC B = +
For the second step, carrying out the reaction at 298.15 K, we have
2 ,298.15 K
o o
rxn
H H =
Finally, for the 3rd step, bringing the products back to the initial temperature, we have
total
3 ,products
298.15 K
T
o
p
H C dT =
So, for the overall process, we have
298.15 K
total total
, ,reactants ,298.15 K ,products
298.15 K
T
o o o
overall rxn T p rxn p
T
H H C dT H C dT = = + +
Remembering that switching the limits of an integral (changing the direction of integration) just
changes its sign, we can combine the two integrals to get
( )
total total
, ,298.15 K ,products ,reactants
298.15 K
T
o o
rxn T rxn p p
H H C C dT = +
That is, the heat of reaction at some other temperature is equal to the heat of reaction at the
standard temperature of 298.15 K plus the integral from the standard temperature to the
temperature of interest of the difference in heat capacity between the reactants and the products.
Just like the enthalpy of reaction is the difference between the enthalpy of the products and
reactants, we can define the change in heat capacity of reaction as the difference between the
total heat capacity of the products and the reactants. So, for the generic reaction
aA + bB lL + mM
For which the heat of reaction is
H
rxn
= lH
f
(L) + mH
f
(M) - aH
f
(A) - bH
f
(B)
The difference in heat capacity on reaction would be
C
p
= lC
p
(L) + mC
p
(M) - aC
p
(A) - bC
p
(B)
where all of the heat capacities can be temperature dependent, so their difference is also
temperature dependent. The heat of reaction at some arbitrary temperature T is then
, ,298.15 K
298.15 K
T
o o
rxn T rxn p
H H C dT = +
We could further generalize this by writing the general chemical reaction with an arbitrary
number of reactants and products as done by SVA on p. 137
1 1 2 2 3 3 4 4
... ... A A A A + + + +
where
i
is the stoichiometric coefficient of species A
i
. The A
i
(A
1
, A
2
, etc.) are just the names
of the species participating in the reaction. The convention we will use here (and that most
widely used) is that the stoichiometric coefficients are positive for products and negative for
reactants. Then, the above equation can be formally written as:
1 1 2 2 3 3 4 4
... 0 A A A A + + + + =
or simply
0
i i
i
v A =
With this notation, the heat of reaction can be written as
( )
o o
rxn i f i
i
H v H A =
The change in heat capacity can be written as
, p i p i
i
C v C =
and so on. This is very handy for doing things in an automated way on the computer.
If the individual heat capacities of the species are written in the usual polynomial form that we
used last time
, 2 2 p i
i i i i
C
A BT C T DT
R
= + + +
Then we can write
( )
2 2
,
2 2
p i p i i i i i i
i i
p
C v C R v A BT C T DT
C
A BT CT DT
R
= = + + +
= + + +
where we define
, , ,
i i i i i i i i
i i i i
A v A B v B C v C D v D = = = =
Using this notation with the heat capacity integral to find the heat of reaction at some arbitrary
temperature T, we have
( )
( ) ( )
( )
( )
( )
, ,298.15 K
298.15 K
2 2
, ,298.15 K
298.15 K
2 3
2 3
, ,298.15 K
1 1
298.15 K 298.15 K 298.15 K
2 3 298.15 K
T
o o
rxn T rxn p
T
o o
rxn T rxn
o o
rxn T rxn
H H C dT
H H R A BT CT DT dT
B C
H H R A T T T D
T
= +
= + + + +
( | |
= + + +
| (
\ .
As was the case for the usual heat capacity integral, SVA have given a defined function and an
example computer program for this in the appendices of the text. Similarly, we might
use a spreadsheet . This example spreadsheet is set up to evaluate the heat of reaction for
methanol synthesis from synthesis gas, as shown in example 4.6 in SVA.
The final section in chapter 4 of SVA is on 'Heat Effects of Industrial Reactions', which we will
study by working a couple of the examples from that section, which also illustrate the
other things we've been looking at today.
Unsolved Problems
Prob No. 4.8 and 4.19
The first law of thermodynamics says that energy is conserved, and does not differentiate
between energy added to or removed from a system as heat and energy added to or removed
from a system as work. In the first law, heat and work are equivalent. However, our everyday
experience, as well as centuries of scientific studies and engineering experience shows that heat
and work are not equivalent in all ways. Work is spontaneously converted to heat all around us,
and we can find many examples where work is fully converted to heat. The classical example of
this is the historic experiments of Joule where he stirred fluids and measured the amount of work
done and the temperature rise of the fluids. However, there are no examples of energy in the
form of heat being spontaneously and completely converted to work, despite centuries of effort
directed at constructing a device that would do so. This suggests that there is some sense in
which work is a higher or more valuable form of energy than heat, and in fact if that were not the
case, there would not be much need to differentiate between the two. This idea is formalized and
quantified in the second law of thermodynamics.
There are many ways to state the second law of thermodynamics. Given one of them, it is
possible (sometimes by simple means, sometimes by more elaborate means) to derive the others.
Two possible statements of it (from SVA, p. 156) are:
No device can operate such that its only effect is to convert heat absorbed by a system
completely into work done by the system.
Or
No process is possible that consists solely of the transfer of heat from one temperature level to a
higher one.
We will provide more mathematical statements of the second law shortly. This will all be done
quickly, on the assumption that it is review of material recently learned in EAS 204.
Heat Engines, Carnot Engines, and so forth:
A heat engine is a device that converts heat to work. Much of classical thermodynamics was
developed in the context of trying to understand and improve steam engines, which powered
locomotives, steamboats, etc. in times past. In these and in other heat engines, heat is absorbed
by the system at some high temperature and partially converted to work and partially transferred
out of the system at some lower temperature. For example, in a steam engine, water flows into a
boiler where it is vaporized and superheated (absorbing heat at a high temperature), then the
steam expands through a mechanical device like a turbine, converting part of this heat to work.
The steam then condenses (transferring heat out of the system at a low temperature) and is
returned to the boiler, so that the overall process is cyclic. The working fluid of the heat engine
(i.e. the steam) absorbs some heat |Q
H
| at a high temperature, does some net amount of work |W|
on the surroundings, then rejects some heat |Q
C
| at low temperature. Since the overall process is
THE SECOND LAW OF THERMODYNAMICS
cyclic, the overall change in the state of the system (and therefore in the internal energy of the
working fluid) is zero. So, the first law tells us that
|W| = |Q
H
| - |Q
C
|
That is, the net work done is equal to the difference between the heat flow in at high temperature
and the heat flow out at low temperature. Note that for the example of the steam engine, the net
work is some relatively large amount of work done by the expanding steam minus some
relatively small amount of work that must be done to pump the water up to the high pressure of
the boiler. The thermal efficiency of a heat engine is simply defined as the fraction of the heat
input that is converted to work. That is
net work output
1
heat absorbed
H C C
H H H
W Q Q Q
Q Q Q
= = =
If we could just eliminate the rejection of heat at from the system (make |Q
C
| zero), we could
achieve unity (100%) efficiency. However, the second law of thermodynamics forbids this.
In fact, the most efficient possible engine is one that operates in an entirely reversible fashion
(with no irreversibilities). Such an idealized engine is called a Carnot engine, after N.L.S.
Carnot, who first described it. In such a (hypothetical) device, the following four steps make up
the engine cycle:
1. Reversible, adiabatic compression from temperature T
C
to temperature T
H
.
2. Reversible, isothermal heat transfer during which |Q
H
| is absorbed from the surroundings.
3. Reversible, adiabatic expansion from temperature T
H
to T
C
.
4. Reversible, isothermal heat transfer during which |Q
C
| is rejected to the surroundings,
bringing the working fluid back to its original state.
Any engine that operates in a fully reversible way is considered a Carnot engine. In SVA, they
demonstrate that (a) no engine can have a higher thermal efficiency than a Carnot engine (for the
same high and low temperature levels) and (b) the thermal efficiency of a Carnot engine depends
only on the high and low temperature levels.
Thermodynamic Temperature Scales:
The fact that the efficiency of a Carnot engine depends only on temperatures and not on any
properties of the working fluid allows us to establish a thermodynamic temperature scale that is
independent of any material properties. Conveniently, this will work out to be the same as the
ideal gas temperature scale (and the Kelvin scale) that we have already been using. Assuming
that this was well covered in EAS204, we will skip the derivations. In SVA, they demonstrate
that if temperature is defined by the ideal gas temperature scale, then the thermal efficiency of a
Carnot engine is
1
C
H H
W T
Q T
= =
and the heat absorbed and heat rejected are related by
H
H
C C
Q
T
Q T
=
To get 100% efficiency, we would have to either have a cold reservoir at 0K, or a hot reservoir at
infinite temperature. Obviously, neither of these is achievable.
Entropy:
From the above discussion of Carnot engines, we get (among other things) a relationship
between the heat absorbed at T
H
and the heat rejected at T
C
in a Carnot cycle:
H
H
C C
Q
T
Q T
=
or
H C
H C
Q Q
T T
=
If we take the working fluid as our system and use our convention that heat flow into the system
is positive, then Q
H
is positive and Q
C
is negative, and we have
C H
H C
Q Q
T T
=
or
0
C H
H C
Q Q
T T
+ =
Thus, for a complete cycle, the two quantities Q/T associated with heat transfer into and out of
the system sum to zero. A key characteristic of a thermodynamic property (like temperature,
pressure, internal energy, etc.) is that for any cyclic process, its final value is the same as its
initial value. This suggests that there may be another thermodynamic property of the system
whose changes are given by Q/T. In SVA they approach this by pointing out that an arbitrary
reversible, cyclic process can be constructed from many Carnot cycles in which an infinitesimal
amount of heat dQ
H
is absorbed at temperature T
H
and another infinitesimal (but different)
amount of heat dQ
C
is rejected at temperature T
C
. From such an analysis, one can conclude that
0
rev
dQ
T
=
That is, the integral over any cyclic, reversible process of dQ/T is zero. This is then used to
define a new property entropy, denoted by the symbol S
t
, as
t rev
dQ
dS
T
=
or
t
rev
dQ TdS =
The superscript t is used here to indicate that this is the total entropy of the system (an extensive
quantity) not the molar or specific entropy (an intensive quantity). We can show that entropy is a
state function, since we could evaluate the change in entropy from state A to state B by doing the
integral
1
A to B via path 1
t rev
dQ
S
T
=
along one reversible path, and then do the integral from state B back to A along another
reversible path
2
B to A via path 2 A to B via path 2
t rev rev
dQ dQ
S
T T
= =
Since the total integral over the cyclic path from A to B and back to A must be zero, we have
1 2
0
t t
S S + =
Since we are free to pick an arbritrary reversible path for either part 1 or part 2, the entropy
change must be independent of the path selected, and
1 2
A to B by any
reversible path
t t t t rev
B A
dQ
S S S S
T
= = =
If the system goes from state A to state B by an irreversible process, then the entropy change will
still be the same, but it will not be given by the integral, of dQ/T along the irreversible path. So,
even for irreversible processes, we can always compute the entropy change along some
hypothetical reversible path that connects the same initial and final states as the actual
irreversible path. If a process is both reversible and adiabatic, then dQ
rev
will be zero throughout
the process, and dS
t
will be zero throughout, and finally, S
t
will be zero. So, a reversible,
adiabatic process is also called an isentropic process.
The discussion of entropy so far can be summarized as:
There exists a property that we will call entropy and denote by S. It is related to observable
properties of the system. For a reversible process, changes in this property are given by
t rev
dQ
dS
T
=
The change in entropy for any process that takes a system from state A to B (whether the actual
process is reversible or not) is given by
A to B by any
reversible path
t rev
dQ
S
T
=
Mechanically reversible processes are processes where all driving forces for change in the state
of the system are infinitesimal except for temperature differences for heat transfer between the
system and the surroundings. For mechanically reversible processes, the entropy change can still
be calculated as
t
dQ
S
T
=
Although we used a mechanically reversible process to derive this, the final two equations relate
only properties of the system (state functions) and are therefore independent of path and apply to
all processes, reversible or irreversible.
For an ideal gas with constant heat capacity, the above equation is even simpler
0 0
ln ln
p
T P
S C R
T P
=
For a gas in which the heat capacity is defined by the sort of polynomial that we have been
using:
2 2 p
C
A BT CT DT
R
= + + +
then
( )
( ) ( )
0
3
0
2 2 2 2
0 0 0
0 0
ln
ln ln
2 2
T
T
S A D P
B CT dT
R T P
T
S T C D P
A B T T T T T T
R T P
= + + +
= + +
( )
( ) ( )
0
2 2 2 2
0 0 0
0
ln
2 2
T
p
T
C
T C D
ICPS dT A B T T T T T T
RT T
= + +
As they did for enthalpy and internal energy changes, SVA also express this in terms of a mean
heat capacity and provide a defined function and example computer program for it in
the appendices. We can also do this calculation in a spreadsheet that evaluates the
first part of this:
S
total
0 for all processes
That is, the total entropy of the universe (the system plus the surroundings) cannot decrease. For
reversible processes, it does not change. For irreversible processes, it increases. This is, in
essence, a statement that things always happen in a certain direction. Heat always flows from
hot to cold because that increases the total entropy. Or, from a different perspective, we define
entropy such that its total value increases when heat flows from hot to cold. The fact that
spontaneous heat transfer always occurs from hot to cold is then consistent with the version of
the second law as written above. If heat leaves a hot object (at T
H
) and flows into a cooler object
(at T
C
) then the entropy of the hot object decreases by Q/T
H
while that of the cooler object
increases by Q/T
C
. So, the total change of entropy of the universe for this process is Q(1/T
C
1/T
H
), which is positive. If the temperatures were only infinitesimally different, the process
would be reversible, and this difference would approach zero.
Notice that while the first law of thermodynamics is an equality stating that the total amount of
energy in the system plus the surroundings remains constant the second law of
thermodynamics is an inequality stating that the total amount of entropy in the system plus the
surroundings can only remain constant or increase.
Entropy balances for open systems:
We can write an entropy balance for an open system in the same way that we wrote an energy
balance and a mass balance except that entropy is not conserved so we must include generation
of entropy as well as its flow in and out of the system. Mathematically, we can write the entropy
balance as:
( )
( )
0
t
surr cv
g
fs
d mS
dS
Sm S
dt dt
+ + =
The first term is the net rate of entropy transport out of the system via flow. It is the difference
between the entropy carried out by streams flowing out of the control volume and that carried in
by streams flowing into the control volume. The second term is the rate of accumulation of
entropy within the control volume. The third term is the rate of change of the entropy of the
surroundings due to the process of interest.
g
S
Substituting this, the entropy balance becomes
( )
( )
0
j
cv
g
fs
j j
d mS Q
Sm S
dt T
+ =
If the system is at steady state, this simplifies to
( ) 0
j
g
fs
j j
Q
Sm S
T
=
If there is a single inlet and a single outlet from the system, with equal mass flowrates, we can
write
0
j
out in g
j j
Q
S S S
T
=
In this final equation, the heat flow and entropy generation terms are per unit mass of material
flowing through the system, rather than the total heat flows and total entropy generation used
in the previous equations.
Examples 5.5 and 5.6 from SVA
Unsolved Problems
Prob 5.8, 5.10, 5.12, 5.19, 5.29, 5.35 and others, u can solve for practice
The reversible work done on the closed system is
dW
rev
= -Pd(nV)
which is our usual expression developed in the first chapter of SVA.
The reversible heat flow is (as developed in the previous lecture and chapter 5 of SVA)
dQ
rev
= Td(nS)
Combining these last 3 equations gives
d(nU) = Td(nS) - Pd(nV)
or dU
t
= TdS
t
- PdV
t
These last two equations give a general relationship between the primary thermodynamic
variables for the fluid - U, S, T, P and V. Although we derived it for a reversible process, the
final expressions relate only state variables that are path-independent. Therefore, this
relationship holds for any process occurring in a closed system between equilibrium states. The
system can be made up of multiple phases, and it can undergo reactions, as long as no material
enters or leaves it and the initial and final states that we wish to relate are equilibrium states.
For convenience, we define three more thermodynamic properties related to the internal energy.
These are the enthalpy (which we have been using)
H U PV +
the Helmholtz energy (or Helmholtz free energy)
A U TS
and the Gibbs energy (or Gibbs free energy)
G H TS
In lectures 3 through 6 of this course, we looked fairly
intensively at how to compute the volumetric properties (i.e. volume given T
and P) for fluids from equations of state and from correlations based on the theorem
of corresponding states. After that, we looked at heat effects, and computed enthalpy
and internal energy changes when the temperature of a substance is changed
using the heat capacity. Then we defined entropy, and computed entropy changes when the
temperature of a substance was changed, again using heat capacity data. Now, we want to
bring this all together to see how, in general, not just for ideal gases, we can compute changes
in the thermodynamic properties of a fluid (internal energy, enthalpy, and entropy) for a
change in state of the fluid, using PVT and heat capacity data. We will also
define and use two new thermodynamic quantities, the Helmholtz free energy and the
Gibbs free energy.
Property Relations for Homogeneous Phases
First, we will combine the first and second laws to derive a fundamental
property relationship between U, S, T, P and V - all of the primary thermodynamic variables.
We can write the first law for a closed system containing n moles of fluid as:
d(nU) = dQ + dW
For the special case of a reversible process, this is
d(nU) = dQ
rev
+ dW
rev
THERMODYNAMIC PROPERTIES OF FLUIDS,
For each of these, we can transform the thermodynamic property relation for internal energy
using the definitions above to get a new thermodynamic property relation where internal energy
is replaced by enthalpy, Helmholtz energy, or Gibbs energy.
For enthalpy, we have
nH = nU + nPV
so
d(nH) = d(nU) + Pd(nV) + nVdP
or
d(nU) = d(nH) - Pd(nV) - nVdP
equating this to the fundamental thermodynamic property relationship for d(nU)
d(nU) = d(nH) - Pd(nV) - nVdP = Td(nS) - Pd(nV)
so
d(nH) = Td(nS) + nVdP
or
dH
t
= TdS
t
+ V
t
dP
Likewise, differentiating the definition of A gives
d(nA) = d(nU) - Td(nS) - nSdT
so
d(nU) = d(nA) + Td(nS) + nSdT = Td(nS) - Pd(nV)
so
d(nA) = -nSdT - Pd(nV)
or
dA
t
= -S
t
dT - PdV
t
Finally, for the Gibbs energy,
d(nG) = d(nH) - Td(nS) - nSdT
so
d(nH) = d(nG) +Td(nS) + nSdT = Td(nS) +nVdP
so
d(nG) = -nSdT + nVdP
or
dG
t
= -S
t
dT + V
t
dP
The above are all written for the total mass of a closed system. What we are often interested in is
the corresponding relationships on a unit mole basis. For one mole of a homogeneous fluid of
constant composition, they are
dU = TdS - PdV
dH = TdS
+ VdP
dA = -SdT - PdV
dG = -SdT + VdP
These are the fundamental property relationships for a homogeneous fluid, and we will use them
a lot in the next few weeks.
The Maxwell Relations:
We can derive another set of important relationships from the above relationships by taking a
second derivative of them. From the fundamental property relationships and the definitions of
the partial derivatives, we see that
V
S
U
T
S
U
P
V
=
If we take the derivative of the first of these with respect to V at constant S, and take the
derivative of the second one with respect to S at constant V, we have
2
2
V S
S
S V
V
U U T
V S S V V
U U P
S V S V S
= =
= =
Since the order in which we take the two partial derivatives doesn't matter, the two expressions
above have to be equal, and we have
S V
T P
V S
=
Doing the same thing with the fundamental property relationship for enthalpy:
P
T
H
T
S
H
V
P
=
and
2
2
P S
S
S P
P
H H T
P S S P P
H H V
S P S P S
= =
= =
so
S P
T V
P S
=
Doing this with the relationship for Helmholtz energy
V
T
A
S
T
A
P
V
=
and
2
2
V T
T
T V
V
A A S
V T T V V
A A P
T V T V T
= =
= =
so
T V
S P
V T
=
Finally, doing this for the Gibbs energy gives
P
T
G
S
T
G
V
P
=
and
2
2
P T
T
T P
P
G G S
P T T P P
G G V
T P T P T
= =
= =
so
P T
V S
T P
=
These four relationships between partial derivatives:
S V
T P
V S
=
S P
T V
P S
=
T V
S P
V T
=
P T
V S
T P
=
are known as the Maxwell Relations. SVA call them Maxwell's Equations.
Relationships for Determining Enthalpy and Entropy as Functions of T and P:
As chemical engineers, we most often encounter problems where temperature and pressure are
known and we want to compute changes in enthalpy and entropy. We have already done this for
ideal gases, for which enthalpy is independent of pressure and entropy depends on pressure in a
very simple way. We can derive a more general relationship from the fundamental property
relationships. To relate enthalpy and entropy to temperature and pressure, we use expressions
like:
P T
P T
H H
dH dT dP
T P
S S
dS dT dP
T P
= +
= +
where we need to know the partial derivatives in these expressions in terms of quantities (like
heat capacity and PVT properties) that we can look up in a table or compute from an equation of
state.
By definition,
p
P
H
C
T
=
, so that one is no problem.
To get the partial derivative of S with respect to T at constant P, we start from the fundamental
property relation for enthalpy:
dH = TdS
+ VdP
and take the differential changes (dH, dS, and dP) to be changes with respect to a differential
change in T at constant P (or 'divide through' by dT and restrict the change to constant P)
P P P P
H S P S
T V T
T T T T
= + =
Replacing
P
H
T
with C
p
, this gives
p
P
C
S
T T
=
We get the partial derivative of entropy with respect to pressure at constant temperature directly
from the Maxwell relation:
T P
S V
P T
=
and
P
V
T
is a quantity that we can evaluate from PVT data or from an equation of state.
Finally, we can get the partial derivative of H with respect to P at constant T by starting from the
fundamental property relation for enthalpy:
dH = TdS
+ VdP
and taking the differential changes (dH, dS, and dP) to be changes with respect to a differential
change in P at constant T (or 'dividing through' by dP and restricting the change to constant T)
T T T T
H S P S
T V T V
P P P P
= + = +
At first, this doesn't seem like an improvement, since we can't measure
T
S
P
any more than we
could measure
T
H
P
. However, we can use the Maxwell relation
P T
V S
T P
=
to get
T P
H V
T V
P T
= +
and
P
V
T
can be evaluated from an equation of state or directly from PVT data.
Putting this all together, we have expressions that relate changes in enthalpy and entropy to
changes in temperature and pressure in terms of measurable quantities (heat capacity and PVT
data):
p
P
p
P
V
dH C dT V T dP
T
C
V
dS dT dP
T T
= +
To find the change in enthalpy of a substance going from P
1
, T
1
to P
2
, T
2
, we can integrate these
at constant pressure from P
1
, T
1
to P
1
, T
2
then integrate at constant temperature from P
1
, T
2
to P
2
,
T
2
. The first of those two steps will be like what we have done for ideal gases. The second of
the two steps is similar, but with respect to pressure.
The pressure dependence of internal energy:
We can derive the pressure dependence of internal energy by taking the derivative of the
definition of enthalpy and using our recently-derived expression for the dependence of enthalpy
on pressure:
H = U + PV
So
T T T T T T
H U V P U V
P V P V
P P P P P P
= + + = + +
Setting this equal to the expression for
T
H
P
derived above:
T T T P
H U V V
P V T V
P P P T
= + + = +
so
T T P
U V V
P T
P P T
=
Simplifications for Ideal Gases:
For an ideal gas, we can very easily evaluate the PVT derivatives in the above expressions. To
evaluate the enthalpy expression we use
, so
ig ig
ig
P
RT V R V
V
P T P T
= = =
Substituting this into
p
P
V
dH C dT V T dP
T
= +
gives
ig
ig ig ig ig
p p
V
dH C dT V T dP C dT
T
= + =
As we have already been assuming, the enthalpy of an ideal gas is independent of pressure.
Similarly, substituting
ig
P
V R
T P
into the expression for entropy changes gives
ig ig
ig
p p
ig
P
C C
V R
dS dT dP dT dP
T T T P
= =
which is the expression for entropy changes in an ideal gas that we came up with in the previous
lecture.
Alternative forms for liquids:
For liquids, we might want to use the isothermal compressibility and volume expansivity to
describe the PVT behavior. We remember from earlier in the semester that:
Volume expansivity () is defined as
1
P
V
V T
Isothermal compressibility () is defined as
1
T
V
V P
Substituting these into
p
P
p
P
V
dH C dT V T dP
T
C
V
dS dT dP
T T
= +
gives
( ) ( ) 1
p p
p
dH C dT V T V dP C dT T VdP
C
dS dT VdP
T
= + = +
=
and for the pressure dependence of internal energy, we get
( )
T T P
U V V
P T PV TV P T V
P P T
= = =
Internal Energy and Entropy as functions of V and T:
Similar to what we did above to get an expression for the dependence of enthalpy and entropy on
P and T in terms of measurable quantities, we would like to express internal energy and entropy
as functions of V and T, so that we can analyze constant-volume or specified-volume problems
where V and T are the natural independent variables.
V T
V T
U U
dU dT dV
T V
S S
dS dT dV
T V
= +
= +
The first partial derivative is, by definition, C
v
v
V
U
C
T
so we don't have to do anything further to obtain it.
To get the partial derivative of S with respect to T at constant V, we start from the fundamental
property relation for internal energy:
dU = TdS - PdV
and take the differential changes (dU, dS, and dV) to be changes with respect to a differential
change in T at constant V (or 'divide through' by dT and restrict the change to constant V)
V V V V
U S V S
T V T
T T T T
= + =
Replacing
V
U
T
with C
v
, this gives
v
V
C S
T T
=
We get the partial derivative of S with respect to V at constant T directly from the Maxwell
relation:
T V
S P
V T
=
Finally, we can get the partial derivative of U with respect to V at constant T by starting from the
fundamental property relation for internal energy:
dU = TdS
- PdV
and taking the differential changes (dU, dS, and dV) to be changes with respect to a differential
change in V at constant T (or 'dividing through' by dV and restricting the change to constant T)
T T T T
U S V S
T P T P
V V V V
= =
Using the Maxwell relation
T V
S P
V T
=
we get
T V
U P
T P
V T
=
and
V
P
T
can be evaluated from an equation of state or directly from PVT data.
So, now we have all of the needed partial derivatives, which we can put together to get U and S
as functions of T and V:
v
V
v
V
P
dU C dT T P dV
T
C P
dS dT dV
T T
= +
= +
These are general equations for a homogeneous substance of constant composition.
The Gibbs Energy as a Generating Function
The fundamental property relations that we derived at the beginning of this lecture were
dU = TdS - PdV
dH = TdS
+ VdP
dA = -SdT - PdV
dG = -SdT + VdP
The form of these suggests that there is a 'natural' set of variables for each of the four
thermodynamic functions U, H, A, and G. The natural independent variables for U are S and V.
For H they are S and P. For A they are T and V. For G, they are T and P. In each case, these
natural independent variables are known as the 'canonical' variables for the corresponding
thermodynamic function. Of course, as we have done above, we can write each of the
thermodynamic quantities in terms of different pairs of independent variables.
The Gibbs energy, G, is special because it has as its natural variables, T and P, which are the
most readily measured and controlled of the four possible combinations of independent variables
(clearly S is the least measurable and controllable, and V is usually only controllable for closed
systems).
We will now demonstrate that if we know G as a function of its canonical variables (T and P),
then we can determine all of the other thermodynamic properties by simple differentiation and
algebraic manipulation. G is said to serve as a generating function for the other thermodynamic
properties.
To do this, we will work with G/(RT), which is dimensionless. Using the rules of calculus,
2
1 G G
d dG dT
RT RT
RT
=
Substituting
dG = -SdT + VdP
we have
2 2
G S V G V G ST
d dT dP dT dP dT
RT RT RT RT
RT RT
+
= + =
Recognizing that G + ST is, by definition, H, we have
2
G V H
d dP dT
RT RT
RT
=
This tells us that
T
V G
RT P RT
=
and
2
P
H G
T RT
RT
=
So, if we know G as a function of P and T we can get H and V as functions of P and T directly.
Given that, we can get S, U, and whatever other thermodynamic quantities are of interest form
the defining equations, i.e.
S H G
R RT RT
=
and
U H PV
RT RT RT
=
R ig
M M M
So, for example, the residual volume is
( ) 1
R ig
RT RT
V V V V Z
P P
= =
Equations that relate molar properties linearly can generally also be written for residual
properties. For example, the definition of the Gibbs energy in terms of the enthalpy and energy
can be written in general as
G = H - TS
and for an ideal gas as
G
ig
= H
ig
- TS
ig
and taking the difference between the two gives
G
R
= H
R
- TS
R
Likewise, in the last lecture notes, we developed the equation
2
G V H
d dP dT
RT RT
RT
=
Writing this for an ideal gas, we have
2
ig ig ig
G V H
d dP dT
RT RT
RT
=
Subtracting the ideal gas version from the original version of this equation gives
2 2
ig ig ig
G G V V H H
d d dP dP dT dT
RT RT RT RT
RT RT
=
or
2
R R R
G V H
d dP dT
RT RT
RT
=
this equation is called the fundamental property relation for residual properties. It shows us that
if we know the residual Gibbs energy as a function of temperature and pressure, we can
determine all of the other thermodynamic properties. Like the equations from which it was
derived, it applies to a unit mole of a homogeneous fluid of constant composition.
From it, we see that
R R
P
V G
RT P RT
As the pressure goes to zero, all gases become ideal, so the residual volume at P = 0 is zero.
Likewise, the residual Gibbs energy is zero at P = 0. Thus, we could integrate the above
Residual Properties
For compressible fluids, a natural framework for describing thermodynamic properties is to use
the ideal gas state as a reference. Thus, we will define residual properties as
the difference between the actual value of a property of a real material
and the corresponding ideal gas property. That is, for the generic molar property M
(which could be V, S, H, G, etc.), we define
equation with respect to pressure from P = 0 to some pressure of interest to get the residual
Gibbs energy at that pressure.
( )
0 0 0
1 1
1
P P P
R R
G V RT Z
dP Z dP dP
RT RT RT P P
= = =
(at constant T)
If we have an equation of state, correlation, or table of experimental data from which we can
obtain Z, we can do this integral to get the Gibbs energy. Thus, we can obtain the Gibbs energy
from experimentally accessible quantities.
Also from the equation
2
R R R
G V H
d dP dT
RT RT
RT
=
we see that
2
R R
P
H G
T RT
RT
or
R R
P
H G
T
RT T RT
Taking the partial derivative with respect to temperature (at constant pressure) of the integral
with respect to pressure (at constant temperature) that we derived above gives
0 0
1 1
P P
R R
P
P
P
H G Z Z
T T dP T dP
RT T RT T P T P
= = =
The residual entropy is obtained from
G
R
= H
R
- TS
R
or
R R R
S H G
R RT RT
=
Substituting in the two integrals derived above,
0 0
1 1
P P
R
P
S Z Z
T dP dP
R T P P
=
These residual properties can thus all be obtained from an equation of state or PVT data or
correlations.
Once we have computed residual properties, the corresponding total properties are obtained by
simply adding the residual properties to the ideal gas properties. We have already learned how
to compute enthalpy and entropy changes of ideal gases and have practiced doing so. All that is
required to compute the ideal gas entropy and enthalpy are the ideal gas heat capacities. All that
is required to compute residual thermodynamic properties is PVT data or an equation of state that
represents it. So, if we have the ideal gas heat capacity of a substance and PVT data from low
pressure (where the substance is an ideal gas) to the pressure of interest at the temperature of
interest, we can compute all of the thermodynamic properties of the substance at this temperature
and pressure.
These residual property relations are valid for gases, liquids, supercritical fluids, etc., but
have their greatest utility for gases and gas-like supercritical fluids where the residual
properties are small compared to the ideal gas properties, so that the residual properties are a
small correction to the basic ideal-gas-like behavior. For liquids, the residual properties include
the big property changes that occur upon condensation, and are therefore not small
corrections to the ideal gas properties.
Two-Phase Systems:
When a substance changes phase, say from liquid to vapor, most of its thermodynamic properties
change, including V, S, H, U, etc. However, the Gibbs energy does not change. At the beginning
of the previous lecture, we arrived at the fundamental property relation for the Gibbs energy:
d(nG) = -nSdT + nVdP
which applies to any constant composition (non-reacting) closed system. We can apply this to a
system of n
l
moles of liquid and n
v
moles of vapor (with Gibbs energy G
l
and G
v
, respectively).
The total number of moles is n = n
l
+ n
v
, and the total Gibbs energy is nG = n
l
G
l
+ n
v
G
v
, or nG =
n
l
G
l
+ (n - n
l
)
G
v
. Substituting this into the fundamental property relation gives
d(n
l
G
l
+ (n - n
l
)
G
v
) = -nSdT + nVdP
Now, consider evaporation of some of the liquid at constant temperature and pressure. If the
pressure and temperature are constant, then dT = dP = 0. So, we have
d(n
l
G
l
+ (n - n
l
)
G
v
) = 0
Furthermore, we know that the molar Gibbs energy is a function only of temperature and
pressure, so at constant temperature and pressure, the Gibbs energy (of either phase) cannot
change, that is,
dG
l
= dG
v
= 0 at constant temperature and pressure
so we can take the Gibbs energies out of the derivative:
G
l
dn
l
+ G
v
d(n - n
l
) = 0
Finally, we know that the total number of moles, n does not change in this process, so dn = 0,
And we have
G
l
dn
l
- G
v
dn
l
= (G
l
- G
v
) dn
l
= 0
This shows us that for arbitrary temperature and pressure (thus for all conditions where vapor
and liquid are in equilbrium) G
l
- G
v
= 0. The Gibbs energy of the liquid and vapor, in
equilibrium, are equal.
In general, for any two phases and of a pure species in equilibrium,
G
= G
and dG
= dG
for arbitrary changes in conditions, provided that the two phases remain present.
From this, we can derive the Clayperon equation, with which we became acquainted in a
previous chapter. For the two phases, we write the fundamental property relation
dG = -SdT + VdP
for each of the two phases (which are at the same T and P, since that is a criterion for their being
in equilibrium)
-S
dT + V
dP
sat
= -S
dT + V
dP
sat
or
(S
- S
) dT = (V
- V
) dP
sat
or
sat
dP S S S
dT
V V V
= =
The entropy change S
=
And by definition, the latent heat for the phase change is the heat required to reversibly transfer
one mole of phase to phase . So
H
S
T
=
Substituting this into the equation derived above gives
sat
dP H
dT
T V
which is the Clapeyron equation.
This is most often of use for liquid to vapor phase transitions (which have large changes in
volume and enthalpy) for which we write it as
sat lv
lv
dP H
dT
T V
Thus, a model of how the vapor pressure depends on temperature provides, implicitly, a way to
compute the heat of vaporization. The most commonly used model is the Antoine equation,
which we already encountered early in the semester:
ln
sat
B
P A
T C
=
+
Antoine coefficients (A, B, and C) are available for many compounds. One must be careful with
these because different sources use different units for them, and some sources use the natural
logarithm (as shown here) while other use the base 10 logarithm.
When we have a 2-phase, pure component system, the two phases will have different values of
intensive properties (other than G, T, and P), and the extensive properties of the 2-phase system
will be the sum of the contributions from the two phases. Thus, for example, the total volume
will be
nV = n
l
V
l
+ n
v
V
v
and the molar volume of the 2-phase mixture will be
( ) ( )
1
l v
l v l l v v v l v v l v v l
n n
V V V x V x V x V x V V x V V
n n
= + = + = + = +
where x
l
and x
v
have been defined as the fractions of the mixture in the liquid and vapor phases,
respectively (which can be on either a mass or molar basis).
The fraction of the fluid that is vapor (x
v
) is often called the quality, particularly when referring
to water. Thus, water (steam) with a quality of 1 is pure (saturated) vapor, while steam with a
quality of 0 is pure (saturated) liquid. Relationships like the above relationship for volume can
be written for any thermodynamic property (V, U, H, S, etc.) as
( ) ( )
1
v l v v l v v l l v lv
M x M x M M x M M M x M = + = + = +
r
Thermodynamic Diagrams and Tables:
Thermodynamic data are often presented in various diagrams. Usually pressure vs.
enthalpy, temperature vs. entropy, or enthalpy vs. entropy are chosen as the axes. Phase
boundaries and lines representing constant values of other thermodynamic variables (those
not chosen for the axes) are plotted. In these diagrams, two-phase regions are areas
(rather than lines, as on a pressure vs. temperature plot) and the triple point is a line. Within
the two-phase regions, lines of constant quality (constant phase fractions) can be plotted.
Examples of these are given on p. 220 and 220 of SVA. Also, PH diagrams for methane and
the refrigerant HFC-134a are in the appendices, and the Mollier diagram (entropy vs.
enthalpy) for steam is inside the back cover. Historically, such figures were used for
design calculations. While we are now much more likely to use a computer database for
such purposes, sketching processes on diagrams like these is still an excellent way to get
an overall idea how thermodynamic properties are changing during a process and how the
changes in the various variables are related.
For very well studied materials, tables of thermodynamic properties are available. When these
are available, they are generally more accurate than equations of state or
generalized correlations. The most complete, familiar, and widely used such table is the steam
tables (in the back of SVA), but similar tables for some refrigerants and other
commonly encountered materials are available in handbooks. These can be used with
manual interpolation for hand calculations or can be included as look-up tables
(databases) in process simulators and other automated property calculation tools.
Unsolved Problems 6.7 to 6.10 and 6.30 to 6.35
Mass fraction of species
i i
i
m m
i x
m m
= = =
Likewise, the mole fraction of species i in a mixture is the number of moles of species i divided
by the total number of moles.
Mole fraction of species
i i
i
n n
i x
n n
= = =
Vapor/Liquid Equilibrium (VLE),
So far, we have learned how to compute properties of pure fluids and mixtures
of fixed composition, and how to use these properties with the laws of thermodynamics to
make process calculations on pure fluids. That is, we have learned how to compute how
properties change with conditions of temperature and pressure (or temperature and volume),
how much work and heat are required to make such changes happen, and whether a
proposed process is possible (doesn't violate the 2nd law of thermodynamics). These are
the tools needed to do things like analyzing heat engines, refrigeration cycles, compressors,
pumps, etc. Chapters 7, 8, and 9 in SVA (which we are skipping) discuss some of the
specifics of these calculations.
However, the tools of thermodynamics are very often applied by chemical engineers to
multi-component, multi-phase systems and reacting systems where we deal with mixtures of
varying composition. Thermodynamics tells us what the equilibrium state (composition
of phases in phase equilibrium, extent of reaction in reaction equilibrium, etc.) will be for such
systems, and thereby allows us to predict and analyze the performance of reactors and
separation equipment (distillation columns, extraction columns, flash tanks, etc.) that operate near
equilibrium.
We will begin by looking at some simple treatments of vapor/liquid equilibrium in the
next couple of lectures, and then go on to consider more detailed (and accurate)
models for the thermodynamics of liquid mixtures, and then come back to equilibrium
calculations with these more accurate models.
What do we mean by equilibrium?
In an isolated system, equilibrium is a condition in which no changes in the macroscopic state of
the system occur with time. This means that all potentials that can cause change (like differences
in temperature, differences in pressure, etc.) are balanced. An open system can be in steady state
(not changing in time) but not in equilibrium. If the conditions within the open system would not
change appreciably in time if we isolated it (turned off all heat, work, and material flows into and
out of it) then its contents (and therefore the streams leaving it) can be considered to
be in equilibrium. Given enough time, any isolated system will come to equilibrium.
However, the time required may be longer than time scales of interest, particularly for
the equilibrium of chemical reactions.
Review of ways to measure composition:
The most common measures of the composition of a mixture are mole fractions, mass fractions,
and molar concentrations. The mass fraction of substance i in a mixture is exactly what it sounds
like it would be - the mass of substance i in the mixture divided by the total mass of the mixture.
SVA uses the same notation, x, for both mass fractions and mole fractions. The molar
concentration of a substance in a mixture is the number of moles of it per unit volume. The total
number of moles per unit volume is 1/V where V is the molar volume of the mixture, so
Molar concentration of species
i
i
x
i C
V
= =
The mass fractions and mole fractions, by their definitions, have to sum to 1, so we have
1 1
1 and
i
i i i
i i i i
x
x C x
V V V
= = = =
The Phase Rule and Duhem's Theorem:
Early in the semester, we presented the phase rule without deriving or proving it. Now, we'll
present it with a little more evidence, and also discuss a related rule - Duhem's theorem.
We recall that the phase rule was
F = 2 - + N
which says that the number of degrees of freedom of a system at equilibrium (F) is equal to 2
minus the number of phases present () plus the number of species N. This follows directly from
counting the number of things we have to specify to specify the state of a system, and then
subtracting the number of equations that we have relating these things. The intensive state of the
system is described by the temperature and pressure plus the mole fractions of each species in
each phase. Thus, we need N mole fractions in each phase, plus the temperature and pressure of
the whole system (a total of 2 + N variables) to describe the state of the system. However, we
can't specify all of these independently, because we have (or soon will have) equations relating
many of them. First of all, we have the equations that the mole fractions in each phase must sum
to 1
1
i
i
x =
in each phase
This gives us equations (1 in each phase) relating the mole fractions. We will soon develop
equations that relate the composition in two phases in equilibrium. That is, if we are given the
temperature, pressure, and the composition of phase and we know that phase is in
equilibrium with it, then we can compute the composition of phase . That is, we have
relationships like
( ) 1 2 3 4
, , , , , ,...
i
x f T P x x x x
=
and there are N such relationships for each phase that is in equilibrium with the first phase (that
we've called ). Thus, we have N( - 1) such equations. So, the number of variables that we can
independently specify (the number of degrees of freedom) is
F = 2 + N - - N( - 1) = 2 - + N
Each additional phase that we have in equilibrium adds only N - 1 new degrees of freedom, but
adds N new equations.
A related rule is Duhem's theorem, which applies to the extensive state of a closed system. To
describe the extensive state of the system, we need to know all of the intensive properties of each
phase, plus the number moles (or mass) of each phase that is present. So, we have more
variables (the number of moles of each phase) than we had when we were only specifying the
intensive state of the system. Thus, if we did not know anything else, there would be 2 + N
variables that could be specified to set the extensive state of the system. However, if the system
is prepared from known amounts of each of the species, so that N
i
tot
is known for each species,
then we have N additional relationships like
tot j
i i
j
N x N =
where N
j
is the total number of moles of species j present. In that case, we have N more
equations, leaving only 2 degrees of freedom. Thus, Duhem's theorem can be stated as
For any closed system formed initially from given masses of chemical species, the
equilibrium state is completely determined when 2 independent variables are specified.
Definitions and Qualitative Description of VLE:
We can most easily visualize multicomponent VLE for the simplest possible case, where we
have 2 components and 2 phases (of course, if it is VLE we always have two phases, vapor and
liquid, but we could also have VLLE, with 2 liquid phases, etc.). If we have two components
and two phases, then the number of degrees of freedom remaining is
F = 2 - + N = 2 - 2 + 2 = 2
The intensive state of the two-component two-phase system is described by its temperature,
pressure, liquid phase mole fractions and vapor phase mole fractions. Specifying any two of
these specifies the state of the system. Thus, at given T and P, the mole fractions x
1
, x
2
, y
1
, and
y
2
are fixed. Here, and in most VLE descriptions and calculations, we will use x for the liquid
phase mole fractions and y for the vapor phase mole fractions. We could equally well specify
other pairs of variables (x
1
and y
1
for example) and the remaining variables would be fixed. Of
course we cannot specify x
1
and x
2
independently, because they have to sum to 1. So, the pairs
of variables that we could pick to specify are (T, P), (T, x
1
), (T, y
1
), (T, x
2
), (T, y
2
), (P, x
1
), (P, y
1
),
(P, x
2
), (P, y
2
), (x
1
, y
1
), (x
1
, y
2
), (x
2
, y
1
), or (x
2
, y
2
).
Since specifying two variables implies the values of all of the others, we can plot any of the
intensive variables as a function of any of the pairs of variables listed above. We usually do not
plot or x
2
and y
2
separately from x
1
and y
1
, since they are simply given by x
2
= 1 - x
1
and y
2
= 1 -
y
1
. A plot of one variable as a function of two others will be a surface in 3-dimensional space. If
we take a 'slice' of this surface at a constant value of one of the independent variables, we get a
curve. Since x
1
and y
1
both have the same range (from zero to 1), we can plot them both on the
same axis. Thus, we could plot x
1
and y
1
both vs. T and P, and we would have two surfaces in
three-dimensional space. The figure on the cover of the textbook is a plot of P vs. (x
1
, T) and P
vs. (y
1
, T) on the same plot. The two surfaces look like an envelope, so this sort of construction
is often called the 'phase envelope'. The plot also has superimposed on it a selection of curves
that are 'slices' through the 3-d plot at constant P or T (I think - it's a little hard to see which
direction they are supposed to be going). I've constructed a plot in Matlab from VLE 'data'
predicted by HYSIS for n-pentane and n-heptane using the Peng-Robinson equation of state for
both species. It is shown on the following page, and by clicking here, you can download a
Matlab figure file that you could open in a Matlab plotting window where you could rotate it,
etc.
50
100
150
200
250
0
0.2
0.4
0.6
0.8
1
10
2
10
3
Temperature (celsius)
Heptane/Pentane VLE from PengRobinson
Mole Fractions
P
r
e
s
s
u
r
e
(
k
P
a
)
red = vapor, blue = liquid
In the above figure, the red curves give the pressure vs. temperature for the saturated vapor phase
at fixed heptane mole fractions. The blue curves give the pressure vs. temperature for the
saturated liquid phase at fixed mole fractions. The two surfaces that contain the red curves and
the blue curves, respectively, form the phase envelope.
If we took a mixture of fixed composition (say 50% pentane, 50% heptane) at a fixed pressure
(say 1000 kPa) and heated it starting from a temperature where it is a single phase liquid, we
would observe the following. When we reached the temperature corresponding to a blue curve
in the above figure, the first bubble of vapor would appear. For the above data set this would
happen at about 154 C. This point is called the bubble point. Since a negligibly small amount
of vapor has formed, the liquid phase has the same composition as the overall composition (i.e.
50% of each in this example). The first tiny bit of vapor formed would have a composition that
lies on the surface containing the red lines at the same temperature and pressure. For this
example, this would be 74.5% n-pentane and 25.5% n-heptane. As we continued heating, the
compositions and amounts of the liquid and vapor would change until the last bit of liquid was
vaporized. At that point, the vapor composition would be the same as the overall composition
(50% each), and we would be on a red curve in the above diagram. This point is called the dew
point, and for this example, the dew point temperature (for P = 1000 kPa, and y
1
= y
2
= 0.5)
would be about 173 C. The last bit of liquid to be vaporized would have a composition that lies
on the surface made up of the blue curves at a temperature of 173 C and a pressure of 1000 kPa.
For this example, that would be about 27.3% pentane and 72.7% heptane.
The figures and discussion in SVA on pages 332-341 show and describe a variety of 'slices' at
constant pressure, temperature, or composition through such phase envelopes for different
systems, and we will spend some time looking through these in class.
( ) for all species 1, 2,...,
sat
i i i
y P x P i N = =
where x
i
is the liquid phase mole fraction, y
i
is the vapor phase mole fraction, P
i
sat
is the vapor
pressure of pure component i, and P is the total pressure. In words, Raoults law says that the
partial pressure of each species in the vapor phase is equal to its mole fraction in the liquid phase
times its pure-component vapor pressure.
Bubble Point and Dew Point Calculations using Raoults Law:
The most straightforward, and perhaps the most commonly encountered, types of VLE
calculations are bubble point and dew point calculations. There are 4 types of these, depending
on which conditions are known. These are
Bubble Point Pressure calculation (BUBL P): compute {y
i
} and P given {x
i
} and T.
Dew Point Pressure calculation (DEW P): compute {x
i
} and P given {y
i
} and T.
Bubble Point Temperature calculation (BUBL T): compute {y
i
} and T given {x
i
} and P.
Dew Point Temperature calculation (DEW T): compute {x
i
} and T given {y
i
} and P.
The phase rule tells us that we must fix
F = 2 - + N = N
independent intensive variables to specify the state of the system. Specifying the composition
(mole fractions) in one of the phases sets N 1 independent intensive variables (the Nth mole
fraction doesnt count because it depends on the others since the mole fractions have to sum to
1). Thus, to specify N total intensive variables, we can specify the composition of either the
liquid or the vapor, plus either the temperature or the pressure, leading to the four combinations
listed above. We could take a huge number of other combinations (like specifying both T and P
and N 2 mole fractions), but the four combinations listed are the most common specifications
that are encountered in practice.
Bubble Point Pressure calculations:
Vapor/Liquid Equilibrium
Raoults Law:
The simplest model that allows us do VLE calculations is obtained when we assume that
the vapor phase is an ideal gas, and the liquid phase is an ideal solution. We have not talked
about ideal solutions yet, but we will do so in the next few lectures. Basically, an ideal
solution is a mixture of liquids in which the interactions between molecules of different species
are the same as the interactions between molecules of the same species. As a result, the energy
and enthalpy of the mixture are just the sum of the mole fractions times the energy
and enthalpy of the components, just as they are for a mixture of ideal gases. This is only a
good model for mixtures of things that are very chemically similar, like different
isomers of the same compound. However, it allows us set up vapor liquid
equilibrium calculations with a simple, easy-to- understand model that we can then
extend to use more realistic models of the liquid phase behavior.
Mathematically, Raoults law is expressed as
In a bubble point pressure calculation, we calculate the composition of the first (infinitesimally
small) bubble that would form as we decrease the pressure of a liquid mixture of specified
composition at constant temperature. Since the amount vaporized at that point is very small, the
liquid composition is known, as well as the temperature, and the unknowns are the pressure
where the first bit of vapor forms and its composition. This calculation is straightforward and
explicit. Since the saturation pressure is only a function of temperature, we know all of the
values of P
i
sat
as well as all the x
i
. We first sum Raoults law over all the species to get
1 1
1 1
1
N N
sat
i i i
i i
N N
sat
i i i
i i
N
sat
i i
i
y P x P
P y x P
P x P
= =
= =
=
=
=
=
where we have used the fact that the gas phase mole fractions have to sum to 1. Knowing all of
the x
i
and P
i
sat
values, we can now compute each of the mole fractions directly from Raoults law
written separately for each species:
sat
i i
i
x P
y
P
=
Dew Point Pressure Calculations:
Here we compute the composition of the first tiny droplet of liquid that would form when we
compress a gas mixture of specified composition at fixed temperature. Since the amount
condensed at that point is very small, the vapor composition is known, as well as the
temperature, and the unknowns are the pressure where the first bit of liquid forms and its
composition. This calculation is also straightforward and explicit. Since the saturation pressure
is only a function of temperature, we again know all of the values of P
i
sat
. This time we know all
of the y
i
rather than all the x
i
values. This time, we will divide Raoults law for each species by
the species saturation pressure and then sum the results over all the species to get the total
pressure:
( )
1 1
1
1
for all species 1, 2,...,
1
1
1
i
i
sat
i
N N
i
i sat
i i
i
N
i
sat
i i
N
i
sat
i i
y P
x i N
P
y P
x
P
y
P
P
P
y
P
= =
=
=
= =
= =
=
=
Once we know P, we can go back and compute the liquid phase mole fractions of each species
from Roults law as
i
i sat
i
y P
x
P
=
Bubble Point Temperature and Dew Point Temperature calculations:
In a bubble point temperature calculation, we compute the temperature at which the first tiny bit
of vapor forms when a liquid mixture of specified composition is heated at constant pressure, as
well as the composition of that first bit of vapor. Since the vapor pressures of the components
depend on temperature in some way that we have not yet specified, we cant necessarily solve
explicitly for the temperature. One approach to this is to start from an initial guess for the
temperature and then do a bubble point pressure calculation, compare the computed total
pressure to the specified total pressure, and then change the temperature (iterate) until the
computed pressure matches the specified pressure. Once we know the temperature and pressure,
we can compute the vapor mole fractions from
sat
i i
i
x P
y
P
=
Similarly, in a dew point temperature calculation, we compute the temperature at which the first
tiny bit of liquid forms when a vapor mixture of specified composition is cooled at constant
pressure, as well as the composition of the liquid. Again, we have to do this iteratively, because
the saturation pressures depend on the temperature. We guess a temperature, do a dew point
pressure calculation, compare the pressure to the specified pressure, and iterate until they match.
Once we know the temperature and pressure, we can compute the liquid phase mole fractions
from
i
i sat
i
y P
x
P
=
Example10.1 in SVA.
Henrys Law:
To apply Raoults law, or extensions of it based on the same idea, we must have the saturation
vapor pressure of each species. Thus, it cant be applied to species above their critical
temperature (where there is no saturation pressure). Thus, for example if we have a container
containing water and nitrogen at room temperature, Raoults law can apply to the water (which is
below its critical temperature) but not to the nitrogen (which has a critical temperature of 126.2
K). To compute the mole fraction of water vapor in the vapor phase, we could assume that the
liquid phase mole fraction of water is almost 1 and write
2 2 2
2
sat sat
H O H O H O
H O
x P P
y
P P
=
However, we might also want to know how much nitrogen can dissolve in the water. We cant
write Raoults law for it, because its vapor pressure is not defined above its critical temperature.
Henrys law is devised for just such a situation. It simply says that for a species present as a very
dilute solute in a liquid phase, the mole fraction in the liquid phase is directly proportional to its
partial pressure in the vapor phase (just as it is in Raoults law, but with a different
proportionality constant). That is
i i i
y P x H =
where H
i
is the Henrys law constant for species i in that particular solution. Henrys law
constants must generally be determined experimentally. Some are given in Table 10.1 on page
348 of SVA.
Example 10.2 in SVA.
Modified Raoults Law formulations of VLE:
There are a wide range of situations where the pressure is low enough that the vapor phase is
nearly ideal (the assumption of an ideal gas mixture in the vapor phase is good), but the liquid
phase is not an ideal solution. Thus, much more realistic VLE calculations can often be done
using a modified version of Raoults Law that can be stated as
( ) for all species 1, 2,...,
sat
i i i i
y P x P i N = =
where
i
is called the activity coefficient of species i in the solution, and generally depends on
both temperature and the solution composition. The activity coefficients must be determined
from experiment, usually via an activity coefficient model fitted to experimental data. This will
be discussed in great detail in upcoming lectures on solution thermodynamics. For the moment,
we will assume that we know the activity coefficients.
Then, bubble point pressure and dew point pressure calculations can be done just as we did with
Raoults law, summing over all the species to get
1 1
1 1
1
N N
sat
i i i i
i i
N N
sat
i i i i
i i
N
sat
i i i
i
y P x P
P y x P
P x P
= =
= =
=
=
=
=
for the bubble point pressure calculation and
( )
1 1
1
1
for all species 1, 2,...,
1
1
1
i
i
sat
i i
N N
i
i sat
i i
i i
N
i
sat
i i i
N
i
sat
i i i
y P
x i N
P
y P
x
P
y
P
P
P
y
P
= =
=
=
= =
= =
=
=
for dew point pressure calculations.
As was the case for Raoults law without the activity coefficients, for bubble point temperature
calculations and dew point temperature calculations we will usually want to use an iterative
solution strategy in which we perform a series of bubble point pressure or dew point pressure
calculations until our computed pressure matches the specified pressure.
Example 10.3 in SVA
K-values as a description of VLE:
The K-value of a substance in a vapor/liquid system is defined as
i
i
i
y
K
x
This number provides a convenient relative measure of the lightness of a component. Things
with K-values greater than one favor the vapor phase, while those with K-values greater than 1
favor the liquid phase. For a system that obeys Raoults law, we have
sat
i i
i
i
y P
K
x P
=
and for a system that obeys the modified form of Raoults law discussed above, we have
sat
i i i
i
i
y P
K
x P
=
Charts of K-values for mixtures of light hydrocarbons (where use of these is most common) are
given in SVA on pages 356 and 357. To use these, you use a straight-edge to connect the
pressure and temperature of interest and then read off the K-values where the straight-edge
crosses the curve for the species of interest. Generally only one or the other of T and P is known,
so this requires iterative graph reading/straight-edge use.
Flash Calculations:
Another important type of calculation is the flash calculation, in which we specify the
temperature and pressure and total amounts of each species and want to compute the composition
and total amounts of each phase. We know from Duhems theorem that specifying 2 intensive
variables plus the total amount of each species in the system determines the state of the system.
We will call the set of (known) overall mole fractions {z
i
}, and call the liquid phase fraction L
and the vapor phase fraction V. Then we have the following equations:
L + V = 1
z
i
= x
i
L + y
i
V (for i = 1,2,,N)
as well as Raoults law (or modified Raoults law) for each species and the requirement that the
mole fractions in each phase sum to 1.
If we substitute x
i
= y
i
/K
i
into the above and then solve for y
i
, we get
i i
i i i
i i
i i
i
i
y L VK
z L yV y
K K
z K
y
L VK
+
= + =
=
+
Then substituting L = 1 V
( ) 1 1 1
i i i i
i
i i
z K z K
y
V VK V K
= =
+ +
Summing this over all the species gives
( )
1 1
1
1 1
N N
i i
i
i i
i
z K
y
V K
= =
= =
+
This is a single equation in which the only unknown is V. After solving it for V, we could use
the preceding equations to find L and all of the mole fractions.
Example 10.5 in SVA
( ) ( ) ( ) d nG nV dP nS dT =
This was derived for a closed, nonreacting system where the number of moles and composition
cannot change. Now, we will apply it to an open system where the number of moles and
composition of the system can change. In this case, the total Gibbs energy, G
t
or nG is a
function of pressure, temperature, and the number of moles of each species in the system. That
is,
( )
1 2
, , , ,...,
N
nG g P T n n n =
so
( )
( ) ( ) ( )
1
, , , ,
j i
N
i
i
i
T n P n T P n
nG nG nG
d nG dP dT dn
P T n
=
= + +
where the partial derivatives with respect to T and P are taken at fixed composition, and the
partial derivatives with respect to the number of moles of species i are taken at fixed T, P, and
numbers of moles of all other species. The partial derivative of the total free energy with respect
to the number of moles of a species is so important that we give it its own name and symbol. It
is defined to be the chemical potential of species i in the mixture and given the symbol .
( )
, ,
j i
i
i
T P n
nG
n
The chemical potentials of the species tell us how the total free energy changes when the mixture
composition changes.
Putting this definition into the equation, and replacing the other two partial derivatives with their
values for the closed-system fundamental property relation (which we can do, because these
derivatives are taken with the numbers of moles held constant, and when the numbers of moles
are held constant, we have the closed system again), we have
( ) ( ) ( )
1
N
i i
i
d nG nV dP nS dT dn
=
= +
This is the fundamental property relation for single-phase fluid systems of constant or variable
mass and constant or variable composition. It is the basis upon which we will develop solution
Solution Thermodynamics: Theory
In lectures 11 and 12 (chapter 6 of SVA), we learned how to compute thermodynamic property
changes for real (non-ideal) fluids of constant composition. In lectures 13 and 14 (chapter 10 of
SVA) we got an introduction to phase equilibrium and vapor-liquid equilibrium
(VLE) in particular. To do more realistic phase equilibrium calculations (such that they would
be useful in predicting the behavior of real systems) we need improved models of
the thermodynamic properties of fluid mixtures and of how they vary with composition. In the
next several chapters, we will apply the laws of thermodynamics to fluid mixtures
and look at models for the thermodynamic behavior of fluid mixtures with varying
composition.
The Fundamental Property Relation applied to Mixtures:
The fundamental property relation for Gibbs energy (which we developed several lectures ago,
and in chapter 6 of SVA) was
thermodynamics. We can write it for the special case of 1 mole of fluid, so that n = 1, and n
i
= x
i
to get
1
N
i i
i
dG VdP SdT dx
=
= +
Notice that although all of the mole numbers and changes in them (n
i
and dn
i
) were independent,
the mole fractions are not all independent. They are required to satisfy
1
1
N
i
i
x
=
=
so
1
0
N
i
i
dx
=
=
Looking at the above fundamental property relationship, we identify
,
,
P x
T x
G
S
T
G
V
P
=
and just as we had for constant composition systems, G serves as a generating function for all of
the thermodynamic properties. This means that if we know G as a function of T, P, and
composition, we can find all the other thermodynamic properties. To do this, we can use the
above expression for S, then use H = G + TS and so on.
The Chemical Potential and Phase Equilibria:
We showed a few lectures back, that two pure component phases were in equilibrium when the
Gibbs energy of each phase was equal. We will now show that for multicomponent phases, two
phases are in equilibrium when the chemical potential of each species is the same in both phases.
To do so, we write the extensive version of the fundamental property relation for each phase:
( ) ( ) ( )
( ) ( ) ( )
1
1
N
i i
i
N
i i
i
d nG nV dP nS dT dn
d nG nV dP nS dT dn
=
=
= +
= +
where the two phases are denoted by the superscripts and . T and P are taken to be uniform
throughout the system, because if they were not, the system could not be in thermal and
mechanical equilibrium. The total system value of each of the extensive properties of the system
is the sum of the values for the two phases, i.e.
( ) ( )
( ) ( )
( ) ( )
nG nG nG
nV nV nV
nS nS nS
= +
= +
= +
So, adding the fundamental property relations for the two phases gives:
( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( )
1 1
1 1
N N
i i i i
i i
N N
i i i i
i i
d nG d nG nV dP nS dT dn nV dP nS dT dn
d nG nV dP nS dT dn dn
= =
= =
+ = + + +
= + +
Since the overall 2-phase system is closed and has constant (overall) composition, we can also
write for the overall system the fundamental property relation for a closed, constant composition
system with which we started this lecture:
( ) ( ) ( ) d nG nV dP nS dT =
Comparing these last two equations (or subtracting the last one from the previous one), we see
that
1 1
0
N N
i i i i
i i
dn dn
= =
+ =
Since the overall composition of the two-phase system is fixed, we have for each species
i i i
n n n
+ =
where n
i
is the fixed total number of moles of species i in the system. So
0
i i
dn dn
+ =
or
i i
dn dn
=
Substituting this into
1 1
0
N N
i i i i
i i
dn dn
= =
+ =
, gives
( )
1
0
N
i i i
i
dn
=
=
Since the dn
i
Thus, the chemical potential is also the partial molar Gibbs energy. It is the only partial molar
property that is so special that it gets its own name and symbol. Other partial molar properties
are
( )
, ,
j i
i
i
T P n
nV
V
n
partial molar volume
( )
, ,
j i
i
i
T P n
nH
H
n
partial molar enthalpy
( )
, ,
j i
i
i
T P n
nS
S
n
partial molar entropy
and so on.
The above equations tell us how to determine these partial molar properties from the properties
of the mixture. However, we often want to do the reverse that is, we want to find the
properties of the mixture from the partial molar properties of the solution. We start by noting
that any thermodynamic property M of a homogeneous phase is a function of temperature,
pressure, and the number of moles of the species that make up the phase. That is
( )
1 2
, , , ,....,
N
nM T P n n n = M
So, we can write the total differential of nM as
( )
( ) ( ) ( )
1
, , , ,
j i
N
i
i
i
T n P n P T n
d nM d nM d nM
d nM dP dT dn
dP dT dn
=
= + +
Since the first two derivatives are taken with the number of moles of each species held constant,
the total number of moles is also constant, and we can take it out of the derivatives. The
derivatives with respect to the numbers of moles of each species are, by definition, the partial
molar properties. So, we can write the above as
( )
1
, ,
N
i i
i
T x P x
dM dM
d nM n dP n dT M dn
dP dT
=
= + +
where we have used the subscript x to denote differentiation at constant mole fractions which
(for constant total number of moles) is the same as differentiation at constant numbers of moles
of each species. We will now re-write the dn
i
in terms of the mole fractions instead of the total
number of moles. Since n
i
= x
i
n, we have dn
i
= x
i
dn + ndx
i
, so
( ) ( )
1
, ,
N
i i i
i
T x P x
dM dM
d nM n dP n dT M ndx x dn
dP dT
=
= + + +
Likewise, we can replace d(nM) with d(nM) = ndM + Mdn to get
( )
1
, ,
N
i i i
i
T x P x
dM dM
ndM Mdn n dP n dT M ndx x dn
dP dT
=
+ = + + +
Now, lets move everything to the left hand side of the equation, and group together the terms
with n in them and the terms with dn in them (note that each term has either a n or a dn, but not
both). Doing that
1 1
, ,
0
N N
i i i i
i i
T x P x
dM dM
n dM dP dT M dx M M x dn
dP dT
= =
+ =
Next, we realize that n, the total number of moles in the system, and dn, the change in number of
moles in the system are independent and arbitrary. That is, we can pick whatever n and whatever
dn we want, and the above still has to be true. The only way that can be the case is if each of the
groups in parenthesis is zero. We have magically turned one equation into two:
1
, ,
1
0
0
N
i i
i
T x P x
N
i i
i
dM dM
dM dP dT M dx
dP dT
M M x
=
=
=
=
The second equation can be written as
1
N
i i
i
M M x
=
=
or multiplied by the total number of moles to get
1
N
i i
i
nM M n
=
=
It says that the total property of the system is the sum of the number of moles of each species
times the partial molar property of that species. The partial molar properties provide a
convenient way to divide up the total property into contributions that we assign to each of the
species. However, we should make a point of remembering that the partial molar properties,
while associated with a particular species for convenience, are properties of the whole system
and depend on the system composition. That is, there is no such thing as the partial molar
volume of water, only the partial molar volume of water in a mixture of . Without the
mixture, there are no partial molar properties.
From the other piece of our equation:
1
, ,
0
N
i i
i
T x P x
dM dM
dM dP dT M dx
dP dT
=
=
we can derive the Gibbs/Duhem equation an equation so important to solution thermodynamics
that it gets its own special name. It will show us that the partial molar properties cannot change
arbitrarily, but that they are constrained such that if changes in a partial molar property of all but
one species are specified, then the change in the partial molar property of the last species is
already determined and cant be specified independently of the others. This is like the situation
with mole fractions in a way. If we specify all but one mole fraction, we cant specify the last
one because it is already determined by the fact that the mole fractions have to sum to one.
Writing
1
N
i i
i
M M x
=
=
This last equation is called the Gibbs/Duhem equation. It is often written for changes in
composition taking place at constant T and P, for which it is just
1
0
N
i i
i
x dM
=
=
Partial Properties for Binary Solutions:
For a binary mixture (which is the simplest possible mixture and therefore something on which
we will spend significant time and effort) the general summability relationship
1
N
i i
i
M M x
=
=
is
simply
1 1 2 2
M M x M x = +
so
1 1 2 2 1 1 2 2
dM M dx M dx x dM x dM = + + +
The Gibbs/Duhem relationship for a binary system at constant T and P is
1 1 2 2
0 x dM x dM + =
and using this in the previous equation gives us
1 1 2 2
dM M dx M dx = +
Since the mole fractions must sum to one, x
2
= 1 x
1
, and dx
2
= -dx
1
. Substituting this in, we get
( )
1 1 2 1 1 2 1
dM M dx M dx M M dx = =
so
1 2
1
dM
M M
dx
=
or
2 1
1
dM
M M
dx
=
Substituting this into the summability relationship that we started with (
1 1 2 2
M M x M x = + ) gives
( )
1 1 1 2 1 2 1 2
1 1
1 2
1
dM dM
M M x M x x x M x
dx dx
dM
M M x
dx
= + = +
=
and finally
1 2
1
dM
M M x
dx
= +
Similarly, we could eliminate
1
M from the equations to get
2 1
1
dM
M M x
dx
=
These equations tell us how to get the partial molar properties of a binary solution directly from a
measurement of the total solution properties as functions of composition.
See examples 11.2, 11.3, and 11.4 in SVA.
Relationships among Partial Properties:
As shown on pages 380 and 381 of SVA:
For every equation that provides a linear relation among thermodynamic properties of a
constant-composition fluid, the same relation among partial molar properties of the species in
solution also applies.
This includes all the relationships between different thermodynamic properties like
H = U + PV
Which for partial molar properties of species i is just the same:
i i i
H U PV = +
The fundamental property relationship for G for a constant-composition solution was
dG VdP SdT =
and for the corresponding partial molar properties of species i we have
i i i
dG VdP S dT =
So, we can re-use many equations that we derived for constant-composition systems without re-
deriving them!
1
exp
sat
i
P
sat sat
i i i i
P
f P V dP
RT
Usually, the molar volume of a liquid (not too near its critical point) is nearly constant, and we
can approximate the above using
( )
sat
i
P
sat sat
i i i
P
V dP V P P
in which case
( )
exp
sat sat
i i
sat sat
i i i
V P P
f P
RT
=
The exponential part is called the Poynting factor.
See example 11.5
( )
ln
i i i
T RT f = +
where we have used a little hat to denote that this is for a species in solution rather than for a
pure species. We see that for this definition, if the chemical potential of a substance is the same
in different phases, its fugacity must be the same in different phases as well. That is, our
criterion for phase equilibrium in terms of chemical potential
...
i i i
= = = , for i = 1,2,,N
implies the same criterion in terms of fugacities
...
i i i
f f f
= = = , for i = 1,2,,N
We can state this criterion for phase equilibrium as: Multiple phases at the same T and P are in
equilibrium when the fugacity of each constituent species is the same in all phases.
This is the equilibrium criterion most often applied in solution of phase equilibrium problems.
For the specific case of vapor/liquid equilibrium,
l v
i i
f f = , for i = 1,2,,N
In chapter 6, we defined residual properties as the difference between actual properties of a
substance and the properties of the substance in a hypothetical ideal gas state
M
R
= M - M
ig
Likewise, we define partial molar residual properties as
R ig
i i i
M M M =
Thus, the partial molar residual Gibbs energy is defined as
R ig ig
i i i i i
G G G = =
where in the second equality we have just replaced the partial molar Gibbs energy by the
chemical potential (since they are the same by definition).
In terms of fugacities, the two chemical potentials are
( )
ln
i i i
T RT f = +
and ( ) ln
ig
i i i
T RT y P = +
Subtracting these,
ln
R ig i
i i i
i
f
G RT
y P
= =
We define the fugacity coefficient of species i in the mixture (again with a hat on it to
differentiate it from its pure-component counterpart) as
i
i
i
f
y P
FUGACITY AND FUGACITY COEFFICIENT FOR SPECIES IN SOLUTION:
For a species in solution, we define the fugacity in a way parallel to our
definition for pure species fugacities. For species i in a mixture, the fugacity definition is
so
ln
R
i i
G RT =
The Fundamental Residual Property Relations:
Now, we want to take the fundamental property relations for mixtures and extend them to
residual properties. That is, we want to write relations like
( ) ( ) ( )
1
N
i i
i
d nG nV dP nS dT dn
=
= +
in terms of residual properties. To do so, we will first write an expression for d(nG/RT) instead
of just d(nG):
( )
2
1 nG nG
d d nG dT
RT RT RT
=
Next, we substitute into this
( ) ( ) ( )
1
N
i i
i
d nG nV dP nS dT dn
=
= +
and G = H TS
to get
( ) ( )
1
2
2
1
N
i i
i
N
i
i
i
nV dP nS dT dn
nG nH nTS
d dT
RT RT RT
nG nV nH
d dP dT dn
RT RT RT RT
=
=
+
=
= +
in this equation, H rather than S shows up on the right-hand-side, and all of the terms have units
of moles. Writing this equation for an ideal gas, we get
2
1
ig ig ig ig N
i
i
i
nG nV nH
d dP dT dn
RT RT RT RT
=
= +
Subtracting this from the general equation gives
2
1
ig R R R N
i i
i
i
nG nV nH
d dP dT dn
RT RT RT RT
=
= +
we can write
ig
i i
as
ig R
i i i
G G G = to get
2
1
R R R R N
i
i
i
G nG nV nH
d dP dT dn
RT RT RT RT
=
= +
or, we can introduce the fugacity coefficient,
ln ln
R ig i
i i i i
i
f
G RT RT
y P
= = =
to get
2
1
ln
R R R N
i i
i
nG nV nH
d dP dT dn
RT RT RT
=
= +
Thus, if we know the residual Gibbs energy as a function of temperature, pressure, and
composition, we can get the residual volume, residual enthalpy, and fugacity coefficients from
equations like
, ,
,
R R R
T n T x
R R
T x
d nG d G nV
n
dP RT dP RT RT
d G V
dP RT RT
= =
=
and
2
, ,
2
,
R R R
P n P x
R R
P x
d nG d G nH
n
dT RT dT RT RT
d G H
dT RT RT
= =
=
and
, ,
ln
j i
R
i
i
T P n
d nG
dn RT
=
ln
ig ig ig
i i i i
G G RT y = +
Similarly, we define an ideal solution as one for which the chemical potentials (or partial molar
Gibbs energies) of the components
ln
id id
i i i i
G G RT x = +
where we have used x
i
for the mole fraction because we usually use this description for liquid
solutions. All of the other partial molar properties follow from the above definition, and they
come out to be
ln
id
i i i
id
i i
id
i i
S S R x
V V
H H
=
=
=
For the special case of an ideal solution, the summability relation is
1
N
id id
i i
i
M M x
=
=
Which, combined with the previous equations gives us the properties of an ideal mixture in terms
of the properties of the pure components of which it is composed:
1 1
1 1
1
1
ln
ln
N N
id
i i i i
i i
N N
id
i i i i
i i
N
id
i i
i
N
id
i i
i
G x G RT x x
S x S R x x
V xV
H x H
= =
= =
=
=
= +
=
=
=
The volume and enthalpy of the mixture are just the weighted average of those of the
components. That is, the enthalpy of mixing and the volume change of mixing of an ideal
solution are zero. The entropy of mixing is
1
ln
N
i i
i
R x x
=
and the Gibbs energy of mixing
The ideal solution
For gases, we often use the ideal gas as a reference state, and then work in terms of
residual properties that re deviations from the ideal gas reference state. For liquid
mixtures, we will similarly use an ideal solution as a reference state. For the ideal solution, the
components are not ideal, but the mixture properties vary with composition in the same
way as in an ideal gas mixture. For a mixture of ideal gases, we had the chemical
potential or partial molar Gibbs energy of component i given by
is
1
ln
N
i i
i
RT x x
=
. Note that the entropy of mixing of an ideal soluation is always positive and the
Gibbs energy of mixing is always negative.
The Lewis/Randall Rule:
The above definition of an ideal solution also provides us with a means to compute fugacities of
components of an ideal solution. Our definition of the fugacity was
( )
ln
i i i
T RT f = +
so for an ideal solution, we have
( )
ln
id id id
i i i i
G T RT f = = +
and for the pure component, we have
( ) ln
i i i
G T RT f = +
and our definition of the ideal solution was that
ln
id id
i i i i
G G RT x = +
Combining these, we have
( )
( )
( ) ( )
( )
ln ln ln
ln ln ln ln
id id
i i i i i i i
id
i i i i i
id
i i i
G G T RT f T RT f RT x
RT f RT f RT x RT x f
f x f
= + + =
= + =
=
The fugacity of a component in an ideal solution is just the pure component fugacity times the
mole fraction of the component. This equation is called the Lewis/Randall rule. We can also
write it in terms of the fugacity coefficient, which we remember was
i
i
i
f
x P
for a component in a mixture, and
i
i
f
P
for a pure component,
so, we have
id
id i i i i
i i
i i
f x f f
x P x P P
= = =
The fugacity coefficient of a component in an ideal solution is the same as its pure species
fugacity coefficient (in the same phase and at the same conditions).
Excess Properties
Just as we defined residual properties as the difference between properties in the real state of a
system and properties of the same substance in a hypothetical ideal gas state at the same
conditions, we can define excess properties as the difference between the properties of a real
solution and the corresponding ideal solution of the same composition at the same conditions.
That is, we define the excess property M
E
by
E id
M M M
We can derive all of the same equations for excess properties that we had for ideal properties. In
particular, all of the partial molar properties have excess partial molar counterparts. That is, we
also have
E id
i i i
M M M =
The excess partial molar property of a species in a mixture is the difference between the partial
molar property of the species and the partial molar property of the species in an ideal mixture of
the same composition and at the same conditions.
There is a fundamental excess-property relation that is just like the fundamental residual-
property relation that we derived earlier, and that we could (but wont) derive here in the same
way:
2
1
E E E E N
i
i
i
G nG nV nH
d dP dT dn
RT RT RT RT
=
= +
See example 11.10 and table 11.1 in SVA
The Excess Gibbs Energy and the Activity Coefficient:
A quantity of particular interest is the excess partial molar Gibbs energy. In general, we defined
the fugacity of a component in a mixture from
( )
ln
i i i i
G T RT f = = +
and for an ideal mixture, it is
( ) ( ) ( )
ln ln
id id id
i i i i i i i
G T RT f T RT x f = = + = +
Subtracting the second of these from the first, we have
( )
( )
( ) ( ) ( )
ln ln
ln
E id
i i i i i i i i
E i
i
i i
G G G T RT f T RT x f
f
G RT
x f
= = + +
=
We define the activity coefficient of species i in the mixture as
i
i
i i
f
x f
So, the partial molar excess Gibbs energy is
ln
E
i i
G RT =
For an ideal solution, the activity coefficients are 1 (so that the partial molar excess Gibbs energy
is zero).
We can, if we like, re-write the fundamental excess property relation in terms of the activity
coefficients as
( )
2
1
ln
E E E N
i i
i
nG nV nH
d dP dT dn
RT RT RT
=
= +
Knowing the excess Gibbs energy as a function of temperature, pressure, and composition, we
can get the excess volume, excess enthalpy, and activity coefficients from equations just like we
had for the corresponding residual properties:
, ,
,
E E E
T n T x
E E
T x
d nG d G nV
n
dP RT dP RT RT
d G V
dP RT RT
= =
=
and
2
, ,
2
,
E E E
P n P x
E E
P x
d nG d G nH
n
dT RT dT RT RT
d G H
dT RT RT
= =
=
and
, ,
ln
j i
E
i
i
T P n
d nG
dn RT
=
This last equation shows us that RT ln
i
is the partial molar excess Gibbs energy of species i, and
therefore it obeys the summability relation and the Gibbs-Duhem relation and all other
relationshnips that apply to partial molar properties. As a result, we get
1
ln
E N
i i
i
G
x
RT
=
=
(from the summability relation)
and ( )
1
ln 0
N
i i
i
x d
=
=
The amount of reaction that has occurred (relative to some initial state) can be characterized by a
single reaction progress variable or reaction coordinate. As defined in SVA, this is an extensive
quantity (that might also be called an extensive extent of reaction). It is denoted by and
defined such that for a single reaction
3 1 2
1 2 3
, .
i
i
dn dn dn dn
d etc
= = = =
So, the change in number the number of moles of each species is given by dn
i
=
i
d. Or, if the
initial number of moles (before reaction) is n
i,o
, then the number of moles after some extent of
reaction (at reaction coordinate ) is
n
i
= n
i,o
+
i
The total number of moles is
( )
, ,
1 1 1 1
N N N N
i i o i i o i
i i i i
n n n n
= = = =
= = + = +
or simply n = n
o
+ , with the definitions
phase changes, separation processes, and phase equilibria is that of chemical reactions
and chemical reaction equilibria. A large fraction of chemical reactions carried out
commercially are run such that the reactor effluent is very close to the thermodynamic
equilibrium distribution of products. Even for other reactions, which are limited by the
rate of reaction and do not reach equilibrium, the equilibrium composition gives us an upper
limit on the amount of reaction that can occur. In the next couple of lectures, we will learn
how to predict the equilibrium composition of a reacting mixture from the thermodynamic
properties (heat of formation, standard entropy, heat capacity, etc.) of the components. This
is one of the first steps in the development of almost any new chemical process,
since these thermodynamic limits tell us what is possible.
Reaction Stoichiometry and the Reaction Coordinate:
We will begin with a brief review of notation used to systematically write chemical reactions and
to keep track of their progress using the minimum number of variables. We can write a
general chemical reaction in the form
Chemical Reaction Equilibria,
A topic of perhaps even more central importance to chemical engineers than
,
1
1
N
o i o
i
N
i
i
n n
=
=
=
=
So, the mole fractions of the species are given by
, i o i
i
i
o
n
n
y
n n
+
= =
+
So, for the above example of
CH
4
+ 2 O
2
CO
2
+ 2 H
2
O
we have
= -1 2 + 1 + 2 = 0
and
4 4
2 2
2 2
2 2
,
,
,
,
2
2
o
CH CH o
O O o
CO CO o
H O H O o
n n
n n
n n
n n
n n
=
=
=
= +
= +
and the mole fractions are given by
4
4
2
2
2
2
2
2
,
,
,
,
2
2
CH o
CH
o
O o
O
o
CO o
CO
o
H O o
H O
o
n
y
n
n
y
n
n
y
n
n
y
n
=
+
=
+
=
We can further generalize all of this to the case of multiple reactions. If
ij
is defined as the
stoichiometric coefficient of species i in reaction j, then we can write M total reactions involving
N total species as
( )
1
0, 1, 2,..,
N
ij i
i
A j M
=
= =
and we will call the reaction coordinate of the j
th
reaction
j
. Then the number of moles of each
species is
,
1
M
i i o ij j
j
n n
=
= +
The total number of moles is
,
1 1 1 1
N N M N
i i o j ij
i i j i
n n n
= = = =
= = +
or, defining the total stoichiometric number for each reaction as
1
N
j ij
i
=
=
we have
1
M
o j j
j
n n
=
= +
And finally, the mole fractions are given by
,
1
1
M
i o ij j
j
i
i M
o j j
j
n
n
y
n
n
=
=
+
= =
+
So, for example, if we had the pair of reactions
CH
4
+ 2 O
2
CO
2
+ 2 H
2
O (reaction 1)
And CO
2
+ H
2
H
2
O + CO (reaction 2)
Then,
1
= -1 2 + 1 + 2 = 0 and
2
= -1 1 + 1 + 1 = 0, and
4 4
2 2
2 2
2 2
2 2
, 1
, 1
, 1 2
, 1 2
, 2
, 2
2
2
o
CH CH o
O O o
CO CO o
H O H O o
H H o
CO CO o
n n
n n
n n
n n
n n
n n
n n
=
=
=
= +
= + +
=
= +
and the mole fractions are
4
4
2
2
2
2
, 1
, 1
, 1 2
2
CH o
CH
o
O o
O
o
CO o
CO
o
n
y
n
n
y
n
n
y
n
=
+
=
2
2
2
2
, 1 2
, 2
, 2
2
H O o
H O
o
H o
H
o
CO o
CO
o
n
y
n
n
y
n
n
y
n
+ +
=
=
+
=
The number of independent quantities needed to determine all of the mole fractions (give the
initial mole fractions) is only 2 (the two reaction coordinates, or two mole fractions from which
we could compute the reaction coordinates), even though there are 6 species and therefore 6
mole fractions.
The Criteria for Equilibrium of Chemical Reactions:
Things are, once again, a little bit out of order, so you will just have to trust me (or SVA) that a
general criterion for a closed system to be in thermodynamic equilibrium is that its Gibbs energy
(the total Gibbs energy of the system) is a minimum with respect to all allowed changes
(including chemical reactions). Hopefully, we will demonstrate this in subsequent lectures based
on material in chapter 14 of SVA. So, to find the equilibrium composition of a reacting system,
we want to find the minimum in the total Gibbs energy with respect to the reaction coordinate (or
reaction coordinates for multiple reactions). Of course, the system must also be at uniform
temperature and pressure (criteria for thermal and mechanical equilibrium) and if multiple phases
are present, the chemical potential of each species must be the same in every phase (criteria for
phase equilibrium).
The Standard Gibbs Energy Change of Reaction and the Equilibrium Constant
For the moment, we will consider a homogeneous, single-phase system. Then (as derived in
chapter 11 of SVA) the changes in total Gibbs energy are given by:
( ) ( ) ( )
1
N
i i
i
d nG nV dP nS dT dn
=
= +
If this is a closed, reacting system, undergoing a single reaction, then the changes in the number
of moles of species are only due to that reaction and are given by
i i
dn d =
So we have
( ) ( ) ( )
1
N
i i
i
d nG nV dP nS dT d
=
= +
And therefore
( )
1
,
N
i i
i
T P
nG
=
=
or
1
,
t N
i i
i
T P
G
=
=
That is, the partial derivative of the total Gibbs energy with respect to the reaction coordinate is
given by the sum over all of the species of the species chemical potential times its stoichiometric
coefficient in the reaction. When the Gibbs energy is a minimum with respect to the reaction
coordinate, this partial derivative is zero. Thus, a necessary condition for the reaction to be in
equilibrium is that
1
0
N
i i
i
=
=
In order to apply this equation, we need to write the chemical potentials of the species in terms of
things that we can relate to the mixture composition (which is the thing we ultimately want to
compute). To do so, we first write the chemical potential in terms of the fugacity as we did in
chapter 11:
( )
ln
i i i
T RT f = +
We can also write the Gibbs energy of the species in its standard state (like the standard states
for which the enthalpies of formation, standard entropies, etc. are tabulated) as
( ) ln
o o
i i i
G T RT f = +
Subtracting the second of these from the first, we have
ln
o i
i i
o
i
f
G RT
f
=
or
ln
o i
i i
o
i
f
G RT
f
= +
Substituting this into the equilibrium criterion above,
1 1
ln 0
N N
o i
i i i i o
i i i
f
G RT
f
= =
= + =
We can rearrange this to get
1 1
1 1
1 1
ln
ln
exp
i
i
o N N
i i i
i o
i i
i
o N N
i i i
o
i i i
o N N
i i i
o
i i i
f G
f RT
f G
f RT
f G
f RT
= =
= =
= =
=
=
=
where
1
i
N
i
o
i i
f
f
means that we take the product over all of the N species. We define the
standard Gibbs energy change of reaction as
1
N
o o
i i
i
G G
=
and we define the equilibrium constant of the reaction as
1 1
exp exp
i
o o N N
i i i
o
i i i
G f G
K
RT RT f
= =
= =
Since the free energies, G
i
o
are functions only of temperature (they are defined to be in a
standard state at a fixed pressure), the equilibrium constant is also a function only of
temperature. Note that the standard Gibbs energy change of reaction is evaluated at standard
pressure and state, but at the actual temperature of interest. The standard state chosen for
evaluating G
i
o
must be the same as that used for the fugacity, f
i
o
. Other standard property
changes of reaction are defined in the same way as the standard Gibbs energy change of reaction.
That is, in general, we define
1
N
o o
i i
i
M M
=
For example, the standard enthalpy change of reaction is
1
N
o o
i i
i
H H
=
Since these property changes of reaction are sums and differences (linear combinations) of
individual species properties, any linear relationship that we have among the species properties
also applies to these property changes of reaction. For example, for species i, we have
2
o
o i
i
G d
H RT
dT RT
=
Summing this up over the species in a reaction times their stoichiometric coefficients,
2 2 1
1 1
N
o
i
N N
o
o i i
i
i i
G
G d d
H RT RT
dT RT dT RT
=
= =
= =
or
2
o
o
d G
H RT
dT RT
=
Effect of temperature on the equilibrium constant:
Just above, we derived that
2
o
o
d G
H RT
dT RT
=
or
2
o o
d G H
dT RT RT
=
and the equilibrium constant is given by
exp
o
G
K
RT
Combining these, we have
( )
2
ln
o
d K
H
dT RT
=
or simply
( )
2
ln
o
d K
H
dT RT
=
which tells us the temperature dependence of the equilibrium constant. For exothermic reactions,
the enthalpy change of reaction is negative, so the equilibrium constant decreases with increasing
temperature. That is, as the temperature goes up, the equilibrium shifts toward the reactants.
The converse is also true; for endothermic reactions, increasing the temperature shifts the
equilibrium toward the products.
If the enthalpy change of reaction were independent of temperature (or we can take it to be
approximately independent of temperature over some temperature range), then we can integrate
the above relationship from temperature T to temperature T to get
1 1
ln ln ' ln
' '
o
K H
K K
K R T T
= =
That is, a plot of ln(K) vs. 1/T will be approximately a straight line with a slope given by
o
H
R
So applying this to the corresponding property changes:
( ) ( )
( ) ( )
o
o
T
o o o
o p
T
o T
p o o
o
T
H T H T C dT
C
S T S T dT
T
= +
= +
So, putting everything together, we end up with
( ) ( ) ( )
( ) ( ) ( )
o o
o o o
o T T
p o o o o
o p o
T T
G T H T T S T
C
G T H T C dT T S T T dT
T
=
= +
Substituting into this
( )
( ) ( )
o o
o o o
o
o
H T G T
S T
T
=
we have
( ) ( )
( ) ( )
( ) ( ) ( ) 1
o o
o o
o o o T T
p o o o o o
o p
o T T
o T T
p o o o o
o o p
o o T T
C H T G T
G T H T C dT T T dT
T T
C
T T
G T G T H T C dT T dT
T T T
= +
= + +
Dividing this through by RT, we get
( ) ( ) ( ) 1
1
o o
o o o o T T
p o o o
p
o o T T
C G T G T H T T
C dT dT
RT RT RT T RT RT
= + +
Or, since exp
o
G
K
RT
( ) 1
ln ln 1
o o
o o T T
p o o
o p
o T T
C H T T
K K C dT dT
RT T RT RT
= +
Or, exponentiating both sides of the equation:
( ) 1
exp 1 exp
o o
o o T T
p o o
o p
o T T
C H T
T
K K C dT dT
RT T RT RT
= +
SVA point out that this is the product of 3 terms with some physical meaning. We could write it
as
1 2 o
K K K K =
where
( )
exp
o
o
o
o
G T
K
RT
=
is the value of the equilibrium constant at the reference temperature
( )
1
exp 1
o
o
o
H T
T
K
RT T
=
is the main part of the temperature dependence, and is the change in the equilibrium constant that
we would have if the enthalpy change of reaction were constant, and
2
1
exp
o o
o T T
p o
p
T T
C
K C dT dT
RT RT
= +
is the smaller temperature dependence due to the fact that the heat capacity of
reactants and products is (in general) different, and therefore the enthalpy
of reaction changes with temperature. We can evaluate these heat capacity integrals
in the same way that we have done in the past for evaluating heat capacity integrals to determine
enthalpies of individual species. SVA give defined functions for this,
Work Through Example 13.4 in SVA.