CHAPTER 1: THERMODYNAMIC SYSTEMS: BASIC CONCEPTS

1.1 Introduction
The word Thermodynamics originates from its Greek roots (therme, heat; dynamis, force). As a
subject it is concerned with quantification of inter-relation between energy and the change of

ld
state of any real world system. The extent of such change of state due to transfer of energy to or
from the system is captured through the basic equations of thermodynamics which are derived
starting from a set of fundamental observations known as Laws of Thermodynamics. The laws

or
are essentially postulates that govern the nature of interaction of real systems and energy. They
are products of human experiential observations to which no exceptions have been found so far,
and so are considered to be laws. The scope of application of the laws of thermodynamics

W
ranges from the microscopic to the macroscopic order, and indeed to cosmological processes.
Thus, all processes taking place in the universe, whether in non-living or living systems, are
subject to the laws of thermodynamics.
Historically speaking, thermodynamics, is an extension of Newtonian mechanics which
considered mechanical forces (or energy) as the agent of change of state of a body (anything
TU
possessing mass), the state being defined by its position and momentum with respect to a frame
of reference. With the discovery steam power which propelled the so-called Industrial
Revolution of the 18th century, it became evident that not only the direct application of
mechanical energy can change the state of a system, but that fluids themselves can act as
reservoir of energy, which can be harnessed to effect changes in the real world to human
JN

advantage. It was this observation that laid the foundations of thermodynamics, which now
constitutes a generalized way of understanding and quantifying all changes that occur during
processes taking place in the universe as a result of application of energy in any form.
ll

1.2 Thermodynamic System: Select Definitions

It may be evident from the foregoing introduction, that for the purpose of any thermodynamic
A

analysis it is necessary to define a system. A system, in general, is any part of the universe
which may be defined by a boundary which distinguishes it from the rest of the universe. Such a
thermodynamic system is usually referred to as control volume as it would possess a volume and
would also contain a definite quantity of matter. The system boundary may be real or imaginary,
and may change in shape as well as in size over time, i.e., increase or decrease.
A system can either be closed or open. A closed system does not allow any transfer of
mass (material) across its boundary, while an open system is one which does. In either case
energy transfer can occur across the system boundary in any of its various forms; for example,

ld
heat, work, electrical / magnetic energy, etc. However, for most real world systems of interest to
chemical engineers the primary forms of energy that may transfer across boundaries are heat and
work. In contrast to closed or open systems, a system which is enclosed by a boundary that

or
allows neither mass nor energy transfer is an isolated system.
All matter external to the system constitutes the surroundings. The combination of the
system and surroundings is called the universe. For all practical purposes, in any thermodynamic

W
analysis of a system it is necessary to include only the immediate surroundings in which the
effects are felt.
A very common and simple example of a thermodynamic system is a gas contained in a
piston-and-cylinder arrangement derived from the idea of steam engines, which may typically
TU
JN

Fig. 1.1 Example of simple thermodynamic system

ll

exchange heat or work with its surroundings. The dotted rectangle represents the control
volume, which essentially encloses the mass of gas in the system, and walls (including that of
A

the piston) form the boundary of the system. If the internal gas pressure and the external pressure
(acting on the moveable piston) is the same, no net force operates on the system. If, however,
there is a force imbalance, the piston would move until the internal and external pressures
equalize. In the process, some net work would be either delivered to or by the system, depending
on whether the initial pressure of the gas is lower or higher than the externally applied pressure.
In addition, if there is a temperature differential between the system and the surroundings the
former may gain or lose energy through heat transfer across its boundary.

ld
This brings us to a pertinent question: how does one characterize the changes that occur in
the system during any thermodynamic process? Intuitively speaking, this may be most readily
done if one could measure the change in terms of some properties of the system. A

or
thermodynamic system is, thus, characterized by its properties, which essentially are descriptors
of the state of the system. Change of state of a system is synonymous with change in the
magnitude of its characteristic properties. The aim of the laws of thermodynamics is to establish a

W
quantitative relationship between the energy applied during a process and the resulting change in
the properties, and hence in the state of the system.

Thermodynamic properties are typically classified as extensive and intensive. A property

which depends on the size (i.e., mass) of a system is an extensive property. The total volume of a
TU
system is an example of an extensive property. On the other hand, the properties which are
independent of the size of a system are called intensive properties. Examples of intensive
properties are pressure and temperature. The ratio of an extensive property to the mass or the
property per unit mass (or mole) is called specific property. The ratio of an extensive property to
the number of moles of the substance in the system, or the property per mole of the substance, is
JN

called the molar property.

Specific volume (volume per mass or mole) V = V t / M ..(1.1)

Molar Volume (volume per mole) V = V t / N

where,
ll

V t = total system volume (m3 )

A

M = total system mass (kg)

N = total moles in system (kg moles)
1.3 Types of Energies associated with Thermodynamic Processes:
We know from the fundamentals of Mechanics, that the energy possessed by a body by virtue of
its position or configuration is termed potential energy (PE). The potential energy of a body of

ld
mass m which is at an elevation z from the earths surface (or any particular datum) is given by:
PE = mgz ..(1.2)

or
Where, g is the acceleration due to gravity (= 9.81 m/s2).

The energy possessed by a body by virtue of its motion is called the kinetic energy (KE). For a
body of mass m moving with a velocity u, the kinetic energy of the body is given by:

W
1
KE = mu 2 ..(1.3)
2

It follows that, like any mechanical body, a thermodynamic system containing a fluid, in
TU
principle may possess both PE and KE. It may be noted that both PE and KE are expressed in
terms of macroscopic, directly measurable quantities; they, therefore, constitute macroscopic,
mechanical forms of energy that a thermodynamic system may possess. As one may recall from
the basic tenets of mechanics, PE and KE are inter-convertible in form.

It may also be noted that PE and KE are forms of energy possessed by a body as a whole
JN

by virtue of its macroscopic mass. However, matter is composed of atoms /molecules which have
the capacity to translate, rotate and vibrate. Accordingly, one ascribes three forms intra-molecular
energies: translational, rotational and vibrational. Further, energy is also associated with the
motion of the electrons, spin of the electrons, intra-atomic (nucleus-electron, nucleus-nucleus)
interactions, etc. Lastly, molecules are also subject to inter-molecular interactions which are
ll

electromagnetic in nature, especially at short intermolecular separation distances. All these forms
of energy are microscopic in form and they cannot be readily estimated in terms of
A

macroscopically measurable properties of matter. It needs to be emphasized that the microscopic

form of energy is distinct from PE and KE of a body or a system, and are generally independent
of the position or velocity of the body. Thus the energy possessed by matter due to the
microscopic modes of motion is referred to as the internal energy of the matter. The microscopic
variety of energy forms the principal consideration in case of transformations that occur in a
thermodynamic system. Indeed, as mentioned earlier, it is the realization that matter or fluids
possessed useful form of microscopic energy (independent of macroscopic KE or PE) that formed
the basis of the 18th century Industrial Revolution.

ld
As we will see later, the majority of practical thermodynamic systems of interest are the
ones that do not undergo change of state that entails significant change in its macroscopic

or
potential and kinetic energies. Thus, it may be intuitively obvious that in a very general sense,
when a thermodynamic system undergoes change of state, the attendant change in the internal
energy is responsible for the energy leaving or entering the system. Such exchange of energy

W
between a thermodynamic system and its surroundings may occur across the system boundary as
either heat or work or both.

Thermodynamic Work:

Work can be of various forms: electrical, magnetic, gravitational, mechanical, etc. In general
TU
work refers to a form of energy transfer which results due to changes in the
external macroscopic physical constraints on a thermodynamic system. For example, electrical
work results when a charge moves against an externally applied electrical field. As we will see
later, it is mechanical work that is most commonly encountered form in real thermodynamic
systems, for example a typical chemical plant. In its simplest form, such work results from the
JN

energy applied to expand the volume of a system against an external pressure, or by driving a
piston-head out of a cylinder against an external force. In both the last examples, work transfer
takes place due to the application of a differential (or finite) force applied on the system
boundary; the boundary either contracts or expands due to the application of such a force. In
effect this results in the applied force acting over a distance, which results in mechanical energy
ll

transfer.

Consider the system in fig.1.2, where a force F acts on the piston and is given by pressure
A

x piston area. Work W is performed whenever this force translates through a distance.
 ld
or
Fig. 1.2 Illustration of Thermodynamic Work

Thus for a differential displacement dx of the piston the quantity of work is given by the

W
equation:

dW = Fdx ..(1.4)

Here F is the force acting along the line of the displacement x. If the movement takes place over a
TU
finite distance, the resulting work is obtained by integrating the above equation. By convention,
work is regarded as positive when the displacement is in the same direction as the applied force
and negative when they are in opposite directions. Thus, for the above example, the equation 1.4
may be rewritten as:
JN

dW = PAd (V t / A) ..(1.5)

If the piston area A is constant, then:

dW = PdV t ..(1.6)

As may be evident from eqn. 1.6, when work is done on a system (say through compression) the
ll

volume decreases and hence the work term is positive. The reverse is true when the system
performs work on the surroundings (through expansion of its boundary).
A
Heat
We invoke here the common observation that when a hot and a cold object are contacted, the hot
one becomes cooler while the cold one becomes warmer. It is logical to argue that this need be
due to transfer of something between the two objects. The transferred entity is called heat.
Thus, heat is that form of energy that is exchanged between system and its surrounding owing to

ld
a temperature differential between the two. More generally, heat is a form of energy that is
transferred due to temperature gradient across space. Thus heat always flows down the gradient
of temperature; i.e., from a higher to a lower temperature regions in space. In absence of such

or
temperature differential there is no flow of heat energy between two points. Heat flow is regarded
to be positive for a thermodynamic system, if it enters the latter and negative if it leaves.
Like work, heat is a form of energy that exists only in transit between a system and its

W
surrounding. Neither work nor heat may be regarded as being possessed by a thermodynamic
system. In a fundamental sense, the ultimate repositories of energy in matter are the atoms and
molecules that comprise it. So after transit both work and heat can only transform into the kinetic
and potential energy of the constituent atoms and molecules.
TU
1.4 Thermodynamic Equilibrium
In general change of state of a thermodynamic system results from existence of gradients of
various types within or across its boundary. Thus a gradient of pressure results in momentum or
convective transport of mass. Temperature gradients result in heat transfer, while a gradient of
JN

concentration (more exactly, of chemical potential, as we shall see later) promotes diffusive mass
transfer. Thus, as long as internal or cross-boundary gradients of any form as above exist with
respect to a thermodynamic system it will undergo change of state in time. The result of all such
changes is to annul the gradient that in the first place causes the changes. This process will
continue till all types of gradients are nullified. In the ultimate limit one may then conceive of a
ll

state where all gradients (external or internal) are non-existent and the system exhibits no further
changes. Under such a limiting condition, the system is said to be in a state of thermodynamic
equilibrium. For a system to be thermodynamic equilibrium, it thus needs to also satisfy the
A

criteria for mechanical, thermal and chemical equilibrium.

Types of Thermodynamic Equilibrium
A thermodynamic system may exist in various forms of equilibrium: stable, unstable and
metastable. These diverse types of equilibrium states may be understood through analogy with a
simple mechanical system as depicted in fig. 1.3 a spherical body in a variety of gradients on a

ld
surface.

or
W
Fig. 1.3 Types of Mechanical Equilibrium

Consider the body to be initially in state I. If disturbed by a mechanical force of a very small
TU
magnitude the body will return to its initial state. However, if the disturbance is of a large
magnitude, the body is unlikely to return to its initial state. In this type of situation the body is
said to be in unstable equilibrium. Consider next the state II; even a very small disturbance will
move the body to either positions I or III. This type of original equilibrium state is termed
metastable. Lastly, if the body is initially in state III, it will tend to return to this state even
JN

under the influence of relatively larger disturbances. The body is then said to be in a stable
equilibrium state. If E is the potential energy of the body and x is the effective displacement
provided to the body in the vertical direction, the three equilibrium states may be described by the
following equations:

E 2 E
ll

Stable Equilibrium:
= 0; >0 ..(1.7)
x x 2
A

E 2 E
Unstable Equilibrium:
= 0; <0 ..(1.8)
x x 2
 E 2 E
Metastable Equilibrium:
= 0;= 0 ..(1.9)
x x 2

The above arguments may well be extended to understand equilibrium states of thermodynamic
systems, which are relatively more complex in configuration. The disturbances in such cases

ld
could be mechanical, thermal or chemical in nature. As we shall see later (section 6.3), for
thermodynamic systems, the equivalent of (mechanical) potential energy is Gibbs free energy.
The considerations of change of Gibbs free energy are required to understand various complex

or
behaviour that a thermodynamic system containing multiple phases and components (either
reactive or non-reactive) may display under the influence of changes brought about by exchange
of energy across its boundary.

W
1.5 The Phase Rule
Originally formulated by the American scientist Josiah Willard Gibbs in the 1870s, the phase
rule determines the number of independent variables that must be specified to establish the
intensive state of any system at equilibrium. The derivation of the general phase rule is shown in
TU
chapter 6, but here we state it without proof:
F = 2 + N r ..(1.10)
Here, F = degrees of freedom of the thermodynamic system in question; N = Number of
components; = number of co-existing phases, and r = number of independent reactions that may
occur between the system components. For a non-reactive system, the phase rule simplifies to:
JN

F =2 + N ..(1.11)
In the most general sense a thermodynamic system may be multiphase and multi-
component in nature. A phase is a form of matter that is homogeneous in chemical composition
and physical state. Typical phases are solids, liquids and gases. For a multiphase system,
interfaces typically demarcate the various phases, properties changing abruptly across such
ll

interfaces. Various phases can coexist, but they must be in equilibrium for the phase rule to
apply. An example of a three-phase system at equilibrium is water at its triple point (~ 00C, and
A

0.0061 bar), with ice, water and steam co-existing. A system involving one pure substance is an
example of a single-component system. On the other hand mixtures of water and acetone have
two chemically independent components.
 The intensive state of a system at equilibrium is established when its temperature,
pressure, and the compositions of all phases are fixed. These are therefore, regarded as phase-rule
variables; but they are not all independent. The degrees of freedom derivable from the phase rule
gives the number of variables which must be specified to fix all other remaining phase-rule
variables. Thus, F means the number of intensive properties (such as temperature or pressure),

ld
which are independent of other intensive variables. For example, for a pure component gaseous
system, phase rule yields two degrees of freedom. This implies that if one specifies temperature
and pressure, all other intensive properties are then uniquely determined these two variables.

or
Similarly for a biphasic system of a pure component say water and steam there is only one
degree of freedom, i.e., either temperature or pressure may be specified to fix all other intensive
properties of the system. At the triple point the degrees of freedom is zero, i.e., any change from

W
such a state causes at least one of the phases to disappear.

1.6 Zeroth Law of Thermodynamics and Absolute Temperature

Thermometers with liquid working fluids are usually used for measurement of temperature.
TU
When such a device is brought in contact with a body whose temperature is to be measured, the
liquid column inside the thermometer expands due to heat conducted from the body. The
expanded length can be said to represent the degree of hotness in a somewhat quantitative
manner.
The Zeroth Law of Thermodynamics states that if two bodies are in thermal equilibrium
JN

with a third body, then the two given bodies will be in thermal equilibrium with each other. The
zeroth law of thermodynamics is used for measurement of temperature. In the Celsius
temperature scale, two fixed points ice point and steam point are used to devise the scale.
Thus, the freezing point of water (at standard atmospheric pressure) is assigned a value of zero,
while the boiling point of pure water (at standard atmospheric pressure) denoted as 100.
ll

However for introducing detail, the distance between the two end points of the liquid column
marks is arbitrarily divided into 100 equal spaces called degrees. This exercise can be extended
both below zero and above 100 to expand the range of the thermometer.
A

The entire exercise can be carried out with any other substance as the thermometric fluid.
However, for any specific measured temperature the extent of expansion of the liquid column
will vary with the thermometric fluid as each fluid would expand to different extent under the
influence of temperature. To overcome this problem, the ideal gas (see next section) has been
arbitrarily chosen as the thermometric fluid. Accordingly, the temperature scale of the SI system
is then described by the Kelvin unit (T0K). Its relation to the Celsius (t0C) scale is given by:

ld
( 0 K ) 273.15 + t ( 0C )
T=
Thus the lower limit of temperature, called absolute zero on the Kelvin scale, occurs at
273.150C.

or
1.7 The Ideal Gas
In the foregoing discussions we have pointed out that a thermodynamic system typically encloses

W
a fluid (pure gas, liquid or solid or a mixture) within its boundary. The simplest of the intensive
variables that can be used to define its state are temperature, pressure and molar volume (or
density), and composition (in case of mixtures). Let us consider for example a pure gas in a
vessel. As mentioned above, by phase rule the system has two degrees of freedom. It is an
experimentally observed phenomenon that in an equilibrium state the intensive variables such as
TU
pressure, temperature and volume obey a definitive inter-relationship, which in its simplest form
is expressed mathematically by the Boyles and Charless laws. These laws are compositely
expressed in the form of the following equation that is said to represent a behaviour termed as
Ideal Gas Law:
PV = RT ..(1.12)
JN

Where, P = system pressure (say, Pa = N/m2), T = system temperature (in 0K), V = gas molar
volume (mol/m3) and, R = universal gas constant (= 8.314 J / mol 0 K ). The above relation is said
to represent an equation of state, and may alternately be written as:
PV t = nRT ..(1.13)
ll

Where, Vt = total system volume; n = total moles of gas in the system. Units of typical
thermodynamic variables and that of the gas constant are provided in Appendix I.
The equations (1.10) and (1.11) are also termed Equations of State (EOS) as they relate the
A

variables that represent the thermodynamic state of a system in the simplest possible manner. It is
obvious that the EOS indicates that if one fixes temperature and pressure the molar volume is
automatically fixed as well, i.e., the latter is not an independent property in such a case.
The ideal gas law is a limiting law in the sense that it is valid primarily for gaseous
systems at low pressure, strictly speaking at pressure far below the atmospheric. However, for
practical purposes it is observed to remain valid at atmospheric pressures as well. As we shall see

ld
later, the ideal gas law serves as a very useful approximation as well as a datum for estimation of
both the volumetric (chapter 2) as well as all other real fluid thermodynamic properties of
practical interest (chapter 5, for example).

or
1.8 State and Path Dependent Thermodynamic Variables
Consider a gas at a certain temperature and a pressure within a piston-cylinder assembly (for

W
example, fig. 1.2), which for arguments sake we may assume to be isolated. If the piston
position is held fixed at this point the gas state is said to be characterized by the temperature and
the pressure and its corresponding volume. In its simplest form the relationship between these
intensive variables may be described by (say) eqn. 1.12. Consider next that the gas is compressed
TU
by application of an extra force on the piston so that it moves inwards into the cylinder. This
motion will continue till it reaches a point when the internal gas pressure equals the externally
applied pressure on the piston. If there is no further increase in the force applied to the piston, the
gas will also attain a new equilibrium state wherein the pressure and temperature would attain a
new set of values. If, on the other hand the extra applied pressure is removed and the gas reverts
JN

to the earlier state the original temperature and pressure (and, of course volume) is restored.
Extending this argument, in general, if the gas is heated or cooled, compressed or expanded, and
then returned to its initial temperature and pressure, its intensive properties are restored to their
initial values. It is evident, therefore, that such properties do not depend on the past history of the
fluid or on the path by which it reaches a given state. They depend only on present state,
ll

irrespective of how they are attained. Such quantities are thus defined as state variables.
Mathematically, this idea may be expressed as follows:
A

T2 P2 V2
T1
dT =
T2 T1 =T ; dP =
P1
P2 P1 =P; dV =
V1
V2 V1 =V ..(1.14)

The changes in the above intensive properties depend only on the initial and final states of the
system. They constitute point functions and their differentials are exact.
Let us next consider the case of thermodynamic work as defined by eqn. 1.6. It may be readily
evident that if one can depict the exact variation of pressure and volume during a change of state
of a system on a two-dimensional P-V graph, the area under the curve between the initial and
final volumes equal the work associated with process. This is illustrated in fig. 1.4.

ld
or
W
Fig. 1.4: Depiction of thermodynamic work on P-V plot
As shown in the above figure the work associated with a thermodynamic process clearly in
TU
dependent on the path followed in terms of P and V. It follows that if one were to go from state
1 to 2 by path X and then return to 1 by path Y the work in the two processes would differ
and so one would not be giving and taking work out of the system in equal measure. An entity
such as P-V work is, therefore, described as a path variable, and therefore is not directly
dependent on the state of the system. This is obviously distinctive from the case of state variables
JN

such as P and V (and T). Thus, for quantifying work, one cannot write an equation of the same
type as (1.12). The more appropriate relation for such variables may be written as:
2
1
W = W12 ..(1.15)

It may be pointed out that the notation is used to depict differential quantum of work in order to
ll

distinguish it from the differential quantity of a state variable as in eqn. 1.14. We demonstrate in
chapter 3 that, like P-V work, heat transferred between a system and the surrounding is also a
A

path variable and so one may also write:

2
1
Q = Q12 ..(1.16)
Heat and work are therefore quantities, and not properties; they account for the energy changes
that occur in the system and surroundings and appear only when changes occur in a system.
Although time is not a thermodynamic coordinate, the passage of time is inevitable whenever
heat is transferred or work is accomplished.

ld
1.9 Reversible and Irreversible Thermodynamic Processes
We have seen above that in absence of any gradients (or motive forces) a thermodynamic system
continues to remain in a state of equilibrium. Obviously, if a disturbance (i.e., mechanical,

or
thermal or chemical potential gradient) is impressed upon such a system it will transit from its
initial state of equilibrium. However, as it moves away from its initial state the originally applied
gradients will diminish progressively in time, and ultimately when they are reduced to

W
infinitesimal levels the system will attain a new equilibrium state. A question arises here as to the
nature of the process of change: if the initially impressed disturbances are reversed in direction
(not magnitude) can the system return to its first equilibrium state back through the same
intermediate states as it went through during the first phase of change? If that happens we depict
TU
the process as reversible, if not, then the process is termed irreversible.
It is necessary to understand the concept of reversibility of thermodynamic process more deeply
as it is an idealized form of process of change and without that consideration it is not possible to
represent or understand real thermodynamic processes, which are generally irreversible in nature.
What makes a thermodynamic process reversible? To answer the question let us again
JN

take the example of the simple gas-in-piston-and-cylinder system as shown in figure 1.5.
ll
A
 Fig. 1.5 Illustration of Reversibility of Thermodynamic Process

The system initially contains a pure gas whose pressure equals that exerted externally (due to
piston weight), and its temperature is the same as that of the environment. Thus it is at
equilibrium (say state A) as there are no mechanical, thermal or chemical concentration

ld
gradients in the system. Now a ball of a known weight is transferred on to the piston, whereupon
the external pressure exceeds the gas pressure and the piston moves down to attain a new lower
position at which point the gas has been compressed and its pressure once again equals that

or
applied externally. At the same time if any differentials in temperature (within or across the
system boundary) and internal concentration distribution of the gas molecules result due to the
applied mechanical imbalance, heat and mass transfer will take place simultaneously until these

W
gradients are also annulled and the system eventually comes to rest at a new equilibrium point
(say, B). We say that the system has undergone a process due to which its state has changed
from A to B. Note that this process can be continued as long as desires by sequentially
transferring more and more balls individually onto the piston and impelling the system to change
in steps till say the end point state X. The question that one may pose: is the process A-X
TU
reversible? That is, if one reversed all the initial steps of sequentially moving each ball off the
piston so as to reach from state X back to A would all the interim states of the system as
defined by temperature, pressure and volume at any point be identical to those obtained during
the process of going from A to X?
To answer this question we need to understand the process occurring in the system a little
JN

more deeply. Consider first that a mass m o is suddenly moved onto the piston from a shelf (at the
same level). The piston assembly accelerates downwards, reaching its maximum velocity at the
point where the downward force on the piston just balanced by the pressure exerted by the gas in
the cylinder. However, the initial momentum of the plunging piston would carry it to a somewhat
lower level, at which point it reverses direction. If the piston were held in this position of
ll

maximum depression brought about by transfer of the mass m 0 , the decrease in its potential-
energy would very nearly equal the work done on the gas during the downward movement.
A

However, if unrestrained, the piston assembly would oscillate, with progressively decreasing
amplitude, and would eventually come to rest at a new equilibrium position at a level below its
initial position.
 The oscillation of the piston assembly cease because it is opposed by the viscosity of the
gas, leading to a gradual conversion of the work initially done by the piston into heat, which in
turn is converted to internal energy of the gas.
All processes carried out in finite time with real substances are accompanied in some
degree by dissipative effects of one kind or another. However, one may conceive of processes

ld
that are free of dissipative effects. For the compression process depicted in Fig. 1.4, such effects
issue from sudden addition of a finite mass to the piston. The resulting imbalance of forces acting
on the piston causes its acceleration, and leads to its subsequent oscillation. The sudden addition

or
of smaller mass increments may reduce but does not eliminate this dissipative effect. Even the
addition of an infinitesimal mass leads to piston oscillations of infinitesimal amplitude and a
consequent dissipative effect. However, one may conceive of an ideal process in which small

W
mass increments are added one after another at a rate such that the piston movement downwards
is continuous, with minute oscillation only at the end of the entire process.
This idealized case derives if one imagines of the masses added to the piston as being
infinitesimally small. In such a situation the piston moves down at a uniform but infinitesimally
TU
slow rate. Since the disturbance each time is infinitesimal, the system is always infinitesimally
displaced from the equilibrium state both internally as well with respect to external surroundings.
Such a process which occurs very slowly and with infinitesimal driving forces is called a quasi-
static process. To freeze ideas let us assume that the gas in the system follows the ideal gas law.
Thus the pressure, temperature and volume at any point during the process are related by eqn.
JN

1.12 (or 1.13). Now imagine that the process of gradual compression is reversed by removing
each infinitesimal mass from the piston just as they were added during the forward process. Since
during the expansion process also the system will always be differentially removed from
equilibrium state at each point, the pressure, temperature and volume will also be governed by the
relation 1.12. Since the latter is an equilibrium relationship and hence a unique one, each interim
ll

state of the system would exactly converge during both forward and backward progress of system
states. Under such a condition the process of compression is said to be thermodynamically
reversible. Both the system and its surroundings are ultimately restored to their initial conditions.
A

In summary, therefore, if both the system and its surroundings can be restored to their respective
initial states by reversing the direction of the process, then the process is said to be reversible. If a
process does not fulfill this criterion it is called an irreversible process.
It need be emphasized that a reversible process need be a quasi-static process, and that the
origin of irreversibility lie in the existence of dissipative forces in real systems, such as viscosity,
mechanical friction. These forces degrade useful work irreversibly into heat which is not re-

ld
convertible by simply reversing the direction of the process, since during a reverse process a
fraction of useful work will again be lost in the form heat in overcoming the dissipative forces.
Thus, in the above example system if there were no viscous or frictional forces opposing the

or
motion of the piston the processes of compression and expansion would be reversible, provided
of course all changes occur under infinitesimal gradients of force. The argument in the last
sentence may be extended to state that if changes are brought about by finite gradients (in this

W
case finite difference in force across the piston, associated with addition of finite mass to the
piston), the process would necessarily be irreversible. This is because finite gradients will force
the system to traverse through non-equlibrium interim states, during which the pressure,
temperature and volume will not be constrained by a unique relationship such as eqn. 1.12, which
TU
holds for equilibrium states. Indeed, it would not be possible to define the non-equilibrium states
in terms of a single temperature or pressure, as there would be internal gradients of these
variables during processes induced by finite force imbalances across the system boundary. These
very same considerations would apply for the reverse process of expansion as well, if it occurs
under finite mechanical gradients. So in general during such processes it would not be possible to
JN

ascribe unique intensive properties to interim states during a change, and hence the forward and
reverse paths would not coincide as they would if the process occurs under quasi-static
conditions.
An additional point that obtains from the above considerations is that only under
reversible conditions can one calculate the thermodynamic work by integrating eqn. 1.6, since at
ll

all points during the process the variables P, V and T are always uniquely related by the eqn.
1.12. Clearly if the process were occurring under irreversible conditions no such relation would
hold and hence the calculation of the thermodynamic work would not be possible through a
A

simple integration of eqn. 1.6.

 The foregoing discussion has used the example of a single-phase closed-system, where
compression and expansion processes are induced by gradients of mechanical force across the
system boundary. There are, however, many processes which are occur due to potential gradients
other than mechanical forces. For example, heat flow is induced by temperature differences,
electromotive force gradients lead to flow of electricity, and chemical reactions take place as

ld
there is a difference between the chemical potential of reactants and products. In general, it may
be shown that all such processes brought about by potential gradients of various kinds would tend
to reversibility if the gradients are themselves infinitesimal. For example, heat transfer across the

or
boundary of a thermodynamic system would be reversible if the difference across it is of a
differential amount dT, and so on.
1.10 Significance of Chemical Engineering Thermodynamics: Process Plant Schema

W
Before we conclude the present chapter it would be appropriate to obtain a brief preview of the
scope and utility of the principles of thermodynamics insofar as application to real world
processes is concerned. Although based on relatively abstract principles, the laws of
thermodynamics provide the fundamental constraints under which all real world process take
TU
place. The ultimate application of the knowledge of the core principles of chemical engineering is
in the design of a chemical process plant. Engineering thermodynamics constitutes one of the
principal elements of such knowledge. Typically such a plant converts a set of raw materials to a
desired product through a variety of steps that are schematically represented by Fig. 1.6.
JN
ll
A

Fig. 1.6 Chemical Process Plant Schema

The raw materials most often are mixtures which need to be purified to obtain the right
composition required for conversion to products A wide variety of separation processes are
available for carrying out such purification; examples include distillation, liquid-liquid extraction,
precipitation from solutions, crystallization, etc. Practically all such separation processes involve

ld
generation of two or more phases, in one of which the desired raw material components are
preferentially concentrated, which is then used recover the substances in a relatively purer form.
For a typical large scale chemical plant the separation process equipments may constitute more

or
than half of the total capital investment.
The chemical reactor forms the heart of a chemical plant. It is here that once the feed
materials are available in the right proportions (and compositions) they are reacted to yield the

W
product. Obtaining the desired product requires an optimal choice of conditions under which the
reactor may be operated. However, the product formed is very rarely obtained in a pure form.
This is because typically the feed is never fully converted to product molecules and therefore the
stream exiting the reactor is not a pure substance. In addition it is usually a common phenomenon
TU
that the intended chemical reaction is accompanied by often more than a single side reaction. The
latter leads to the formation of side products, which results in contamination of the final
product. Therefore, it is usually required to subject the reactor exit stream to another round of
purification to obtain a product with the desired specifications of the product
With regards to all such processes of purification and reaction, the laws of
JN

thermodynamics play a very fundamental role: they allow the calculation of the principal entities
that form the basis of design and operation of process plants:
1. The maximum degree of purification that is possible under a given set of processing
conditions
2. The maximum degree of conversion possible under the reaction conditions
ll

3. The optimal operating conditions for separation and reaction processes

4. The total energy required to achieve the intended degree of separation and reaction, and
therefore the plant energy load
A

The calculation of the above parameters tends to constitute 50-70% of the computational load
encountered during the stage of basic process plant design. Thus, the principles of chemical
engineering thermodynamics is one of the mainstays of knowledge needed to realize the goal
of plant design and operation.

ld
or
W
TU
JN
ll
A
Chapter 2: Volumetric Properties of Real Fluids
In the previous chapter we have introduced the concept of ideal gas and the corresponding
equation of state (EOS). However, such a state obtains for a gas only at pressures around and
below atmospheric (and at high temperatures), and, therefore, constitutes a limiting case. As we

ld
know, substances exist also in other forms: solids, liquids, etc. Also more often than not, in
practice (as in process plants) gases (as well as other phases) may exist at substantially higher
pressures (up to several thousands of atmospheres). This necessitates the development of other

or
EOSs not only for gases, but also for relating P-V-T behavior of liquids and solids. Such
mathematical relations, if expressed in suitably generalized forms, provide the added advantage
of being able to quantify P-V-T behaviour for a large number of individual substances.

W
Such volumetric properties form a group of macroscopic thermodynamic state variables
which are most easily measured. As we will see in later chapters, all other intensive
thermodynamic properties can be represented in terms of mathematical expressions which denote
functions of volumetric and a number of other directly measurable state variables. In the last
chapter we have already introduced the ideas of thermodynamic work which clearly can be
TU
calculated if the relation between P and V is known. In chapters 3-5 we will demonstrate that heat
transfer occurring under reversible conditions, as well as a host of other intensive, state variables
(internal energy, enthalpy, entropy, Gibbs free energy, etc) can also be calculated using
volumetric properties of a substance in question. Since work and heat are two principal modes of
energy transfer in most thermodynamic systems of practical interest, it follows that the
JN

knowledge of volumetric properties is fundamental to all such calculations. Finally, as part of

process plant design one needs volumetric properties for the purpose of sizing of process
pipelines and all major process equipments such as reactors, heat exchangers, distillation
columns, and so on. These considerations underscore the precise significance of P-V-T behavior
of substances in all plant design activity.
ll

In the following sections we first describe the general nature of the P-V-T behavior of
pure substances: in gaseous, liquid and solid forms. The various EOSs available to quantify such
A

real fluid behaviour are then considered. Lastly, generalized correlations to relate gas and liquid
behaviour are presented. Such analytical EOSs and generalized correlations allow prediction of
P-V-T values of real fluids and are, therefore, of great value as they can obviate the need for
detailed experimental data.

2.1 General P-V-T Behaviour of Real Fluids

P-V Diagrams

ld
Fig. 2.1 represents the general pure component, real fluid phase behavior that typically obtains
from experimental measurements. Consider first the fig. 2.1a. Let us take a substance at some
temperature T 1 and certain pressure P 1 such that it is in a gaseous state A 1 . Keeping the

or
temperature fixed at T 1 if one pressurizes the gas (say in a piston-cylinder assembly as in fig. 1.2)
its molar volume will decrease along the curve A-B. At point B, any further pressurization leads
to commencement of condensation of the gas into a liquid from. Point B is thus said to

W
correspond to a state where the substance is in a saturated vapour state. Once condensation
begins any attempt at reducing the volume by further
TU
(a)
JN
ll

.
A

(b)

Fig. 2.1 General P-V plots for real fluids

pressurization more of the saturated vapour present at B progressively liquefies until a point X is
reached where all the original gas (or saturated vapour) is fully converted to liquid state. Point X
is described as a saturated liquid state. It follows that at all point between B and X the substance
exists partitioned into two phases, i.e., part vapour and part liquid. As one transits from B to X,
pressure and temperature both remain constant; the only change that occurs is that the fraction of

ld
the original gas at point A (or B) that is liquefied increases, until it is 1.0 at point X. The line B-X
connecting the saturated vapour and liquid phases is called the tie-line. For a given T and P, the
relative amounts of the phases determine the effective molar (or specific) volume at any point

or
within the two-phase region. Any further attempt to pressurize the saturated liquid results in
relatively very little compression, and this is captured by the steep slope of the curve X-Y, which
signifies that the liquid state is far less compressible, compared to the gas state (i.e., points over

W
A-B). Essentially points between X-Y (including Y itself) represent compressed liquid states.
An important point to re-emphasize is that on the two-phase line B-X, the pressure of the
system remains constant at a fixed value. This pressure is termed the saturation pressure ( P sat )
corresponding to the temperature T 1. We recall your attention to the phase rule described in
TU
section 1.5, and eqn. 1.11. By this eqn. the degrees of freedom is one, which is borne by the fact
that if one fixes temperature the system pressure also becomes fixed. However, in both regions A-
B and X-Y the degrees of freedom is two, as pressure becomes fixed only if one defines both
temperature and volume.
In general, the same behviour as detailed above may repeat at another temperature
JN

T (>T 1 ). One can on the one hand connect all the saturated vapour phase points at different
temperatures and on the other connect all the points representing saturated liquid phase, the locus
of such points give rise to the dome-shaped portion X-C-B of the P-V diagram which essentially
signifies that at any pressure and volume combination within this dome, the state of the system is
biphasic (part gas and part liquid). The region right of the dome B-C represents saturated gas
ll

phase while to the left (X-C) the state is saturated liquid. If one continues to conduct the
pressurization at increasingly higher temperatures, one eventually arrives at a temperature for
A

which the tie-line is reduced to a point and the P-V curve turns into an inflexion point to the two-
phase dome. The temperature which such a behavior obtains is called the critical temperature
(T C ), while the pressure at corresponding point of inflexion is termed the critical pressure (P C ).
The molar volume at the point is termed the critical volume, and the state itself the critical point.
A fluid which is at a temperature and pressure above the critical point values is said to be in a
supercritical state; this is indicated by the hatched region in fig. 2b. As has been shown for the P-
V curves for a T > T C , there exists no liquid phase as the curve passes beyond the two-phase
dome region. Thus, the critical temperature is a temperature above which a gas cannot be

ld
liquefied by compressing, as can be below it. Compilation of values of critical properties and
for a large number of substances are available readily from several sources (see:
http://srdata.nist.gov/gateway/gateway). Values of these parameters for some select substances

or
are provided in Appendix II.
In fig. 2.1b the phase behavior depicted in fig. 2.1a is extended and more generalized to
include solid phase as well. Accordingly, not only vapour-liquid region, other two phase regions,

W
i.e., solid-vapour and solid-liquid regions are also displayed. The same arguments as made above
for explaining the nature of co-existence of vapour and liquid phases apply to the other two
biphasic regions.
TU
P-T Diagrams
The phase behaviour described by fig. 2.1 can also be expressed in a more condensed manner by
means of a pressure-temperature (P-T) diagram shown in fig 2.2. Just as P-V curves were
depicted at constant temperature, the P-T diagram is obtained at a constant molar volume. The
two phase regions which were areas in the P-V diagram are reduced to lines (or curves) in fig.
JN

2.2. The P-T curves shown by lines X-Y, Y-Z, and Y-C result from measurements of the vapour
ll
A

Fig. 2.2 Pressure-Temperature Diagram of a Pure Substance

pressure of a pure substance, both as a solid and as a liquid. X-Y corresponds to the solid-vapour
(sublimation) line; X-Y represents the co-existence of solid and liquid phases or the fusion line,
while the curve Y-C displays the vapour-liquid equilibrium region. The pressure at each
temperature on the Y-C curve corresponds to the equilibrium vapour pressure. (Similar
considerations apply for P-T relation on the sublimation curve, X-Y). The terminal point C

ld
represents the critical point, while the hatched region corresponds to the supercritical region. It is
of interest to note that the above three curves meet at the triple-point where all three phases, solid,
liquid and vapour co-exist in equilibrium. By the phase rule (eqn. 1.11) the degrees of freedom at

or
this state is zero. It may be noted that the triple point converts to a line in fig. 2.1b. As already
noted, the two phases become indistinguishable at the critical point. Paths such as F to G lead
from the liquid region to the gas region without crossing a phase boundary. In contrast, paths

W
which cross phase boundary Z-Y include a vaporization step, where a sudden change from liquid
to gas occurs.
A substance in the compressed liquid state is also often termed as sub-cooled, while gas at
a pressure lower than its saturation vapour pressure for a given temperature is said to be
TU
superheated. These descriptions may be understood with reference to fig. 2.2. Let us consider a
compressed liquid at some temperature (T) and pressure (P). The saturation temperature for the
pressure P would be expected to be above the given T. Hence the liquid is said to be sub-cooled
with respect to its saturation temperature. Consider next a pure vapour at some temperature (T)
and pressure (P). Clearly for the given pressure P the saturation temperature for the pressure P
JN

would be expected to be below the given T. Hence with respect to the saturation temperature the
vapour is superheated.
The considerations for P-V and P-T diagrams may be extended to describe the complete
P-V-T phase behaviour in the form of three dimensional diagrams as shown in fig. 2.3. Instead of
two-dimensional plots in figs. 2.1 and 2.2 we obtain a P-V-T surface. The P-V plots are recovered
ll
A
 ld
or
W
Fig. 2.3 Generalized Three-dimensional P-V-T Surface for a Pure Substance
if one takes a slice of the three dimensional surface for a given temperature, while the P-T curve
obtains if one takes a cross-section at a fixed volume. As may be evident, depending on the
volume at which the surface is cut the P-T diagram changes shape.
Fig. 2.4 illustrates the phase diagram for the specific case of water. The data that is
TU
pictorially depicted so, is also available in the form of tables popularly known as the steam
table. The steam table (see http://www.steamtablesonline.com/) provides values of the following
thermodynamic properties of water and
JN
ll
A

Fig. 2.4 Three-dimensional P-V-T Plot for Water

vapour as a function of temperature and pressure starting from its normal freezing point to the
critical point: molar volume, internal energy, enthalpy and entropy (the last three properties are
introduced and discussed in detail in chapters 3 and 4).
The steam tables are available for saturated (two-phase), the compressed liquid and
superheated vapour state properties. The first table presents the properties of saturated gas and

ld
liquid as a function of temperature (and in addition provides the saturation pressure). For the
other two states the property values are tabulated in individual tables in terms of temperature and
pressure, as the degree of freedom is two for a pure component, single state. For fixing the values

or
of internal energy, enthalpy and entropy at any temperature and pressure those for the saturated
liquid state at the triple point are arbitrarily assigned zero value. The steam tables comprise the
most comprehensive collection of properties for a pure substance.

W
2.2 Origins of Deviation from Ideal Gas Behaviour
The ideal gas EOS is given by eqn. 1.12. While this is a relationship between the macroscopic
intensive properties there are two assumptions about the microscopic behaviour of molecules in
TU
an ideal gas state:
i. The molecules have no extension in space (i.e., they posses zero volume)
ii. The molecules do not interact with each other
In particular, the second assumption is relatively more fundamental to explaining deviations from
ideal gas behavior; and indeed for understanding thermodynamic behavior of real fluids (pure or
JN

mixtures) in general. For this, one needs to understand the interaction forces that exist between
molecules of any substance, typically at very short intermolecular separation distances (~ 5 20
0 0
A (where 1 A = 10-8m).

Uncharged molecules may either be polar or non-polar depending on both on their

geometry as well as the electro-negativity of the constituent atoms. If the centre of total positive
ll

and negative charges in a molecule do not coincide (for example, for water), it results in a
permanent dipole, which imparts a polarity to the molecule. Conversely, molecules for which the
A

centres of positive and negative charge coincide (for example, methane) do not possess a
permanent dipole and are termed non-polar. However, even a so-called non-polar molecule, may
possess an instantaneous dipole for the following reason. At the atomic level as electrons
oscillate about the positively charged central nucleus, at any point of time a dipole is set up.
However, averaged over time, the net dipole moment is zero.
When two polar molecules approach each other closely the electric fields of the dipoles
overlap, resulting in their re-orientation in space such that there is a net attractive force between
them. If on the other hand a polar molecule approaches a non-polar molecule, the former induces

ld
a dipole in the latter (due to displacement of the electrons from their normal position) resulting
once again in a net attractive interaction between them. Lastly when two non-polar molecules are
close enough their instantaneous dipoles interact resulting in an attractive force. Due to these

or
three types of interactions (dipole-dipole, dipole-induced dipole, and induced dipole-induced
dipole) molecules of any substance or a mixture are subjected to an attractive force as they
approach each other to very short separation distances.

W
However, intermolecular interactions are not only attractive. When molecules approach to
0
distances even less than ~ 5 A or so, a repulsive interaction force comes into play due to overlap
of the electron clouds of each molecule, which results in a repulsive force field between them.
Thus if one combines both the attractive and repulsive intermolecular interactions the overall
TU
interaction potential U resembles the schematic shown in fig. 2.5.
JN

Fig. 2.5 Schematic of Intermolecular potential energy U for a pair of uncharged molecules
ll

Many expressions have been proposed for the overall interaction potential U [see, J.M.
A

Prausnitz, R.N Lichtenthaler and E.G. Azevedo, Molecular Thermodynamics of Fluid Phase
Equilibria, (3rd ed.), 1999, Prentice Hall, NJ (USA)]. These are essentially empirical, although
their functional forms often are based on fundamental molecular theory of matter. The most
 widely used equation in this genre is the Lennard-Jones (LJ) 12/6 pair-potential function which
is given by eqn. 2.1:
12 6
U ( LJ ) 4
=
r r
..(2.1)
Where, r = intermolecular separation distance; , = characteristic L-J parameters for a

ld
substance.
The r 12 term represents the repulsive interaction, whereas the r 6 term corresponds to the
attractive interaction potential. As already indicated, the domain of intermolecular interactions is

or
limited to relatively low range of separation distances. In principle they are expected to be
operative over r = 0 ; but for practical purposes they reduce to insignificant magnitudes for
separations exceeding about 10 times the molecular diameter.

W
The L-J parameters , are representative of the molecular interaction and size
respectively. Typical values of the L-J equation parameters for various substances may be found
elsewhere (G. Maitland, M. Rigby and W. Wakeham, 1981, Intermolecular Forces: Their Origin
and Determination, Oxford, Oxford University Press.)
TU
Since gases behave ideally at low pressures, intermolecular separation distances therein
are typically much higher than the range over which intermolecular interactions are significant.
This is the reason why such interactions are negligible in case of ideal gas, which essentially is
one of the assumptions behind the definition of ideal gas state. Indeed while the ideal gas EOS is
expressed in macroscopic terms in eqn. 1.12, the same equation may be derived from microscopic
JN

(thermodynamic) theory of matter.

The root of non-ideal gas behavior, which typically obtains at higher pressure, thus is due
to the fact that at elevated pressures, the intermolecular separations tend to lie within the
interactive range and hence the ideal gas assumption is no longer valid. Thus, the ideal gas EOS
is insufficient to describe the phase behavior of gases under such conditions.
ll

Intermolecular interactions also help explain the behavior of fluids in other states. Gases
can condense when compressed, as molecules are then brought within the separations where the
A

attractive forces constrain the molecules to remain within distances typical of liquid phase. It
follows that a pure component liquid phase cannot be ideal in the same sense as a gas phase can
be. Further, the fact that liquids are far less compressible also is due to the repulsive forces that
 operate at close intermolecular distances. Obviously these phenomena would not be observed
unless there were interactions between molecules. Thus, it follows that while properties of the
ideal gas depend only on those of isolated, non-interacting moleclues, those of real fluids depends
additionally on the intermolecular potential. Properties which are determined by the
intermolecular interaction are known as configurational properties, an example of which is the

ld
energy required for vapourization; this is because during the process of vapourization energy has
to be provided so as to overcome the intermolecular attractive force between molecules in the
liquid phase and achieve the gas state where the seprations are relatively larger.

or
2.3 Equations of State for Real Fluids
The generic form of an equation of state (EOS) is: f ( P, V , T ) = 0

W
However, as we have already seen by the phase rule, for a single phase pure component the
degrees of freedom are two. This may be expressed in the form of an EOS equation as follows:
V = V (T , P )
TU
It follows that:

V V
=dV dT + dP ..(2.2)
T P P T

1 V
Defining Volume Expansivity as ..(2.3)
JN

V T P

1 V
Isothermal compressibility as ..(2.4)
V P T

The generic EOS (2.2) may be written as:

ll

dV
= dT kdP ..(2.5)
V
A

2.3.1 EOS for Liquids

For liquids, which are relatively incompressible, the factors and are generally show an weak
dependence on T and P and hence averaged values of these parameters may be used for
estimating the liquid volume at any temperature using the following integrated form of the
equation (2.5):

V2
ln = (T2 T1 ) k ( P2 P1 ) ..(2.6)
V1

ld
For liquids the usual datum volume (i.e., V 1 in eqn.2.6) can be the saturated volume at a given
temperature, which may be obtained from the Rackett equation (H. G. Rackett, J. Chem. Eng.
Data, 1970, vol. 15, pp. 514-517); i.e.,:

or
0.2857
V sat Vc Z c(1Tr )
= ; where
= / TC reduced temperture.
Tr T= ..(2.7)

Where Z C is the critical compressibility factor (see below).

W
--------------------------------------------------------------------------------------------------------------------
Example 2.1
For a liquid A at 350K and 1 bar, = 50x10-6 bar-1. (i) To what pressure must water be
compressed at 350 K to change its density by 0.5%? Assume that is independent of P.
TU
(Click for solution)

--------------------------------------------------------------------------------------------------------------------
2.3.2 EOS for Gases

In contrast to liquids, gases are relatively far more compressible, and so volume is strongly
JN

dependant on temperature and pressure. Consequently eqn. 2.6 cannot be used easily to estimate
volume at a given T & P, as both and are strong functions of T and P. Thus, various EOSs
have been proposed to describe gas phase volumetric properties. The next section presents select
EOSs that are typically used for the gas phase, ranging from those applicable to moderate
pressure to others which are more accurate at high pressures.
ll

Virial EOS:
Generally applicable to moderate deviations from ideal gas behavior, the virial EOS is given by
A

two alternate forms:

PV B C
=1 + + 2 + ... ..(2.8)
RT V V
 Where B, C = second and third virial coefficients

PV
Or: 1 + BP + C P 2 + ...
= ..(2..9)
RT

It may be shown that the following relations hold further:

ld
B C B2
B= ; and C = ..(2.10)
RT ( RT ) 2

As may be evident, the second and third terms on the right side of eqns. 2.8 and 2.9 constitute

or
corrections for the non-ideal behavior of a gas. The virial coefficients are essentially dependent of
temperature. The more the number of virial coefficients used in the equation the better is the
prediction of gas molar volume. While the estimation of the second virial coefficient is relatively

W
straightforward, that of the third virial coefficient is generally far more complex, and there is
scant data in the literature on its value for different substances. Because of this the virial EOS is
most commonly used in the truncated form, and applies to moderate deviations from ideal gas
behavior:
TU
PV B
= 1+ ..(2.11)
RT V

BP BP P
1+
Alternately, Z = 1+ c r
= ..(2.12)
RT RTc Tr

Where, Pr and Tr are reduced pressure and temperature respectively, given by:
JN

= / PC ; Tr T / TC
Pr P=

A set of generalized correlations have been proposed by Pitzer and co-workers (K. S. Pitzer,
Thermodynamics, 3d ed., App. 3, McGraw-Hill, New York, 1995) for computing the second
virial coefficients whereby:
BPc
ll

B + B
0 1
= ..(2.13)
c
RT
A

The terms B 0 and B1 are given by:

0.422
B 0 0.083
= ..(2.14)
Tr1.6
 0.172
B1 0.139
= ..(2.15)
Tr4.2
The parameter , discussed in a later section, is the acentric factor, which is a fundamental
thermodynamic property of a substance, and is a measure of asphericity of molecular shape (see
Appendix II for values of acentric factor for select substances). The value of acentric factor may

ld
be obtained from the following expression from experimental vapour pressure data of a pure
substance (K. S. Pitzer, Thermodynamics, 3d ed., App. 3, McGraw-Hill, New York, 1995):
1.0 log10 ( Prsat )T =0.7 ..(2.16)

or
r

( Prsat )Tr =0.7 = reduced saturation vapour pressure at a reduced temperature equalling 0.7.

Since the virial EOS applies to only moderate deviations from ideal behaviour several complex

W
extensions of it have been proposed for increasing accuracy of prediction at higher pressures; an
example of such an EOS is the so-called Benedict-Webb-Rubin (BWR) EOS (G.B.Benedict, G.
B. Webb, and L. C. Rubin, J. Chem. Phys., vol. 8, pp. 334-345, 1940) used in the petroleum and
natural-gas industries:
RT B0 RT A0 C0 / T 2 bRT a a
TU
c
P= + 2
+ 3
+ 6 + 3 2 1 + 2 exp 2 ..(2.17)
V V V V V T V V
where A 0 , B 0 , C o , a, b, c, and are all constant for a given fluid.
--------------------------------------------------------------------------------------------------------------------
Example 2.2
Calculate the molar volume for butane at 2.5bar and 298 K using the truncated virial EOS using
JN

the following
= data: TC 425.1
= K ; PC 37.96
= bar ; 0.2.

(Click for Solution)

--------------------------------------------------------------------------------------------------------------------
2.3.3 Cubic EOS for Gases
ll

While the truncated virial EOS applies to moderate pressures, it is often necessary to obtain
volumetric properties of gases at relatively much higher pressures. It has generally proved to be
computationally unwieldy to use extensions of the virial EOS at higher pressure. To obviate this
A

problem an entire range of alternate EOSs, termed cubic EOS, have been proposed by various
workers. The term cubic equation of state implies an equation which, if expanded, would
contain the volume terms raised to the first, second, or third power. Such an EOS provides two
distinct advantages over many other more complex EOS found in the literature; they allow:
Prediction of both gas and liquid (saturated) phase molar volumes
Provide a trade-off between complexity and accuracy of prediction
The first of such cubic EOS was proposed by the Dutch physicist Johannes Diderik van der

ld
Waals in 1873, and has the following form:
RT a
=P 2 ..(2.18)
V b V

or
Where, a and b are characteristic constants for a pure substance. For example, for nitrogen the
values of a and b are 0.1368 Pa.m6 / mol 2 and 3.864x105 m3 / mol respectively. Values of these
parameters for other substances may be found in the literature (see, for example: S.I. Sandler,

W
Chemical, Biochemical and Engineering Thermodynamics, 4th edition, Wiley India, 2006).
Once can take advantage of the phase P-V behaviour at the critical point (see fig. 2.1) to
determine the expressions for a and b. Since the P-V curve offers a point of inflection at the
critical point one may write:
TU
(P / V )TC =
0 ..(2.19)
( 2 P / V 2 )TC =
0
..(2.20)

On applying eqns. (2.19) and (2.20) on 2.18 two independent equations obtain, which may solved
simultaneously to show that for the vdW EOS:
JN

a = (27 / 64) R 2TC2 / PC

b = (1/ 8) RTC / PC

VC = (3 / 8) RTC / P

It follows that the critical compressibility factor is then given by:

Z C PCVC / RT
= = 3=
/ 8 0.375
ll

The vdW EOS suggests that for all pure substances the critical compressibility factor is a
constant (= 0.375). This, however, is contrary to experience. Indeed, although similar in
A

magnitude, the value of Z C is specific to a pure substance. For example, for methane it is 0.286,
while for helium 0.302. This indicates that although an improvement over ideal gas EOS, the
 vdW EOS does not reflect the specificity needed to predict the behaviour of all substances at high
pressures.
To bridge this gap several other cubic EOS of higher accuracy have been advanced, the three
most widely used of which are:
Redlich-Kwong (RK) EOS (Otto Redlich and J. N. S. Kwong, Chem. Rev.,vol. 44, pp.

ld
233-244, 1949)
Soave-Redlich-Kwong (SRK) EOS (G. Soave, Chem. Eng. Sci., vol. 27, pp. 1197-1203,
1972.)

or
Peng-Robinson (PR) EOS (D. Y. Peng and D. B. Robinson, Ind. Eng. Chem. Fundam.,
vol. 15, pp. 59-64, 1976.)
All the above EOSs including that due to van der Waals may be expressed by a single equation of

W
the following form:
RT a
=P ..(2.21)
V b V + ubV + wb
2 2

As part of a generalized approach to representing non-ideal gas behavior parameter Z, termed the
TU
compressibility factor, may be defined. This is discussed further in the following section.
However, here we provide the expression for it:
Z = PV/RT ..(2.22)
Using eqn. 2.20, eqn. 2.19 may be rewritten as:
Z3 +Z 2 + Z + =0 ..(2.23)
JN

Where:
A = aP / ( RT ) 2

B = bP / RT
=1 B + uB
=A + wB uB uB
2 2

= AB wB wB
ll

2 3

Expressions for all the above parameters for each EOS are provided in Table 2.1 below.
A
 Table 2.1 Values of u , w , , , for various cubic EOSs

EOS u w a b

vdW 0 0 1B A AB 27 R 2TC2 RTC

64 PC 8 PC

ld
RK 1 0 1 A B B2 AB 0.42748R 2TC2 RK 0.08664RTC
PC PC

SRK 1 0 1 A B B2 AB 0.42748R 2TC2 SRK 0.08664RTC

or
PC PC

PR 2 1 1+B A 2B 3B2 AB + B2 +B3 0.45724R 2TC2 PR 0.07779RTC

PC PC

W
The parameters RK , SRK and PR given by:

RK = Tr1/2

SRK =
[1 + (0.48 + 1.574 0.176 2 )(1 Tr1/2 )]2
TU
PR =
[1 + (0.37464 + 1.5422 0.62992 2 )(1 Tr1/2 )]2

As may be evident from table 2.1 all the parameters are determinable from the critical properties,
and the temperature at which the molar volume or pressure needs to be calculated. Once these
are computed the eqn. 2.21 may be solved to yield the value of Z, which in turn may be used to
JN

obtain the molar volume. Eqn. 2.21 may be solved using a suitable iterative algorithm or more
readily by the general analytical solution for cubic algebraic equations, which is outlined below.
______________________________________________________________________________
ll
A
 For solving eqn. 2.21, substitute, Z= X ( / 3) which gives:
X + pX + q =
3
0
where:
p= ( / 3) 2

q= (2 / 27) ( / 3) +
2

ld
Now,=let D (q / 4) + ( p / 27)2 3

Then the roots of eqn. (2.21) are as follows:

Case 1:

or
If D > 0, then only one real root exists:

1/3 1/3
q q
Z = + D + D
2 2 3

W
Case 2 :
If D = 0, then there are three real roots, of which two are equal:
q q
1/3 1/3

2 ; Z =
Z = Z =

2 3 2 3
1 2 3

Case 3 :
TU
If D < 0, then there are three unequal real roots:

=Z 2r cos( / 3)
1/3

3
1

2 +
=Z 2r cos
1/3

3 3
2
JN

4 +
=Z 2r cos
1/3

3 3
3

1/2
q 27
where: cos =

2 p 3

p 3

r=
ll

27

______________________________________________________________________________
A

The true advantage of use of cubic EOS derives from the fact that not only can it provide gas
phase molar volumes, but also that of saturated liquid at a given temperature (note that this is not
feasible by use of virial EOS). To understand this point we refer to fig. 2.6, which is a simplified
 ld
or
Figure 2.6 P-V plots at constant temperatures (as obtained from a cubic equation of state)

W
version of fig. 2.1.
For given values of V and T one can calculate P (using any cubic EOS) as a function of V
for various values of T. Figure 2.4 is a schematic P-V curves for three different temperatures. The
variations shown by T = T C and T > T C have already been discussed in section 2.1. We draw
TU
your attention here to the plot at T < T C . In this case the pressure decreases rapidly in the
subcooled liquid region with increasing V; after crossing the saturated-liquid line, it goes through
a minimum (1-2-3), rises to a maximum (3-4-5), and then decreases, crossing the saturated-vapor
line and continuing downward into the superheated-vapor region. However, experimentally
obtained plots of this kind (at T < T C ) do not show this manner of transition from saturated liquid
JN

to saturated vapor. As already discussed in section 2.1, in the two phase region within the dome,
the curve connecting the saturated vapor and liquid states constitutes a tie-line. Along this
horizontal segment the two phases coexist in varying proportions at the saturation vapor pressure.
Clearly then the true behavior of the substance cannot be exactly reproduced by an analytical
cubic EOS in the two phase region.
ll

Nevertheless, the P-V behavior that obtains in the two-phase region using the cubic EOS
is not entirely unrealistic. Indeed the two points of intersection of the analytical P-V curve (at a
A

given temperature) with the boundary of the dome occurs (approximately) at the points where the
experimental saturated vapour and liquid state volumes obtain (i.e., points 1 and 5). Cubic
equations of state have three volume roots, of which two may often be complex conjugates.
Physically relevant values of V are always real, positive, and greater than constant b which
appears in the cubic EOSs. For T > T C , at any pressure the solution for eqn. 2.21 yields one real
root. This is also true at any pressure on the T = T C curve, except for P = P C ; at the latter
condition three real roots obtain all equal to V C . For T < T C , in general there could be one or
three real roots. If the pressure is different from the saturation vapour pressure (Psat, at the given

ld
temperature) the roots are not realistic. If, however, the pressure is taken to be Psat, three real
roots are found. The highest amongst these correspond to the saturated vapour phase volume,
while the lowest one approximates the saturated liquid phase volume; the third root, however, is

or
not physically meaningful.
--------------------------------------------------------------------------------------------------------------------
Example 2.3

W
For methane at 298K and 2 MPa compute the molar volume using SRK equation. For methane,
T c = 190.7 K, P c = 46.41 bar, = 0.011.
(Click for solution)
--------------------------------------------------------------------------------------------------------------------
TU
2.3.4 Principle of Corresponding States
The compressibility factor Z has already been defined in eqn. 2.20. A typical plot of Z as a
function of T & P for methane is shown in fig.2.7. Experimentally measured values of Z for
different fluids display similar dependence on reduced temperature and pressures, i.e., Pr and Tr .
JN
ll
A
 Fig. 2.7 Variation of Compressibility-factor with pressure and temperature (Source: J.M. Smith,
H.C. Van Ness and M.M. Abbott, Introduction to Chemical Engineering Thermodynamics, 6th ed.,
McGraw-Hill, 2001)

This observation has been generalized to formulate the two-parameter theorem of corresponding

ld
states which is stated as follows: All fluids at the same reduced temperature and reduced
pressure have approximately the same compressibility factor, and all deviate from ideal-gas
behavior to about the same degree. Fig. 2.8 presents select experimental data which support this

or
observation.

W
TU
Fig. 2.8 Experimental compressibility factors for different fluids as a function of the reduced
JN

temperature and pressure. [Source: H. C.Van Ness and M.M Abbott (1982) based on data from
G.-J. Su (1946). Ind. Engr. Chem. 38, p 803.]

While this theorem applies fairly reliably to the simple fluids (argon, krypton, and xenon),
for more complex fluids the deviations are significant. To address this gap Pitzer and coworkers
ll

introduced a third corresponding-states parameter, characteristic of molecular structure, more

particularly the degree of sphericity of the molecule; the most widely used one is acentric
A

factor (already utilized in eqn. 2.13 and for computation for cubic EOS parameters in table
2.1) (K. S. Pitzer, Thermodynamics, 3d ed., App. 3, McGraw-Hill, New York, 1995). The
expression for was provided in eqn. 2.16. As is evident, it can be computed for any substance
using critical properties and a single vapor-pressure measurement made at T r = 0.7.
By definition (see J.M. Smith, H.C. Van Ness and M.M. Abbott, Introduction to
Chemical Engineering Thermodynamics, 6th ed., McGraw-Hill, 2001) is zero for the simple
fluids argon, krypton, and xenon, which are generally regarded as spherical molecules. For other

ld
substances, the greater the deviation molecular sphericity, the larger is the departure of its
corresponding from zero. For example for methane it is 0.012, while for butane it is 0.2, and so
on. Experimentally determined values of Z for the three simply fluids coincide if measured at

or
identical reduced temperature and pressures. This observation forms the basis for extending the
two-parameter theorem (stated above) to the three-parameter theorem of corresponding states:
The compressibility factor for all fluids with the same value of , when compared at the same

W
reduced temperature and pressure are approximately the same, and hence the deviation from
ideal-gas behavior is nearly the same.
This theorem leads to vary convenient approach involving generalized correlations for
computing not only the volumetric properties but for estimating a wide variety of other
TU
thermodynamic properties.

Generalized Compressibility factor Approach to EOS: Pitzer Correlations

For prediction of volumetric properties (using the compressibility factor Z or the second virial
coefficient) the most commonly used correlations are those due to Pitzer and coworkers (op. cit.).
JN

According to this approach, compressibility factor is decomposed as follows:

Z Z 0 + Z 1
= ..(2.24)
Where Z 0 and Z 1 are both functions of Tr and Pr only. When = 0, as for the simple fluids, the

second term disappears. Thus the second term generally accounts for relatively small contribution
to the overall Z due to the asphericity of a molecule. As noted earlier, the value of may be
ll

computed using the following equation:

=
1.0 log( Prsat )T =0.7 ..(2.16)
A

Fig. 2.9 shows a plot of reduced vapor pressures for select substances as a function of reduced
temperature. At a value of about ~ 0.7 for the reduced temperature (T r ) of the typical simple
fluids (argon, krypton, xenon) the logarithm of reduced pressure is 1. For other molecules the
greater the departure from sphericity of the structure the lesser is the value of reduced pressure at
Tr = 0.7. This is indicative of lower volatility of the substance, which suggests relatively stronger

intermolecular interactions in the condensed phase. Stronger interactions result from higher
polarity of molecules, which is in turn originates from the asymmetry of the molecular structure.
Equation 2.16 indicates that the difference between the reduced pressures at the common reduced

ld
temperature (Tr = 0.7) is the measure for the acentric factor.
Based on the Pitzer-type correlations, Lee-Kesler (B.I. Lee and M. G. Kesler, AIChE J.,

or
vol. 21, pp. 510-527,1975) has developed generalized correlations using a variant of the BWR-
EOS (eqn. 2.17) for computing Z 0 and Z 1 as function of Tr and Pr . The values of Z 0 and Z 1 are

W
TU
JN

Fig. 2.9 Plot of P r vs. T r for select substances of varying polarity (Source: J.W. Tester and M.
Modell, Thermodynamics and its Applications, 3rd ed., Prentice Hall, 1999).

available in the form of tables from where their values may be read off after due interpolation
wherever necessary, or in the form of figures (see figs 2.10a and 2.10b). The method is presented
in the detail in Appendix1.1 of this chapter.
ll
A
 ld
or
W
(a)
TU
JN
ll

(b)

Fig. 2.10a & 2.10b Z0 and Z1 contributions to generalized corresponding states correlation
developed by Pitzer and coworkers (1955) [Source: Petroleum Refiner, April 1958, Gulf
A

Publishing Co].

--------------------------------------------------------------------------------------------------------------------
Example 2.4
A rigid 0.5-m3 vessel at 25oC and 2500kPa holds ethane; compute the number of moles of ethane
in the vessel. For ethane: T c = 305 K; P c = 48.72 bar, = 0.1
(Click for solution)
--------------------------------------------------------------------------------------------------------------------

ld
Example 2.5
Compute the saturate liquid phase molar volume for methane at 150K. For methane T c = 190.7 K,
P c = 46.41 bar, V C = 98.6cm3/mol, Z C = 0.286, = 0.011.

or
(Click for solution)
--------------------------------------------------------------------------------------------------------------------
2.4 Volumetric Properties of Mixtures

W
The EOSs discussed in the preceding sections may be applied to mixtures as well, through use of
what are called mixing rules. Such rules help re-define the fundamental parameters of each
type of EOS in terms of those corresponding to pure species and overall composition. The
principle behind these rules is that the mixture parameters are equivalent to those of a
TU
hypothetical pure species, which would display the same behaviour as the mixture. They are
essentially semi-empirical in nature, in that they contain parameters which although grounded in
molecular theory are difficult to predict fully. Nevertheless, such rules have proved reasonably
reliable for prediction of mixture properties for most practical, engineering applications.
JN

Virial EOS
For a gas mixture the virial EOS is exactly the same as for a pure species (eqn. 2.12):

BP
Z = 1+
RT

However, for the mixture the second virial coefficient B is dependent not only on temperature
ll

but also on the mixture composition. Its exact composition dependence is derivable from the
relations provided by statistical mechanics, and takes the following form:
A

Bm = y y B
i j
i j ij ..(2.26)
 Where, y i = mole fractions in a gas mixture. The indices i and j identify species, and both run
over all species present in the mixture. The virial coefficient B ij characterizes a bimolecular
interaction between molecule i and molecule j, and therefore B ij = B ji . The summation in eqn.
2.26 accounts for all possible bimolecular interactions. For a binary mixture i = 2 and j = 2; the
expansion of eqn.2.23 yields:

ld
y12 B11 + 2 y1 y2 B12 + y22 B22
B= ..(2.27)
Since they correspond to pure species, the calculation of the parameters B 11 and B 22 can be made

or
directly employing eqns. 2.13 2.15. For any cross-coefficient B ij (where i and j are not same)
pseudo-critical parameters need to be employed in general as shown below:
= (i + j ) / 2

W
ij

=Tcij (TciTcj )1/2 (1 kij )

1 8(VciVcj )1/2 / (Vci1/3 + Vcj1/3 )3
kij =
Often kij is set to zero for simplicity
Pcij = Z cij RTcij / Vcij
Z= ( Z ci + Z cj ) / 2
TU
cij

Vcij [(Vci1/3 + Vcj1/3 ) / 2]3

=
Now, Trij = T / Tcij
Bij0 0.083 (0.422 / Trij1.6 ); =
= Bij1 0.139 (0.172 / Trij4.2 )

Finally: Bij [( RTcij ) / Pcij ][ Bij0 + ij Bij1 ]

= ..(2.28)
JN

Next B m is computed by eqn. 2.26, which is used in eqn. 2.12 for computing the mixture molar
volume.
--------------------------------------------------------------------------------------------------------------------
Example 2.6

Estimate the second virial coefficient for an equimolar mixture of propane and n-pentane at 500K
ll

and 10 bar.
T c (K) Pc ( V c x 103 (m3/mol) Zc yi
A

bar)
Propane (1) 369.9 42.57 0.2 0.271 0.153 0.5
 Pentane (2) 469.8 33.75 0.311 0.269 0.269 0.5

(Click for solution)

--------------------------------------------------------------------------------------------------------------------

ld
Cubic EOS:
The parameters in all cubic EOSs are principally: a, b and . It follows that for computing the
molar volume of a mixture these parameters need to be re-defined using mixing rules. For a
binary mixture they are:

or
For a binary mixture (m):
y12 a1 + 2 y1 y2 (a1a2 )1/ 2 + y22 a2
am = ..(2.29)

W
=
bm y1b1 + y2b2 ..(2.30)
m y11 + y22
= ..(2.31)
The values of each parameters can be computed after those of the individual species are
calculated using the expressions that apply to each type of cubic EOS (see table 2.1).
TU
Generalized Correlations
Generalized correlations presented above for pure species may also be conveniently extended to
prediction of volumetric properties of gas phase mixtures as well. For the mixture, the apparent
critical properties are computed using the following set of linear relations:
JN

TC ,m = yT
i
i C ,i ; PC ,m = y P
i
i C ,i m = yii
i
..(2.32)

The subscript i runs over all the species present in the mixture m. The above properties are
designated as pseudo-critical as they do not represent the true critical properties of a mixture;
indeed the latter are most often difficult to obtain.
The pseudo-reduced temperature and pressure, are then determined by: Tr ,m = T / TC ,m ;
ll

Pr ,m = P / PC ,m . As for pure components, the compressibility factor for the mixture is next
A

obtained using standard functions of Z 0 (Tr , Pr ) and Z 1 (Tr , Pr ), which are then used in the

following equation: Z=
m Z m0 + Z m1 ..(2.33)
--------------------------------------------------------------------------------------------------------------------
Example 2.7
Calculate the molar volume of an ethylene and propylene mixture comprising 70 mole percent

ld
ethylene and 30 mole percent propylene at 600 K and 60 bar. Assume that the mixture follows
the Redlich-Kwong equation of state.

or
T c (K) P c ( bar)
Ethylene (1) 283.1 51.17
Propylene (2) 365.1 46.0

W
(Click for solution)
--------------------------------------------------------------------------------------------------------------------
2.5 Property Representation of 2-phase systems:
Consider a system consisting of liquid and vapor phases of a pure component coexisting in
TU
equilibrium. The total value of any extensive property of the two-phase system is the sum of the
total properties of the phases. Let the total volume of the system be Vt; nL = mass (or moles) of
liq, nV = mass (or moles of vapour.
Then:
V=t
= n LV L + nV V V
nV
JN

=
n n L + nV
xL = nL / n
xV = nV / n
1
x L + xV =
V= (1 xV )V L + xV V V =+
x LV L + xV V V = V L xV V LV
= V V V L
ll

V LV
Let M = Any Molar Thermodynamic Property

Thus, one may write: =

M M L + xV M LV ..(2.34)
A

Where, M LV = M V M L ..(2.35)
It may be noted that M is not a state variable, but an effective molar property of the system. The
actual phase rule state variables are MV and ML. The parameter xV the mass fraction of vapour
phase in the system if called the quality.
--------------------------------------------------------------------------------------------------------------------
Example 2.8

ld
Find the molar volume and internal energy of a system containing water and steam at 50%
quality at 2000C.
(Click for solution)

or
--------------------------------------------------------------------------------------------------------------------
2.6 Heat Effects of Phase Change
A change of phase is generally accompanied by heat release or absorption. The heat effect during

W
the change of phase from liquid to gas, for example, is called the latent heat (or enthalpy of
vapourization), while for a transition from soild to liquid the effect is termed heat of fusion. Such
enthalpy changes typically quantify the amount of heat need to change the phase of one mole of a
substance and are usually obtained experimentally. However, correlations have been developed
TU
for their prediction as well. The Reidel equation (L. Riedel, Chem. Ing. Tech., vol. 26, pp. 679-
683, 1954) provides a suitable expression for the enthalpy of vapourization at the normal boling
point (H nvap ) of a pure substance:

H nvap 1.092(ln PC 1.013)

= ..(2.36)
Tn (0.093 Trn )
JN

Where, n = normal boiling point, P C (bar).

The Watson equation may be used (K. M. Watson, Ind. Eng. Chem., vol. 35, pp. 398-406, 1943)
for computing the enthalpy of vaporization of a pure liquid at any temperature using its value at a
chosen datum temperature:
ll

0.38
H 2vap 1 Tr 2
= ;=
H ivap Heat of vapourization at Ti ..(2.37)
A

H 1 1 Tr 1
vap

--------------------------------------------------------------------------------------------------------------------
Example 9
Using Riedels correlation, estimate the enthalpy of vaporization of water at its normal boiling
point and compare the result with that given in steam tables.

(Click for solution)

--------------------------------------------------------------------------------------------------------------------

ld
Example 2.10

The enthalpy of vaporization of water at 100oC is 2256.94 kJ/kg. Determine the value at 150oC,

or
and compare the value with that listed in the steam tables.

(Click for solution)

W
--------------------------------------------------------------------------------------------------------------------

Appendix: Lee-Kesler and Pitzer Methods

Lee Kesler Method:

TU
The method used a modified Bendict-Webb-Rubin equation using Pitzers acentric factor ( ).
The compressibility factor of a real fluid is related to the properties of a simple fluid (= 0) and
those of a reference fluid; for the latter n-Octane is used... The compressibility factor Z for the
fluid is given by:

R (
ZR Z0)
JN

Z0 +
Z= (A.2.1)

Where Z = compressibility factor for the fluid
Z0 = compressibility factor for the simple fluid
ZR= compressibility factor for the reference fluid
= acentric factor for the fluid
ll

R = acentric factor for the reference fluid (n-octane) = 0.3978.

A

According to the theory of corresponding states = 0, for simple fluids: i.e. whose molecules
are considered spherical; examples include noble gases, such as argon, krypton and xenon.
Sphericity of a molecule tends to reduce its polarity, as the centre of negative and positive
charges approach each other, leading to low dipole moments. This is also true for relatively linear
but symmetric molecule. For example, carbon dioxide has virtually no dipole moment; however,
it possesses a quadrupole moment. The acentric factor, therefore, is an indirect measure of the
dipole moment of a molecule, and hence of the deviation of the intermolecular potential of real
fluids from that of simple fluids.

ld
The compressibility factors for the simple fluid (Z0) and the reference fluid (ZR) are
generated from the following equation which is a variant for the BWR EOS (eqn. 2.17):

or
PV 0 B C D C4
Z 0 =r r =+
1 + + + + exp ..(A.2.2)
Tr Vr0 (V 0 )2 (V 0 )5 T 3 (V 0 )2 (V )
0 2 (V 0 )2
r r r r r r

W
PV B C D C4
Z =
R r r
1+ R +
= + + + exp ..(A.2.3)
(VrR ) (VrR ) Tr3 (VrR ) (VrR ) (VrR )
2 5 2 2 2
Tr Vr

b2 b3 b4
Where B =b1 ..(A.2.4)
Tr Tr2 Tr3
c2 c3
TU
C =c1 + ..(A.2.5)
Tr Tr3
d2
D= d1 + ..(A.2.6)
Tr
Where P r and T r are the reduced pressure and temperature, respectively, at which the volume of
the real fluid is required.
JN

Further, Vr0 = PV
c
0
/ RTc ..(A.2.7)
Where, V0 = molar volume of the simple fluid (at the given T & P). The Lee-Kesler constants for
the simple fluid and the reference fluid to be used with Eqns. 2-6 are enumerated in Table A.1.1
below.
Table A.1.1 Lee-Kesler Constants for use with Eqns.A.2.2-A.2.6
ll

Constant Simple Reference Constant Simple Reference

fluid fluid fluid fluid
A

b1 0.1181193 0.2026579 c3 0.0 0.016901

b2 0.265728 0.331511 C 0.042724 0.041577
 b3 0.154790 0.027655 d 1 x 104 0.155488 048736
b4 0.030323 0.203488 d 2 x 104 0.623689 0.0740336
c1 0.0236744 0.0313385 0.65392 1.226

c2 0.0186984 0.0503618 0.060167 0.03754

ld
The method of calculating the volume at a given T r and P r is as follows:
Step 1: using the critical properties of the real fluid find Vr0 by eqn. A.2.7

or
Step 2: using Vr0 , calculate Z0 by eqn. A.2.1

Step 3: using the given values of T r and P r , Eqn.A.2.2 is solved again, but with reference
fluid constants from above table. The solution obtained for Vr0 is now called VrR

W
Step 4: ZR, the compressibility for the reference fluid, is next computed with: given by

r r / Tr
Z R = PV R
(A.2.8)

Step 5: With the above values of Z0 and ZR, the compressibility factor Z for the fluid in
question can be calculated from Eqn.A.2.1.
TU
The Generalized Lee-Kesler Charts essentially present the values of Z0 and Z1 obtained by the
above procedure as function of T r and P r (as). A variety of other thermodynamic variables can
also be computed in an identical manner. Examples include residual enthalpy and entropy, and
gas fugacity. These properties are presented either in the form of figures or charts (see, for
example, Appendix E of J.M. Smith, H.C. Van Ness and M.M. Abbott, Introduction to Chemical
JN

Engineering Thermodynamics, 6th ed., McGraw-Hill, 2001).

Pitzer Method:
The Pitzer method is nearly identical to that of Lee-Kesler; it assumes that the compressibility
factor is linearly dependent on the acentric factor. Thus, eqn. A.2.1 is reformulated using the
ll

compressibility factors of both the simple (1) and reference fluid (2), whence:
R1
Z R1 + R 2 ( Z R 2 Z R1 ) ..(A.2.9)
A

Z=
R1
Any two fluids may be used as the reference fluids. The method of computing the values of
Z R 2 and Z R1 and hence, Z for the fluid of interest follows the same procedure described above
for Lee-Kesler.
--------------------------------------------------------------------------------------------------------------------
Assignment-Chapter 2

ld
or
W
TU
JN
ll
A
CHAPTER 3: First Law of Thermodynamics
3.1 Concept of Internal Energy
We have noted in chapter 1 that the two most common modes of energy exchanged by a
thermodynamic system and its surroundings are work and heat. The interconvertibility between

ld
these two forms of energy was first demonstrated by the British scientist James P. Joule during
1840s by a series of carefully executed experiments. The experimental setup he used is shown as
a schematic in fig. 3.1. Known quantities of a set of fluids (water, oil, and mercury) were placed

or
in an

W
TU
Fig. 3.1 Schematic of Joules Experimental System
insulated, rigid vessel and stirred by means of rotating shaft provided with vanes. The amounts of
work done on the fluid by the stirrer were measured in terms that needed to lower or raise a
weight, and the resultant change in the temperature of the fluid was recorded. The key
JN

observation made by Joule was that for each fluid a fixed amount of work was required per unit
mass for every degree of temperature rise caused by the rotating paddle wheel. Further, the
experiments showed that the temperature of the fluid could be restored to its initial value by the
transfer of heat by bringing it in contact with a cooler object. These experimental findings
demonstrated for the first time that inter-convertibility exists between work and heat, and
ll

therefore the latter was also a form of energy.

Joules observation also provided the basis for postulation of the concept of internal
energy (introduced briefly in section 1.3). Since work are heat are distinctly different forms of
A

energy how is it possible to convert one into another? The question can be answered if one
assumes that although these two types of energies are distinct in transit across a thermodynamic
system boundary, they must eventually be stored within a thermodynamic system in a common
1|Page
form. That common form is the so-called internal energy. As we have already discussed in
section 1.3, such a form of energy can only repose at the microscopic level of atoms and
molecules, essentially in the form of translational, vibrational and rotational energies. To this
may be added the potential energy of intermolecular interactions (as introduced in section 2.2).
On a sub-molecular scale energy is associated with the electrons and nuclei of atoms, and with

ld
bond energy resulting from the forces holding atoms together as molecules. With these
considerations one is in a position to rationalize the observation that while a system may receive
energy in the form of work done on it, it may part with it also in the form of heat to another body

or
or surroundings and be restored to its state prior to receipt of work. This is possible as in the
interim between these two processes all energy may be stored in the form of internal energy.
As may be evident from the foregoing discussion, the addition of heat or work from an

W
external source can lead to enhancement of the microscopic form of systemic (internal) energy.
As also noted in chapter 1 the terminology internal is applied mainly to distinguish it from the
mechanical potential and kinetic energies that a thermodynamic system may also possess by
virtue of its position and velocity with respect to a datum. The latter two may then be thought of
TU
as external forms of energy.
It is important to note that like other intensive, macroscopic variables such as pressure,
temperature, mass or volume, internal energy is a state variable as it is wholly dependent on the
energy states that its atoms / molecules. Thus any change in the (say, specific) internal energy
due to a process would only depend on the initial and final states, and not on the path followed
JN

during the change. Thus as for changes in P, V or T, one may write:

U2
U1
dU =U =
U 2 U1

However, unlike P, V, T or mass, U is not a directly measurable property. Besides, in common

with potential and kinetic energies, no absolute values of internal energy are possible. However,
this is not of particular significance as in thermodynamic processes one is always interested in
ll

changes in energies rather than their absolute values

A

2|Page
3.2 The First Law of Thermodynamics
The empirical conclusion that heat and internal energy belong to the general category of energies,
help extending the law of conservation of mechanical energy, which states that potential and
kinetic energies are fully inter-convertible. As already discussed in chapter 1, a thermodynamic
system may possess any other forms of energy such as surface energy, electrical energy, and

ld
magnetic energy, etc. Thus one may arrive at an extended postulate that all forms are energies are
inter-convertible. This constitutes the basis of the First Law of Thermodynamics, which may be
stated as follows:

or
Energy can neither be destroyed nor created, when it disappears in one form it must re-appear at
the same time in other forms.

W
It must be said that there is no formal proof of the first law (or indeed of other laws of
thermodynamics) is possible, but that no evidence have been found to date that violates the
principle enunciated by it.
TU
For any thermodynamic process, in general one needs to account for changes occurring
both within a system as well as its surroundings. Since the two together forms the universe in
thermodynamic terms, the application of the first law to a process leads to the following
mathematical form:
(Total energy of the universe) = 0
JN

(Total energy of the system) + (Total energy of the surroundings ) = 0

Where finite change occurring during the thermodynamic process

3.2 Application of the First Law to Closed Systems

In general, a thermodynamic system in its most complex form may be multi-component as well
ll

as multiphase in nature, and may contain species which react chemically with each other.
Thermodynamic analysis tends to focus dominantly on the energy changes occurring within such
A

a thermodynamic system due to change of state (or vice versa), and therefore it is often
convenient to formulate the first law specifically for the system in question. Here we focus on
closed systems, i.e., one that does not allow transfer of mass across its boundary. As already

3|Page
pointed out work and heat may enter or leave such a system across its boundary (to and fro with
respect to the surrounding) and also be stored in the common form of internal energy. Since in a
system may also possess potential and kinetic energies, one may reframe the first law as follows.
Using the notations U , EK , EP for specific internal, kinetic and potential energies, respectively:

[U total + Ekinetic potential ]system = Total energy input to system in all forms
+ E total

ld
total

If the energy transfer across the system boundary takes place only the form of work and heat:
[U total + Ekinetic
total
potential ]system
+ E=
total
Q total (heat input ) + W total ( work input )

or
The above relation may be written per unit mass / mole of closed system, i.e.,:
U + EK + EP= Q + W
The above equation may also be written in a differential form:

W
dU + dEK + dEP = Q + W ..(3.1)

If there is no change in potential and kinetic energies for the system or it is negligible as is
usually true for most thermodynamic systems of practical interest the above equation reduces
to:
TU
= Q + W
dU ..(3.2)
One of the great strengths of the mathematical statement of the first law as codified by eqn. 3.2 is
that it equates a state variable (U) with two path variables (Q, W). As a differentiator we use the
symbol to indicate infinitesimal work and heat transfer (as opposed to d used state variables).
The last equation potentially allows the calculation of work and heat energies required for a
JN

process, by simply computing the change in internal energy. As we shall see later (chapters. 4 &
5) changes in internal energy can be conveniently expressed as functions of changes in state
properties such as T, P and V.
In the above equation the term W represents any form of work transfer to or from the
system. In many situations of practical interest the thermodynamic work for closed systems is
ll

typically the PdV work (eqn. 1.6). Hence in such cases one may reframe eqn. 3.2 as follows:
V2

dU = Q + W = Q PdV ..(3.3)
A

V1

In keeping with the definition of work above, we adopt the following convention:

4|Page
W > 0, if work is done on system
W < 0, if work is done by system
Q > 0, for heat addition to system
Q < 0, for heat removal from system

The process of change in a thermodynamic system may occur under various types of constraints,

ld
which are enlisted below:
Constant pressure (isobaric)
Constant volume (isochoric)

or
Constant temperature (isothermal)
Without heat transfer (adiabatic)
The mathematical treatment of each of these processes is presented below.

W
For a constant pressure process (fig. 3.2), we may write:
=Q dU + PdV
=Q d + Ud ( PV )
=Q d (U + PV )
or Q = dH
TU
where, H= U + PV
The term H is termed enthalpy. It follows that like U, H is also a state variable. On integrating the
differential form of the equation above one obtains for the process:
2 2

1
Q = dH
1
JN

Or: Q = H ..(3.4)
ll
A

Fig. 3.2 Schematic of an Isobaric Path

5|Page
On the other hand if the process occurs under isochoric (const. V) conditions (shown in fig. 3.3)
the first law leads to:

ld
or
W
Fig. 3.3 Schematic of an Isochoric Path
Q = dU
2 2
1
Q = dH
1

Q = U ..(3.5)
TU
We have already seen that a body can retain heat in the form of internal energy. This gives rise to
the concept of heat capacity C and is mathematically defined as:
Q
C
dT
It follows that using eqns. 3.4 and 3.5 two types of heat capacity may be defined:
JN

Constant pressure heat capacity CP such that:

Q H
=
C = ..(3.6)
dT T P
P

Constant volume heat capacity CV such that:

ll

Q U
=
C = ..(3.7)
dT T V
V
A

Thus, using eqns. (3.6) and (3.7) one may rewrite eqns. (3.4) and (3.5) as follows:
T2
Q=H = CP dT T1
(At const P.) ..(3.8)

6|Page
 T2
Q=U = CV dT
T1
(At const V.) ..(3.9)

With the addition of heat to a system, the translation, vibrational and rotational (as well as
subatomic) energies of the molecules are enhanced and so it may be expected that the specific
heats would be dependent on temperature. On the other hand when substances are compressed

ld
intermolecular interactions begin to contribute to internal energy (as hence to enthalpy) and
therefore specific heats are rendered pressure dependent. For the case of ideal gases, however, the
specific heats are independent of pressure as there are no intermolecular interactions; they are

or
only temperature dependent. Values of specific heats of ideal gases (ig), say at constant pressure,
are available from experimental measurements and are typically expressed in the form of
polynomials such as:

W
CPig D
=A + BT + CT 2 + 2
R T ..(3.10)
Where, A, B, C and D are characteristic constants for a substance, and R is the universal gas
constant. Values of the constants in eqn. (3.10) are readily available for a large number of pure
TU
substances (see Appendix III for values of select gases). Fig. 3.4 shows typical dependence of
const, pressure specific heat for select substances with temperature.
JN
ll

Fig. 3.4 Variation of C P ig/R vs. Temperature (Source: J. M Smith, H.C. Van Ness, M.M. Abbott,
A

Introduction to Chemical Engineering Thermodynamics, 6th ed., McGraw-Hill,

2001)

7|Page
 Values of the coefficients of similar specific heat capacity polynomials for liquids and solids are
available elsewhere (see for example, J.M. Smith, H.C. Van Ness and M.M. Abbott, Introduction
to Chemical Engineering Thermodynamics, 6th ed., McGraw-Hill, 2001).

Applications to Ideal Gases

ld
The other two types of thermodynamic processes isothermal and adiabatic in closed systems
are conveniently understood by applying the first law to a system comprised of an ideal gas. For
such a case the relationship between U and H may be rewritten using the EOS:

or
H=U + PV =U + RT
Or: H U =
RT
Since both H and U are only temperature dependent for ideal gases we write:

W
dH dU
=
R
dT dT
Now using eqns. 3.6 and 3.7 it follows that:
CP CV =
R ..(3.11)
TU
(It may be noted that since the molar volume V is relatively small for both liquids and solids, one
may write: H U ; hence: CP CV ).

Let us now consider the relationships that obtain for an isothermal process for an ideal gas. Using
eqn. 3.2, since dU= 0:
Q = W
JN

Or, Q = W
If we consider only P-V work, the work term is calculable if the process is carried out reversibly,
as the ideal gas EOS relate the P and V at all points of change, hence:
V2 V2 RT V P
PdV =
W= RT ln 1 =
dV = RT ln 2 ..(3.12)
V1 V1
V V2 P1
ll

Thus:
A

V P
= =
Q RT ln 2 RT ln 1
V1 P2 ..(3.13)
--------------------------------------------------------------------------------------------------------------------
8|Page
Example 3.1
A gas initially at 1 MPa, 500C is contained in a piston-cylinder arrangement of initial volume of
0.1 m3. The gas expanded isothermally to a final pressure of 100 kPa. Determine the work.
(Click for solution)
--------------------------------------------------------------------------------------------------------------------

ld
V2
For isobaric process:Q =H = CP dT ; W = PdV =P(V2 V1 ) =R(T2 T1 )
V1 ..(3.14)

For adiabatic process dQ = 0

or
Hence : dU = W
RT
Or :CV dT =
PdV =
dV
V

W
dT R dV
=
T Cv V ..(3.15)

The ratio of heat capacities is defined as:

= CP / CV
TU
Or:
CV + R R
= = 1+
CV CV ..(3.16)
On integrating eqn. 3.15 and using the relationship provided by eqn. 3.16, the following set of
results may be derived easily:
1
JN

T2 V1
=
T1 V2
..(3.17)
( 1)/
T2 P2
=
T1 P1
..(3.18)

and PV
=1 1 =
PV
2 2 = const.
PV ..(3.19)
ll

Further: dW = CV dT
= dU

Or = = CV dT
W dU
A

Since, W = PdV

9|Page
 Using eqn. 3.19 in the expression for work in the last equation:

2 2 PV
PV
W= 1 1
1 ..(3.20)
Or:

ld
PV P ( 1)/ RT P ( 1)/
=W 1 1
=2
1 1
2 1 ..(3.21)
1 P1 1 P1

or
All the above processes discussed can be captured in the form of a single P-V relation, which is
termed a polytropic equation as it can be reduced to yield all forms of processes. The polytropic
relations are written by generalizing eqns. 3.17 3.19, as follows:

W
PV = constant ..(3.22)
TV 1 = constant ..(3.23)
TP (1 )/ = constant ..(3.24)

The schematic of polytropic process is shown in fig. 3.5.

TU
JN

Fig. 3.5 Schematic of Polytropic Processes

ll

As may be seen the various values of reproduce the isothermal, isochoric, isobaric and adiabatic
A

processes. In line with eqn. 3.21, for such the generalized expressions for work and heat transfer
may be shown to be:

10 | P a g e
 ( 1)/
RT1 P2
=W 1
1 P1
..(3.25)
Further:
( 1)/
( ) RT1 P2
1
( 1)( 1) P1
Q

ld

..(3.26)

--------------------------------------------------------------------------------------------------------------------

or
Example 3.2
Helium gas expands from 125 kPa, 350 K and 0.25 m3 to 100 kPa in a polytropic process with
= 1.667. How much work does it give out?

W
(Click for solution)
--------------------------------------------------------------------------------------------------------------------

3.3 Application of the First Law to Open Systems

While the last section addressed processes occurring in closed systems, the wider application of
the first law involves formulating the energy balance differently in order to accommodate the fact
TU
that most thermodynamic systems, i.e., equipments, in continuous process plants are essentially
open systems: they allow mass transfer across their boundaries (i.e., through inlet and outlet).
Examples include pumps, compressors, reactors, distillation columns, heat exchangers etc. Since
such open systems admit both material and energy transfer across their boundaries the
thermodynamic analysis necessarily involves both mass and energy balances to be carried out
JN

together. Also such systems may in general operate under both steady (during normal plant
operation) and unsteady states (say during startup and shutdown). As we will see the former state
is a limiting case of the more general situation of unsteady state behavior.
ll

Mass Balance for Open Systems:

For generality consider an open system with which has multiple inlets (1, 2) and outlets (3, 4).
The volume enclosed by the physical boundary is the control volume (CV). The general mass
A

balance equation for such a system may be written as:

11 | P a g e
 ld
or
Fig. 3.6 Schematic of an open system

dmcv
m m

W
= ..(3.27)
dt inlet outlet

The mass flow rate m is given by:

m = uA ..(3.28)
Where, = fluid density; u = fluid velocity; and A = aperture cross-sectional area
TU
The first term on the left side of the eqn. (3.27) denotes sum of all flow rates over all inlets, while
the second term corresponds to the summation over all outlet flow rates. For the system shown in
fig. 3.5 the eqn. 3.27 may be written as:
dmcv
= m1 + m 2 m3 m 4 ..(3.28)
dt
JN

The last equation may be reframed in a general way as follows:

dmcv
+ (m) fs = 0 ..(3.29)
dt
Where the symbol oulet-inlet; and the subscript fs stands for "flow streams".
Equation 3.29 may be recast as:
ll

dmcv
+ ( uA) fs = 0 ..(3.30)
dt
A

The above equation simplifies under steady flow conditions as the accumulation term for the
dmcv
control volume, i.e., =0
dt

12 | P a g e
In such a case for the simplest case of a system with one inlet and outlet (say: 1 and 2
respectively), which typically represents the majority of process plant equipments, the mass
balance equation reduces to:

m = constant ..(3.31)
Or: 1u1 A1 = 2u2 A2 u1 A1 / V1 =u2 A2 / V2

ld
Where V = specific volume

or
Energy Balance for Open Systems:
Consider the schematic of an open system as shown in fig. 3.7. For simplicity we assume one

W
TU
Fig. 3.7 Schematic of an open system showing flow and energy interactions

inlet and one exit ports to the control volume. The thermodynamic states at the inlet i and exit e
are defined by the P, V, T, u (average fluid velocity across the cross section of the port), and Z,
JN

the height of the port above a datum plane. A fluid element (consider an unit mole or mass) enters
the CV carrying internal energy, kinetic and potential energies at the inlet conditions (P i , T i , with
molar volume as V i ) and leaves values of these energies at the exit state conditions (P e , T e , with
molar volume as V e ). Thus the total specific energy of the fluid at the two ports corresponds to
u2
the sum of specific internal, potential and kinetic energies, given by: U + + gZ . In addition,
ll

2

the CV exchanges heat with the surroundings at the rate Q , and say a total work (in one or more
A

forms) at the rate of W total

13 | P a g e
 In the schematic we, however, have shown a specific work form, shaft work, that is
delivered to or by the system by means of rotatory motion of a paddle wheel which, as we will
see later in the section, is implicated in many typical process plant units. As with material
balance one may write a total energy balance equation for the control volume as follows:
d (mU )cv 1 1

ld
= U + (u 2 ) + zg ) m U + (u 2 ) + zg ) m + Q + W total
dt 2 i 2 e

Or:

or
d (mU )cv 1
+ U + (u 2 ) + zg ) m = Q + W total
dt 2 fs
..(3.32)
The general total work term should include all forms of work. We draw the readers attention to

W
the fact that the total work interaction also should include that needed to push fluid into the CV as
well as that implicated in pushing it out of CV. The fluid state at the inlet or exit is characterized
by a set of state properties, U, V, H, etc. Consider a unit mass (or mole) of fluid entering the CV.
This fluid element obviously needs to be pushed by another that follows it so as to make the
formed enter the CV. In essence a fluid element of (specific) volume V is pushed into the CV at a
TU
pressure P. This is akin to a P-V form of work (as in the case of a piston-in-a-cylinder system)
that is done on the CV and so may be quantified as P i V i . The same considerations apply at the
exit in which case in pushing out a similar fluid element at exit conditions, i.e., P e V e . Thus,
eqn. 3.32 may be rewritten as follows:
d (mU )cv 1
JN

+ U + (u 2 ) + zg ) m = Q + W [ PV m]i + [ PV m]e ..(3.33)
dt 2 fs

W [ PV m]i + [ PV m]e
Where, Wtotal =

The term W represents sum of all other forms of work associated with the process occurring
within the CV. This residual work term may include the shaft work, P-V work resulting from
ll

expansion or contraction of the CV, electrical work, and so on. As the last two work terms on the
left side of the eqn. 3.33 are associated with the flow streams we may rewrite the equation as
A

follows:

14 | P a g e
d (mU )cv 1
+ U + (u 2 ) + zg ) m = Q + W [ PV m] fs ..(3.34)
dt 2 fs

On rearranging:

d (mU )cv u2
+ (U + PV ) + + zg m = Q + W ..(3.35)
2

ld
dt fs

d (mU )cv u2
Or: + H + + zg m = Q + W

or
..(3.36)
dt 2 fs

It may be noted that eqn. 3.34 assumes that the CV is fixed in space and therefore no overall
potential of kinetic energy terms depicting these mechanical energies for the control volume is

W
included. This, of course, is valid for all process plant applications. In addition, For many cases
of practical importance (though not all) in a chemical plant the kinetic and potential energy
changes between the inlet and exit streams may not be significant, whence the last equation may
be simplified as: an equipment may be neglected; hence:
TU
d (mU )cv
+ H m = Q+ W
dt fs
..(3.37)
Further, for the special case where the only shaft work is involved, the above equation may be
simplified to:
d (mU )cv
+ H m = Q+ W s
JN

dt fs
..(3.38)
For steady state applications the eqn. 3.34 reduces to:
u2
H + + zg m =Q + W
2 fs ..(3.39)
ll

Further, if the kinetic and potential energy changes associated with the flow streams are

insignificant, it follows that:

A

H m = Q+ W ..(3.40)
fs

Since under steady state m is constant we may write:
15 | P a g e
H = Q + W ..(3.41)
Where Q and W denote the work and heat interactions per unit mass of mole of fluid flowing
through the system flowing through system. Once again if W s is the only form of work
interaction between the system and the surrounding then:
H = Q + W s ..(3.42)

ld
Examples of process plant units to which eqn. 3.42 applies are: pumps, compressors, turbines,
fans, blowers, etc. In all cases a rotatory part is used exchange work between the system and
surrounding.

or
--------------------------------------------------------------------------------------------------------------------
Example 3.3
A chiller cools liquid water (Sp. Ht = 4.2 J/gmK) for air-conditioning purposes. Assume 2.5 kg/s

W
water at 20oC and 100 kPa is cooled to 5oC in a chiller. How much heat transfer (kW) is needed?
(Click for solution)
--------------------------------------------------------------------------------------------------------------------
Example 3.4
TU
A piston-cylinder assembly contains 0.1 kg wet steam of quality 0.75 at 100 kPa. If 150 kJ of
energy is added as heat while the pressure of the steam is held constant determine the final state
of steam.
(Click for solution)
--------------------------------------------------------------------------------------------------------------------
JN

Example 3.5
An adiabatic compressor operating under steady-state conditions receives air (ideal gas) at 0.1
MPa and 300 K and discharges at 1 MPa. If the flow rate of air through the compressor is 2
mol/s, determine the power consumption of the compressor. Constant pressure specific heat for
air = 1kJ/kg.
ll

(Click for solution)

--------------------------------------------------------------------------------------------------------------------
A

Example 3.6
An insulated piston-cylinder system has air at 400kPa & 600K. Through an inlet pipe to the
cylinder air at certain temperature T(K) and pressure P (kPa) is supplied reversibly into the

16 | P a g e
cylinder till the volume of the air in the cylinder is 4 times the initial volume. The expansion
occurs isobarically at 400kPa. At the end of the process the air temperature inside the cylinder is
450K. Assume ideal gas behaviour compute the temperature of the air supplied through the inlet
pipe.
(Click for solution)

ld
--------------------------------------------------------------------------------------------------------------------
3.4 Measurement of Enthalpy and Internal Energy using Flow Calorimeter
The use of the first law for open or closed systems necessitates the experimental determination

or
(or, estimation from thermodynamic relations) of internal energy and enthalpy, both being state
properties. The flow calorimeter (fig. 3.8), readily allows measurement of enthalpy, which in
turn can be used to compute the internal energy at the same conditions of (say) temperature and

W
pressure. The fluid whose properties are to be measured is pumped through a constant
temperature bath so that it attains a desired temperature (say T 1 ) prior to entry into the vessel
provided with an electric heater. The corresponding pressure (P 1 ) may also be recorded. Heat is
next provided to the passing fluid at a pre-determined and fixed rate over a period of time until
TU
the temperature and pressure at the exit of the vessel attains steady values (say, T 2 , & P 2 ). At
such a condition the calorimeter is under steady state, hence eqn. 3.41 may be applied to the
heating vessel. Since there is no work transfer of any kind, the resultant energy balance yields:
H = Q ..(3.40)
Or: H 2 H 1 = Q
JN

So: H 2 = H 1 + Q ..(3.41)
ll
A

17 | P a g e
 Fig. 3.8 A schematic of a flow calorimeter

Clearly, if we chose the enthalpy H 1 to be a datum state and arbitrarily assign it a zero value,
then:

ld
H2 = Q
..(3.42)
The last equation then allows one to uniquely determine the value of H 2 at any condition

or
achieved at state 2 by applying a known quantity of heat through the electric heating system. The
internal energy at the same state can next be determined using the relation:
U = H PV ..(3.42)

W
Further we can measure the density (in terms of mass or mole) at the same state, and one may

rewrite the last equation as: U = H P/ ..(3.43)

Thus using the above relations, U and H may be obtained experimentally at any P, T (and/or V)
for any substance. The steam tables discussed at the concluding portion of section 2.1 constitutes
TU
such a tabulation of values of standard thermodynamic properties (saturation vapour pressure,
internal energy, enthalpy (and entropy) of water obtained empirically over a wide range
temperature and pressure.
---------------------------------------------------------------------------------------------------------------------
Assignment-Chapter 3
JN
ll
A

18 | P a g e
Chapter 4: Second Law of Thermodynamics
4.1 Heat Engines and Second Law Statements
The First Law provides a constraint on the total energy contained in a system and its
surroundings. If it disappears in one form from the system during any thermodynamic process of
change, it must reappear in another form either within the system or in the surroundings.

ld
However, a pertinent question that one may often need to answer is: Is the process of change
aimed at feasible? As may be evident, the first law provides no constraint on the possible
direction a process may take place. Nevertheless, in the real world such constraints do exist. For

or
example, heat always flows from a high temperature body to one at a lower temperature.
Momentum flow is always prompted in the direction of a pressure gradient, and molecules
always migrate from a region of higher to lower chemical potential. These observations clearly

W
are indicative of the existence of a constraint on natural processes, which have never been found
to be violated.
Further, it is common observation that work is readily transformed into other forms of
energy, including heat. But all efforts to develop a device that may work in a continuous manner
TU
and convert heat completely into work or any other form of energy have proved impossible.
Experimental observations show that typically no more than 40-50% of the total heat available
may be converted to work or other energy forms. This finding has led to the conclusion that heat
is a lower form of energy in that while it may be feasible to degrade work to heat, it is
impossible to upgrade heat completely into work.
JN

Heat may be seen as a more primitive form of energy, as it always has to be made
available from matter (say by combustion) and subsequently converted to work for carrying out
activities useful to humans. In this sense one never derives work directly from the energy locked
in matter as enthalpy. This prompts the natural question: what determines the efficiency of such a
conversion of heat to work? Evidently one needs a limiting principle that may help answer this
ll

question. These considerations provide the basis for formulating the Second Law of
Thermodynamics.
A

1|Page
2.1 Second Law Statements

It need be mentioned that the second law is a product of experiential observations involving heat
engines that powered the Industrial Revolution of the 19th century. A heat engine is a machine
that produces work from heat through a cyclic process. An example is a steam power plant in
which the working fluid (steam) periodically goes through a series of steps in a cyclic manner as

ld
follows:
Step 1: Liquid water at ambient temperature is pumped into a boiler operating at high

or
pressure
Step 2: Heat released by burning a fossil fuel is transferred in the boiler to the water,
converting it to steam at high-temperature and pressure

W
Step 3: The energy contained in the steam is then transferred as shaft work to a turbine;
during this process steam temperature and pressure are reduced.
Step 4: Steam exiting the turbine is converted to water by cooling it and transferring the
heat released to the surroundings. The water is then returned to step 1.
TU
Like the steam power plant all heat engines absorb heat at a higher temperature body (source)
and release a fraction of it to a low temperature body (sink), the difference between the two
quantities constitutes the net work delivered during the cycle. The schematic of a heat engine (for
example: steam / gas power plant, automotive engines, etc) is shown in fig. 4.1. As in the case of
the steam cycle, a series of heat and work exchanges takes place, in each case a specific hot
JN

source and a cold sink are implicated. A schematic of such processes is suggested inside the
ll
A

Fig. 4.1 Schematic of Heat Engine

2|Page
yellow circle between the hot and cold sources. The opposite of a heat engine is called the heat
pump (refrigerators being an example of such device) is shown in fig. 4.2. There are indeed a
large number other types of practical heat engines and power cycles. Select examples include:
Ericsson Cycle, Stirling cycle, Otto cycle (e.g. Gasoline/Petrol engine, high-speed diesel engine),
Diesel cycle (e.g. low-speed diesel engine), etc. The Rankine cycle most closely reproduces the

ld
functioning of heat engines that use steam as the process fluid function (fig. 4.3); such heat
engines are most commonly found in power generation plants. In such plants typically heat is
derived from nuclear fission or the combustion of fossil fuels such as coal, natural gas, and oil.

or
Detailed thermodynamic analysis of the various heat engine cycles may be found in a number of
textbooks (for example: J.W. Tester and M. Modell, Thermodynamics and its Applications, 3rd
ed., Prentice Hall, 1999).

W
TU
Fig. 4.2 Comparison of Heat Engine and Heat Pump
JN
ll
A

3|Page
 ld
or
W
Fig. 4.3 Schematic of a Power Plant (Rankine) Cycle
As evident, the operation of practical heat engines requires two bodies at constant differential
temperature levels. These bodies are termed heat reservoirs; they essentially are bodies with
theoretically speaking infinite thermal mass (i.e., mCP ) which therefore do not undergo a
TU
change of temperature due to either release or absorption of heat. The above considerations may
be converted to a set of statements that are equivalent descriptors of the second law (R. Balzheiser,
M. Samuels, and J. Eliassen, Chemical Engineering Thermodynamics, Prentice Hall, 1972):

Kelvin-Planck Statement: It is impossible to devise a cyclically operating device, the sole

JN

effect of which is to absorb energy in the form of heat from a single thermal reservoir and to
deliver an equivalent amount of work.

Clausius Statement: It is impossible to devise a cyclically operating device, the sole effect of
which is to transfer energy in the form of heat from a low temperature body to a high temperature
body.
ll

2.2 Carnot Heat Engine Cycle and the 2nd Law

A

In theory we may say that a heat engine absorbs a quantity of heat QH from a high temperature
reservoir at T H and rejects QC amount of heat to a colder reservoir at T H . It follows that the net
work W delivered by the engine is given by:

4|Page
=
W QH QC ..(4.1)

Hence the efficiency of the engine is:

W QH QC

= = ..(4.2)
QH QH

ld
QC
= 1 ..(4.3)
QH

or
Of the various forms of heat engines ideated, the Carnot engine proposed in 1824 by the French
engineer Nicholas Leonard Sadi Carnot (1796-1832), provides a fundamental reference concept
in the development of the second law. The so-called Carnot cycle (depicted in fig. 4.4) is a series

W
of reversible steps executed as follows:
Step 1: A system at the temperature of a cold reservoir T C undergoes a reversible
adiabatic compression which raises it temperature to that of a hot reservoir at T H .
Step 2: While in contact with the hot reservoir the system absorbs QH amount of heat
TU
through an isothermal process during which its temperature remains at T H .
Step 3: The system next undergoes a reversible adiabatic process in a direction reverse of
step 1 during which its temperature drops back to T C .
Step 4: A reversible isothermal process of expansion at T C transfers QC amount of heat

to the cold reservoir and the system state returns to that at the commencement of step 1.
JN
ll
A

Fig. 4.4 Carnot Cycle Processes

5|Page
The Carnot engine, therefore, operates between two heat reservoirs in such a way that all heat
exchanges with heat reservoirs occur under isothermal conditions for the system and at the
temperatures corresponding to those of the reservoirs. This implies that the heat transfer occurs
under infinitesimal temperature gradients across the system boundary, and hence these processes
are reversible (see last paragraph of section 1.9). If in addition the isothermal and adiabatic

ld
processes are also carried out under mechanically reversible (quasi-static) conditions the cycle
operates in a fully reversible manner. It follows that any other heat engine operating on a
different cycle (between two heat reservoirs) must necessarily transfer heat across finite

or
temperature differences and therefore cannot be thermally reversible. As we have argued in
section that irreversibility also derives from the existence of dissipative forces in nature, which
essentially leads to waste of useful energy in the conversion of work to heat. It follows therefore

W
the Carnot cycle (which also comprises mechanically reversible processes) offers the maximum
efficiency possible as defined by eqn. 4.3. This conclusion may also be proved more formally
(see K. Denbigh, Principles of Chemical Equilibrium, 4th ed., Cambridge University Press,
1981).
TU
We next derive an expression of Carnot cycle efficiency in terms of macroscopic state
properties. Consider that for the Carnot cycle shown in fig. 4.4 the process fluid in the engine is
an ideal gas. Applying the eqn. 3.13 the heat interactions during the isothermal process may be
shown to be:
Vx
QH = RTH ln ..(4.4)
JN

Vw

Vy
And QC = RTC ln ..(4.5)
Vz
Further for the adiabatic paths xy and zw using eqn. 3.17 one may easily derive the following
equality:
ll

Vx Vy
ln = ln ..(4.6)
Vw Vz
A

Using eqns. 4.4 4.5 we may write:

6|Page
 Vx
TH ln
QH Vw
= ..(4.7)
QC Vy
TC ln
Vz
Finally applying eqn. (4.6) on obtains:

ld
QH TH
= ..(4.8)
QC TC

TC
Hence, using eqn. 4.3: = 1 ..(4.9)

or
TH

Also, =
W QH QC ..(4.10)

Eqn. (4.8) may be recast as:

W
QH Q
= C ..(4.10)
TH TC
As heat Q H enters the system it is positive, while Q C leaves the system, which makes it negative
in value. Thus, removing the modulus use, eqn. 4.10 may be written as:
TU
QH Q
= C ..(4.11)
TH TC
QH QC
Or: + 0
= ..(4.12)
TH TC
JN

Q
Or: =0 ..(4.13)
T
Consider now eqn. 4.9. For the Carnot efficiency to approach unity (i.e., 100%) the following
conditions are needed: TH ; or TC 0. Obviously neither situation are practicable, which

suggests that the efficiency must always be less than unity. In practice, the naturally occurring
ll

bodies that approximate a cold reservoir are: atmospheres, rivers, oceans, etc, for which a
representative temperature T C is ~ 3000K. The hot reservoirs, on the other hand are typically
A

furnaces for which T H ~ 6000K. Thus the Carnot efficiency is ~ 0.5. However, in practice, due to
mechanical irreversibilities associated with real processes heat engine efficiencies never exceed
40%.

7|Page
--------------------------------------------------------------------------------------------------------------------
Example 4.1
An inventor claims to have devised a cyclic engine which exchanges heat with reservoirs at 27C
and 327C, and which produces 0.6 kJ of work for each kJ of heat extracted from the hot
reservoir. Is the claim believable? If instead he claimed that the delivered work would be 0.25kJ /

ld
kJ of extracted heat, would the engine be feasible?
(Click for solution)
-------------------------------------------------------------------------------------------------------------------

or
It is interesting to note that the final step of thermodynamic analysis of Carnot cycle (i.e., eqn.
Q

W
4.13) leads to the conclusion that there exists a quantity which add up to zero for the complete
T
cycle. Let us explore extending the idea to any general reversible cycle (as illustrated in fig.4.5)
run by any working fluid in the heat engine.
TU
JN

.
Fig. 4.5 Illustration of an arbitrary cycle decomposed into a series of small Carnot cycles
The complete cycle may, in principle, be divided into a number of Carnot cycles (shown by
dotted cycles) in series. Each such Carnot cycle would be situated between two heat reservoirs. In
ll

the limit that each cycle becomes infinitesimal in nature and so the number of such cycles
infinity, the original, finite cycle is reproduced. Thus for each infinitesimal cycle the heat
A

absorbed and released in a reversible manner by the system fluid may be written as QH and

QC respectively. Thus, now invoking eqn. 4.12 for each cycle:

8|Page
 QH QC
+ 0
= ..(4.14)
TH TC
Hence, applying eqn. 4.13 to sum up the effects of series of all the infinitesimal cycles, we arrive
at the following relation for the entire original cycle:
Qrev

ld
=0 ..(4.15)
T
Clearly, the relation expressed by eqn. 4.15 suggests the existence of a state variable of the form
Qrev
as its sum over a cycle is zero. This state variable is termed as Entropy (S) such that:

or
T
Qrev
dS = ..(4.16)
T

W
Thus: dS = 0
Thus, for a reversible process: dQrev = TdS ..(4.17)

If applied to a perfectly reversible adiabatic process eqn. 4.16 leads to the following result:
dS = 0. Thus, such a process is alternately termed as isentropic.
TU
Since entropy is a state property (just as internal energy or enthalpy), even for irreversible
process occurring between two states, the change in entropy would be given by eqn. 4.16.
However since entropy is calculable directly by this equation one necessarily needs to construct a
reversible process by which the system may transit between the same two states. Finite changes
of entropy for irreversible processes cannot be calculated by a simple integration of eqn. 4.17.
JN

However, this difficulty is circumvented by applying the concept that regardless of the nature of
the process, the entropy change is identical if the initial and final states are the same for each type
of process. This is equally true for any change of state brought about by irreversible heat transfer
due to finite temperature gradients across the system and the surroundings. The same
consideration holds even for mechanically irreversibly processes.
ll

With the introduction of the definition of entropy the Carnot engine cycle may be redrawn
on a temperature-entropy diagram as shown in fig. 4.6.
A

9|Page
 ld
or
Fig. 4.6 Representation of Carnot cycle on T-S diagram
2.3 Entropy Calculations for Ideal Gases

W
For a closed system, the first law provides:
= dQ + dW
dU
If the process of change occurs under reversible conditions then the above equation becomes:

d ( H PV ) =TdS PdV
TU
dH PdV VdP = TdS PdV ..(4.18)

However for an ideal gas: dH = CP dT

ig
..(4.19)

Thus combining eqns. 4.18 and 4.19:

JN

CPig =
dT TdS + VdP

dS CPig dT
= d ln P ..(4.20)
Or: R R T

Integrating between an initial (T 0 , P 0 ) and any final (T, P):

ll

S
T
CPig dT P
=
R T R T ln P0
0
A

T
dT P
Or: S
= C R ln ..(4.21)
P ig
P P
0
T0
T P0

10 | P a g e
The last equation provides a direct expression for computing entropy change between two states
for an ideal gas.

For the special case of a reversible adiabatic dQ = 0; hence dS= 0. Thus:

S
T
CPig dT P
= T R T ln P=0 0

ld
R 0

R / C ig
For constant CPig , it follows = P / P0 ) P ( P / P0 )( 1) /
that: T / T0 (= ..(4.22)

or
One may note that eqn. 4.22 provides the same relation as obtained by the Fist Law analysis as in
eqn. 3.18.

W
2.4 Mathematical Statement of the 2nd Law:

Consider a mechanically irreversible adiabatic between two equilibrium states X and Y. In fig.
4.7 this path is reproduced as a broken line between points. Next the fluid is returned to the
original state by means two sequential processes: (i) a mechanically reversible adiabatic process
Y-Z, and then (ii) a reversible isobaric compression Z-X. Assuming that the mechanically
TU
irreversible process X-Y leads to an entropy change, heat transfer must occur during process Z-X
(since none occurs on Y-Z). This is because for the reversible return path the same entropy
change must occur as between X to Y. Since the return path is reversible we must get
dQrev = TdS . Thus, integrating this expression for the return path:
JN
ll
A

4.7 Cycle containing an irreversible adiabatic process A to B

11 | P a g e
 X X dQrev
S t =
Y
dS t = S Xt SYt =
Y T
..(4.23)

0. Thus, net total work during the cycle may be written

For the entire cycle we must have U t =
as: W t = Qrev . However, since (as depicted by the P-V diagram) a net work would need to be

done on the system over the cycle, it follows that W t > 0; hence Qrev < 0. Therefore, using eqn.

ld
X dQrev
4.23, it follows that: S Xt =
SYt
Y T
< 0, that is: S Xt < SYt . This implies that the entropy

or
change for the original irreversible step A to B, we have: S t = SYt S Xt > 0. Therefore, we
conclude that the original irreversible step is accompanied by a positive change of entropy. It may
be additionally shown that if the original process occurred through irreversible heat transfer

W
process an increase of entropy would result likewise.
It may be pointed out during the return reversible process, since the heat transfer (during
the isobaric path) occurs under reversible conditions it must occur under infinitesimal
temperature gradient between the system and surrounding. Thus for the system and the
surrounding the net entropy change (gives by: Qrev / T ) is equal and opposite. Consequently the
TU
total entropy change of the universe (system + surrounding) is zero during this step. During the
initial irreversible path X to Y, while entropy increases for the system, for the surrounding there
is no change. Nevertheless, for the universe as such there is a positive change in entropy.
Since the above conclusion is not premised on any specific internal nature of the system,
the above result may be generalized: all irreversible processes are accompanied by a positive
JN

change of entropy of the universe. This leads to a mathematical statement of the second law:
S t 0. ..(4.24)
At this point we can revert to the question that was stated at the beginning of this chapter:
what change in a system is thermodynamically feasible? The second law provides answer to this,
i.e., a process can only proceed in a direction that results in a positive change in the total entropy
ll

of the universe, the limiting value of zero being attained only by a fully reversible process. The
corollary to this is: no process is possible for which the total entropy of the universe decreases.
A

---------------------------------------------------------------------------------------------------------------------
Example 4.2

12 | P a g e
A rigid vessel of 0.06 m3 volume contains an ideal gas, C V = (5/2)R, at 500 K and 1 bar. (a) If 15
kJ of heat is transferred to the gas, determine its entropy change. (b) If the vessel is fitted with a
stirrer that is rotated by a shaft so that work in the amount of 15 kJ is done on the gas, what is the
entropy change of the gas if the process is adiabatic? What is Stotal?
(Click for solution)

ld
---------------------------------------------------------------------------------------------------------------------
Example 4.3
An ideal gas, Cp = (7/2)R, is heated in a steady-flow heat exchanger from 70C to 190C by

or
another stream of the same ideal gas which enters at 320C. The flow rates of the two streams are
the same, and heat losses from the exchanger are negligible. Calculate the molar entropy changes
of the two gas streams for both parallel and countercurrent flow in the exchanger. What is Stotal

W
in each case?
Click for solution
---------------------------------------------------------------------------------------------------------------------
2.5 Entropy Balance for Open Systems
TU
As with energy balance for open systems, once can extend the equation 4.24 to a generalized
entropy balance equation that may be written for the system shown in fig. 3.6. There is, however,
an important point of departure from the first law. Unlike energy, entropy is not a conserved
quantity for real world processes in which both mechanical and thermal irreversibilities are

inevitable. Hence for such processes are attended by a positive entropy generation rate, S G . This
JN

conclusion may be expressed in mathematical terms as follows:

ll
A

13 | P a g e
 Rate
Net of
rate Rate

change Total
of of
of rate

entropy entropy
+ _ entropy + = of

ld
change change change entropy
through of
in generation
flow control surrounding
streams

or
volume

d (mS )CV dS surr

t
( S m) fs + + SG 0
= ..(4.25)
dt dt

W
As with eqn. 4.24, for eqn. 4.25 too, the left side reduces to zero if the processes occurring in the
open system are totally reversible; that is, both with respect to the system as well to the
surroundings. Or else there is a net entropy generation.
A process is said to be internally reversible if all the processes occurring within the
TU
system are mechanically reversible, that is they are not subject to dissipative forces. External
reversibility, on the other hand, signifies that all heat transfer between the system and
surrounding occur under infinitesimal gradients and are therefore thermally reversible. In
principle reversible heat transfer is possible if the surroundings have heat reservoirs with
temperatures equal to those of the control surface or if reversible Carnot engines operate between
JN

the control-surface temperatures and the heat-reservoir temperatures.

We next expand eqn. 4.25 to a further level of detail. Let there be heat transfer at the rate

Q j at a particular part of the control surface of the open system where the surrounding
temperature is given by T , j .

ll

t
dS surr Qj
Thus: = ..(4.26)
dt j T , j
A

Here, j runs over all the heat reservoirs associated with the system. The negative sign is used for
the entropy term for the surroundings as heat transfer terms by convention are associated with the
system.

14 | P a g e
Putting eqn. 4.26 in eqn. 4.25 one obtains:

d (mS )CV Qj
( S m) fs + SG 0
= ..(4.27)
dt j T , j

For steady flows through the control volume eqn. 4.27 reduces to:

ld

Qj
( S m) fs SG 0
= ..(4.28)
j T , j

or
Further, for the simplest case of one inlet and exit, and one surrounding temperature:

Q
m(S ) fs = SG 0 ..(4.29)
T

W
---------------------------------------------------------------------------------------------------------------------
Example 4.4
Ten kmol per hour of air is throttled from upstream conditions of 25C and 10 bar to a
downstream pressure of 1.2 bar. Assume air to be an ideal gas with C P = (7/2)R. (a) What is the
TU
downstream temperature? (b) What is the entropy change of the air in J/molK? (c) What is the
rate of entropy generation in W/K? (d) If the surroundings are at 20C, what is the lost work?
Click for solution
---------------------------------------------------------------------------------------------------------------------
JN

2.6 Ideal and Lost Work for Flow Systems

We next derive the expressions for work exchange between system and surroundings for an open
system operating under steady state by incorporating the idea of irreversibility. As we have
discussed earlier, mechanical irreversibilities lead to loss of work due to dissipative conversion to
ll

heat. Thus, if work is to be delivered by an open system the maximum work obtains when the
processes are mechanically reversible, we call that ideal work. Conversely, when work is done on
A

the system the ideal work provides the minimum work needed to change the fluid state between
the inlet and the exit. This is because an extra work would need to be provided beyond the ideal
work against mechanical dissipative forces. From the considerations in the last section it may be

15 | P a g e
evident that ideal work obtains when the processes associated with the open system are both
internally and externally reversible. For such a case one may write eqn. 4.29 as follows:

Q
m(S ) fs = SG = 0 ..(4.30)
T

ld

Thus:
= Q T m(S ) fs ..(4.31)

From the 1st Law applied to the ideal case:

or
1 2
: H + (u / g c ) + z ( g / g c ) m =Q + W ideal ..(4.32)
2 fs
Putting eqn. 4.31 into 4.32 and simplifying (assuming negligible kinetic and potential energy

W


changes): W ideal = H m T ( S m) fs ..(4.33)
fs

Or: W ideal = m[H T S fs ]
..(4.34)
In real processes however, actual work involved will depend on the extent of associated
TU
irreversibilities; hence we define lost work as follows:

=
W lost W actual W ideal ..(4.35)
Now from eqn. 4.32 for the simplest case where kinetic and potential energy changes are
negligible:
JN

H m = Qactual + W actual ..(4.36)
fs
Using eqns. 4.34 and 4.35:

T ( S m) fs Qactual
W lost = ..(4.37)
ll

Q
Applying eqn. 4.30, to a real process: m(S ) fs actual =
SG
T ..(4.38)
A

Or: Q actual = T m(S ) fs T SG ..(4.39)

Thus from eqns. 4.37 and 4.39 it follows that:

16 | P a g e

W lost = T SG ..(4.40)

Eqn. 4.40 suggests that greater the entropy generation rate SG due to process irreversibility,
greater is the lost work. Since irreversibilities implicit in a process cannot be calculated
theoretically, it is indirectly expressed by a process efficiency factor, . The expression for such

ld
efficiency is as follows.

For work done by system, = Wactual / Wideal . .(4.41)

or

For work done on a system, = Wideal / Wactual ..(4.42)

---------------------------------------------------------------------------------------------------------------------

W
Example 4.5
A steady-flow adiabatic turbine (expander) accepts gas at conditions T 1 = 500 K, P 1 = 6 bar, and
discharges at conditions T 2 = 371 K, P 2 = 1.2 bar. Assuming ideal gases, determine (per mole of
gas) W actual , W ideal , W lost , and entropy generation rate. T surr = 300 K, C P /R = 7/2.
Click for solution
TU
---------------------------------------------------------------------------------------------------------------------
Example 4.6
A steam turbine operates adiabatically at a power level of 3500 kW. Steam enters the turbine at
2400 kPa and 500C and exhausts from the turbine at 20 kPa. What is the steam rate through the
turbine, and what is the turbine efficiency?
JN

Click for solution

-------------------------------------------------------------------------------------------------------------------
Appendix: 4.1
Entropy: A Statistical Thermodynamic Interpretation
As with internal energy the concept of entropy may be provided a microscopic, molecular
ll

interpretation. However, it may be noted that unlike internal energy, molecules do not possess
entropy. Entropy is purely a measure thermodynamic probability associated with the distribution
A

of various positional and energy states that molecules of a substance may exist under given
conditions of temperature and pressure. To understand this, consider a simple system comprised
of a rigid, insulated vessel divided into two compartments of equal volume by a partition (of

17 | P a g e
negligible volume, fig. A.4.1). Initially, the left compartment contains gas at a temperature T1 and

pressure P1 , and the volume of each compartment is V1t ; the right compartment is fully evacuated.
The system is at equilibrium under the above conditions, and is described by the ideal

ld
or
Fig. A.4.1 Ideal Gas Expansion into an evacuated chamber

gas EOS. At a certain point of time the partition is ruptured and the gas contained in the left

W
compartment allowed to fill the entire system volume. Eventually the gas would reach a new
equilibrium state (2) that is characterized by a pressure P1 / 2 throughout. For the above process,
applying the first law:
U = Q + W
TU
But : Q
= W= 0
Hence, U =
0 ..(A.4.1.1)
This implies that for the gas initially on the left side:=
T 0; or
= T2 T1

The entropy change per mole of the gas then is given by eqn. 4.21:
dT dP dP
Thus, S =CP
JN

R = R
T P P
P2 V
S = R ln =R ln 2 =R ln 2 ..(A.4.1.2)
P1 V1
From a microscopic point of view the molecular state variables are: position and energy. Since
there is no change of temperature during the expansion process, the internal energy of each
ll

molecule does not change. There, is, however, a possible change in the position occupied a
molecule. After the expansion a molecule may be located either on the left half or in the right half
A

of the compartment at any point of time. We may say, in a simplified sense, there are two
possible positions that it may have. Considering one mole of gas as basis, if there are, in general,
j such states (either of position or energy) that a molecule may exist in, the total number of ways

18 | P a g e
 that an Avogadro number of molecules ( N A = 6.023 x1023 ) may be distributed over all the

available states is given by (with n j molecules in the jth state):

NA !
= ; N A= n1 + n2 + ... + n j ..(A.4.1.3)
(n1 !)(n2 !)...( n j !)

The notation used for the right side of the above equation, is called the thermodynamic

ld
probability. Thus for the system in question, in the initial state all the molecules are located in the
left half of the compartment, so there is only one positional state; thus:

or
NA !
=1 = 1 ..(A.4.1.4)
( N A !)(0!)
However, in the final state as there are two possible positional states, the corresponding maximum
thermodynamic probability is:

W
NA !
2 = ..(A.4.1.5)
[( N A / 2)!][( N A / 2)!]
The description maximum is important, as the exact number of molecules that may be
distributed between the two compartments at any point of time need not be as the same, as
TU
assumed in eqn. A.4.1.5. In general there could be many possible bifurcations of the total number
molecules into the two compartments, each characterized by a corresponding thermodynamic
probability. However, simple calculations suggest that 2 has the highest possible value if each
compartment contains equal number of molecules, i.e., N A /2. Indeed, intuitively speaking, one
would expect that to be the most likely distribution in the second equilibrium state. This implies
JN

that the system would tend to move to an equilibrium state that has the highest of all the possible
values of the thermodynamic probability that one may attribute to the system. It was Ludwig
Eduard Boltzmann (18441906) the Austrian physicist, who is credited with the development of
the statistical concept of nature (now known as statistical mechanics), first proposed the relation
between entropy and thermodynamic probability:
ll

S k ln
= ..(A.4.1.6)
k = Boltzmann constant = R / N A
A

If one applies the equation to the expansion process, one obtains the following relation:
2
S = S 2 S1 = k ln 2 k ln 1 = k ln ..(A.4.1.7)
1
19 | P a g e
 NA !
Thus: S =k ln
[( N A / 2)!][( N A / 2)!]
Or:
= S k [ ln N A ! 2 ln( N A / 2)!] ..(A.4.1.8)

By the well-known Stirlings formula, or large values of an integer N, one has:

ld
ln N ! N ln N N
= ..(A.4.1.9)
On applying eqn. A.4.1.9 in A.4.1.8 and simplifying the following result obtains:
NA
S kN A ln =
= kN= ln 2 R ln 2 ..(A.4.1.10)

or
NA / 2
A

Thus we see that the results of both eqns. A.4.1.2 and A.4.1.10 are identical. There is thus a
convergence between the macroscopic and microscopic description of the concept of entropy.

W
It is particularly significant that the above result obtains if one assumes the highest value
of 2 as its most likely value. It may be argued that the highest value of 2 corresponds to the

most disordered state at the molecular level. Indeed extending the argument, since 1 =1 the
initial state is more ordered than state 2, as in the former all molecules are characterized by a
TU
single position coordinate, and therefore the probability of finding a molecule in that position is
unity; in other words one has complete (or certain) information about position of any molecule.
In contrast in the second state the probability of finding a molecule at any one position is half.
This obtains if each compartment contains exactly half the total number of molecules. Any other
distribution would be less disordered than this state, and also less probable! This is the reasoning
JN

which lies at the root of the popular description of entropy as a measure of disorder at the
molecular level. It follows that a system at non-equilibrium state (say, at the point of rupture of
the partition) always tends to the most disordered molecular state when it attains a new
equilibrium. Since one has less microscopic information about a system with a greater degree of
disorder, information is also described as negentropy.
ll

-------------------------------------------------------------------------------------------------------------------
Assignment- Chapter 4
A

20 | P a g e
Chapter 5 Thermodynamic Properties of Real Fluids
It has already been demonstrated through the first and second laws, that the work and heat
interactions between a system and its surroundings may be related to the state variables such as
internal energy, enthalpy and entropy. So far we have illustrated the calculations of energy and

ld
entropy primarily for pure (component) ideal gas systems. However, in practice this is an
exception rather than a rule as one has to deal with not only gases removed from ideal gas state,
but also with liquids and solids. In addition, mixtures rather than pure components are far more

or
common in chemical process plants. Therefore, computation of work and heat interactions for
system comprised of real fluids requires more complex thermodynamic formulations. This
chapter is devoted to development of such relations that can help calculate energy requirements

W
for given changes of state for real systems. As in the case of ideal gases the goal is to correlate
the energy and entropy changes for real fluids in terms of their volumetric and other easily
measurable macroscopic properties.

5.1 Thermodynamic Property Relations for Single Phase Systems

TU
Apart from internal energy and enthalpy, two other ones that are particularly useful in depiction
of thermodynamic equilibrium are Helmholtz free energy (A) and Gibbs free energy (G). We
defer expanding upon the concept of these two types of energies to chapter 6; however, we state
their definition at this point as they are instrumental in the development of property correlations
for real fluids.
JN

Specific Helmholtz free energy: A= U TS ..(5.1)

Specific Gibbs free energy: G= H TS ..(5.2)
For a reversible process in a closed system the first law gives:

= dQ + dW
dU
ll

Or:
A

= TdS PdV
dU ..(5.3)

Using H= U + PV and taking a total differential of both sides:

1|Page
 dH =dU + PdV + VdP ..(5.4)

Putting eqn. 5.3 in 5.4 we get:

= TdS + VdP
dH ..(5.5)

ld
In the same manner as above one may easily show that the following two relations obtain:
dA =
SdT PdV ..(5.6)

=
dG VdP SdT ..(5.7)

or
Equations 5.3 to 5.7 comprise the fundamental energy relations for thermodynamic systems
where there is a single phase with constant composition. In principle, they may be integrated to

W
compute the energy changes for a system transiting from one equilibrium state to another.

5.2 Maxwell Relations

All the four types of energy relations above satisfy the mathematical condition of being
continuous variables, as they are themselves functions of state variables. One can thus apply of
TU
the criterion of exact differential for these functions.
For a function of the form P = P ( X , Y ) one can write the following total differential:

P P
dP =
dX + dY =
MdX + NdY ..(5.8)
X Y Y X
P P
Where: M =
= and N ..(5.9)
JN

X Y Y X
M P N P
2 2

Further,
= = and ..(5.10)
Y X Y X X Y X Y
M N
It follows: = ..(5.11)
Y X X Y
ll

Applying eqn. 5.11 to 5.3, 5.5, 5.6 and 5.7 one may derive the following relationships termed
Maxwell relations:
A

T P
= ..(5.12)
V S S V
T V
= ..(5.13)
P S S P
2|Page
 P S
= ..(5.14)
T V V T
V S
= ..(5.15)
T P P T

ld
5.4 Relations for Enthalpy, Entropy and Internal Energy
One may conveniently employ the general energy relations and Maxwell equations to obtain
expressions for change in enthalpy and entropy and internal energy for any process, which in turn

or
may be used for computing the associated heat and work interactions.
Let H = H (T , P)

W
H H
Then: dH
= dT + dP
T P P T

H
But = CP
T P
TU
H
Thus:=
dH CP dT + dP ..(5.16)
P T

H S
Using dH = TdS + VdP =T +V ..(5.17)
P T P T

S V
From Maxwell relations as in eqn. 5.15: = ..(5.18)
JN

P T T P

Thus using eqns. 5.17 and 5.18 in 5.16 we get:

V
dH= CP dT + V T dP ..(5.20)
T P
ll

In the same manner starting from the general = (T , V ) an dS S (T , P )

function: U U= and
applying appropriate Maxwell relations one may derive the following general expressions for
A

differential changes in internal energy and entropy.

P
dU =
CV dT + T P dV ..(5.21)
T V
3|Page
 dT V
=
dS CP dP ..(5.22)
T T P

dT P
Or, alternately:
= dS CV + dV ..(5.23)
T T V

ld
Thus, eqns. 5.20 to 5.23 provide convenient general relations for computing enthalpy, internal
energy and entropy changes as function of volumetric properties and specific heats. If a fluid is
described by a suitable EOS, these equations may be conveniently integrated to obtain analytical

or
expressions for energy and entropy changes.
---------------------------------------------------------------------------------------------------------------------
Example 5.1

W
Derive an expression for enthalpy change of a gas during an isothermal process assuming using the
following EOS: P (V b) =
RT
(Click here for solution)
---------------------------------------------------------------------------------------------------------------------
5.5 Residual Property Relations
TU
An alternate method of computing energy and entropy changes for real gases involves the
definition of residual property. The specific residual property M R is defined as follows:
M R M (T , P ) M ig (T , P )
= ..(5.24)

Where, M (T , P) is the specific property of a real gas at a given T & P, and M ig (T , P ) is the
JN

value of the same property if the gas were to behave ideally at the same T & P. Thus for example:
V R= V V ig
Using the generalized compressibility factor Z:
=V R ( ZRT / P) ( RT / P)
V=
R
( Z 1) RT / P ..(5.25)
ll

The residual properties are usually used for gases only. Using such a property for a liquid (or
solid) is inconvenient as then it would also include the property change of vapourization (and
A

solidification) which generally are large in magnitude. This detracts from the advantage of
working with the residual property as a measure of small corrections to ideal gas behaviour. Thus
the use of residual functions is restricted to prediction of real gas behaviour only. To exploit the

4|Page
concept of residual properties we take advantage of the Gibbs free energy as it can be used as a
generating function for other thermodynamic properties.
Derivation of Residual functions:

We start from the generic equation: G= H TS ..(5.2)

ld
And =
dG VdP SdT ..(5.7)

Taking the total differential for the function G / RT :

G 1

or
G
d=
dG dT ..(5.26)
RT RT RT 2

Substituting eqn. 5.2 and 5.7 in 5.26 we get:

W
G V H
d = dP dT ..(5.27)
RT RT RT 2

One may write the same equation specifically for an ideal gas, whence:

G V
ig ig
H ig
d = dP dT ..(5.28)
TU
RT RT RT 2

Subtracting eqn. 5.28 from 5.27:

G V
R R
HR
d = dP dT ..(5.29)
RT RT RT 2

Thus we may write the following further generative relations:

JN

V R (G R / RT )
= ..(5.30)
RT P
T

HR (G / RT )
R

= T ..(5.31)
RT T P
ll

S R H R GR
And further: = ..(5.32)
RT RT RT
A

G G
R R
P GR GR
From eqn. 5.29: 0
d =
RT RT P RT P =0
; but
RT P =0
=0

5|Page
 GR P V
R

Thus: = dP ..(5.33)
0
RT RT

GR P dP
Putting eqn. 5.25 in 5.33: =
RT 0
( Z 1)
P
..(5.34)

Differentiating eqn. 5.34 w.r.t T in accordance with 5.31 gives:

ld
HR P
Z dP
= T ..(5.35)
RT 0
T P P

or
Finally using eqns 5.32, 5.34, and 5.35:

SR P
Z dP P dP
T
= ( Z 1) ..(5.36)
R 0
T P P 0
P

W
The last two equations may be expressed in alternative forms in terms of reduced temperature and
pressure:

HR Pr Z dPr
= Tr2 ..(5.37)
RTC 0
Tr Pr Pr
TU
SR Pr Z dPr Pr dP
Tr
= ( Z 1) r ..(5.38)
R 0
Tr Pr Pr 0 Pr

---------------------------------------------------------------------------------------------------------------------
Example 5.2
JN

Derive an expression for enthalpy change of a gas during an isothermal process assuming using
the following EOS: Z = 1 + AP r / T r
(Click here for solution)
---------------------------------------------------------------------------------------------------------------------
5.6 Residual Property Calculation from EOS
ll

From Virial EOS:

Using 5.35 and 2.12 one obtains:

A

H R P B dB
= ..(5.39)
RT R T dT

6|Page
Next, on substituting eqns. 2.13 2.15 in 5.37 the following relations result:

HR 1.097 0.894
= Pr 0.083 1.6 + 0.139 4.2 ..(5.40)
RTC Tr Tr

Similarly using 5.36 and 2.12 we get:

ld
SR P dB
= ..(5.40)
R R dT

or
Finally employing eqns. 2.13 2.15 in 5.38:

SR 0.675 0.722
Pr 1.6 + 4.2
= ..(5.41)
R Tr Tr

W
From Cubic EOS:

One may use the following form of the cubic EOS presented in chapter 2:
TU
RT a
=P 2 ..(2.21)
V b V + ubV + wb 2

Or equivalently: Z 3 + Z 2 + Z + =0 ..(2.23)

While the eqns. 5.35 and 5.36 are useful for volume explicit EOS, they are unsuitable for cubic
EOS which are pressure explicit. For the latter type of EOS one may show that the appropriate
JN

equations for residual enthalpy and entropy are (see S.I. Sandler, Chemical, Biochemical and
Engineering Thermodynamics, ch. 6, 4th Edition, Wiley India, 2006):

HR 1 V P
RT
= Z 1+
RT V = T
T V
P dV

..(5.42)

1 V P R
ll

SR
ln Z +
= dV ..(5.43)
R R V = T V V
A

The above relations may be applied to the various cubic EOSs to obtain the necessary residual
property relations. The final results are shown below.
From RK-EOS:

7|Page
 HR 3a Z+B
= ( Z 1) ln ..(5.44)
RT 2bRT Z

SR a Z+B
= ln( Z B ) ln ..(5.47)
R 2bRT Z

ld
0.42748R 2TC2 RK (Tr ) 0.08664RTC
=where, a = ;b
PC PC
RK (Tr ) = Tr1/2

or
For SRK-EOS:

a
1
T a
T Z+B
R
H
= ( Z 1) + ln (5.48)

W

RT RT bRT Z

SR 1 a Z + B
= ln( Z B ) ln ..(5.49)
R bRT T Z

a a
=
(0.48 + 1.574 0.176 2 ) ..(5.51)
TU
T SRK (Tr )TTC

0.42748R 2TC2 SRK (Tr ) 0.08664RTC

=where, a = ;b
PC PC
and , SRK (Tr ) =
[1 + (0.48 + 1.574 0.176 2 )(1 Tr1/2 )]2
JN

For PR-EOS:

a
T a
T Z + B(1 + 2)
R
H
= ( Z 1) + ln ..(5.52)

RT 2 2bRT Z + B (1 2)
ll

SR 1 a Z + B(1 + 2)
= ln( Z B) ln ..(5.53)
R 2 2bRT T Z + B(1 2)
A

a a
=
(0.37464 + 1.5422 0.62992 2 ) ..(5.54)
T PR (Tr )TTC

8|Page
 0.45724R 2TC2 PR (Tr ) 0.07779RTC
=where, a = ;b
PC PC
and , PR (Tr ) =
[1 + (0.37464 + 1.5422 0.62992 2 )(1 Tr1/2 )]2

As part of computing the residual enthalpy or entropy (and internal energy) at any set of
temperature and pressure, the molar volume or correspondingly Zvap needs to be first computed

ld
employing the usual algorithm for solving cubic EOS described in section 2.3.3.
---------------------------------------------------------------------------------------------------------------------

or
Example 5.3
Derive expressions for HR, SR from RK-EOS.
(Click here for solution)

W
--------------------------------------------------------------------------------------------------------------------

5.7 Generalized Correlations for computing dH and dS for a real gas:

The approach based on the use of the compressibility factors can be applied to the present
instance to evolve generalized correlations for computing enthalpy and entropy changes for
TU
gases. We start with the Pitzer-type expression for the compressibility factor:
Z Z 0 + Z1
= ..(2.25)

Differentiating with respect to the reduced temperature we get:

Z Z 0 Z 1
= + ..(5.55)
Tr Pr Tr Pr Tr Pr
JN

Thus using eqn. 5.54 we may recast eqns. 5.37 and 5.38 as follows:

2 Z 2 Z
P0 1 P
HR r r
dPr
Tr
= Tr ..(5.56)
RTc 0
Tr P 0
Tr Pr
r Pr
ll

Pr dP Pr
SR Z 0 Z 1 1 dPr
And: Tr
= + Z 0
1 r
0 Tr
+ Z ..(5.57)
R 0 Tr P Pr Pr
Pr
A

Both the above equations may be rewritten individually as follows:

9|Page
HR
=
HR
+
( )
( H R )1
0

..(5.58)
RTc RTc RTc

(H )
0
R
Z 0
Pr

0 Tr
2
Where, = T r
RTc ..(5.59a)

ld
Pr

And:

( H R )1 2 Z
r 1 P
dPr
= Tr

or

RTc 0
Tr P Pr ..(5.59b)
r

Similarly:

W
(S )
0
R
SR ( S R )1
= + ..(5.60)
R R R

(S )
0
R Pr dP
Z 0
TU
Tr
= + Z 0
1 r ..(5.61a)
0
Pr
r Pr
R T

(S )
1
R Pr
Z 1 1 dPr

= + Z
R 0 Tr P Pr ..(5.61b)
r

(H )
JN

0
R
( H R )1
The term in eqn. 5.59a constitutes the first order enthalpy departure, and (in eqn.
RTc RTc
5.59b) the second order (with respect to simple fluids) enthalpy departure at specified T r and P r .
The same is true for the corresponding entropic terms provided by eqns. 5.61a and 5.61b.
The evaluation of the integrals in eqns. 5.59 to 5.62 may be carried out assuming an EOS.
ll

The most widely used approach is that of Lee and Kesler who employed a modified form of the
BWR EOS (eqn. 2.17) to extend their generalized correlation to residual property estimation.
A

Figs. 5.1 and 5.2 respectively provide values of ( H R )0 / RTC , an ( Hd R )1 / RTC respectively.

Similar plots for the entropy terms ( S R )0 / R, an ( Sd R )1 / R are also available; however, here the

plot of the entire entropy term ( S R ) as function of T r and P r is shown in fig. 5.3. Table of
10 | P a g e
values of all the above (eqns. 5.59 to 5.62) are also available as functions of T r and P r at discrete
intervals (see for example: J.M. Smith, H.C. Van Ness and M.M. Abbott, Introduction to
Chemical Engineering Thermodynamics, 6th ed., McGraw-Hill, 2001.

ld
or
W
TU
Fig. 5.1 Correlation of H R0 RTC drawn from tables of Lee-Kesler (Source: AIChE J., pp. 510-527,
1975)
JN
ll
A

11 | P a g e
 ld
or
W
Fig. 5.2 Correlation of H R1 RTC drawn from tables of Lee-Kesler (Source: AIChE J., pp. 510-527,
1975)
TU
JN
ll
A

12 | P a g e
 ld
or
W
TU
Figure 5.3 Generalized entropy departure functions using corresponding states.
JN

[Source: O.A. Hougen, K.M. Watson, and R.A. Ragatz, Chemical Process
Principles Charts, 2nd ed., John Wiley & Sons, New York, 1960]

5.8 Computation of H and S for a Gas using Generalized Departure Functions

The residual function equations presented in the last section are particularly useful for estimating
finite changes in enthalpy and entropy for real gases undergoing change in either closed or open
ll

system processes. We consider that a pure fluid changes state from (T1 , P1 ) to (T2 , P2 ); shown
A

schematically in fig. 5.4.

13 | P a g e
 ld
or
Fig. 5.4 Schematic of a General Thermodynamic Process on P T co-ordinates

Since the departure functions H R and S R capture deviations from ideal gas behaviour at the same

W
temperature as the real gas, one can conceive of the pathway between states 1 and 2 to be
decomposed into following steps (see fig 5.5):
a) Real gas state at (T1 , P1 ) to ideal gas state (ig) at (T1 , P1 )

b) Ideal gas state at (T1 , P1 ) to ideal gas state at (T2 , P2 )

TU
c) Ideal gas state at (T2 , P2 ) to real gas state at (T2 , P2 )
JN
ll

5.5 Pathway for calculating H and S for Real Gases

A

For step a the change of enthalpy is given by: H1ig H1 =

H1R
For step b the change of enthalpy is given by: H 2ig H1ig =
H ig

For step c the change of enthalpy is given by: H 2 H 2ig =

H 2R
14 | P a g e
Therefore, the overall change of enthalpy is given by:
T2
H
= H 2 H=
1 H 2R H1R + H ig C igp dT
T1

T2
Using eqn. 3.8: H 2 H1 = H 2R H1R + C igp dT ..(5.63)
T1

ld
The same considerations apply for computing the change of entropy between the two states:
S= S 2 S=
1 S 2R S1R + S ig
T2
Using eqn. 4.21: S 2 S1 = S 2R S1R + C igp dT / T R ln( P2 / P1 ) ..(5.64)

or
T1

Generalized residual property relations may be used for calculation of change in internal energy
i.e. U 2 U1 for a process in the following manner:

W
U 2 U1 = ( H 2 PV
2 2 ) ( H1 PV
1 1)

Or: U 2 U1 = ( H 2 H1 ) ( PV 1 1)
2 2 PV ..(5.65)

The term ( H 2 H1 ) can be calculated using eqn. 5.63, while the term ( PV 1 1 ) may be
2 2 PV

computed after obtaining V1 and V2 applying the generalized compressibility factor approach.
TU
One may, however, also use the generalized residual property charts for internal energy for the
same purpose (fig. 5.6).
---------------------------------------------------------------------------------------------------------------------
Example 5.5
Estimate the final temperature and the work required when 1 mol of n-butane is compressed
JN

isentropically in a steady-flow process from 1 bar and 50oC to 7.8 bar.

(Click here for solution)
---------------------------------------------------------------------------------------------------------------------
ll
A

15 | P a g e
 ld
or
W
TU
JN

Figure 5.6 Generalized internal energy departure functions using corresponding states
[Source: O.A. Hougen, K.M. Watson, and R.A. Ragatz (1960), Chemical Process
Principles Charts, 2nd ed., John Wiley & Sons, New York]
ll

5.9 Extension to Gas Mixtures

The generalized equations developed for H and S in the last section may be extended to
A

compute corresponding changes for a real gaseous mixture. The method used is the same as that

16 | P a g e
 developed in section 2.4 through the definition of pseudo-critical mixture properties using linear
mixing rules:
TC ,m = yT
i
i C ,i PC ,m = y P
i
i C ,i m = yii
i
..(2.32)

Using the above equations pseudo-reduced properties are computed:

ld
Tr ,m = T / TC ,m and Pr ,m = P / PC ,m

Further calculations of changes in internal energy, enthalpy, and entropy follow the same
principles developed in the last section.

or
---------------------------------------------------------------------------------------------------------------------
Example 5.6
Calculate the changes in enthalpy and entropy per mole when a mixture of 70 mole % ethylene

W
(1) and 30 mole% propylene (2) at 323K and 10 bar is taken to 60 bar and 600 K using the
generalized compressibility factor approach.
4.196 + 154.565 x103 T 81.076 x106 T 2 + 16.813 x109 T 3 ;
C igp1 =

3.305 + 235.821x103 T 117.58 x106 T 2 + 22.673 x109 T 3

C igp2 =
TU
(Click here for solution)
---------------------------------------------------------------------------------------------------------------------
5.10 Relations for H and S for Liquids

One starts with the generic equations for dH and dS developed in section 5.4.
JN

V
dH= CP dT + V T dP ..(5.20)
P T

dT V
=
dS CP dP ..(5.22)
T T P

As discussed in section 2.3, for liquids it is often simpler to use volume expansivity and
ll

isothermal compressibility parameters for computing thermodynamic properties of interest. Thus

from eqn. 2.3:
A

1 V
Volume Expansivity ,
V T P
Using eqn. 2.3 in 5.20 and 5.22 the following relations obtain:
17 | P a g e
 = CP dT + [1 T ] dP
dH ..(5.66)

dT
=
dS CP VdP ..(5.67)
T

The above equations may also be used to compute the properties of a compressed liquid state.

ld
Since the volumetric properties of liquids are very weakly dependent of pressure one can often
use the saturated liquid phase properties as reference points and integrate the eqns. 5.66 to 5.67
(at constant temperature) to obtain enthalpy and entropy respectively. The relevant equations are:

or
P
H isat + sat Vi (1 iT )dP
Hi = ..(5.68)
Pi

P
Si Sisat
= iVi dP ..(5.69)
Pi sat

W
In the last two equations, the molar volume V i may be set equal to Vi sat (liq.) , and the volume
expansivity approximated to that at the saturated liquid point at the given temperature.

5.11 Applications to real fluid processes in process plant equipments

TU
In a typical process plant one encounters a variety of flow devices such pumps, compressors,
turbines, nozzles, diffusers, etc. Such devices are not subject to heat transfer by design as are heat
exchangers, condensers, evaporators, reactors, etc. However, the flow devices typically are
subject to mechanical irreversibility owing to existence of dissipative forces such as fluid
viscosity and mechanical friction, which results in reduction of their efficiency. In addition such
JN

devices may be subject to thermal irreversibility as their operation may not be truly adiabatic.
Therefore, it is necessary to compute the efficiency of such devices in relation to a perfectly
reversible (isentropic) process between their inlet and outlet.
The performance of a flow device is expressed in terms of isentropic efficiency in which
the actual performance of the device is compared with that of an isentropic device for the same
ll

inlet conditions and exit pressure. For example, the isentropic efficiency T of a turbine (which

essentially converts fluid enthalpy to shaft work, fig. 5.7) is defined as:
A

Power output of the actual turbine H He

T = = i ..(5.69)
Power out put of the turbine,if it were isentropic H i H es

18 | P a g e
 ld
Fig.5.7 Schematic of a Turbine

or
Where, H i = enthalpy of the fluid at the inlet of the turbine, H e = enthalpy of the fluid at the exit
of the actual turbine, H es = enthalpy of the fluid at the exit of the turbine, if it were isentropic.

W
Similarly, the isentropic efficiency c of a compressor (fig. 5.8) or a pump p (which

convert applied shaft work to fluid enthalpy) is given by:

(=
Ws )
s
H es H i
c (or=
P ) ..(5.70)
(Ws ) He Hi
TU
Where, H i = enthalpy of the fluid at the inlet to the compressor (or pump), H e = enthalpy of the

fluid at the exit of the actual compressor (pump), and H es = enthalpy of the fluid at the exit of an

isentropic compressor (or pump), and Ws represents the shaft work per mole (or mass) of fluid in
the two situations.
JN
ll

Fig.5.8 Schematic of a Turbine

A

19 | P a g e
 The isentropic efficiency N of a nozzle, which is used to achieve high fluid velocity at
exit (by conversion of enthalpy to kinetic energy) is given by:
(V22 ) / 2 Kinetic energy of gas leaving the actual nozzle
=N = 2 s
..(5.71)
(V2 ) / 2 Kinetic energy of gas leaving the nozzle, if it were isentropic

ld
---------------------------------------------------------------------------------------------------------------------
Example 5.7
A certain gas is compressed adiabatically from 293 K and 135 KPa to 550 KPa. What is the work

or
needed? What is the final T 2 ? Assume ideal gas behavior. Compressor = 0.8.
For the gas: C p ig = 1.65 + 8.9 x 10-3T 2.2 x 10-6 T2
---------------------------------------------------------------------------------------------------------------------

W
Assignment - Chapter 5
TU
JN
ll
A

20 | P a g e
Chapter 6: Solution Thermodynamics and Principles of Phase Equilibria
In all the preceding chapters we have focused primarily on thermodynamic systems comprising pure
substances. However, in all of nature, mixtures are ubiquitous. In chemical process plants the
ultimate domain of application of the principles of chemical engineering thermodynamics matter is

ld
dominantly processed in the form of mixtures. Process streams are typically comprised of multiple
components, very often distributed over multiple phases. Separation or mixing processes necessitate
the use of multiple phases in order to preferentially concentrate the desired materials in one of the

or
phases. Reactors very often bring together various reactants that exist in different phases. It follows
that during mixing, separation, inter-phase transfer, and reaction processes occurring in chemical
plants multi-component gases or liquids undergo composition changes. Thus, in the thermodynamic

W
description of such systems, in addition to pressure and temperature, composition plays a key role.
Further, whenever multiple phases are present in a system, material and energy transfer occurs
between till the phases are in equilibrium with each other, i.e., the system tends to a state wherein the
all thermal, mechanical and chemical potential (introduced earlier in section 1.4) gradients within and
across all phases cease to exist.
TU
The present chapter constitutes a systematic development of the concept of a new class of
properties essential to description of real mixtures, as well of the idea of the chemical potential
necessary for deriving the criterion of phase and chemical reaction equilibrium. Such properties
facilitate the application of the first and second law principles to quantitatively describe changes of
internal, energy, enthalpy and entropy of multi-component and multiphase systems.
JN

Of the separate class of properties relevant to multi-component and multi-phase systems, the
partial molar property and the chemical potential are particularly important. The former is used for
describing behaviour of homogeneous multi-component systems, while the latter forms the fundament
to description equilibrium in multi-phase, as well as reactive systems.
As in the case of pure gases, the ideal gas mixture acts as a datum for estimating the properties
ll

of real gas mixtures. The comparison of the properties of the real and ideal gas mixtures leads to the
introduction of the concept of fugacity, a property that is further related to the chemical potential.
A

Fugacity may also be expressed as a function of volumetric properties of fluids. As we will see, the
functional equivalence of fugacity and the chemical potential provides a convenient pathway for
relating the temperature, pressure and phase composition of a system under equilibrium.
 In the last chapter it was demonstrated that residual properties provide very suitable means of
estimating real gas properties. But as pointed out its usage for description of liquid states is not
convenient. This difficulty is overcome by the formulation of a concept of ideal solution behaviour,
which serves as a datum for estimating properties of real liquid solutions. The departure of the
property of a real solution from that of an ideal one is termed as excess property. In other words, the

ld
excess property plays a role similar to that of residual property. In the description of solution behaviour
at low to moderate pressures, we employ yet another property, the activity coefficient; which originates

or
from the concept of fugacity. The activity coefficient may also be related to the excess Gibbs energy. It
is useful not only as a measure of the extent of non-ideality of a real solution but also, more
significantly in describing phase equilibria at low to moderate pressures.

W
6.1 Partial Molar Property
We consider first the case of a homogenous (single-phase), open system that can interchange matter
with its surroundings and hence undergo a change of composition. Therefore, the total value of any
extensive property M t ( M V , U , H , S , A, G ) is not only a function of T and P, but also of the actual
TU
number of moles of each species present in the system. Thus, we may write the following general
property relation:
M t = nM = M (T , P, n1 , n2 ..ni .nN ) ..(6.1)

Where N total number of chemical species in the system

JN

N
n =
total number of moles in the system n
i
i ..(6.2)

Taking the total derivative for both sides of eqn. 6.1:

(nM ) (nM ) (nM )
d ( nM ) = dP + dT + dni ..(6.2)
P T ,n T P ,n ni T , P ,n
j i

Where, subscript n indicates that all mole numbers are held constant and subscript n j i that all mole
ll

numbers except ni are held constant. This equation has the simpler form:
M M
d ( nM ) = n dP + n dT + M i dni
A

..(6.3)
P T , x T T , x i

(the subscript x denotes differential at constant composition)

 (nM )
Where: M i = ..(6.4)
ni P ,T ,n ji

Eqn. 6.4 defines the partial molar property M i of species i in solution. It represents the change of total

property nM of a mixture resulting from addition at constant T and P of a differential amount of

ld
species i to a finite amount of solution. In other words it also signifies the value of the property per
mole of the specific species when it exists in solution. In general, the partial molar property of a
substance differs from the molar property of the same substance in a pure state at the same temperature

or
and pressure as the mixture or solution. This owing to the fact that while in a pure state the molecules
interact with its own species, in a solution it may be subjected to different interaction potential with
dissimilar molecules. This may render the value of a molar property different in mixed and pure states.

W
=
Now, ni xi n;=
dni xi dn + n dxi

Or=
: dni xi dn + n dxi ..(6.5)

Also: d ( nM
= ) ndM + Mdn ..(6.6)

Substituting eqns. 6.5 and 6.6 in eqn. 6.3 leads to:

TU
M M
ndM +=
Mdn n dP + n dT + M i ( xi dn + n dxi ) ..(6.7)
P T , x T T , x i

On re-arranging:
M M
dM dP dT + M i dxi n + M M i xi dn =
0 ..(6.8)
P T
JN

T , x T , x i i

While deriving eqn. 6.8 no specific constraints on the values of either n or dn have been applied. This
suggests that the equation is valid for any arbitrary values of these two variables. Thus n and dn are
independent of each other. Therefore, eqn. 6.8 can only be valid if the coefficients of these two
variables are identically zero. On putting the coefficients to zero the following equations obtain:
M M
ll

dM = dP dT + M i dxi
P T , x T T , x i
..(6.9)
A

M = M i xi ( at constant T & P)
i
..(6.10)
From eqn. 6.10, it follows, that nM = n xi M i (6.11)
i

And also: dM = xi dM i + M i dxi (6.12)

i i

Equating (6.19) and (6.21) yields the well-known Gibbs-Duhem equation (GDE):

ld
M M
dP dT xi dM i =
0
P T , x T T , x i
..(6.13)
The GDE must be satisfied for all changes in P, T, and in M i caused by changes of state in a

or
homogeneous phase. For the important special case of changes at constant T and P, it simplifies to:

x dM
i
i i =0 ..(6.14)

W
dM i
Or, taking any arbitrary species j: x
i
i
dx j
=0 (at const T, P) ..(6.15)

Equations (6.14 and 6.15) implies that the partial molar properties of the various species ( M i ' s are not
independent. Some key properties of partial molar properties are defined as follows:
lim M i = M i
TU
xi o

lim M i = M i
xi 1

lim M = M i
xi 1

Select additional relations among partial properties are demonstrated in the Appendix 6.1.
Based on the foregoing considerations one may define an isothermal molar property change of mixing
JN

M mix as follows:

M mix=
(T , P) M (T , P) xi M i (T , P ) ..(6.16)

Or: M mix = M xi M i = xi M i xi M i = xi ( M i M i ) ..(6.17)

(M can be = V, U, H, S, A, G)
ll

Typical examples of molar volume change of mixing for a number of binary solutions are shown in
figs. 6.1 and 6.2. Clearly then there can be substantial variation of this property depending upon the
A

nature of the constituent molecules.

 ld
or
W
Fig. 6.1 Molar volume change of mixing for solutions of cyclohexane (1) with some other C6
hydrocarbons (Source: H.C Van Ness and M. M. Abbott, Perrys Chemical Engineers Handbook (7th
ed.), McGraw Hill, 1997.
TU
--------------------------------------------------------------------------------------------------------------------------
Example 6.1
Consider a solution of two species S1/S2 at 25oC such that x1 = 0.4. If V1 = 40 x 10-6 m3/mol, find V2 .
The solution specific gravity is = 0.90, and the molecular weights of the species are 32 and 18
respectively.
(Click for solution)
JN

-------------------------------------------------------------------------------------------------------------------------
ll
A
 ld
or
W
Fig. 6.2 Property changes of mixing at 50C for 6 binary liquid systems: (a) chloroform(1)/n-
heptane(2); (b) acetone(1)/methanol(2); (c) acetone(1)/chloroform(2); (d) ethanol(1)/n-heptane(2);
(e) ethanol(1)/chloroform(2); ( f) ethanol(1)/water(2). (Source: H.C Van Ness and M. M. Abbott,
Perrys Chemical Engineers Handbook (7th ed.), McGraw Hill, 1997.
TU
6.2 Partial Properties for Binary Solutions
The partial property is generally not amenable to ab initio computation from theory, but may be
conveniently determined by suitably designed experiments that help obtain isothermal molar property
of mixing (eqn. 6.17). Here we illustrate a set of results that derive for a binary mixture, and which
may be applied to compute the relevant partial molar properties at any composition. Applying eqn.
JN

6.10:
=
M M 1 x1 + M 2 x2 ( at constant T & P) ..(6.18)

Or: dM = (x1dM 1 + M 1dx1 ) + ( x2 dM 2 + M 2 dx2 ) ..(6.19)

Applying the Gibbs-Duhem equation (6.14) gives:
ll

x1dM 1 + x2 dM 2 =
0 ..(6.20)

Now, since x1 + x=
2 1, dx=
1 dx2 ..(6.21)
A

Using Eqns. 6.19 6.21:

dM = M 1dx1 M 2 dx1 ..(6.22)

dM
Or: = M1 M 2 ..(6.23)
dx1
On solving eqns. 6.22 and 6.23 simultaneously:
dM
M=
1 M + x2
dx1 ..(6.24)
dM
M=
2 M x1 ..(6.25)
dx1

ld
If the molar property of the mixture M is available either from experiments or in analytical form, the
partial molar properties may be estimated by applying the last two equations. Alternately, the

or
experimental values of the mixture molar property (at a given T & P) may be plotted as a function of
x1 as shown in fig.6.3.

W
TU
Fig. 6.3. Graphical method of determination of partial molar properties for a binary solution

For determining the partial molar properties at a given concentration one may draw a tangent to the M
JN

vs. x1 curve, and M 1 and M 2 obtain as the right and left intercepts on the y-axis. This may be evident on
comparing with the equations 6.24 and 6.25.
Using eqn. 6.17, the molar property M of a mixture is also written as:
=M xM i i + M mix ..(6.26)
ll

In general the most common form of analytical relation (obtained by fitting a polynomial to the
experimentally determined values of isothermal M mix as a function of composition) is the well-known
A

Redlich-Kister equation, which for a binary solution is given by:

M mix
= x1 x2 [ A + B( x2 x1 ) + C ( x2 x1 ) 2 + ...] ..(6.27)
Where, A, B, C are temperature dependent and are determined from experimental measurements of
M mix . For many practical applications the above equation is usually truncated to include only the
terms corresponding to the parameters A and B. Fig. 6.4 shows the relation between the mixture
enthalpy, the enthalpy change of mixing, the pure component and the partial molar enthalpies for a

ld
representative binary system.

or
W
TU
Fig. 6.4. Schematic showing pure component and partial molar enthalpies for a binary solution

--------------------------------------------------------------------------------------------------------------------------
Example 6.2
The molar enthalpy of a binary solution is given by:
JN

V = 500 x1 + 1000 x2 + x1x2 (50 x1+40x2) cm3/mol. Find the expressions for: V1 , and V1 .
(Click for solution)
---------------------------------------------------------------------------------------------------------------------------

6.3 Criteria of Thermodynamic Equilibrium

The nature of thermodynamic equilibrium has been introduced in section 1.4. As we know, it involves
ll

simultaneous thermal and mechanical equilibrium within a system. Apart from these constraints, the
system must also be in a state of chemical equilibrium. In the most general sense, chemical equilibrium
A

subsumes the following restrictions: (i) the various phases that may exist within the system are in
equilibrium with each other in that there is no mass transfer of any chemical species between the
phases; (ii) all reactions occurring in the system are also equilibrated, i.e., there is no further progress
of the reaction in terms of conversion of the reactants to products. The last two criteria are better
understood in terms of the property called chemical potential, which we shall introduce in the
following section. Here we focus on the overall, general criterion that must be in obeyed for any
chemical system in equilibrium regardless of whatever species, phases or reactions that defines it.
Consider a closed system, which can be either homogeneous or heterogeneous, and which
exists in a state of thermal and mechanical equilibrium with its surroundings. However, we assume that

ld
it is not under equilibrium with respect to possible inter-phase transfer of the chemical species or
reactions between them. If the latter conditions prevail, all inter-phase transfer processes or reactive

or
transformation of species must continue to occur till the point when the system is also at chemical
equilibrium. In all real systems such changes are induced by finite gradients and are therefore are
irreversible in nature. Applying the first law equation for all such changes for the system:
dU=t
dQ + dW

W
..(6.28)
We next consider that the system is under thermal and mechanical equilibrium with the surroundings
(surr). Under such a situation:
dW = PdV t ..(6.29)
dQsurr dQ t
TU
And also: dS surr = = (note that dQsurr = dQ t for system) ..(6.30)
Tsurr T

But by the second law: dS surr + dS t 0 ..(6.31)

On combining eqns. 6.30 and 6.31 we get:
dQ t TdS t ..(6.32)
JN

Combining eqns. 6.28, 6.29 and 6.32:

dU t + PdV t TdS t 0 ..(6.33)
It follows that for all incremental changes within the system, which take it closer to the final
thermodynamic equilibrium, the property changes must satisfy the constraint imposed by eqn. 6.33. If
the changes internal to the system occur under reversible conditions, the equality sign is valid; on the
ll

other hand, for irreversible processes the inequality condition holds.

Equation 6.33 can be used to generate alternate criteria of thermodynamic equilibrium, namely:
A

(dU t )V t , S t 0 ..(6.34)

(dS t )U t ,V t 0 ..(6.35)
The other two criteria which are most apt in relation to thermodynamics of phase equilibria involve the
use of Helmholtz and Gibbs free energy. If a process takes place under the constraints of constant
temperature and volume then:
dU t d (TS t ) t 0
T ,V

ld
Or:
d (U t TS t ) t 0
T ,V

(dAt )T ,V t 0

or
..(6.36a)
And, if the process occurs under constant temperature and pressure one may write:
dU t + d ( PV t ) d (TS t ) 0
T ,P

W
Or: dG t 0 ..(6.36b)
T ,P

Equation 6.36b provides the most practical of the three versions of general criteria of approach to
equilibrium, as temperature and pressure are the most easily measurable of all the thermodynamic
properties. As we have argued at the early part of this section at total thermodynamic equilibrium not
TU
only are thermal and mechanical gradients non-existent, there can be no further change in either the
composition of any of the phases, or that of the reactive species. If there are any such incremental,
infinitesimal changes of composition variables at the state of complete equilibrium the system once
again must return to its stable state. This is exactly akin to the concept of equilibrium for mechanical
systems presented in section 1.4. Therefore one may write the following equation to characterize
JN

thermodynamic equilibrium:
dG t = 0 ..(6.37)
T ,P

In summary, therefore, eqn. 6.37 constitutes a generalized description of thermodynamic equilibrium,

which may be stated as follows:
ll

The equilibrium state of a closed system is that state for which the total Gibbs energy is a minimum
A

with respect to all possible changes at the given T and P.

This criterion of equilibrium can be employed for determination of equilibrium states of a system in
terms of its T, P and compositions. In principle, one first expresses Gt (at a given temperature and
 pressure) as a function of the numbers of moles of each chemical species present in the various phases.
Next one makes a partial differential operation on Gt with respect to moles of each species in the
phases, and sets each such differential to zero to obtain the set of values for the mole numbers that
minimizes Gt, subject to the constraints of conservation of mass. This procedure can be applied to
problems of phase, chemical-reaction equilibria; and, of course, the most complex of chemical

ld
thermodynamic problems, where the criteria of phase and chemical-reaction equilibrium are valid
simultaneously. As we will see, eqn. 6.37 forms the foundation for developing more specific criteria

or
that can help describe both phase (section 6.3) and chemical reaction equilibria (section 8.3).

6.3 The Chemical Potential

In this section we focus on the properties of partial molar Gibbs free energy, which as we observed at

W
the beginning of the chapter, is used for the description of phase and chemical reaction equlibria. The
application of eqn. 6.3 to the molar Gibbs free energy of a mixture gives:
G G
d ( nG ) = n dP + n dT + Gi dni ..(6.38)
P T , x T T , x i
TU
By definition the partial molar Gibbs free energy is termed the chemical potential of species i in the
mixture, i.e.,:

(nG )
i =
ni
T , P , n j i
..(6.39)
JN

Or: i = Gi ..(6.40)
Using the result in eqn. 5.7 the first two partial derivatives in eqn. 6.28, may be replaced by
( nV ) and ( nS ) . Eqn. 6.38 then becomes:

=d ( nG ) ( nV ) dP ( nS ) dT + i dni ..(6.41)
i
ll

With n =1 , and so ni = xi , eqn. 6.41 becomes:

=dG VdP SdT + i dxi ..(6.42)

A

We next consider the use of chemical potential for obtaining a general criterion of thermodynamic
equilibrium. Consider a closed system consisting of two phases and which are in equilibrium with
each other. These phases could be vapour and liquid, solid and liquid, solid and vapour etc. Each phase
 in the system may be treated as an open system, the interface between the two phases acting as the
boundary across which material may be transferred. Thus we can apply 6.41 to both phases
individually:

d ( nG ) ( nV ) dP ( nS ) dT + i dni

= ..(6.43)
i

ld
d ( nG ) ( nV ) dP ( nS ) dT + i dni

= ..(6.44)
i

In general we should denote the temperature and pressure of each phase also with the superscript in

or
order to distinguish them. However, for the present purpose we assume thermal and mechanical
equilibrium to prevail, i.e.

T =T

W

P = P
The total Gibbs free energy of the system changes with mass transfer between the two phases. The
change in the total Gibbs energy of the two-phase system is the sum of the changes in each phase. The
total volume and entropy of each phase is expressed by the following equations.
TU
( nV ) + ( nV )

=
nV ..(6.45)

( nS ) + ( nS )

=
And nS ..(6.46)

Summing eqns. 6.43 and 6.44 and using eqns. 6.55 and 6.46 we get:
d ( nG ) = ( nV ) dP ( nS ) dT + i dni + i dni = 0 ..(6.47)
i i
JN

Since the mass transfer of each species takes place between the two phases in question, their change of
mass in each phase must be equal and opposite: dni = dni
In which case eqn. 6.47 becomes:
d ( nG ) = ( nV ) dP ( nS ) dT + ( i i )dni = 0 ..(6.48)
i
ll

If the system is considered to be already under thermal and mechanical equilibrium no changes in
temperature and pressure may occur, then the last equation simplifies to:
A

( )dn
i
i i i =0 ..(6.49)
We note that in the above sequence of equations no explicit constraint has been placed on the
individual dni , which then are independent of each other. Thus for eqn. 6.49 to have general validity

the coefficient of each dni has to be identically zero. Thus:

i = i (i = 1, 2, N). ..(6.50)

ld
The above proof has been simplified by assuming identical temperature and pressure for each phase.
However, a more rigorous mathematical derivation of the phase equilibrium criterion leads to the result
that if a system is under thermodynamic equilibrium, the temperature and pressures of all the phases

or
are the same. If there is third phase in the system we have considered, a second equation of the type
(6.50) obtains:
i = i

W
..(6.51)
Thus, if there are a total of phases in the system, one can generalize the result as follows:
By considering successive pairs of phases, we may readily generalize to more than two phases the
equality of chemical potentials; thus for phases:
i= i= ...= i (i = 1, 2, ..N) ..(6.52)
TU
The above result allows us to advance a more general statement of the phase equilibrium criterion:

For a system under thermodynamic equilibrium, along with equality of the temperature and pressures
of all phases, the chemical potential of each species is identical across all the phases.
JN

6.4 Ideal Gas Mixtures and Liquid Solutions

We next explore the development of a quantitative definition of the chemical potential in terms of the
volumetric properties and composition of mixtures. We have observed earlier that just as ideal gas state
is a reference for real gas properties, ideal gas mixtures play the same role with respect to real gas
mixtures. Therefore, it is instructive to establish the property relations for ideal gas mixture first.
ll

Consider the constitution of an ideal gas mixture (containing N species) at a given temperature
(T) and pressure (P). To obtain n moles of the total mixture we need to bring together ni moles of each
A

N
species (n = ni ) at temperature T but at a pressure pi which corresponds to the partial pressure that
i

each species would exert in the final mixture. If Vt is the total volume of the mixture, the following set
of relations hold.
 =P nRT / V t
= pi ni RT / V t
Or pi / P =ni / n =yi

n
ni + n j
n= Hence : =
1
j
i T , P , n j i
n

ld
(nV ig ) (nRT / P) n
=Vi =
ig
= ( RT /=
P) RT / P
ni T , P ,n ji ni T , P , n j i ni T , P ,n ji ..(6.53)
But the molar volume of the i th species V ig = RT / P

or
Hence it follows:
Vi ig = V ig ..(6.54)

W
The last result indicates that the molar volume for a species does not change between its pure state and
in an ideal gas mixture at the same T & P. It may then be concluded that for an ideal gas mixture the
properties of each species are independent of that of the other ones. This may be easy to appreciate as
the concept of an ideal gas is premised on the idea that the intermolecular interaction is non-existent in
such a state. This conclusion leads to the well-known Gibbs theorem:
TU
Except for volume all other partial molar property of a species in an ideal-gas mixture is equal to the
corresponding molar property of the species as a pure ideal gas at a temperature same as that of the
mixture, but at a pressure equal to its partial pressure in the mixture.
JN

In mathematical terms : M iig (T , P) = M iig (T , pi ) ..(6.55)

As an example let us consider the case of enthalpy of an ideal gas mixture. By Gibbs theorem:
H iig (T , P ) = H iig (T , pi ) ..(6.56)
But, as the enthalpy of an ideal gas is independent of pressure it follows that:
H iig (T , pi ) = H iig (T , P )
ll

..(6.57)

It follows: H iig (T , P ) = H iig (T , P ) ..(6.58)

A

By the standard definition, the enthalpy of the mixture is:

ig
H mix (T , P ) = yi H iig (T , P ) ..(6.59)
i

Thus, using eqns. 6.56 6.59 we get:

 ig
H mix (T , P ) = yi H iig (T , P ) ..(6.60)
i

It follows: H=
ig
mix
ig
H mix yi H
= i
ig
0 ..(6.61)

Employing the same reasoning: U=

ig
mix
ig
U mix yiU iig ..(6.62)

ld
The molar entropy of mixing of ideal gas mixture, however, is not zero. As stated above, the
formation of 1 mole of mixture results from bringing together yi moles each species at T and partial
pressure pi to form a mixture at T and P, Hence for isothermal mixing, yi moles of each species goes

or
from (T, pi) to (T, P). Therefore:
Siig (T , P ) Siig (T , pi ) =
R ln( P / pi ) =
R ln yi ..(6.63)

On transposing: Si=
ig
(T , pi ) Siig (T , P ) R ln yi ..(6.64)

W
But: S ig (T , P ) = yi Siig (T , P ) ..(6.65)
i

So: S ig (T , P ) = yi Siig (T , pi ) ..(6.66)

i

Using eqns. 6.63 - 6.67, one obtains:

TU
=
S ig (T , P ) yS
i
i i
ig
(T , P ) R yi ln yi
i
..(6.67)

On applying the partial molar property operation (as given by eqn. 6.4) on 6.58, it may be shown that:
S=ig
i (T , P ) Siig (T , P ) R ln yi ..(6.68)

=
It further follows: S mix
ig
S ig (T , P ) yi Si=
ig
(T , P ) R yi ln yi ..(6.69)
JN

For Gibbs free energy relation we start from: G= H TS

For an ideal gas mixture: G=
ig
H ig TS ig ..(6.70)
Taking the partial molar property derivative: G=
i
ig
H iig TSiig ..(6.71)

On putting eqns. 6.58 and 6.59 into 6.71 we get the following relation for the chemical potential of
ll

each species in an ideal gas mixture:

Giig = iig Giig + RT ln yi ..(6.72)
A

Using eqns. 6.51 and 6.58, it may be also shown that:

=ig
Gmix yG
i
i i
ig
+ RT yi ln yi
i
..(6.73)

---------------------------------------------------------------------------------------------------------------------------
Example 6.3
What is the change in entropy when 0.6 m3 of CO2 and 0.4 m3 of N2, each at 1 bar and 25oC blend to
form a gas mixture at the same conditions? Assume ideal gases.
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------

ld
The Ideal Solution:
We have already seen that owing to the fact that pure ideal gases and mixtures are not subject to

or
intermolecular interactions the partial molar properties (apart from volume) of each species is the same
as that of the pure species at the same temperature and pressure. In other words each species sees no
difference in their environment in pure or mixed state. One can conceptually extend this idea to posit
an ideal solution behaviour which may serve as a model to which real-solution behavior can be

W
compared. Consider a solution of two liquids, say A and B. If the intermolecular interaction in the pure
species, (i.e., A-A and B-B) is equal to the cross-species interaction A-B, neither A nor B type
molecules will see any difference in their environment before and after mixing. This is in a sense the
same condition as one obtains with idea gas mixtures. Hence an identical set of ideal solution property
TU
relations may be constructed based on the model of ideal gas mixture. By convention while describing
properties of liquid solutions mole fractions yi are replaced by xi. The following relations therefore,
derive for ideal (liquid) solution properties (denoted by a superscript id):
H id = xi H i and, V id = xiVi ..(6.74)
i i

S id = xi S i R xi ln xi
JN

..(6.75)
i i

=
Hence G id x G + RT x ln x
i
i i
i
i i ..(6.76)

Lastly, iid= Giid= Gi + RT ln xi ..(6.77)

As we will see later the ideal solution model can also serve to describe the behaviour of mixtures of
ll

real gases or solids.

A

6.5 Excess Properties

Unlike for real gases (pure or mixtures) the EOS based approach to calculation of thermodynamic
properties of real liquid solutions have not proved very successful. However, as molar residual
property is defined for real gases, for real liquid solutions one may formulate a different departure
function called the molar excess property that quantify the deviation from ideal solution property. The
mathematical formalism of excess properties is, therefore, analogous to that of the residual properties.
If M represents the molar (or unit-mass) value of any extensive thermodynamic property(e.g.,
V,U, H, S, G, etc.), then an excess property M E is defined as the difference between the actual property

ld
value of a solution and the value it would have as an ideal solution at the same temperature, pressure,
and composition. Thus:
M E M M id ..(6.78)

or
The excess property bear a relationship to the property change of mixing. One may take the example of
excess Gibbs free energy to illustrate the point. Thus:
G E= G G id ..(6.79)

W

Or: G =
E
G xi Gi + RT xi ln xi ..(6.80)
i i

Thus: G =
E
Gmix RT xi ln xi ..(6.81)
i

Other relations include:

TU
HE =
H H id =
H mix ..(6.82)
S mix + R yi ln yi
SE = ..(6.83)
i

Also: G = H TS
E E

The non-ideality of real liquid solutions are depicted well by use of excess properties, especially
JN

through the behaviour of G E , H E and SE . The excess Gibbs energy is typically obtained from low

pressure vapour-liquid equilibrium data, while H E is obtained by measuring isothermal enthalpy

change of mixing. Lastly SE is derived using the following relation:
HE GE
S =
E
..(6.84)
T
ll

Fig. 6.5 shows the variation of each of the excess property as a function of liquid mole fraction for a
number of binary solutions.
A
 ld
or
W
Fig. 6.5 Excess properties at 50C for 6 binary liquid systems: (a) chloroform(1)/n-heptane(2); (b)
acetone(1)/methanol(2);(c) acetone(1)/chloroform(2); (d) ethanol(1)/n-heptane(2); (e)
ethanol(1)/chloroform(2); ( f) ethanol(1)/water(2). (Source: H.C Van Ness and M. M. Abbott, Perrys
Chemical Engineers Handbook (7th ed.), McGraw Hill, 1997.
TU
6.6 Fugacity of pure substances
It has been shown in section 6.3 that the chemical potential provides a fundamental description of
phase equilibria. As we shall further see in chapter 8, it also proves an effective tool for depicting
chemical reaction equilibria. Nevertheless, its direct usage is restricted, as it is not easy to directly
relate the chemical potential to thermodynamic properties amenable to easy experimental
JN

determination, such as the volumetric properties. The definition of a new function called fugacity,
itself related to the chemical potential, helps bridge the gap.
The concept of fugacity is advanced based on the following thermodynamic relation for an ideal gas.
For a single component closed system containing an ideal gas we have (from eqn. 5.7):
=
dG VdP SdT
ll

At constant temperature, for a pure ideal gas i the above equation reduces to:
=
dGiig V=ig
i dP =
RTdP / P RTd ln P ..(6.85)
A

Gi ig =
i (T ) + RTd ln P [Where, i (T ) is the constant of integration]
Utilizing the essential simplicity of eqn. 6.85 we apply it a real fluid but by replacing pressure with
fugacity (since it is not valid for a real fluid):
dGi = Vi dP (At const. T)

Thus, dGi = RTd ln fi

ld
Hence, Gi =
i (T ) + RTd ln fi ..(6.86)
Since fi has the units of pressure, it is often described as a fictitious pressure. It may be noted
that the definition of fugacity as provided by eqn. 6.86 is completely general in nature, and so can be

or
extended to liquids and solids as well. However, the calculation of fugacity for the latter will differ
from that for gases. This equation provides a partial definition of fi, the fugacity of pure species i.
Subtracting eqn. 6.85 form 6.86 gives:

W
fi
G G ig = GiR = RT ln
P ..(6.87)
The dimensionless ratio fi / P is termed fugacity coefficient ( ).

Thus: GiR = RT ln i ..(6.88)

Where, i fi / P
TU
..(6.89)
Clearly for an ideal gas the following relations hold:=
GiR 0;= =
fi ig P; accordingly, i 1.
GR dP

P
However, by eqn. 5.38: = ( Z 1)
RT 0
P
Thus, using the last relation in eqn. 6.78:
JN

P
dP
=
ln i (Z
0
i 1)
P
(At const. T) ..(6.90)

6.7 Fugacity-based phase equilibrium criterion for pure component system

The general criterion of thermodynamic equilibrium has been defined by eqn. 6.38. Applying it to, for
example, a vapour (V) and liquid (L) system of a pure component i we have:
ll

iV = iL ..(6.91)

However, for a pure component system: lim M i = M i .

xi 1
A

Thus: iV GiV =
GiV and iL GiL =
GiL ..(6.92)
Thus, using eqn. 6.91 and 6.92 we have:
GiV = GiL ..(6.93)
The above equation may be generalized for any other types of phases. However, the eqn. 6.93 is
rendered more easily applicable if the chemical potential is replaced by fugacity. Thus integrating eqn.
6.86 between vapour and liquid states of a pure component:
V V
L
dGi = RT
L
d ln fi ..(6.94)

ld
GiV GiL =
RT ln( fiV / fi L ) ..(6.95)

Now applying eqn. 6.93 to 6.95 it follows RT ln( fiV / fi L ) = 0

or
Or:
f=
i
V
f=
i
L
fi sat ..(6.96)
In eqn. 6.86 f i sat indicates the value for either saturated liquid or saturated vapor, this is because the

W
coexisting phases of saturated liquid and saturated vapor are in equilibrium. Since under such
condition the pressure is Pi sat , we can write:

f iV
iV =
Pi sat
TU
fi L
iL =
Pi sat

f i sat
isat =
Pi sat

Thus, employing eqn. 6.86 again, it follows:

JN

=
V
i =
i
L
isat ..(6.97)
Both eqns. 6.93, 6.96 and 6.97 represent equivalent criterion of vapor/liquid equilibrium for pure
species.
ll

6.8 Fugacity expressions for pure gases

Fugacity coefficient (and hence fugacity) of pure gases may be conveniently evaluated by applying
A

eqn. 6.80 to a volume-explicit equation of state. The truncated virial EOS is an example of the latter
type, for which the compressibility factor of pure species (i) is given by:
Bii P
Zi = 1 +
RT
 B P
Or Z i 1 = ii
RT
Bii P
Thus, on using eqn. 6.80: ln i =
RT 0
dP (at const T)

Bii P
Hence, ln i = ..(6.98)

ld
RT
Eqn. 6.80 is, however, not amenable to use for obtaining expressions using cubic EOSs. The general
equation for such purposes is relatively more involved and we derive it below.

or
--------------------------------------------------------------------------------------------------------------------------
Example 6.4
Estimate the fugacity of ethane at 122.2 K and 5 bar using the truncated virial EOS. For ethane Tc =
305.4K, Pc = 48.84 bar, = 0.099

W
(Click for solution)
-------------------------------------------------------------------------------------------------------------------------
========================================================================
Derivation of fugacity coefficient expression for cubic EOS:
TU
Starting from eqn. 5.38:
GR f 1 P RT
RT
ln =
= ln=
P RT
0
V

dP
P
..(6.99)

Now:=
VdP d ( PV ) PdV ..(6.100)

Using Eqn. 6.100 in 6.99:

JN

1 P V PV d ( PV ) V dV

ln=
RT
( PV RT )
V= RT
dV
PV =RT PV
+
V= V
PV V dV V P
= ( Z 1)
PV =
RT
d ln( PV ) +
V= V

V=
RT
dV
On simplifying:
ll

V 1 V
ln = ( Z 1) ln Z + ln
V RT V =
PdV ..(6.101)

=
Here: V molar volume at P 0,
= or V ideal gas volume
A

Therefore: V / V = Z ..(6.102)
Using eqn. 6.102 in 6.101 one arrives at:
1 V
ln = ( Z 1)
RT
V =
PdV ..(6.103)
===============================================================================
Equation 6.103 is a generalized expression for obtaining pure component fugacity from a pressure
explicit EOS. We show below the expressions for fugacity coefficients that derive on application of the
above equation to various cubic EOSs. (The reader may refer to section 2.3.3 for various forms of
cubic EOS).

ld
VdW EOS:
P (V b) a
ln = Z 1 ln
RT RTV ..(6.104)

or
RK-EOS:
a Z +B bP
ln = Z 1 ln( Z B ) ln ; B =
bRT Z RT ..(6.105)

W
SRK EOS:
a Z +B bP
ln = Z 1 ln( Z B ) ln ; B =
bRT Z RT ..(6.106)
TU
PR-EOS:
a Z + B (1 + 2) bP
ln = Z 1 ln( Z B ) ln ;B =
2 2bRT Z + B (1 2) RT
..(6.107)
----------------------------------------------------------------------------------------------------------------------------
Example 6.5
JN

Estimate the fugacity of ammonia vapor at 4.0 MPa and 321K assuming that it obeys the RK equation
of state.
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------
6.9 Generalized Correlations for the Fugacity Coefficient
ll

The generalized correlation approach which has been presented earlier for calculation compressibility
factor Z (section 2.3), and the residual enthalpy and entropy of gases (section 5.7), can also be applied
A

for computing fugacity coefficients for pure gases and gaseous mixtures. Equation (6.90) can be
rewritten in a generalized form as follows:
P P=
C Pr ; hence, dP Pc dPr
 Pr
d Pr
=
Hence, an alternately, ln (Z 1)
O
Pr
(at const. Tr) ..(6.109)

= Z 0 + Z1 , gives :
Substitution for Z
Pr Pr
d Pr dP

ln = ( Z O 1) + Z1 r ..(6.110)

ld
O
Pr O
Pr

The last equation may be written in alternative form:

ln ln 0 + ln 1
= ..(6.111)

or
Or: = ( 0 )( 1 ) ..(6.112)
Pr
d Pr
ln 0
Where, = (Z 1)
O
..(6.113)

W
O
Pr
Pr
d Pr

ln = Z 1
1
Pr
O ..(6.114)
For obtaining the values of the integrals in the last two equations, the tabulated data for Z 0 and Z 1 for
TU
various values of Tr and Pr may be used for numerical or graphical computations. Yet another
approach is due to Lee-Kesler which employsa variant of the BWR-EOS (eqn. 2.17) to evaluate the
expressions for ln 0 and ln 1 for a wide range of values of Tr and Pr and are available in the form of
extended tables. In either case the fugacity coefficient is finally computed using eqn. 6.112. (see figs
6.6 and 6.7).
JN

-------------------------------------------------------------------------------------------------------------------------
Example 6.6
Estimate the fugacity of methane at 32C and 9.28 bar. Use the generalized correlation approach.
(Click for solution)
-------------------------------------------------------------------------------------------------------------------------
ll
A
 ld
or
W
TU
Fig. 6.6 Correlation of 0 drawn from tables of Lee-Kesler (Source: AIChE J., pp. 510-527, 1975)
JN
ll
A
 ld
or
W
FIG 6.7 Correlation of 1 drawn from tables of Lee-Kesler (Source: AIChE J., pp. 510-527, 1975)
TU
The above set of equations may well be used for estimating the overall fugacity coefficient of a
gaseous mixture. As in earlier instance (sections, 2.3 and 5.7) one makes use of pseudo-critical
properties that are estimated from the following set of linear relations:
TC ,m = yT PC ,m = y P m = yii
JN

i C ,i ; i C ,i ..(2.32)
i i i

The subscript i runs over all the species present in the mixture m. The above relations allow the
mixture to be treated as single substance; the methodology outlined above for pure species may then be
used to compute the fugacity coefficient of the mixture.
ll

6.10 Expression for Fugacity of a Pure Liquid

Since the representation of liquid state by EOS is generally difficult, the calculation of fugacity of a
A

compressed (or sub-cooled) liquid is based on the saturated liquid state as a reference state. One starts
with the two generic relations (that apply to any pure real fluid i) already introduced in the section
6.7, namely:
dGi = Vi dP (at const. T)
dGi = RTd ln fi
The two equations above may be combined to yield:
Vi dP = RTd ln fi ..(6.115)
As shown (eqn. 6.96) at the saturation condition for co-existing vapour and liquid phases:

ld
f=
i
V
f=
i
L
fi sat

Thus for a compressed liquid state at a given pressure one can write for an isothermal change of
pressure from Pi sat to the compressed liquid pressure P :

or
P
Gi Gisat =Vi dP sat
..(6.116)
Pi

W
fi
Alternately, Gi Gisat =
RT ..(6.117)
fi sat

On equating the last two equations:

P
fi 1
ln
fi sat
=
RT V dP
i ..(6.118)
Pi sat
TU
Since Vi, the liquidphase molar volume, is a relatively weak function of pressure at temperatures well
below the critical temperature, one may approximate Vi Vi l , where, Vi l is the molar volume of the
saturated liquid at the temperature of interest.
Using fi sat = isat Pi sat in the integrated form of eqn. 6.108 gives:
JN

Vi l ( P Pi sat )
fi (T , P ) = isat Pi sat exp[ ] ..(6.119)
RT
The exponential term on the left side of the last equation is known as a Poynting factor. It may be
noted that the calculation of the term isat can be made based on any EOS suitable for gases at the given

temperature T and pressure Pi sat .

ll
A
----------------------------------------------------------------------------------------------------------------------------
Example 6.7
Estimate the fugacity of cyclopentane at 110 C and 275 bar. At 110 C the vapor pressure of
cyclopentane is 5.267 bar.
(Click for solution)

ld
----------------------------------------------------------------------------------------------------------------------------
6.11 Fugacity and Fugacity Coefficient of Species in Mixture

or
The definition of the fugacity of a species in a mixture is similar to the definition of the pure species
fugacity. For an ideal gas mixture the chemical potential is given by eqn. (6.63):
iig Giig + RT ln yi ..(6.63)

W
iig dGiig + RTd ln yi
Or: d= (at const. T) ..(6.120)

=
But dGi
ig
V=
i dP RTd ln P (at const. T)

Thus: d iig = RTd ln P + RTd ln yi = RTd ln( yi P ) ..(6.121)

Or: d iig = RTd ln pi (where pi = partial pressure of i th species) ..(6.122)

TU
Using the same idea implicit in eqn. 6.86, we extend eqn. 6.122 to define a similar expression for
fugacity of species in a real mixture:
d i = RTd ln fi ..(6.123)

Here, fi is the fugacity of species i in the mixture, which replaces the partial pressure pi of eqn. 6.122.
JN

As we will demonstrate later, fi is not a partial molar property, and is therefore denoted by a circumflex

rather than by an overbar, as are partial properties. On integrating eqn. 6.123 between the any two
multi-component phases and in equilibrium with each other:
i
fi

i
d i = RT d ln fi
fi
..(6.124)

( )
ll

Hence, i i =
RT ln fi / fi ..(6.125)

But by eqn. 6.42, for equilibrium: i = i

A

=
It follows, therefore, i i RT ln= (
fi / fi 0 )
Or: fi / fi = 1

That is: fi = fi ..(6.126)

Therefore, for an arbitrary number of phases in equilibrium with each other:
fi= fi= ...= fi (i = 1,2,..N) ..(6.127)
Thus, eqn. 6.126 offers a criterion of equilibrium corresponding to that provided by eqn. 6.42. The
great advantage this equivalence offers is that, the fugacity coefficients of species in a mixture can be

ld
related to the volumetric properties of the mixtures, which facilitates the solution of phase and
chemical reaction equilibria problems.
On comparing eqns. 6.122 and 6.123, the following limiting condition obtains:

or
lim fi = yi P ..(6.128)
P 0

It follows that a fugacity coefficient i (dimensionless) of a species in a real gas mixture may be

W
defined as:
fiV
iV ..(6.129)
yi P
Although the derivation above defines the fugacity coefficient with respect to gaseous mixture the
definition may be extended to represent fugacity coefficient of a species in a real liquid solution as
TU
well. Accordingly, in that case it is defined in the following manner:
fi L
i
L
..(6.130)
xi P

6.12 Relation between residual property and species fugacity coefficients in mixtures
JN

As we know, for any molar property (M), the definition of residual property M R is given by:
MR =
M M ig nM R =
nM nM ig
It follows that:
(nM R ) (nM ) (nM ig )
= ..(6.131)
ni T ,P ,n ji ni T ,P ,n ji ni T ,P ,n ji
ll

Or: M=
i
R
M i M iig ..(6.132)
A

For the specific case of molar Gibbs free energy:

GiR =Gi Giig =i iig ..(6.133)

From eqns. 6.122 and 6.123 d iig = RTd ln pi ; and d i = RTd ln fi

On integrating between the real and ideal states:

f f
=
i
RT ln= RT ln = RT ln
ig
i
i i
i
..(6.134)
p yP i i

On comparing eqns. 6.131 and 6.122:

ld
GiR = RT ln i ..(6.135)

But by definition: G R = yi GiR ..(6.136)

i

or
Thus, on inspecting 6.133 and 6.135:
(nG R / RT )
ln i =

..(6.137)
ni T , P ,n j i

W
It follows that:
GR
= yi ln i ..(6.138)
RT i

Since ln i is a partial molar property, the Gibbs-Duhem relation (eqn. 6.14) applies, that is:
TU
x d ln = 0
i
i i (at const. T and P) ..(6.139)

6.13 Expressions for gas phase i using EOS

In this section we derive the expressions for computing the fugacity coefficients for individual species
JN

in a real gas mixture. As in the case of pure real gases (section 6.8) we consider the truncated virial
EOS and the various cubic EOSs. From eqn. (6.90):
P dP
ln
= 0
( Z 1)
P
Further by eqn. (6.137):
(nG R / RT )
ll

ln i = .
ni T , P ,n j i
A

From these two relations we obtain the following equations:

nG R P dP
=
RT 0
(nZ n)
P
..(6.140)

P ( nZ n ) dP
ln i = ..(6.141)
ni T , P ,n j i P
0

ld
(nZ ) n
= Z=
Now, ; and 1
ni ni
i

(Z 1)
P dP
ln i
Hence: = ..(6.142)

or
i
0
P

Species Fugacity Coefficients (in a gas mixture) from the Virial EOS:

W
The expression for computing the mixture virial coefficient corresponds to eqn. 2.25. One may rewrite
the expression for a binary mixture as follows:
B = y1 B11 + y2 B22 + y1 y212 ..(6.143)

Where, 12 = 2B12 B11 B22 ..(6.144)

TU
nBP
Using (6.140 / 6.141): nZ= n +
RT
nBP
Or: nZ n = ..(6.145)
RT
Therefore:
nBP
n + RT
JN

(nZ ) P (nB )
Z1
= =
1+
n1 T .P ,n2 n1 RT n1 T ,n
2

T .P ,n2
..(6.146)
Using (6.148) in (6.142):
P ( nB ) P (nB )
( ) dP 1
P
ln 1 = 0 Z 1 = dP = ..(6.147)
ll

n1 T ,n2 RT n1 T ,n
1
P RT 0
2

Now putting (6.133) it may be shown that:

A

(nB )
= B11 + y2 12
2
..(6.148)
n1 T , n2

Therefore putting (6.148) in (6.147), one obtains:

 P
ln 1
= ( B11 + y2212 ) ..(6.149)
RT
In a similar manner, one may show that:
P
ln 2
= ( B22 + y1212 ) ..(6.150)
RT

ld
----------------------------------------------------------------------------------------------------------------------------
Example 6.8
For the following system compute the species fugacity coefficients for an equimolar mixture at 20 bar

or
and 500K.
Tc (K) Pc ( bar) Vc X 103 (m3/mol) Zc yi
Propane (1) 369.9 42.57 0.2 0.271 0.153 0.4

W
Pentane (2) 469.8 33.75 0.311 0.269 0.269 0.6

(Click for solution)

----------------------------------------------------------------------------------------------------------------------------
TU
Fugacity coefficient expression for species in a mixture by pressure explicit EOS:
The eqn. 6.130 is useful for application to volume explicit EOSs only. For pressure explicit EOS
(cubic EOSs) the expression for obtaining the species fugacity coefficient in a mixture is relatively
complex. For the present purpose we provide below the generic equation used for pressure explicit
EOS, while its proof is provided at the end of the present chapter.
JN

V
P RT

RT ln i = dV RT ln Z ..(6.153)
V = n
i T ,V ,n j
V

We enlist below the resultant final expressions that obtain for species fugacity in a gaseous mixture
corresponding to various forms of cubic EOSs.
For illustration we take the RK-EOS:
ll

RT a
=P ..(6.154)
V b V (V + b)
Next one writes the parameters a and b for the mixture using typical mixing rules described by
A

eqns. 2.29 to 2.31:

a y y
i j
i j ai a j (1 kij )

kij = binary interaction parameter

where, k=
ii k=
jj 0; kij (i j ) is very often also put to zero, as it is difficult to obtain.
b = yi bi

ld
i

= yii
i

Thus the RK-EOS may be written as:

or
RT
(n n i j / n 2 ) ai a j (1 kij )

i j
P
V (ni / n)bi V (V + (ni / n)bi )
i i

W
P
Using the last expression one can derive the expression for and use it in (6.153) to obtain:
ni T ,V ,n j

bi a b a Z+B bm P
i
ln= ( Z 1) ln( Z B ) + m ( i 2 i ln( );=
B
bm bm RT bm am Z RT
..(6.155)
TU
Similarly for SRK EOS:
bi a bi ai Z + B bm P
i
ln= ( Z 1) ln( Z B) + m ( 2 ln( );=
B ..(6.156)
bm bm RT bm am Z RT

For PR EOS:
JN

bi am bi ai Z + B (1 + 2) bm P
i
ln= ( Z 1) ln( Z B) + 2 ln ;=
B
bm 2 2bm RT bm am Z + B (1 2) RT
..(6.157)
ll
A
----------------------------------------------------------------------------------------------------------------------------
Example 6.9
Calculate the fugacities of ethylene and propylene in a mixture of 70 mole percent ethylene and 30
mole percent propylene at 600 K and 60 bar. Assume that the mixture follows the Redlich-Kwong
equation of state.

ld
Tc (K) Pc ( bar)
Ethylene (1) 283.1 51.17

or
Propylene (2) 365.1 46.0
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------

W
6.14 The Lewis/Randall Rule:
A simple equation for the fugacity of a species in an ideal solution follows from the following
equations. In general, for any solution: i =
i (T ) + RT ln fi ..(6.158)
Applying to ideal solution, iid =
Giid =
i (T ) + RT ln fi id ..(6.159)
TU
However we know (eqn. 6.77) that: iid= Giid= Gi + RT ln xi ..(6.160)
Also, from eqn. 6.86: Gi =
i (T ) + RT ln fi

Thus, iid =
Giid =
i (T ) + RT ln xi fi ..(6.161)

on comparing eqns. 6.158&6.160: fi id = xi fi ..(6.162)

The last relation is known as the Lewis/Randall rule, and applies to each species in an ideal solution at
JN

all conditions of temperature, pressure, and composition. It shows that the fugacity of each species in
an ideal solution is proportional to its mole fraction; the proportionality constant being the fugacity of
pure species i in the same physical state as the solution and at the same T and P.

6.15 Dependence of GE on temperature and pressure

ll

The idea of the molar excess property has been introduced in section 6.5 with the following definition:
A

ME = M Mid ..(6.163)

Typically the property M can be an intensive property such as: V, U, H, S, G, A, etc. The relations

between the molar excess properties of V, H and G is particular significant for description of non-ideal
solution thermodynamics. One may start from the generic total derivative of the Gibbs free energy for

a real solution and an ideal solution in the same way as eqns. 5.27 and 5.28:
G V H
d = dP dT ..(6.164)
RT RT RT 2

ld
One may write the same equation specifically for an ideal solution, whence:
G V
id id
H id
d = dP dT ..(6.165)
RT RT RT 2

or
Thus subtracting eqn. 6.165 from 6.164:
G V
E E
HE
d = dP dT ..(6.166)
RT RT RT 2

W
Thus we may write the following further generative relations:
V E (G E / RT )
= ..(6.167)
RT P
T

HE (G / RT )
E

= T ..(6.168)
RT T
TU
P

S E H E GE
And further: = ..(6.169)
RT RT RT
The sensitivity of the excess Gibbs free energy to changes in temperature and pressure may be
estimated to show the effect of pressure and temperature on liquid phase properties. For example, for
an equimolar mixture of benzene and cyclohexane at 298K and 1 bar are (source: J.M. Smith, H.C.
JN

Van Ness and M.M. Abbott, Introduction to Chemical Engineering Thermodynamics, 6th ed.,
McGraw-Hill, 2001):
V E = 0.65cm3 / mol
H E = 800cm3 / mol
Thus it follows:
ll

(G E / RT ) VE 0.65
= = = 2.62 x105 bar 1
P T , x RT 83.14 x 298
A

(G E / RT ) HE 800
=
=
1.08 x103 K 1
=
T P,x RT 2
83.14 x 298 2

= 1 K) G E(for P
G E (forT = 40 bar)
The above calculations suggest that to effect the same change in excess Gibbs free energy brought
about a change of 1K, one needs to change the pressure to change by about 40bar. Hence the excess
Gibbs free energy exhibits a relatively weak dependence on pressure.

6.16 The Activity Coefficient

ld
While we have defined fugacity coefficients of individual species in a liquid solution by 6.153, we may
define yet another parameter called activity coefficient in order to describe the non-ideality of a liquid
solution, especially at low to moderate system pressure.

or
We have by eqn. 6.155: Gi =
i (T ) + RT ln fi

And from eqn. 6.158: Giid =

i (T ) + RT ln xi fi

W
fi
Using the above equations: Gi G iid= G iE = RT ln ..(6.170)
xi fi

The left side of this equation is the partial excess Gibbs energy Gi E ; the dimensionless ratio fi / xi fi

appearing on the right is the activity coefficient of species i in solution, represented by the symbol i .
TU
Thus, by definition:
fi
i
xi f i

fi = xi i fi ..(6.171)

Whence, GiE / RT = ln i ..(6.172)

JN

But G E = xi GiE ..(6.173)

i

On comparing the last two equations we conclude that i is a partial molar property with respect to
G E . Thus, we have:

(nG E / RT )
ln i =
ll

ni T , P ,n j 1
..(6.174)
A

It follows from the definition of activity coefficient (eqn. 6.163) that for an ideal solution its
value is unity for all species as GE = 0. For a non-ideal solution, however, it may be either greater or
less than unity, the larger the departure from unity the greater the non-ideality of the solution. The
derivatives of the activity coefficient with respect to pressure and temperature can be correlated to the
partial molar excess volume and enthalpy respectively.

Using eqn. 6.167 and 6.168 in conjunction with 6.173/6.174 the following results are obtained:
Vi E ln i

ld
= ..(6.175)
RT P T ,x

H iE ln i
=
RT 2 T P ,x

or
..(6.176)
(G E / RT )
As we have seen in the last section that the value of the function is relatively larger
T P,x

W
(G E / RT )
compared to ; thus comparing eqns. 6.167 and 6.168 with 6.175 and 6.176
P T , x
respectively, one may conclude that the activity coefficients are far more sensitive to changes in
temperature than to changes in pressure. For this reason for phase equilibria computations at low to
TU
moderate pressures, the activity coefficients are assumed invariant with respect to pressure.

Since the activity coefficients are partial molar properties, they are related by the Gibbs Duhem

equation (at constant temperature and pressure) as follows:

x d ln
i
i i =0
..(6.177)
JN

The above equation may be used to validate or check the consistency of experimental data on

isothermal activity coefficients for a binary system. The following equation may be derived from eqn.

6.177 for this purpose (by assuming negligible effect of pressure on the liquid phase properties):
ll

x1 =1 1
x1 =0 2 dx1 = 0
ln
..(6.178)
A

Thus, if the function ln (1/ 2) is plotted over the entire range of x1, (fig. 6.8) the two areas above and
below the x-axis in the resulting curve must add up to zero, if the activity coefficients are consistent.
Representative values and the
 ld
or
W
Figure 6.8 Thermodynamic consistency tests for activity coefficients in binary mixtures.
nature of variation in the magnitude of activity coefficients is shown in fig. 6.9; they correspond to the

same systems for which excess property variations were depicted in fig. 6.5.
TU
JN
ll

Fig. 6.9 Activity coefficients at 50C for 6 binary liquid systems: (a) chloroform(1)/n-heptane(2); (b)
acetone(1)/methanol(2); (c) acetone(1)/chloroform(2); (d) ethanol(1)/n-heptane(2); (e)
A

ethanol(1)/chloroform(2); ( f) ethanol(1)/water(2). (Source: H.C Van Ness and M. M. Abbott, Perrys

Chemical Engineers Handbook (7th ed.), McGraw Hill, 1997.
6.17 Use of VLE data for generation of activity coefficient models
In general during the process design of chemical plants, especially for distillation and liquid-liquid
separation systems, it is necessary to predict the values of activity coefficients for carrying out phase
equilibrium calculations. However, experimental data on activity coefficients of various types of liquid

ld
solutions are often not available at conditions (of temperature and/or composition) of interest. In such
cases activity coefficient models - which relate GE to solution composition (at a given T&P) - are

or
useful. Most of such models are based on semi-empirical considerations. These models are all
characterized by a set of parameters that are temperature dependent. Vapour-liquid equilibrium (VLE)
data at low pressures are readily used for computing the value of such parameters. The more general
form of the VLE relationship is discussed in the next chapter. The VLE equation is essentially a

W
relation that relates the equilibrium vapour and liquid phase compositions at a given temperature and
pressure. For the purpose of demonstrating the method of generating the activity coefficient models,
we state here without proof, the simple VLE relation that applies to low pressure systems containing an
ideal gas phase, but a non-ideal liquid phase. At low pressures it is of the following form:
TU
y i P = xi i Pi sat ..(6.179)

Where, i = 1, 2N; y i , xi , P, & Pi sat denote the vapour phase compos

yi equilibrium vapour phase mole fraction of i th species

xi equilibrium liquid phase mole fraction of i th species
JN

i activity coefficeint of i th species in the liquid phase; i.e., at the given xi

Pi sat = saturation vapour pressure of i th species at the equilibrium temperature

And P = equilibrium pressure

yi P
It follows from eqn. 6.179 that: i = ..(6.180)
xi Pi sat
ll

In a typical isothermal VLE data generation experiment, the equilibrium pressure and the
corresponding vapour and liquid phase compositions are measured and the activity coefficients of each
A

species estimated at each set of liquid phase composition using eqn. 6.180. For freezing ideas let us
consider a binary system. Once the different sets of activity coefficient values for various liquid phase
compositions, ideally ranging over xi = 0 t o 1, the excess molar Gibbs free energy is computed at each
composition using the following equation (obtained by combining eqns. 6.172, and 6.173):
G E / RT = xi ln i (At const. T) ..(6.181)
i

The values of GE/RTare in turn used to obtain a set of values of a new function: (GE/x1x2RT) are
generated which is then fitted using a polynomial function of the following form:
G E / x1 x2 RT =A + Bx1 + Cx12 + ... ..(6.182)

ld
The final form of the polynomial is generally dependent on the nature of the constituent species of the
liquid phase as well as the temperature. Representative plots for this form of data reduction for
generating the (GE/x1x2RT) function are shown in figures 6.10 and 6.11.

or
W
TU
Fig. 6.10 Representative plot of experimental P-x-y, activity coefficient and molar excess Gibbs free
energy data for systems exhibiting positive deviation from Raoults Law
JN
ll
A

Fig. 6.11 Representative plot of experimental P-x-y, activity coefficient and molar excess Gibbs free
energy data for systems exhibiting negative deviation from Raoults Law
Once the values of the parameters A, B, C etc., have been calculated from an isothermal set of VLE
data, the actual function for the activity coefficients may be derived by applying eqn. 6.174.

6.18 Activity Coefficient Models

ld
In this section we present the forms of the various activity coefficient models commonly used for VLE
calculations. Such models may be divided into two major groups depending upon their applicability to

or
various types of solutions:
1. Margules/Van Laar/Regular Solution Models
2. Wilson/NRTL (Non-Random Two Liquid)/UNIQUAC (Universal Quasi Chemical) Model.
The models in Group 1 are termed as "Homogeneous Mixture" models, while those in Group 2

W
are termed "Local Composition" models. The reason for such distinction is as follows. The non-ideal
behaviour of liquid solutions derives from two sources: (i) difference in molecular size/shape of
constituent (ii) and the difference between the inter-species and intra-species molecular interaction
energies. The Group 1 models generally apply to systems in which the inter-species and intra-species
TU
molecular interaction energies differ from each other, but in a relatively moderate measure. On the
other hand type 2 models are principally useful for describing systems where the constituent molecular
species differ on account of both size/shape as well molecular interaction energies. Thus while group I
models useful for moderate deviation from ideal solution behaviour, those in group 2 represent
strongly non-ideal solutions.
JN

In case the chemical species in the solution do not differ significantly in terms of size/shape,
they tend to distribute uniformly across the entire solution volume. Accordingly Group 1 models are
based on the premise that molecules are homogeneously distributed over the solution volume in that
there is no difference between the overall macroscopic composition and microscopic (local)
composition around a single central molecule. The Group 2 models, however, account for both types of
ll

differences: in the size of the molecules of the chemical species as well as in their interaction energies.
It is reasonable to expect that owing to such differences molecular packing at the microscopic level to
A

be non-homogeneous. As a result, the local composition around a central molecule is likely to differ
from the average macroscopic composition.
All models used for predicting species activity coefficient contain about 2 3 parameters.
Models which employ a larger number of parameters improve the accuracy of prediction but at the
same time are rendered computationally difficult. Additionally a larger number of experimentally
determined values of activity coefficients are needed for fixing the values of the model parameters. In
general, these parameters are much more sensitive to variations in temperature than pressure. This has
already been highlighted while discussing the significance of eqns. 6.175 and 6.176. Thus, if
computations of the activity coefficients occur over reasonably low range of pressures (typically about

ld
10-20 bar) the model parameters may be assumed to remain invariant. The dependence of enthalpic
quantities on temperature is usually strong. Thus, once again inspecting eqn. 6.176, we may conclude

or
that, in general, the dependence of activity coefficients on temperature variations is relatively more
significant. If the temperature variations over which activity coefficients are needed exceed ~ 100C, the
effect of temperature on model parameters need to be considered. Table 6.1 presents the domain of
applicability of the various commonly used activity coefficient models. For simplicity, we provide in

W
the table 6.2 the model expressions for binary solutions and that of the corresponding activity
coefficient formulas. In table 6.3 expressions of select models for higher order liquid mixtures are
presented.

Table 6.1 Applicability of Activity Coefficient Models

TU
System Type Models

Species similar in size and shape One-constant Margules

Moderately non-ideal mixtures Two-constant Margules, Van Laar, Regular Solution
Strongly non-ideal mixtures (for Wilson, NRTL, UNIQUAC
JN

example Alcohols+Hydrocarbons)
Solutions with miscibility gap NRTL, UNIQUAC
ll
A
 ld
Table 6.2: Select Models for the Excess Gibbs Energy and Activity Coefficients for Binary Systems

Model GE/RT Binary ln 1 and ln 2

Name parameters
Two-suffix G E / RT = Ax1 x2 A ln 1 = Ax22

or
Margules
(One-constant) ln 2 = Ax12
Three- =
G E / RT x1 x2 ( A21 x1 + A21 , A12 ln 1 = x22 [ A12 + 2( A21 A12 ) x1 ]
suffix
Margules ln 2 = x12 [ A21 + 2( A12 A21 ) x2 ]

W
(Two-constant)
A12 A21 x1 x2 A21 , A12 A12 x1 2
Van Laar G E / RT = ln 1 A12 /(1 +
= ) ;
A12 x1 + A21 x A21 x2
A x
ln 2 A21 /(1 + 21 2 ) 2
=
A12 x1

Conversely:

TU
x2 ln 2 2
ln 1 (1 +
A12 = ln 2 (1
) ; A21 =
x1 ln 1

GE 12 , 21 12
Wilson = x1 ln( x1 + x2 12 ) ln = ln( x1 + x2 12 ) + x2
x1 + 12 x
RT 1
JN
12
ln = ln( x2 + x1 21 ) + x1
x1 + 12 x
2

G G
2

=
NRTL G E / RT x1 x2 21 21= + 12 12 ln 1 x22 21 G21 + 1
x1 + x2G21 x`2 + x1G1 x1 + x2G21 ( x2 +
ll
A
 Where, G
2
2

ld
exp(12 12 ); G21 =
G12 = exp(12 21 ) ln 2 x 12
= 2 12
+
x2 + x1G12 ( x1 +
1

b b
12 = 12 ; 21 = 21
RT RT

or
W
TU
JN
ll
A
Except for the NRTL model, all others are characterized by two parameters that are essentially
adjustable with variations of temperature. For the Wilson equation in the above table the parameters
are defined as follows:
V2 a V1 a
=
12 exp 12 ;=
21 exp 21 ..(6.183)

ld
V1 RT V2 RT

Here V1 and V2 are the pure component molar volumes at the temperature of the system; and a12 and a21
are constants for a given pair of components. However, as the expressions in 6.183 indicate, the

or
parameters 12 and 21 are temperature dependent. The parameters for NRTL are, , 12 , and 21 , the
latter two are temperature dependent.

W
Table 6.3 Expressions for the Molar Excess Gibbs Energy and Activity Coefficients of
Multi-component Systems obeying Wilson & NRTL Models
Name Molar excess Gibbs energy Activity coefficient for component i
Wilson E N N N x
xi ln x j ij ln i =1 ln x j ij k ki
G
=
k x j kj
j
RT i j
j
TU
N

NRTL ji G ji x j N
N

k xk kj Gkj
N
GE
= x j
ji G ji x j N x j Gij
+ N
N
ij N
G ln i =
RT i
j
ki xk
k Gkj xk
N
k
G k
x
ki k
k
k Gkj xk

A common feature of all the models are that the parameters are basically related to the infinite dilute
JN

activity coefficients for each binary. Table 6.4 provides the relevant relations for each model.

Table 6.4 Relation between the model parameters and infinite dilute activity coefficients
Model Relation between the model parameters and infinite dilute activity coefficients
Two-suffix 1 ln=
ln= 2 A
ll

Margules
Three-suffix ln 1 = A12 ; ln 2 = A21
Margules
A

Van Laar ln 1 = A12 ; ln 2 = A21

Wilson ln 1 = 1 21 ln 12 ; ln 2 = 1 12 ln 21
NRTL 21 + 12 exp( 12 ) ; ln 2 =
ln 1 = 12 + 21 exp( 21 )
----------------------------------------------------------------------------------------------------------------------------
Example 6.10
Methanol (1)-acetone (2) system is described by the Van Laar activity coefficient model. At 600C, the
model parameters= =
are A12 0.47; A21 0.78 . Estimate the activity coefficients for a solution containing

ld
10mole% of methanol.
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------

or
Regular Solution (RS) Model
This activity coefficient model is derived for solutions that show moderate deviations from ideal
solution behaviour, and for which V E and S E are both zero. The model was proposed by Scatchard and

W
Hildebrand (J. H. Hildebrand & R. L Scott. The Solubility of Non-electrolytes, Dover, 1964) and
constitutes one of few models that are derived from theory. Known as the Regular Solution Model, the
molar excess Gibbs free energy function is given by:

( x1V1 + x2V2 ) 1 2 (1 2 )
2
=
G E / RT ..(6.184)
TU
It follows that:

V1 22 (1 2 )
ln 1 =
2
..(6.185)

V2 12 (1 2 )
ln 2 =
2
..(6.186)

xiVi
where, i = volume fraction = ; Vi = molar volume of pure ith species, and Vm = xiVi
Vm
JN

Further: i = solubility parameter for ith species, which is given by:

i =
(U ) (H
i
vap
i
vap
RT ) ..(6.187)
Vi Vi

Where, U i vap , H i vap are molar internal energy, and enthalpy of vapourization of ith species at the
ll

temperature of interest.

The above relations may be extended to multi-component mixtures for which:

A

( )
2
ln i Vi 1
RT = ..(6.188)
=
j
j j ..(6.189)
 x jV j
j = ; (where, j runs over all the species) ..(6.190)
x V
j
j j

The solubility parameters for a set of select substances are provided in Appendix IV.
--------------------------------------------------------------------------------------------------------------------------

ld
Example 6.11
Use of Regular Solution Model to estimate activity coefficients for an equimolar benzene (1) /
cyclohexane (2) solution 350oK. The solubility parameters are: 1 = 9.2 (cal/cm3)1/2; 2 = 8.2

or
(cal/cm3)1/2. The molar volumes: V1L = 88 cm3/mol; V2L = 107 cm3/mol
(Click for solution)
--------------------------------------------------------------------------------------------------------------------------

W
UNIQUAC: Group Contribution Method for Evaluation of Activity Coefficients
The acronym UNIQUAC stands for Universal Quasi Chemical Equation. The principle that the
UNIQUAC method is founded on is that of group contribution. The methodology is based on the
postulate that any thermo-physical property of a pure substance is a weighted sum of the chemical
TU
groups that constitute it. Since the number of functional groups (such as
CH 3 , CH 2 , OH , COOH , etc.) are few (about 50 or so) in comparison to the vast number of
chemical species they can form, the approach based on group contribution constitutes an universal, and
indeed a versatile method of prediction of a wide variety of properties of a very large number of
substances. Properties as diverse as normal boiling point, enthalpy of vapourization, molar volume,
JN

thermal conductivity, viscosity, surface tension, etc may be estimated when no direct experimental data
are available.
While estimating any property, a molecule is decomposed into the functional groups that make
it up. Intermolecular interactions (in a pure fluid or a mixture) are then considered to be approximately
equal to the weighted sum of group-group interactions. Characteristic quantitative values representing
ll

group interactions have been obtained from regression of experimental data for binary systems. The
use of this approach for predicting a wide variety of number of physic-chemical properties has been
A

demonstrated comprehensively by Reid et. al. (J C. Reid, J. M. Prausnitz, and B. E. Polling, The
Properties of Gases and Liquids, 4th ed. McGraw Hill, 1987). A group contribution method is,
however, necessarily approximate as it is based on the simple additivity of contributions from each
 group; further it assumes the contribution made by each group is independent of the nature of the
adjacent group.
In extending the method to prediction of activity coefficients, the molar excess Gibbs free
energy is divided into two parts: combinatorial and residual. The combinatorial part accounts for non-
ideality of a mixture arising from differences in size and shape of constituent molecular species;

ld
whereas the residual part considers the difference between inter-molecular and intra-molecular
interaction energies.

or
We provide there the final UNIQUAC activity coefficient relations for a binary. It is may be
easily extended to represent ternary and higher order mixtures (J.M. Prausnitz, R.N. Lichtenthaler and
E.G. Azevedo, Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed., Prentice Hall, 1998).
The function g G E / RT is comprised of the combinatorial and residual parts as follows:

W
=
g gC + g R ..(6.191)
The combinatorial part represents pure species parameters, while the residual part contains two binary
parameters for each pair of molecules. In general for a system comprised of N molecular species, we
have the following relations:
TU
i N N

=g i xi ln x + 5i qi xi ln i
C
..(6.192)
i i

N
g R = qi xi ln i ji ..(6.193)
i j
JN

Where:
xi ri
i = N ..(6.194)
x rj
j j

xi qi
i = N
..(6.195)
x q
ll

j j
j

ri = relative molecular volume; qi = relative molecular surface area (both are pure species parameters)
A

The binary interaction parameters ji and ij which need to be regressed from binary VLE data are

temperature dependent. They are the UNIQUAC model parameters and are provided by:
 u u u u jj
ji exp ji ii
= ij exp ij
; and,= ..(6.196)
RT RT
The activity coefficient is also divided into two parts:
ln i ln iC (combinatorial ) + ln iR (residual ) ..(6.197)

ld
J J
ln iC =1 J i + ln J i 5qi 1 i + ln i
Li Li
..(6.198)
ij

or
N
ln iR = qi 1 ln si j ; ( j dummy variable)
sj
j ..(6.199)
ri
Ji = N

r x

W
j j ..(6.200)
j

qi
Li = N

q xj
j j ..(6.201)
TU
N
=si
l
l li ; (l dummy index)

The computation of the parameters J i , Li and si is based on the UNIFAC (Universal Functional Group

Activity Coefficients) method (see eqns. 6.202 and 6.203 below), which is an extension of the
UNIQUAC model. The application of the UNIFAC method is premised on the idea that a solution is
JN

composed of the sub-groups rather than molecules themselves. The groups are essentially small, self-
contained chemical units, each designated as k, for which the relative volume and surface areas are
denoted as Rk and Qk respectively. When a molecule may be constructed from more than a single set of
subgroups, the one with the least number of different subgroups is assumed as the correct one. The
values of Rk and Qk for various common subgroups are shown in table 1 of Appendix V. More
ll

exhaustive tables may be found elsewhere (J.M. Prausnitz, R.N. Lichtenthaler and E.G. Azevedo,
Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed., Prentice Hall, 1998).
A

Apart from the dependence on pure species properties species Rk and Qk, the activity
coefficients are functions of interaction between the various subgroups. As shown in table 1 of
Appendix V, a cluster of sub-groups are classified under a main group, which are primarily descriptive.
All subgroups under a main group are identical with respect to group interactions. Accordingly the
interaction between group parameters is identified with pairs of main groups under which the former
are classified. Parametric values of select pairs of main groups are provided in table 2 of Appendix V.
One of the distinct advantages of the UNIFAC model is that it is able to predict activity
coefficients more effectively in mixtures where constituent molecules have large size differences, as in
polymer solutions.

ld
In the UNIFAC method the expression for the combinatorial part of the activity coefficient is
the same as that in eqn. 6.196. However, the residual part takes a somewhat different form.
J J

or
ln iC =1 J i + ln J i 5qi 1 i + ln i
Li Li
..(6.200)

ln iR = qi 1 k ik eki ln ik (k subgroup )

W
k sk sk
..(6.201)

The terms Ji and Li are given by eqns. 6.198 and 6.199. However, the other parameters are defined as:
ri = k(i ) Rk
k
..(6.202)
qi = k(i )Qk
TU
k
..(6.203)
k(i )Qk
eki =
qi
..(6.204)
ik = emi mk
JN

m
..(6.205)
k(i ) number of k th subgroup in i th molecular species

xqe i i ki
k = i

x q j
j j
..(6.206)
ll

sk = m mk
m
..(6.207)
a
A

mk exp mk
=
T ..(6.208)

Thus, in summary the key UNIFAC parameters are RK, QK and amk. Tables containing the values of

these parameters are recorded more exhaustively elsewhere (H. K. Hansen, P. Rasmusen, Aa.
Fredenslund, M.Schiller, and J. Gmehling, IEC Research, Vol 30, pp 2352 2355, 1991; and R.C

Reid, J, M., Prausnitz, and B.E. Poling, in Properties of Gases and Liquids, 4th ed., McGraw-Hill,

1987). All the parameters in the UNIFAC approach are essentially binary and there are effectively two

parameters that need to be determined from experimental VLE data.

ld
----------------------------------------------------------------------------------------------------------------------------
Example 6.12
Use UNIFAC model to estimate activity coefficients for an equimolar n-pentane (1) /acetone (2)

or
solution 320oK.
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------

W
Appendix 6.1
Some relations among partial properties
Every equation that provides a linear relation among thermodynamic properties of a constant-
composition (single-phase) solution can be used to derive a corresponding equation relating the partial
TU
molar properties. As an example, this is demonstrated below for the property enthalpy.
One may write for a constant composition solution: H= U + PV
= nU + P(nV )
It follows that for the entire system: nH
Applying the definition of partial molar properties:
(nH ) (nU ) (nV )
= +P
JN

..(A.6.1.1)
ni T , P ,n ji ni T , P ,n ji ni T , P ,n ji

Hence: H=
i U i + PVi

Similar relations may be derived for other thermodynamic properties. Other forms of relationships are
also derivable as is illustrated below, as an example, using Gibbs free energy relations.
For an open multi-component system: d (nG ) = (nV )dP (nS )dT + Gi dni
ll

..(A.6.1.2)
i

V S
By Maxwell Relations: =
A

T P ,n P T ,n
Also for both P and n = constant, we have: d (nG ) = (nS )dT

(nG )
Or = (nS ) ..(A.6.1.3)
T P ,n
 (nG ) (nS )
Thus: =
ni T P ,n T , P ,n
j i
ni T ,P ,n ji

(nG ) (nS )
Or: =
T ni T , P ,n ji ni T ,P ,n ji
P ,n

ld
G (nS )
Hence i = =
Si ..(A.6.1.4)
T P ,n ni T ,P ,n ji

or
G (nV )
In the same manner, it may be shown that: i = =
Vi ..(A.6.1.5)
P T ,n ni T , P ,n j i

So if one writes at const n, Gi = Gi (T , P ) , it follows:

W
G G
=dGi i dP + i dT ..(A.6.1.6)
P T , xi T P , xi
Using (A.6.1.4) and (A.6.1.5) in (A.6.1.6) one obtains:
TU
=
dGi Vi dP Si dT ..(A.6.1.7)
Equation A.6.1.7 is exploited in the derivation of relations for fugacity of individual species in real
mixtures. As with the Gibbs free energy function; those given by eqns. 5.3 5.5 can be used as starting
JN

points for the corresponding relations of the form given by eqn. A.6.1.7.

Appendix 6.2
Derivation of the expression for species fugacity coefficient for pressure explicit EOS
=
dG VdP SdT
ll

=
It may be shown that: dGi Vi dP Si dT

At constant T: dGi = Vi dP ..(A.6.2.1)

A

Using (A.6.2.1) and (A.6.2.2): Vi dP = RTd ln fi ..(A.6.2.3)

But by defintion: dGi = RTd ln fi ..(A.6.2.2)
We may also write for the pure compoent: dGiig = Vi ig dP ..(A.6.2.4)
dGiig = RT ln fi ig ..(A.6.2.5)
fi
Using ( A.6.2.2 ) and ( A.6.2.5 ) and integrating: Gi Giig =
RT ln
fi ig

f

ld
Or: Gi Giig =
RT ln i
yP ..(A.6.2.6)
i
1
Equivalently, using ( A.6.2.3) and ( A.6.2.4 ) : Gi =
P
Giig
RT
0
(Vi Vi ig )dP

or
1 P RT ..(A.6.2.7)
Or: Gi =
Giig
RT 0 Vi
P
dP

Next we develop the expression for fugacity coefficient using the foregoing relations.

W
Using ( A.6.2.2 ) and ( A.6.2.3) :
ln fi Gi i
RT
P = V= = ..(A.6.2.8)
P T ,ni P T ,ni
i
T ,ni

In the last eqn. ( )n implies that all n i = constant

TU
i

V
Now: i = i
P T ,ni V T ,ni P T ,ni ..(A.6.2.9)

Putting ( A.6.2.9 ) in ( A.6.2.8 ) and rearranging:

JN

ln fi
RT 1 = i
P V V T ,ni
T ,ni
P
T ,ni
ln fi P i
RT
=
P T ,ni V T ,ni V T ,ni
ll

ln fi i
RT
V =
A

T ,ni V T ,ni
..(A.6.2.10)

For a multi-component mixture we may write the following expression for the total Helmholtz free
energy:
 At = At (T , V , n1 , n2 ,...nN )
2 At 2 At
One may write: =
V ni ni V
At At
Or: =
V ni T ,V ,n j ni V T ,n ..(A.6.2.11)

ld
T ,ni i T ,V , n
j

One can also prove that the chemical potential may be expressed in any of the following forms:
U t H t At G t
=i = = =

or
ni S ,V ,n ni S , P ,n
j i
ni T ,V ,n ni T , P ,n
j i j i j i

=
Also, since A A= t t t
(T , V , n1 ,..., nN ); t total
At At At
dAt = t dV + dT + i n t dni

W
V T ,ni T V t ,ni i T ,V ,n j
By comparison to the coresponding closed system expression:
At
PdV S dT +
dA =
t t

t
dni
i ni T ,V , n
j

At A
TU
Thus: t = = P
V T ,ni V T ,ni
Therefore, eqn. (A.6.2.11) may be written as:

i P
= (A.6.2.12)
V T ,ni ni T ,V ,n j
JN

Using ( A.6.2.10 ) and ( A.6.2.12 ) :

ln fi P (A.6.2.13)
RT
V =
T ,ni ni T ,V ,n j

ln PV ln PV ln PV
=
In general: RT RT + RT
V T ,ni V T ,ni V T ,ni
ll

ln P ln V ln P RT
= RT + RT = RT +
V T ,ni V T ,ni V T ,ni V
A

ln P RT ln PV
Thus : RT = RT
V T ,ni V V T ,ni
..(A.6.2.14)
Adding (13) and (14) one gets:
ln( f / P ) RT P
V V ln PV
V =


RT
V
i

dV =V =


V

n

T ,V ,n j
RT
V
dV
T ,ni
T ,ni i

fi fi V

RT P

RT ln RT ln = dV RT ln PV + RT ln RT

ld
P P
V ni T ,V ,n j
V =
V =
..(A.6.2.15)
f yi P
But : RT ln i = RT ln P ; since at V =
, i.e. ideal gas situation, fi =
yi P
P

or
V = V =

Thus eqn. ( A.6.2.15 ) becomes:

f V RT P
V = V n
RT ln i = dV RT ln( PV / RT )

W

yi P i T ,V ,n j

V
P RT

RT ln i = dV RT ln Z ..(A.6.2.16)
V =
i T ,V ,n j
n V
TU
Equation (A.6.2.16) is useful for deriving the analytical expression for the fugacity coefficient of
species in a mixture using typical pressure explicit cubic EOSs. However, for doing that the expression
P
for has to be obtained. This has been presented earlier in section 6.14.
ni T ,V , n
j

-------------------------------------------------------------------------------------------------------------------
JN

Assignment- Chapter 6
ll
A
Chapter 7 Vapour-Liquid Equilibria
7.1 Introduction
Both the general criterion of thermodynamic equilibrium as well as the specific condition of equality
of chemical potential of each species which hold at equilibrium was introduced in the last chapter. We

ld
now develop the detailed relationships connecting the phase variables (T, P and composition) that
originate from the concepts of chemical potential and the fugacity coefficient. The basic principle
employed in all separation processes is that under equilibrium the compositions of phases differ from

or
each other, and therefore it is possible to preferentially concentrate one species over another (or others)
in one particular phase. This feature of phase equilibria is exploited in a wide variety of process
equipments such as distillation, extraction, crystallization, etc.

W
This chapter focuses on the vapour-liquid equilibria (VLE) problem which is depicted
schematically in fig. 7.1. When a multi-component, vapour and liquid phase each (say) containing N
chemical species co-exist in thermodynamic equilibrium at a temperature T and pressure P, the phase
compositions { y1, y2, ... y N 1} and { x1, x2, ...xN 1} remain invariant with time, and are related by a unique

set of relations. If the relations are known as a function of temperature, pressure and compositions
TU
{ yi and xi } for each species, then provided some of these variables are specified the rest may be

calculated. We will derive such relations for real multi-component systems; but as in all cases of
thermodynamic modeling we take the ideal system as our starting point. Next the VLE of systems at
moderate pressures are treated. Finally the relations for VLE at high pressures are presented. In the
JN

following section we derive the relations that hold for pure component VLE before describing those
which apply to multi-component systems.
ll
A

Fig. 7.1 The VLE Problem Description

7.2 Single Component System Phase Equilibria
We start with the general criterion of equality of the chemical potential in the two phases. To
generalize the results we assume that any two types of phases and of a pure component are at
equilibrium. Thus as given by eqn. 6.50:
i = i

ld
..(6.50)

However, for a pure component the chemical potential is reduces to the pure component molar Gibbs
free energy. Therefore:

or
i = G and, i = G ..(7.1)
Thus eqn. 6.50 reduces to:
G = G ..(7.2)

W
On taking a differential:
dG = dG ..(7.3)
Using the generic relationship in eqn. 5.7 we may write in keeping with the fact that for a given
equilibrium temperature, the equilibrium pressure corresponds to the saturation vapour pressure P sat :
TU
V dP sat S dT= V dP sat S dT ..(7.4)
On rearranging:
dP sat S S S
= =
dT V V V ..(7.5)
Additionally using the second law we have:
JN

dS = dQ T
..(7.6)
And that for a constant pressure process:
dQ = dH
..(7.7)
Using eqns. (7.6) and (7.7) we obtain:
ll

H =
T S ..(7.8)
Thus, S =
H / T , and substitution in eqn.7.5 gives:
A

dP sat H
= ..(7.9)
dT TV
The last equation is called the Clapeyron equation. For the specific case of phase transition from liquid
(l) to vapor (v), it translates into:
dP sat H LV
= ..(7.10)
dT T V LV
Noting that liquid phase molar volumes are relatively much lesser than vapour phase volumes, we may
write, V LV = V V V V V v ..(7.11)
Further at low to moderate saturation pressures if we assume ideal vapour phase behaviour, then

ld
RT
Vv
P sat ..(7.12)

or
Eqn. 7.10 then becomes:
dP sat H LV
=
dT RT / P sat
Or:

W
dP sat / P sat H LV
= ..(7.13)
dT / T 2 R
d ln P sat
Whence, H LV =
R[ ] ..(7.14)
d (1/ T )
TU
This approximate equation is known as the Clausius-Clapeyron equation. The assumptions used in the
above derivations have approximate validity only at low pressures. Integrating eqn. 7.14 we have:
B
ln P sat= A ..(7.15)
T
On comparing eqns. 7.14 and 7.15, it follows that: B = H LV / R, while A is the constant of integration.
JN

These are generally regarded as constants for a given species. A plot of experimental values of lnPsat
vs. 1/T generally yields a line that is nearly straight between the triple and critical points. However, the
validity of eqn. 7.15is questionable at relatively high pressures, and certainly in the critical region.
Thus the accuracy of the Clausius-Clapeyron equation reduces at higher pressures. A modified form of
eqn. 7.15, called the Antoine Equation, has proved to be more accurate (including at higher pressures),
ll

has the following form:

B
ln P sat= A ..(7.16)
T +C
A

A, B, and C are readily available for a large number of species. Appendix VI provides values of
Antoine constants for select substances. More complex forms of equations relation temperature and
vapour pressure of pure substances have been reported in the literature, which provide even greater
accuracy. An example of such an equation is the Wagner equation, which is given by:
A + B 1.5 + C 3 + D 6
ln Prsat = , where = 1 Tr ..(7.17)
1
The constants for the Wagner equation for specific substances are available in several reference texts

ld
(see R.C. Reid, J.M. Prausnitz and B.E. Poling, Properties of Gases and Liquids, 4th ed., McGraw-
Hill, 1987).

or
7.3 Derivation of the Phase Rule
The phase rule was introduced in section 1.5 without proof. Here we develop its mathematical form
based on the tenets of solution thermodynamics and phase equilibrium criterion presented in the last

W
chapter. Consider a non-reactive system under equilibrium, with phases each containing N
independent chemical species. The degrees of freedom for the system, i.e., the number of intensive
variables that may vary independently of each other would be given by:
TU
Degrees of freedom = Total number of systemic intensive variables number of independent equations
relating all the variables.

For the system of interest here the above terms are as follows:
I. Total number of systemic intensive variables (also called the phase rule variables)= T, P and
JN

(N-1) species mole fractions for each of the phases

II. Number of independent relations connecting the phase rule variables = ( 1)N

The second relation above follows from the fact that for each of the N species one may use the
chemical potential equality relation across all phases, as described by eqn. 6.52. It follows that for
ll

each component there can be only ( 1) independent relations.

Thus the phase rule may be rewritten as:
A

F =[ 2 + ( N 1) ] [ ( 1) N ] =2 + N (7.18)

It may be noted that the actual mass of each of the species present are not considered as phase
rule variables, as they cannot influence the intensive state of the system. A special case of the phase
rule obtains for closed systems for which the initial mass for each species is fixed. Since no mass can
enter or leave the system, the extensive state of the system is rendered fixed along with the intensive
variables. Therefore, apart from the ( 1)N constraining relations involving the species chemical

potentials, there is an additional [ ( N 1) ] constraint on the mass of each species; this follows from

the fact that if a quantum of a species leaves a phase it must reappear in another or more. Thus the

ld
phase rule eqn. leads to:
F =[ 2 + ( N 1) ] [ ( 1) N ] [ ( N 1) ] =2 ..(7.19)

The above equation is known as the Duhem's theorem. It implies that for any closed system

or
formed initially from given masses of a number of chemical species, the equilibrium state is
completely determined when any two independent variables are fixed. The two independent variables
that one may choose to specify may be either intensive or extensive. However, the number of

W
independent intensive variables is given by the phase rule. Therefore, it follows that when F = 1, at
least one of the two variables must be extensive, and when F = 0, both must be extensive.
'
7.4 Description of General VLE Behavior
Before presenting the mathematical formulation of the multi-component VLE problem it is pertinent to
TU
discuss some key features of typical vapour liquid phase behaviour. The description of phase
behaviour of vapour and liquid phases co-existing under equilibrium can be complex and difficult to
visualize for systems containing a large number of chemical species. Thus, to clarify matters it is
useful to consider a binary system. The considerations for such a system may, in principle, be
generalized to understand the behaviour of multi-component systems. However, we restrict ourselves
JN

to description of VLE of multi-component systems to the corresponding mathematical formulation.

When N = 2, the phase rule yields a degree of freedom F = 4 . In the general VLE problem
since there are at least two phases, the corresponding number of independent intensive variables
become 2. In the case of extractive distillation, there are at least two liquid phases and a vapour phase,
whence the degrees of freedom reduce to 1.In the following discussion we present the phase behaviour
ll

of the simplest case of a binary VLE. The more volatile of the two is designated as component (1). The
phase rule variables for this case are: T, P, y 1 and x 1 . Graphical plots of experimentally obtained phase
A

behaviour can then be expressed as a functionof various combinations of these variables. For example
at a given temperature one may plot two curves, P vs. x 1 (or in short P-x 1 ) and P vs. y 1 . Similarly on
specifying a certain pressure, one may plot T vs. x 1 and T vs. y 1 . In addition if one fixes either x 1 or y 1
one may plot P-T diagrams. The combination of these diagrams lead to a 3-dimensional surface
involving pressure, temperature and phase compositions on the three axes. A typical plot is shown in
fig. 7.2.

ld
or
W
Fig. 7.2 Three dimensional VLE phase diagram for a binary system
TU
Consider first a case of VLE phase behaviour at a constant temperature. This is in shown as the
lens AB. The lower part of the lens corresponds to P-y 1 while upper curve is the P-x 1 plot. The
values of the vapour and liquid compositions at any system equilibrium pressure are found by drawing
a line parallel to the x1 y1 axis. The point of intersection of this line with the upper curve provides the

saturated liquid phase composition ( x1 ) and is termed the bubble pressure line; while that with the
JN

lower curve corresponds to the saturated vapour phase composition ( y1 ) and is termed the dew pressure
line. Extending the description the entire upper face of the three dimensional surface constitutes the
bubble surface. Any point above it corresponds to the state of sub-cooled liquid. In the same manner,
the lower face represents the dew surface. For any point below this face the state is that of a
superheated vapour. The line that connects the phase compositions ( x1* and y1* ) is called the tie line
ll

(shown as T 1 -T 2 ). Such tie lines may be drawn at any other pressure and the same considerations as
above are valid.
A

Similarly, if one considers isobaric plots (shown as the lens DE), for any equilibrium
temperature the vapour and liquid phase compositions are found by drawing a line that passes through
the specific temperature and is parallel to the x1 y1 axis. The upper intersection point provides the
vapour phase composition (dew temperature line), while the lower one corresponds to the liquid phase
composition (bubble temperature line). Finally if one fixes the composition at a point on the x1 y1
axis the intersection of a vertical plane through the composition point with the 3-dimension surface
yields the curve FCG. The upper part of this curve, i.e., FC corresponds to the locus of liquid phase

ld
compositions while the lower one (CG) is that of the vapour phase compositions. The point C is the
meeting point of the two curves and defines the critical point of the mixture at the composition
specified by the original vertical plane at a point on the x1 y1 axis. At the two end points of the

or
composition axis are the pure component P-T plots which terminate at the critical points, C 1 -C 2 , of the
two substances.
If one takes a series of varying isothermal or isobaric lenses, two types of plots result, which

W
are shown in fig. 7.3. Consider the P-x-y plot. The lowest lens (T X ) corresponds to the situation
already described in the last paragraph (i.e., the lens MN). However, the one at T Y corresponds to a
case for which TC1 < TY , but TC 2 > TY . Thus at this temperature the vapour and liquid phases of the pure

component 1 cannot co-exist, and hence the P-x-y plot vanishes as the composition tends to x1 1.

If one moves to a still higher temperature say T Z the P-x-y hangs at both ends of the x1 y1 as both
TU
TC1 < TZ , and TC 2 < TZ . The considerations for the isobaric lenses at PX , PY and PZ are the same as for

the isotherms, i,e., for the highest isobar, both PC1 < PZ , and PC 2 < PZ and so on.
JN
ll
A

Fig. 7.3 (a) P-x 1 y 1 diagrams for three temperatures. (b) T-x 1 -y 1 diagrams for three pressures

If one considers now a series of P-T plots they correspond to the curves shown in fig, 7.4. The lines I-J
and K-L which represent the vapour pressure-vs.-T curves for pure species 1 and 2 respectively. At
other intermediate compositions the upper and lower plots (obtained by intersection of a vertical plane
at constant point on the x1 y1 axis, and the P-T-x-y surface of fig. 7.2) constitute a curve that rounds
off at the true critical point of the mixture (as opposed to the pseudo-critical temperature

ld
or
W
Fig. 7.4 P-T diagram for various compositions
and pressure of mixtures discussed in section 2.4). The critical points of the various mixtures of the
two species thus lie along a line on the rounded edge of the surface between J and L; it is, therefore
TU
composition dependent. Each interior loop represents the P-T behavior of saturated liquid and of
saturated vapour for a mixture of constant composition; the loops differ from one composition to
another. It follows that the P-T relation for saturated liquid is different from that for saturated vapour
of the same composition. This is in contrast with the behavior of a pure species, for which the bubble
and dew lines are the same, as for I-J and K-L.
JN

The above discussion suggests that the phase behaviour of even a simple binary can be
relatively more difficult to interpret in its complete three-dimensional form. Reducing such behaviour
to two-dimensional plots enables easier visualization of the phase behaviour and understanding their
features. However, in many instances, even two dimensional plots can be more complex in nature.
Examples of such curves (as are often encountered with solvent mixtures in industrial practice) are
ll

shown schematically in the form of (P-x-y) in fig. 7.5, for systems at relatively low pressures
( 1 atm).
A
 ld
or
W
TU
Fig. 7.5 Schematic P-x-y plots showing deviation from ideal VLE behaviour

All the four systemic VLE behaviour exhibits deviation from ideal vapour-liquid systems. As
discussed in the next section ideal VLE behaviour is characterized by Raoults Law (RL). Suffice it
mention here that non-ideal VLE systems may exhibit both negative (figs. 7.5a & 7.5b) and positive
JN

deviations (figs. 7.5c and 7.5d) from that which obeys the Raoults law (RL). The RL behaviour is
typified by the dotted P-x 1 lines in each set in fig. 7.5. When the actual P-x 1 line lies below the RL-
line, the system is said to show negative deviation from RL, while if it lies above it is indicative of a
positive deviation.
From a molecular thermodynamic viewpoint, negative deviations occur if the 1-2 type
ll

(between unlike species) of molecular interactions (attraction) are stronger than 1-1 or 2-2 (between
like species) type of interaction. As a result, the molecules of the more volatile component (1) are
A

constrained by those of component 2 from transiting to the vapour phase to a greater extent than in
the case when the former is present in a pure form. This effectively translates into an equilibrium
pressure less than the RL system at the same liquid phase composition; hence the actual system
displays negative deviation in comparison to RL for which 1-1, 2-2, and 1-2 types of interactions are
more or less the same. Appreciable negative departures from P-x 1 linearity reflect strong liquid-phase
intermolecular interaction. The opposite applies to the case where a system shows a positive deviation
from RL; that is, the interaction between the unlike species of molecules is lower than that between
like molecules.
There is an additional complexity evident in figs. 7.5c and 7.5d. Considering the former we see

ld
that it is defined by two distinct types of behaviour on either sides of the point termed as the
azeotrope. At this point the P-x 1 and the P-y 1 curves converge; i.e., the two phases are identical in
composition at this point. On left side of the azeotrope x1 > y1 , while on the other side, x1 < y1. The

or
reverse situation holds for the system depicted in fig. 7.4d. Such systems are not uncommon in the
process industry and always pose a difficulty in purifying a mixture to compositions higher than the

W
azeotropic point. This is because during a distillation process when the mixture composition arrives at
the azeotropic point the two phases become identical in composition, and the liquid composition does
not alter further during evaporation.
The phase behaviour of the same systems as in fig. 7.5 is depicted as T-x-y diagrams in fig. 7.6.
In particular we refer to the figs 7.6b and 7.6d. As expected the P-x curves appear inverted on a T-x
TU
diagram. The first kind of system (fig. 7.6b) is said to show a maximum boiling point azeotrope, while
that depicted by fig. 7.5d shows a minimum boiling point azeotropic behaviour.
JN
ll
A
 ld
or
W
TU
Fig. 7.6 Schematic T-x-y plots showing deviation from ideal VLE behaviour

7.5 Raoults Law for VLE

As mentioned in the concluding part of the last section, vapour-liquid systems which are ideal in
nature display a behaviour corresponding to Raoults Law. In such a system, both the vapour and the
JN

liquid phases essentially behave as ideal mixtures. For describing the VLE for such systems we start by
applying eqn. 6.50, to the vapour and liquid phases:
iV = iL ..(7.20)
For an ideal vapor mixture by eqn. 6.72 we have:
ll

iig Giig + RT ln yi
Similarly for an ideal liquid solution eqn. 6.77 provides:
A

=
id
Giid + RT ln xi
Thus rewriting eqn. 7.20 (using expressions provided by eqns. 6.72 and 6.77):
iig iid ..(7.21)
Giig + RT ln yi =
Gi + RT ln xi ..(7.22)

Or: RT ln( yi / xi=

) Gi Giig ..(7.23)
Since effect of pressure is negligible on liquid properties we assume that:
Gil (T , P) Gil (T , Pi S ) ..(7.24)

ld
Now for the gas phase: dGiig = Vi ig dP (at const T) ..(7.25)
Pi s
) RT dP / =
Thus: Giig (T , Pi s ) Giig (T , P= P RT ln P(i s / P ) ..(7.26)

or
P

Combining eqns. 7.23, 7.24 and 7.26 gives:

RT ln( yi / xi ) =Gil (T , Pi s ) Giig (T , Pi s ) + RT ln( Pi s / P ) ..(7.27)

W
The first two terms on the RHS in equation above correspond to the Gibbs free energy of pure liquid
and vapour phases under equilibrium conditions, i.e., at (T , Pi s ) ; hence, as shown in section 7.1, these
terms equal. Therefore, it follows that:
yi / xi = Pi s / P ..(7.28)
TU
Alternately: yi P = xi Pi s ..(7.29)

Equation 7.29 is known as the Raoults Law.

It may be noted that the conditions for ideal mixture behaviour for the gas and liquid phases are
not the same in general. For the gas mixture to be ideal the pressures need to be close to atmospheric or
less. While a liquid solution is ideal if the interaction between the same molecular species is identical
JN

to that between dissimilar molecules.

The algorithms needed for generating RL phase diagrams are discussed later in this section. But
prior to that, we present examples of typical phase diagrams that obtain from the application of the
Raoults law (RL) equations. Consider again a binary system for which a representative isothermal
plot is depicted in fig. 7.7. (The more volatile of the two components is designated as component 1).
ll

The upper straight line represents the saturated liquid compositions, while the lower curve corresponds
to the saturated vapour compositions. At any pressure, the phase compositions are found at the
A

intersections of a line parallel to the x-y axis with the P-x and P-y curves. The straight line connecting
these compositions is the tie line (such as A 1 -B 1 , A 2 -B 2 , etc.). The portion of the diagram enclosed by
the P-x-y curves corresponds to the two phase region where the vapour and liquid phases co-exist. Any
point lying outside of this two-phase envelope corresponds to a state where only a single phase is
present. Now consider the point L which lies above the phase envelope. At this condition the mixture
exists as a compressed (or sub-cooled) state whose composition corresponds to x1* . Lowering the
pressure at this fixed composition eventually brings the liquid mixture to the point A 3 , where any
further reduction of pressure leads to the formation of a vapour phase whose composition is given by

ld
B 3 . Thus this point is characterized by the formation of the first bubble of vapour, and hence is termed
the bubble point, the corresponding pressure being the bubble pressure, at the given composition.

or
W
TU
Fig. 7.7 Model P-x-y plot for a system obeying Raoults Law
Next consider the point V in the above diagram. At this state the mixture is at the same overall
composition as at L, but the state is one of single phase, superheated vapour. Increasing the pressure at
JN

the same composition eventually brings the mixture to the point B 1 , where any further increase of
pressure leads to the formation of the liquid phase, whose composition is given by the point A 1 . Thus
this point is said to be the dew point corresponding to the vapour phase composition given by y1* ,
while the pressure at this point is termed the dew pressure.
We revert to the discussion on the system state at the bubble point A 3 . If one reduces the
ll

pressure progressively formation of more bubbles of vapour occurs, which coalesce and lead to the
development of a bulk vapour phase. The system eventually reaches the point B 1 where practically the
A

entire mixture exists in the vapour form; further reduction of pressure renders the mixture superheated
and finally one reaches the point V (and beyond). Let us focus on what happens as the system transits
through the two-phase region defined by the end points A 3 and B 1 . Note that the overall composition
of the system remains invariant as the pressure reduces. However, since now the original amounts of
each species need to be distributed across the co-existing vapour and liquid phases the actual
composition in each phase must change in accordance with the following mass conservation equation:
z1 ( x1* or y1* ) =
x1 L + y1V ; where, z1 = overall composition; x1 and y1 are compositions of the liquid and
vapour phases at equilibrium, and L and V are the relative amounts of moles (per mole of the original

ld
mixture) in the liquid and vapour phases, respectively (thus L + V =
1) . The liquid phase composition
progressively changes along the line A 3 to A 1 (bubble curve), while the composition of the vapour
phase in equilibrium with the liquid phase transits from B 3 to B 1 (dew curve). At each pressure between

or
the bubble pressure (P b ) and the dew pressure (P d ) the equilibrium vapour and liquid phase
compositions, as well as the relative amounts of mass in each phase are constrained by the relations:
z1 ( x1* or y1* ) = x1 L + y1V ;and L + V = 1.

W
The associated, isobaric T-x-y plots for the same system are shown in fig.7.7. As is expected
the dew temperature curve lies above the bubble temperature curve. The lens-like region corresponds
to the two-phase states of the system. At the point V the system is in a single phase, super-heated state.
TU
JN

Fig. 7.8 Model T-x-y plot for a system obeying Raoults Law
Progressive reduction of temperature brings it to the point B 3 where the first dew of liquid forms,
ll

whose composition is provided by the point A 3 . Thus this point is referred to as the dew point, and the
corresponding temperature called the dew temperature(T d ) for the given vapour phase composition
A

( y1* ). .On the other hand if one starts from the point L (single-phase, compressed liquid state) gradual
increase of temperature brings the system to A 1 , the bubble point, where the first bubble of vapour
forms. The associated temperature then is the bubble temperature (T b ) for the composition x1* . At any
other temperature intermediate to T d and T b , the system contains co-existing vapour and liquid phases
whose compositions are constrained by the same mass conservation relations provided above, i.e.,
z1 ( x1* or y1* ) =
x1 L + y1V ; and L + V =
1. Note that as in the P-x-y plot the T-x-y plots also are
characterized by horizontal tie lines that connect the compositions of the equilibrated vapour and liquid

ld
phases.
The data in figures 7.7 and 7.8 may be may be alternately displayed in the form of a y-x plot
(fig. 7.9). It shows the equilibrium vapour and liquid phase compositions in a more immediate manner.

or
Note that, as required by the phase rule, each pair of equilibrium y and x values correspond to a
different combination of equilibrium temperature and pressure.

W
TU
JN

Fig. 7.9 Model y-x plot for a system obeying Raoults Law

Multi-component VLE calculations using Raoults Law:

For generality we consider a system containing N chemical species. Then by phase rule, for a VLE
situation, the degrees of freedom are 2N, the phase rule variables being {xi},{yi}, T, and P. The
ll

Raoults Law (eqn. 7.29) provides N constraining relations connecting these variables. Thus, for
solving the VLE problem, N variables need to be specified, so that the values of the other N variables
A

may be determined. Five types of VLE computations are commonly encountered in practice. They are
enumerated below in table 7.1 first and then the relevant algorithms used are presented.
 Table 7.1 Types of VLE calculations
VLE Type Specified Variables Computed Variables
Bubble Pressure T and { xi } P and { yi }

Dew Pressure T and { yi } P and { xi }

ld
Bubble Temperature P and { xi } T and { yi }

Dew Temperature P and { yi } T and { xi }

T , P and { zi } L orV , { xi } and { yi }

or
Flash Distillation

In the above table the notations used signify the following:

W
{ xi } overall liquid phase compoition { x1 , x2 ,...xN 1}

{ yi } overall vapour phase compoition { y1 , y2 ,... yN 1}

{ zi } overall feed composition to flash vessel { z1 , z2 ,...z N 1}
L moles of liquid phase formed per mole of feed to flash vessel
TU
In summary, therefore, one specifies either T or P and either the liquid-phase or the vapor-phase
composition, thus fixing 1+(N-1) or N phase rule variables as required by the phase rule for VLE
calculation. The variants of Raoults Law (eqn. 7.29) which are used are as follows:
yi = xi Pi s / P ..(7.30)
JN

Thus:
N

x P
i
i i
s
/ P =1

Or:
N
P = xi Pi s ..(7.31)
ll

Also:
xi = yi P / Pi s
A

..(7.32)
Thus:
N

yP/P
i
i i
s
=1
Or:
N
P =1 y
i
i / Pi s ..(7.33)

For computation of vapour pressures the Antoine equation (or another suitable equation) may be used:
B

ld
ln Pi s =
Ai i ; where t =
0
K (or 0C )
t + Ci ..(7.34)
Values of Antoine constants for a select group of substances are tabulated in Appendix IV. More

or
exhaustive tabulation is available at:
http://www.eng.auburn.edu/users/drmills/mans486/Diffusion%20Tube/Antoine_coefficient_table.PDF

W
Bubble Pressure:
Given T and { xi } , to calculate P and { yi } :

a) Use eqn. 7.31 to compute P

b) Next use eqn. 7.30 to obtain { yi }
TU
Dew Pressure:
Given T and { yi } , to calculate P and { xi } :

a) Use eqn. 7.33 to compute P

b) Use eqn. 7.32 to obtain { xi }
JN

Bubble Temperature:
Given P and { xi } , to calculate T and { yi }

a) For the given pressure compute {Ti s } using the following form of Antoine eqn.

Bi
=Ti s Ci
ll

Ai ln P
N
b) Initialize the bubble temperature as: Tb (i ) = xiTi s
A

c) Using computed T calculate { Pi s }

d) Use equation 7.30 to compute { yi }

 e) Is y 1 < ?
i
i ( =pre - defined acceptable error for convergence)

f) If yes, Tlast = Tb ( f ) ; where Tb ( f ) = final acceptable bubble temperature

g) At Tb ( f ) compute final { yi }

ld


h) If yi 1 > , then Tlast > Tb ( f ) ; revise to new T as: Tnew = Tlast and return to step (c).
i
i
yi 1

or
i) If y 1 < , then the assumed T < T
i
i b( f ) ; where Tb ( f ) = final acceptable bubble temperature

y 1

W
i
Revise to new T using: Tnew = Tlast i and return to step (c).



Dew Temperature:
Given P and { yi } , to calculate T and { xi }
TU
a) For the given pressure compute {Ti s } using the following form of Antoine eqn.

Bi
=
b) Ti s Ci
Ai ln P
N
c) Initialize the dew temperature as: Td (i ) = xiTi s
JN

d) Using computed T calculate { Pi s }

e) Use equation 7.33 to compute { xi }

f) Is x 1 < ?
i
i ( =pre - defined acceptable error for convergence)
ll

g) If yes, Tlast = Td ( f ) ; where Td ( f ) = final acceptable bubble temperature

h) At Td ( f ) compute final { xi }
A

xi 1
j) If xi 1 > , then Tlast < Td ( f ) ; revise to new T as: Tnew = Tlast i and return to step (c).

i


x 1
i
k) If xi 1 < , then Tlast > Td ( f ) ; revise to new T as: Tnew = Tlast i and return to step
i


(c).

ld
----------------------------------------------------------------------------------------------------------------------------
Example 7.1

or
Consider the ternary system: Acetone (1) / Acetonitrile (2) / Nitromethane (3) for which:
2940.46 2940.46 2972.64
=
ln P1S 14.5463 =
; ln P2S 14.5463 =
; ln P3S 14.2043
t + 237.22 t + 237.22 t + 209.00

W
Pi S ( KPa ); t ( 0C ).
Calculate: (a) P, {yi} for a temperature = 80oC, x1 = 0.3, x2 = 0.3 (b) P, {xi}, for t = 70oC, y1 = 0.5, y2 =
0.3.
(Click for solution)
TU
---------------------------------------------------------------------------------------------------------------------------
Flash Distillation Calculations
This is an operation, often exploited in the chemical industry to achieve the desired enrichment of a
feedstock through a one-step distillation process. A schematic of the process is shown in fig. 7.6. A
liquid at a pressure equal to or greater than its bubble point pressure is introduced into the flash by
JN

passing it through a pressure reduction valve. The abrupt reduction in pressure "flashes" or partially
evaporates the feed liquid, which results in the formation of a vapour and a liquid stream which are
typically assumed to leave the flash vessel in equilibrium each other.
ll
A

Fig. 7.10 Schematic of Flash Distillation Process

One of the common forms of flash calculation typically involves the determination of the liquid and
vapour stream composition that results from process as well also the resultant liquid or vapour phase
mole fractions that obtains per mole of feed. Consider a system containing one mole of mixture of
chemical species with an overall composition represented by the set of mole fractions { zi } . Let L and V

ld
be the moles of liquid and vapour formed per mole of feed. The corresponding stream compositions
are denoted as { xi } and { yi } respectively. The material-balance equations are:

F = L + V =1 ..(7.35)

or
zi=
F xi L + yiV

zi = xi (1 V ) + yiV ; ..(7.36)

Now:=
K i y=
i / xi Pi s / P ..(7.37)

W
Putting xi = yi / K i and using relations in (7.44) one obtains:
zi K i
yi =
1 + V ( K i 1) ..(7.38)

y =1
TU
i
i

Using this condition in eqn. 7.38:

zi K i
1 + V (K
i 1)
=1 (i = 1, 2.N) ..(7.39)
i

Since xi = yi / K i , an alternative equation is:

JN

zi
xi = (i = 1,2,.N) ..(7.40)
1 + V ( K i 1)

zi
It follows that: 1 + V (K 1)
=1
ll

i i ..(7.41)

Subtracting eqn. 7.41 from 7.39and defining a function we get

A

zi ( K i 1)

=
1 + V ( K 1)
i
0 ..(7.42)
i
 d z ( K 1) 2
It follows that: = i i ..(7.43)
i [1 + V ( K i 1)]
2
dV

d
The derivative is always negative; in other words the relation between vs. V is monotonic, and
dV

ld
this makes for convenient application of the well-known Newton-Raphson method of solution (see
Appendix VII); this leads to the following equation for the nth iteration:

or
Vn += Vn ..(7.44)
d
1

dV n

W
d
Where, the values for and may be computed using eqns. 7.42 and 7.43 respectively.
dV
---------------------------------------------------------------------------------------------------------------------------
Example 7.2
A liquid mixture containing equimolar amounts of benzene (1) /toluene (2) and ethylbenzene (3) is
TU
flashed to conditions of T = 110oC, P = 90 kPa, determine the equilibrium mole fractions {xi} and {yi}
of the liquid and vapor phase formed and the molar fraction V of the vapor formed. Assume that
B
Raoults law applies. ln P sat ( Pa )= A
t( K ) + C
0

A B C
JN

Benzene 13.8594 2773.78 -53.08

Ethylbenzene 14.0045 3279.47 -59.95
Toluene 14.0098 3103.01 -53.36

(Click for solution)

---------------------------------------------------------------------------------------------------------------------------
ll

7.6 VLE Algorithms for Low to Moderate Pressures

The next level of complexity in VLE algorithms arise when one has to account for non-ideal behaviour
A

for both the gas and liquid phases. This may obtain at pressures away from atmospheric and if the
constituent molecules form a non-ideal liquid phase. The general approach to VLE of such system
involves correcting both sides of the Raoults law to incorporate the effect of non-ideal behaviour. If
the pressures are moderately high the truncated virial EOS may be used to describe the gas phase
behaviour, whereas the liquid phase non-ideality is defined by a suitable activity coefficient model.
The activity coefficient based approach is preferred for moderate pressures, as under such conditions
the liquid phase properties may be conveniently regarded as independent of pressure, hence only
temperature effects on the activity coefficients need be accounted for. This approach, of course, is
rendered inaccurate at relatively high pressures, where both the gas and liquid phases need to be

ld
described using fugacity coefficients derived typically from a cubic (or a higher order) EOS. This is
dealt with in the next section. Presently the VLE algorithms for low to moderate pressure range are

or
introduced. The starting point is the eqn. (6.126):
fi = fi ..(6.126)
Applying it to VLE:

W
fiV = fi L ..(6.127)
For gas phase, we use eqn. 6.129:
fiV = yii P

For liquid phase (using eqn. 6.164):

TU
fi L = xi i fi

Applying eqn. 6.127:

yii P = xi i fi ..(7.45)
From basic fugacity function for liquid phase (eqn.6.119):
Vi l ( P Pi sat )
JN

fi (T , P ) = isat Pi sat exp[ ]

RT ..(6.119)
Using eqn. 6.119 in 6.128 we may write the phase equilibria relation as:
xi i Pi S ;
yi i P = (i = 1,2, N) ..(7.46)

V L ( P Pi s )
=
Where i (i / is ) exp[ i ]
RT ..(7.47)
ll

One may show that the Pontying (exponential factor) in the last equation is usually ~ 1 for low to
moderate pressure range, hence one may write:
A

i (i / is ) ..(7.48)
For a gas mixture obeying the truncated virial EOS (by eqn. 6.98):
 Bii Pi s
isat = exp[ ] ..(7.49)
RT
Specifically for a binary using eqns. 6.149 and 6.150:
P
ln 1
= ( B11 + y2212 )
RT ..(6.149)

ld
P
=2 exp[ ( B22 + y1212 )]
RT ..(6.150)
Using the last four equations it follows that:

or
B11 ( P P1s ) + Py2212
1 =exp[ ] ..(7.50)
RT
B22 ( P P2s ) + Py1212
And 2 =exp[

W
] ..(7.51)
RT
It may be shown that for a multi-component the general expression for i is provided by:

Bii ( P Pi s ) + P yi yk ( ji jk )
1
2 j k
i =exp ..(7.52)
RT
TU
Now, i = i (T , P, y1 ,..., y N 1 )
And:
i i (T , P, x1 ,..., xN 1 ) i (T , x1 ,..., xN 1 ) ..(7.53)
The approximation made in eqn. 7.53 is a reasonable one, as at low to moderate pressures the
JN

dependence of i i on P may be neglected (as at such conditions the liquid phase properties are not

strongly pressure dependent).

The same five classes as provided in table 7.1 may be solved using this modified form of the
Raoults law. In all cases eqn. 7.46 provides the starting point for calculation, which may be re-written
is two principal alternate forms as follows:
ll

=yi xi i Pi s / i P ..(7.54)

xi= yi i P / i Pi s ..(7.55)
A

Since y i =1

x P i i i
s
/ i P =
1

Or:
=P x P i i i
s
/ i ..(7.56)

Similarly since xi = 1= y P/ P
i i i i
s
; it follows that:

=P 1/ yi i / i Pi s ..(7.57)

We may also re-write eqn. 7.46 in terms of the K-factor (as used for Raoults Law in eqn. 7.37) as

ld
follows:
yi i Pi S
K= =
xi i P ..(7.58)
i

or
Accordingly:
yi = K i xi ..(7.59)

W
Or:
yi
xi = ..(7.60)
Ki
From eqn. 7.59, it follows that:

K x =1
TU
i
i i ..(7.61)

From eqn. 7.60:

yi
K
i
=1
i

Note that when i = i = 1 eqn. 7.46 reduces to the ideal case of Raoults Law.
JN

Bubble pressure:
Given T and { xi } , to calculate P and { yi } :

a) Start with given T, { xi } , Antoine constants, (error value for convergence)

b) Set all { i } = 1.0, Evaluate { Pi s } , { i } , Calculate P using eqn. 7.56

ll

c) Calculate { yi } using eqn.7.54

A

d) Now evaluate { i } , using eqns. 7.52

e) Calculate Pnew using eqn. 7.56

f) Is P < ?
 g) If No, go to step c and calculate new { yi } with last { i }

h) If Yes, end at last P, and { yi }

Dew Point Pressure:

ld
Given T and { yi } , to calculate P and { xi }

a) Start with T and { yi } ; Antoine constants; and (error values for convergence); start with

Raoults law by setting all { i } = 1.0, and all { i } = 1.0; Evaluate { Pi s } , then calculate P using

or
eqn. 7.57; Now evaluate { xi } by eqn. 7.55; Evaluate { i } using appropriate activity coefficient

model Liquid-phase; recalculate P using eqn. (7.65), revise { i } using given { yi } and last P.

W
b) Calculate new set { xi } using eqn. 7.55

xi
c) Normalize { xi } using xi ( n ) = , and use normalized { xi } to compute { i }
xi
d) Use last { i } to calculate P by eqn. 7.57
TU
e) Is P < ?
f) If Yes then Plast = Pd ( f )

Bubble Temperature:
JN

Given P and { xi } , to calculate T and { yi }

a) Solve for T and { yi } first by assuming Raoults Law algorithm for bubble temperature

b) Using solution in a estimate { K i } using eqn. 7.58 with the given values of P and { xi } ;

latest values of T and { yi }

ll

c) Next calculate { K i xi }

d) Calculate all yi = K i xi K x
A

i i
i

e) Using normalized { yi } , recalculate { K i } and K x

i
i i

f) Has K x
i
i i changed? If yes return to step d
 g) If K x
i
i i has not changed between two successive iterations between steps c and d is

K x
i
i i = 1?

h) If yes, the last values of T and { yi K i xi } give the final bubble temperature Tb ( f ) ,and vapour

ld
compositions.

= Tlast
1
i) If no, and last K x > 1, then Tlast > Tb ( f ) ; revise to new T as: Tnew and return to
K i xi
i i

or
i
i
step (c).and return to step b.
j) If no, and last K x i i < 1, then Tlast < Tb ( f ) ; revise to new T as: Tnew = Tlast K i xi and return to

W
i i

step (c).and return to step b.

Dew Temperature:
Given P and { yi } , to calculate T and { xi }
TU
a) Solve for T and { xi } first by assuming Raoults Law algorithm for dew temperature

b) Using solution in a estimate { K i } using eqn. 7.58 with the given values of P and { yi } ;

latest values of T and { xi }

c) Next calculate { yi / K i }
JN

d) Calculate all xi = ( yi / K i ) ( y
i
i / Ki )

e) Using normalized { xi } , recalculate { K i } and yi

i / Ki

f) Has y i / K i changed? If yes return to step d

ll

g) If y
i
i / K i has not changed between two successive iterations between steps c and d is
A

yi
i / Ki = 1 ?

h) If yes, the last values of T and { xi yi / K i } give the final dew temperature Td ( f ) ,and liquid

phase compositions.
 i) If no, and yi / K i > 1 , then Tlast < Td ( f ) ; revise to new T as: Tnew = Tlast xi / K i and return to
i i

step (b).
j) If no, and yi / K i < 1 , then Tlast > Td ( f ) ; revise to new T as: Tnew = Tlast xi / K i and return to
i i

ld
step (b).
---------------------------------------------------------------------------------------------------------------------------
Example 7.3

or
Methanol (1)-acetone (2) forms an azeotrope at 760 Torr with x1 = 0.2, T = 55.70C. Using van Laar
model predict the bubble pressure for a system with for x1 = 0.1 at 55.70C.
log10 P1s =
8.0897 [1582.271 / (t + 239.726)];log10 P2s =
7.1171 [1210.595 / (t + 229.664)]

W
Pi s (torr ); t ( 0C )

(Click for solution)

---------------------------------------------------------------------------------------------------------------------------
Example 7.4
TU
For a binary, the activity coefficients are ln 1 = Ax2 and ln 2 = Ax1 . Show that the system forms an
2 2

azeotrope when A > ln P2 / P1

s
(
s
)
(Click for solution)
---------------------------------------------------------------------------------------------------------------------------
JN

Flash Distillation Calculations

The procedure for non-ideal systems takes a form similar to that adopted for systems obeying Raoults
Law except that one needs to additionally check for existence of both liquid and vapour phases
following flash. The algorithm comprises the following steps.
a) Start with flash T, P and feed composition { zi }
ll

b) At the given T, calculate dew pressure Pd by putting { yi } = { zi }

c) Next calculate bubble pressure Pb by putting { xi } = { zi }

A

d) Is Pd < P < Pb ? If no, the vapour phase has not formed.

Pb P
e) If yes, compute { i } , { i } , and V as
Pb Pd
 f) Use { i } , { i } , to get { K i } using eqn. 7.58

g) Then use eqn. 7.42 and 7.43 to evaluate and d / dV .

h) Using Newton-Raphson method, findV
i) With last V compute { xi } using eqn. 7.40 and { yi } by eqn. 7.38

ld
j) Re-calculate { i } , { i } , and { K i } using eqn. 7.58

k) Check if the change in each parameter xi , yi , and V between steps e and j is within pre-

or
defined error values chosen for convergence.
l) If yes, then the last values of xi , yi , and V constitute the solution

m) If no, return to step f with the last values of xi , yi , and V

W
7.7 High Pressure Vapour Liquid Equilibria
At relatively high pressures the VLE relations used in the last section lose exactness especially with
respect to the activity coefficient-based approach for description of the non-ideal behaviour of the
liquid phase. This is because the assumption that the activity coefficients are weakly dependent on
TU
pressure no longer remains a realistic approximation. In addition, the gas phase P-V-T behaviour can
no longer be described by the truncated virial EOS. Under such conditions a use of a higher order EOS,
which may be applied both to the gas and liquid phase is preferred. As we have seen in chapter 2, the
cubic EOS provides just that advantage; besides they offer a reasonable balance between accuracy and
computational complexity. We start with the general criterion for phase equilibria as applied to vapour-
JN

liquid systems, given by eqn. 6.127:

fiV = fi L (i = 1,2, , N) ..(7.62)
An alternative form of the last equation results from introduction of the fugacity coefficient using eqn.
6.129 and 6.130:
ll

yiiV P = xiiL P =
( i 1, 2, N ) ..(7.63)

The last equation reduces to:

A

yiiV = xiiL ..(7.64)

VLE of pure species

For the special case of pure species i, equation 7.64 reduces to:
iV = iL ..(7.65)

If both iV and iL are expressed in terms of cubic EOS as defined by any of the eqns. 6.104 to 6.107,

for a given T one may obtain the saturation vapour pressure by means of suitable algorithm as shown
by the worked out example below.

ld
---------------------------------------------------------------------------------------------------------------------------
Example 7.5
Estimate the vapour pressure of a substance A using PR-EoS, at T = 428oK. For the substance A: TC

or
= 569.4 K, PC = 2.497 MPa, = 24.97 bar, = 0.398.
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------

W
VLE from K-value Correlations for Hydrocarbon Systems
Using eqn. 7.64 one can write, K i = yi / xi

Alternately: K i = iV / iL ..(7.66)
TU
As evident from eqns. 6.155 to 6.157, the expression for species fugacity coefficients for mixtures
described by cubic EOS are relatively complex, which in turn makes the estimation of the K-factors
difficult as iterative solutions to obtaining T, P and/or compositions are inevitable. As demonstrated in
the last section, this is true even for the fugacity and activity coefficient based formulation of the VLE
problem. The use of cubic EOS for description of fugacity coefficients of species in both phases poses
JN

additional difficulty owing to the intrinsic complexity of the expressions shown in eqns. 6.155 to
6.157.
However, in the case of VLE of light hydrocarbon mixtures a reasonable simplification may be
achieved by assuming ideal solution behaviour for both the phases. This is a relatively practical
approximation as hydrocarbons being non-polar in nature, the intermolecular interactions are generally
ll

weaker than amongst polar molecules. In effect in the case of lighter hydrocarbons (C1-C10) the
interactions between the same species and those between dissimilar species are not significantly
A

different. This forms the basis of assuming ideal solution behaviour for such system. It may be noted
that since equilibrium pressures in light hydrocarbon systems tend to be high (as they are low-
boiling) under practical conditions of distillation processes, ideal solution behaviour yields far more
accurate results than would be possible by ideal gas assumption.
 We develop next the result that obtains owing to the assumption of ideal solution behaviour.
The chemical potential of all species in an ideal solution is given by eqn. 6.77:
iid= Giid= Gi + RT ln xi ..(7.67)

For a real solution: dGi = RTd ln fi ..(7.68)

ld
At the same time for pure species at same T&P: dGi = RTd ln f i ..(7.69)
From eqns. (7.68) and (7.69) it follows that:
Gi Gi =
RT ln( fi / f i ) ..(7.70)

or
From (7.67) Giid Gi =
RT ln xi ..(7.71)
Thus from eqns. (7.70) and (7.71):

W
Gi Giid =
RT ln( fi / xi f i ) ..(7.72)

For an ideal solution LHS of (7.72) is identically zero; hence for such a solution:
fi = xi fi ..(7.73)

For a real gas mixture the fugacity coefficient i is defined by: fi = yii P
TU
In analogy, for a real solution we define i by: fi = xii P

Or: i = fi / xi P ..(7.74)

Using (7.74) for an ideal solution: iid = fi id / xi P ..(7.75)

iid xi fi /=
Using (7.73) in (7.75) it follows:= xi P f=
i /P i
JN

Thus for an ideal solution: i = i ..(7.76)

Now considering the light hydrocarbon systems, the application of eqn. 7.76 in 7.66 gives:
iL (T , P ) fi L (T , P )
=Ki = ..(7.77)
iV (T , P ) PiV (T , P)
ll

Using eqn. 6.119 we substitute for the fugacity fi L (T , P) . Thus:

A

Vi L ( P Pi sat )
fi L = Pi satisat exp[ ]
RT
Where, Vi l is the molar volume of pure species i as a saturated liquid. Thus the K-value is given by:
 Pi satisat (T , Pi sat ) Vi L ( P Pi sat )
Ki = exp[ ] ..(7.78)
PiV (T , P) RT

The advantage of eqn. 7.78 is that it is a function of the properties of the pure species only, and

ld
therefore its dependence on composition of the vapour and liquid phases is eliminated. The K-factor
then is a function of temperature and pressure alone. The terms isat and iV in eqn. 7.78 can in principle
be computed using expression provided by cubic EOS (i.e., eqns. 6.104 6.107) or a corresponding

or
expression from an higher order EOS, including the generalized correlation (section 6.9). This allows
K-factors for light hydrocarbons to be as functions of T and P.
However, it may be noted that the computation of fugacities at high pressures (and/or

W
temperatures) can potentially be rendered difficult as above the critical temperature the liquid state is
necessarily hypothetical, while at pressures higher than the saturation pressure the vapour state is
hypothetical. This is corrected for by some form of extrapolations to those hypothetical states. Various
approaches have been described in the literature (T.E. Daubert, Chemical Engineering
Thermodynamics, McGraw-Hill, 1985). The nomographs of K-factors (see figs. 7.11 and 7.12)
TU
reported by Dadyburjor (D.B. Dadyburjor, Chem. Eng. Progr., vol. 74(4), 85-86, 1978) provide an
example of one such approach.
JN
ll
A
 ld
or
W
TU
Fig. 7.11 K-factors in light hydrocarbon systems (low temperature range) [Source: Dadyburjor; D.B.,
Chem. Eng.Progr., Vol. 74 (4) pp.85-86 (1978)].
JN
ll
A
 ld
or
W
TU
JN

Fig. 7.12 K-factors in light hydrocarbon systems (high temperature range) [Source: Dadyburjor; D.B.,
Chem. Eng.Progr., Vol. 74 (4) pp.85-86 (1978)].

The nomographs may be conveniently used purpose of VLE calculations in hydrocarbon

systems as they the K-factors for each species can be estimated at a given T and P. This is done by
ll

drawing a straight line connecting the given temperature and pressure; the corresponding Ki value is
read off from the point of intersection of this line with the Ki curve for a particular species. For bubble
A

point (either T or P) calculations one uses:

yi
=
i
=
Kx
i
i i 1 ..(7.79)
 For pressure calculation: If K i xi > 1 , assumed pressure is lower than the correct value; if
i

K x
i
i i < 1 the assumed pressure is higher than the correct pressure. Thus, pressure needs to be

revised for the next step of calculation.

Similarly, for temperature calculation: if K i xi > 1 , assumed temperature is higher than the

ld

i

correct value; if K i xi < 1 the assumed temperature is lower than the correct value. Thus,
i

or
temperature needs to be revised for the next step of calculation.

On the other hand the solution for dew point calculations derives from:

W
yi
xi
=
i

=
iK
1 ..(7.80)
i

For pressure calculation: If, ( y

i
i / K i ) > 1 assumed pressure is higher than the correct value; if

( y i / K i ) > 1 the assumed pressure is lower than the correct pressure. Thus, pressure needs to
TU
i

be revised for the next step of calculation.

Similarly, for temperature calculation: if ( yi / K i ) > 1 , assumed temperature is lower than its
i

correct value; if ( yi / K i ) < 1 the assumed temperature is higher than its correct value. Thus,
i
JN

temperature needs to be revised for the next step of calculation.

The use of these equations illustrated below using an example.

---------------------------------------------------------------------------------------------------------------------------
Example 7.6
ll

A vapour mixture contains 20mol% methane (1), 30mol% ethane (2) , and rest propane (3), at 300C.
Determine the dew composition.
A

(Click for solution)

----------------------------------------------------------------------------------------------------------------------------
High Pressure VLE using cubic EOS
This constitutes a generalized approach without any simplifying assumptions such as employed for
light hydrocarbons. The governing relation thus is eqn. 7.66.
K i = iV / iL

ld
The fugacity of each species, either in vapour or liquid phase, is computed using the expressions that
apply to use of cubic EOS (eqns. 6.155 to 6.157). For relevant VLE calculations once again the eqns.
7.79 and 7.80 are employed. The steps for computing (for example) the bubble pressure are enlisted

or
below. The basic principle used for other types of standard calculations (such as discussed for low to
moderate pressure VLE systems, table 7.1) remains the same.

W
Bubble pressure algorithm:
Given T and { xi } , to calculate P and { yi }

a) Solve for P and { yi } first by assuming Raoults Law algorithm for bubble pressure

b) Using solution in a estimate { K i } using eqn. 7.66 with the given values of T and { xi } ;
TU
and the latest values of P and { yi }

c) Next calculate { K i xi } and K x

i
i i

d) Calculate all yi = K i xi K x i i
JN

e) Using normalized { yi } , recalculate { K i } and K x

i
i i

f) Has K x
i
i i changed between steps c and e? If yes return to step d

g) If K x
i
i i has not changed between two successive iterations between steps c and e is
ll

K x
i
i i = 1?

h) If yes, the last values of P and { yi K i xi } give the final bubble temperature Pb ( f ) ,and vapour
A

compositions.
 i) If no, and last K x
i
i i > 1, then Plast < Pb ( f ) ; revise to new P as: Pnew = Plast K i xi and return to
i

step (c).and return to step b.

j) If no, and last K x
i
i i < 1, then Plast > Pb ( f ) ; revise to new P as: Pnew = Plast K i xi and
i

ld
return to step (c).and return to step b.

We illustrate the above methodology by a calculation of bubble pressure for an example binary system

or
below.
----------------------------------------------------------------------------------------------------------------------------
Example 7.7

W
For the system of methane (1) and butane (2) compute the bubble pressure for a liquid phase
composition of x1 = 0.2 at a temperature of 310K, using the PR-EOS.
(Click for solution)
----------------------------------------------------------------------------------------------------------------------------
TU
7.8 Henrys Law
The solubility of gases that are sparingly soluble in solvents constitutes a special application of the
general VLE relations developed in sections 7.3 and 7.4. There are numerous real-life examples of
such situations; for example, the solubilization of oxygen in water, which sustains aqueous life.
Similarly, gases such as nitrogen, carbon dioxide, etc., display relatively low solubility (mole fraction:
JN

105 102 ) in water or many solvents of industrial interest. Further, in many such instances, the
solubility of a gas in a solvent is required at temperatures beyond the critical temperature of the gas.
Application of vapour-liquid phase equilibria relations given by Raoult's law or its modified versions
(discussed in the foregoing sections) to a solute species i (in a solvent) requires the saturation vapour
pressure Pi sat at the temperature of application. Clearly if the temperature of interest exceeds the critical
ll

temperature of the solute, the parameter Pi sat is not definable, and hence such VLE relations presented
A

in sections 7.5 and 7.6 are not appropriate in such cases.

As for any VLE problem the starting point for determining the solubility of a gaseous species i in a
liquid is the equality of the fugacity of the solute species and liquid (liq) phases:
fi gas = fi liq ..(7.81)
Using eqn. 7.45 (considering low to moderate pressures):
yii P = xi i fi ..(7.45)
Denoting the gaseous solute as 1 and the solvent as 2, one may write:
f1gas = f1liq ..(7.82)

ld
And: f2gas = f2liq ..(7.83)
Using eqn. 7.45 the last two equations may be re-written as:
y11 P = x1 1 f1

or
..(7.84)
y22 P = x2 2 f 2 ..(7.85)
If we further assume that the gas is very sparingly soluble in the solvent, the liquid phase is essentially

W
pure solvent and the following relations derive:
1 1
2 1
Therefore, for component 1 we may rewrite the eqn. 7.84 as:
TU
y11 P = x1 1 f1 ..(7.86)
Or alternately:
y11 P = x1 H1 ..(7.87)
Where:
JN

H1 = 1 f1 ..(7.88)

Equation 7.88 is termed the Henrys law, and H1 the Henrys constant, which is defined at the system

temperature. If one plots the value of f1 as a function of the gas mole fraction x1 in the solvent phase

(as shown schematically in fig. 7.13), the parameter H1 corresponds to the slope of the tangent drawn
ll

on the curve at the limiting condition of x1 0.

A
 ld
or
Fig. 7.13 Plot of f1 as a function of the gas mole fraction x1

W
Similarly for component 2 the phase equilibrium equation 7.85 may be rewritten as:
y22 P = x2 H 2 ..(7.89)

Where: H 2 = 2 f 2 ..(7.90)
TU
Since 2 1, it follows that:

Where: H 2 = f 2 ..(7.91)

Thus: f2 = x2 f 2 ..(7.92)

It may be noted that eqn. 7.92 is the same as 6.162 (section 6.15), which describes the Lewis -Randall
JN

rule. Thus when Henrys law is applicable for the solute then Lewis-Randall rule is applicable for the
solvent. Since for a system temperature T > Tc ,1 the fugacity f1 of pure liquid phase for 1 is

hypothetical, it follows that the Henrys law constant H1 (= 1 f1 ) is necessarily a hypothetical quantity

as well. Since solubility of a gas is temperature dependent, it follows that H i is also a function of
ll

temperature. The Henrys law constant for a large number of gases with water as the solvent has been
reported in the literature. For example for acetylene the value is 1350bar, for carbon dioxide 1670bar,
and for air 72950bar). Fig. 7.14 presents the value of Henrys law constant for a number of gases in
A

water as a function of temperature.

----------------------------------------------------------------------------------------------------------------------------
Example 7.8
A concentrated binary solution containing mostly species 2( butX 2 1) 2 (but x2 1) is in equilibrium

ld
with a vapor phase containing both species 1 and 2. The pressure of this two-phase system is 1 bar; the
temperature is 298.0K. Determine from the following data good estimates of x1 and y1. H1 = 200 bar;

P sat = 0.10 bar.

2

or
(Click for solution)
-----------------------------------------------------------------------------------------------------------------------

W
TU
JN
ll

Fig. 7.14 Plot of Henrys Constant vs. Temperature, [ (1 H ) = N (mole fraction) P (atm) ] [Reprinted
A

with permission from O.A. Hougen, K.M. Watson, and R.A. Ragatz (1960), Chemical Process Principles
Charts, 2nd ed., John Wiley & Sons, New York]
-------------------------------------------------------------------------------------------------------------------
Assignment- Chapter 7
Chapter 8:Chemical Reaction Equilibria
8.1 Introduction
Reaction chemistry forms the essence of chemical processes. The very distinctiveness of the chemical
industry lies in its quest for transforming less useful substances to those which are useful to modern

ld
life. The perception of old art of alchemy bordered on the magical; perhaps in todays world its role
in the form of modern chemistry is in no sense any less. Almost everything that is of use to humans is
manufactured through the route of chemical synthesis. Such reactive processes need to be

or
characterized in terms of the maximum possible yield of the desired product at any given conditions,
starting from the raw materials (i.e., reactants). The theory of chemical reactions indicates that rates of
reactions are generally enhanced by increase of temperature. However, experience shows that the

W
maximum quantum of conversion of reactants to products does not increase monotonically. Indeed for
a vast majority the maximum conversion reaches a maximum with respect to reaction temperature and
subsequently diminishes. This is shown schematically in fig. 8.1.
TU
JN

Fig. 8.1 Schematic of Equilibrium Reaction vs. Temperature

ll

The reason behind this phenomenon lies in the molecular processes that occur during a reaction.
Consider a typical reaction of the following form occurring in gas phase: A( g ) + B( g ) C ( g ) + D( g ).
A

The reaction typically begins with the reactants being brought together in a reactor. In the initial
phases, molecules of A and B collide and form reactive complexes, which are eventually converted to
the products C and D by means of molecular rearrangement. Clearly then the early phase of the
reaction process is dominated by the presence and depletion of A and B. However, as the process

1|Page
continues, the fraction of C and D in the reactor increases, which in turn enhances the likelihood of
these molecules colliding with each other and undergoing transformation into A and B. Thus, while
initially the forward reaction dominates, in time the backward reaction becomes increasingly
significant, which eventually results in the two rates becoming equal. After this point is reached the
concentrations of each species in the reactor becomes fixed and displays no further propensity to

ld
change unless propelled by any externally imposed disturbance(say, by provision of heat). Under
such a condition the reaction is said to be in a state of equilibrium. The magnitude of all measurable

or
macroscopic variables (T, P and composition) characterizing the reaction remains constant. Clearly
under the equilibrium state the percentage conversion of the reactants to products must be the
maximum possible at the given temperature and pressure. Or else the reaction would progress further
until the state of equilibrium is achieved. The principles of chemical reaction thermodynamics are

W
aimed at the prediction of this equilibrium conversion.
The reason why the equilibrium conversion itself changes with variation of temperature may be
appreciated easily. The rates of the forward and backward reactions both depend on temperature;
however, an increase in temperature will, in general, have different impacts on the rates of each. Hence
TU
the extent of conversion at which they become identical will vary with temperature; this prompts a
change in the equilibrium conversion. Reactions for which the conversion is 100% or nearly so are
termed irreversible, while for those which never attains complete conversion are essentially reversible
in nature. The fact that a maxima may occur in the conversion behaviour (fig. 8.1) suggests that for
such reactions while the forward reaction rates dominate at lower temperatures, while at higher
JN

temperatures the backward reaction may be predominant.

The choice of the reaction conditions thus depends on the maximum (or equilibrium)
conversion possible. Further, the knowledge of equilibrium conversions is essential to intensification
of a process. Finally, it also sets the limit that can never be crossed in practice regardless of the process
strategies. This forms a primary input to the determination of the economic viability of a
ll

manufacturing process. If reaction equilibria considerations suggest that the maximum possible
conversion over practical ranges of temperature is lower than that required for commercial feasibility
A

no further effort is useful in its further development. On the other hand if the absolute maximum
conversion is high then the question of optimizing the process conditions attain significance.
Exploration of the best strategy for conducting the reaction (in terms of temperature, pressure, rate
enhancement by use of catalytic aids, etc) then offers a critical challenge.

2|Page
 This chapter develops the general thermodynamic relations necessary for prediction of the
equilibrium conversion of reactions. As we shall see, as in the case of phase equilibria, the Gibbs free
energy of a reaction constitutes a fundamental property in the estimation of equilibrium conversion.
The next section presents method of depicting the conversion by the means of the reaction co-ordinate,
which is followed by estimation of the heat effects associated with all reactions. The principles of

ld
reaction equilibria are then developed.

or
8.2 Standard Enthalpy and Gibbs free energy of reaction
From the foregoing discussion it may be apparent that a chemical reaction may be carried out in
diverse ways by changing temperature, pressure, and feed composition. Each of the different
conditions would involve different conversions and heat effects. Thus there is need to define a

W
standard way of carrying out a reaction. If all reactions were carried out in the same standard
manner, it becomes possible to compare them with respect to heat effects, and equilibrium conversion
under the same conditions. In general all reactions are subject to heat effects, whether small or large. A
reaction may either release heat (exothermic) or absorb heat (endothermic). However, it is expected
TU
that the heat effect will vary with temperature. Thus, there is a need to develop general relations that
allow computation of the heat effect associated with a reaction at any temperature.
Consider a reaction of the following form:
1 A1 + 2 A2 3 A3 + 4 A4 ..(8.1)

The reactants (A 1 and A 2 ) and products (A 3 and A 4 ) may be gaseous, liquid or solid. The term i is
JN

the stoichiometric coefficient corresponding to the chemical species A i . For the purpose of
development of the reaction equilibria relations it is convenient to designate the stoichiometric
numbers of the reactants as negative, while those of the products as positive. This is to signify that
reactants are depleted in proportion to their stoichiometric numbers, while the products are formed in
proportion to their stoichiometric numbers. Consider, for example, the following gas-phase reaction:
ll

CH 4 ( g ) + 2O2 ( g ) CO2 ( g ) + 2 H 2O( g )

The stoichiometric numbers are written as follows: CH 4 =

1; O2 =
2; CO2 =
1; H 2O =
A

2.

The standard enthalpy of reaction H To0 at say at any temperature T is defined in the following

manner: it is the change in enthalpy that occurs when 1 moles of A 1 and 2 moles of B 2 in their

3|Page
standard states at temperature T convert fully to form 3 moles of A 3 and 4 moles of A 4 in their

respective standard states at the same temperature T. The standard states commonly employed are as
follows:
Gases: the pure substance in the ideal gas state at 1 bar

ld
Liquids and Solids: the pure liquid or solid at 1 bar
The conceptual schema of a standard reaction is depicted in fig. 8.1. All reactants enter and products
leave the reactor in pure component form at the same temperature T, and at their respective standard

or
states. In the literature, data on the standard enthalpy of reaction is typically reported at a

W
TU
JN

Fig. 8.2 Apparatus in which a gas-phase reaction occurs at equilibrium (van't Hoff equilibrium box)
ll

temperature of 2980K. Using the sign convention adopted above, the standard enthalpy of reaction at
A

any temperature T may be mathematically expressed as follows:

i H i0,T
H T0 =
i
..(8.2)

4|Page
Where, H i0,T is the standard state enthalpy of species i at the temperature T, and the summation is over

all the reactants and products. For example, on expansion the eqn. 8.2 takes the following form for the
reaction depicted in eqn. 8.1:
H=
0
T 3 H 3,0 T + 4 H 4,0 T 1 H1,0T 2 H 2,0 T ..(8.2)

ld
If we further consider that each molecular species i is formed from j elements each, an expression for
the standard enthalpy of formation results:
H 0fi ,T = H i0,T j H 0j ,T ..(8.3)

or
j

Where, the summation is over all j constituent elements that make up the ith molecule, H 0fi ,T is

standard state enthalpy of formation of the ith molecule at T, and H 0j ,T the standard state enthalpy of

W
the jth atomic species. If all H 0j ,T are arbitrarily set to zero as the basis of calculation then eqn. 8.3

simplifies to:
H i0,T = H 0fi ,T ..(8.4)

In such a case eqn. 1 becomes:

TU
H T0 = H
i
i
0
fi ,T ..(8.5)

Values of Standard Enthalpy of formation of select substances are shown in Appendix VIII.
For simplicity in the subsequent equations we drop the subscript T, but implicitly all terms correspond
to temperature T. Now writing H i0 in a differential form:
JN

dH i0 = CP0i dT ..(8.6)

Where CP0i is the specific heat of the ith species corresponding to its standard state. Note that since the

standard state pressure for all substances is 1 bar in terms of pressure, for gases CP0i = CPigi , while for

liquids and solids it is the actual value of the specific heat at 1 bar (CP0i = CPi ) . Since the specific heat
ll

of liquids and solids are weakly dependent on pressure, it helps write eqn. 8.6 in the general form
shown. The following summation may be applied on eqn. 8.6 to give:
A

dH
i
i i
0
= i CP0i dT
i
..(8.7)

Since each i is constant one may write:

5|Page
 d ( H
i
i i
0
) = d i H i0
i
..(8.8)

Or: d i H i0 = i CP0i dT ..(8.9)

i i

iCP0i dT =
Thus: d H 0 = CPo dT ..(8.10)

ld
i

iCPoi
Where, CPo =
i

Thus on integrating eqn. 8.10, between a datum T 0 and any T, we have:

or
T
H To =H 0o + CPo dT ..(8.11)
T0

Note that since the standard state pressure is always at 1 bar, for all species one may write the general

W
form of relation for specific heat capacity:
CPoi =Ai + BiT + CiT 2 + ... ..(8.12)

(The values of and CPoi thus are those shown in Appendix III).

Eqn. 8.12 may be substituted in eqn. 8.11 which leads to:

TU
T
H To =H 0o + (A) + (B)T + (C )T 2 + ...dT ..(8.13)
T0

Where:=
A A=
i
; B B=
; C C=
i i ; D D ; and so on.
i
i i
i
i t
i
i t

The standard enthalpy of reaction is most often reported at 2980K. Using this value as the datum, the
value of the standard heat of reaction at any other temperature can be evaluated using eqn. 8.13. As
JN

evident from eqn. 8.5 the enthalpy of a reaction may be recovered from the enthalpy of formation of
the individual species for a reaction. Values of standard enthalpy of formation for a select list of
compounds are tabulated in Appendix VIII.
In continuance of the foregoing considerations one may also define a standard Gibbs free
ll

energy change of a reaction. As we will see in the later sections, this property is essential to computing
the equilibrium constant for a reaction at any temperature. As with enthalpy of reaction (eqn. 8.2) the
A

standard Gibbs free energy change at any temperature is given by the function:

iGi0,T
GT0 =
i
..(8.14)

6|Page
Thus, GT0 is the difference between the Gibbs energies of the products and reactants when each is in

its standard state as a pure substance at the system temperature and at a fixed pressure. Thus, just as the
standard enthalpy of reaction is dependent only on temperature (the standard state pressure being fixed
by definition), so is the Gibbs free energy change of a reaction. It follows that when the temperature is

ld
fixed GT0 is independent of the reaction pressure or composition. Indeed extending the argument, one

can define any standard property change of reaction by the same expression; all being functions of
temperature alone:

or
i M i0,T
M T0 =
i
..(8.15)

Where: M U , H , S , A, G.

W
In the context of chemical reaction equlibria the relations between the standard enthalpy of reaction
and the standard Gibbs energy change of reaction is of particular significance. Using the form
described by eqn. 5.31, since any standard property change of a reaction is only temperature
dependent, one may write:

(
d Gio,T / RT )
TU
H io,T = RT 2
..(8.16)
dT
Multiplying of both sides of this equation by i and summing over all species one obtains:

d ( G o
/ RT )
i H io,T = RT 2
i i ,T

dT

( )
JN

d GT0 / RT
This may be written as: H T0 =
RT 2
..(8.17)
dT

Or:
(
d GT0 / RT ) = H 0
T
..(8.18)
2
dT RT
Now substituting eqn. 8.13 in 8.18:

( ) =
ll

d GT0 / RT
dT
1
RT 2 {
H 0o + (A) + (B )T + (C )T 2 + ... dT
}
If we know the standard Gibbs free energy change GT00 at a particular temperature T0 (typically,
A

values are reported at 2980K) the above equation may be integrated as follows:

GT0 GT0
0
1
{ }
T
=
RT RT0
T0 RT 2
H 0o + (A) + (B )T + (C )T 2 + ... dT dT

..(8.19)

7|Page
 Or finally:

GT0 GT0
0
1
{ }
T
=
T T0
T0 T2
H 0o + (A) + (B )T + (C )T 2 + ... dT dT

..(8.20)

--------------------------------------------------------------------------------------------------------------------

ld
Example 8.1
Consider the reaction: C 2 H 4 (g) + H 2 O(g) C 2 H 5 OH(g). If an equimolar mixture of ethylene and
water vapor is fed to a reactor which is maintained at 500 K and 40 bar determine the Gibbs free

or
energy of the reaction, assuming that the reaction mixture behaves like an ideal gas. Assume the
following ideal gas specific heat data: C p ig = a + bT + cT2 + dT3 + eT2 (J/mol); T(K).
Species a bx103 cx106 dx109 ex10-5

W
C2H4 20.691 205.346 99.793 18.825 -
H2O 4.196 154.565 81.076 16.813 -
C 2 H 5 OH 28.850 12.055 - - 1.006

(Click for Solution)

--------------------------------------------------------------------------------------------------------------------
TU
8.3 The Reaction Coordinate
Consider again the general chemical reaction depicted in eqn. 8.1:
1 A1 + 2 A2 = 3 A3 + 4 A4

During the progress of the reaction, at each point the extent of depletion of the reactants, and the
enhancement in the amount of product is exactly in proportion to their respective stoichiometric
JN

coefficients. Thus for any change dn i in the number of moles of the ith species for a differential
progress of the reaction one may write:
dn1 dn2 dn dn4
= = =...= 3 ..(8.21)
1 2 3 4
Since all terms are equal, they can all be set equal to a single quantity d , defined to represent the
ll

extent of reaction as follows:

dn1 dn2 dn3 dn4
= d
A

= =...= = ..(8.22)
1 2 3 4
The general relation between a differential change dn i in the number of moles of a reacting species and
d is therefore: dni = i d (i = 1,2, ...N) ..(8.23)

8|Page
 This new variable , called the reaction coordinate, describe the extent of conversion of reactants to
products for a reaction. Thus, it follows that the value of is zero at the start of the reaction. On the
other hand when = 1 , it follows that the reaction has progressed to an extent at which point each
reactant has depleted by an amount equal to its stoichiometric number of moles while each product has

ld
formed also in an amount equal to its stoichiometric number of moles. For dimensional consistency
one designates such a degree of reaction as corresponding to =
1 mole.
Now, considering that at the point where the reaction has proceeded to an arbitrary extent

or
characterized by (such that > 0 ), the number of moles of ith species is n i we obtain the following
relation:
ni
dni = i d ; where, is a dummy variable and ni 0 = initial number of moles of 'i'. Thus:

W
ni 0 0

n=
i nio + i ;(i = 1,2,...,N) ..(8.24)

Thus the total number of moles of all species corresponding to extent of reaction:

=n
= n n
i io + i
..(8.25)
TU
n n0 +
Or: = ..(8.26)
Where:
n0 = n io
..(8.27)
= i ..(8.28)
JN

ni niio + i
Thus, y=i = ..(8.29)
n no +

---------------------------------------------------------------------------------------------------------------------------
Example 8.2
Consider the following reaction: A(g) + B(g) = C(g) + 3D(g).
ll

Intially the following number of moles are introduced in the reactor. Obtain the mole fraction
expressions in terms of reaction coordinate.
A

n0, A = 2 mol, n0,B = 1 mol, n0,C = 1 mol n0,D = 4 mol

(Click for Solution)

--------------------------------------------------------------------------------------------------------------------------

9|Page
The foregoing approach may be easily extended to develop the corresponding relations for a set of
multiple, independent reactions which may occur in a thermodynamic system. In such a case each
reaction is assigned an autonomous reaction co-ordinate j (to represent the jth reaction). Further the

stoichiometric coefficient of the ith species as it appears in the jth reaction is designated by i , j . Since a

ld
species may participate in more than a single reaction, the change in the total number of moles of the
species at any point of time would be the sum of the change due each independent reaction; thus, in
general:

or
dni = j
i, j d j (i= 1,2,...N) ..(8.30)

On integrating the above equation starting from the initial number of moles nio to ni corresponding to

W
the reaction coordinate j of each reaction:
i

ni
ni 0
dni =
0
j
i, j d j (i = 1,2,...,N) ..(8.31)

Or: n=
i nio + i , j j ..(8.32)
j
TU
Summing over all species gives:

= n n +
i
i
i
io
i j
i, j d j ..(8.33)

Now: n = n and, n
i
i
i
io = n0 ..(8.34)
JN

We may interchange the order of the summation on the right side of eqn. (8.33); thus:

i j
i, j d j = i , j d j
j i
..(8.35)

Thus, using eqns. 8.34 and 8.35, eqn. 8.33 may be written as:

=
n n0 + j i
i, j j

..(8.36)
ll

In the same manner as eqn. 8.28, one may write: .

j = i, j ..(8.37)
A

n n0 + j j
Thus eqn. 8.33 becomes: = ..(8.38)
j

Using eqns. 8.32 and 8.38 one finally obtains:

10 | P a g e
 nio + i, j j

yi =
j
(i = 1,2,....,N) ..(8.39)
no + j j
j

--------------------------------------------------------------------------------------------------------------------------
Example 8.3

ld
Consider the following simultaneous reactions. Express the reaction mixture composition as function
of the reaction co-ordinates. All reactants and products are gaseous.

or
A + B = C + 3D ..(1)
A + 2B = E + 4D ..(2)
Initial number of moles: n0, A = 2 mol; n0,B = 3 mol
(Click for Solution)

W
--------------------------------------------------------------------------------------------------------------------------
8.4 Criteria for Chemical Reaction Equilibrium
The general criterion for thermodynamic equilibrium was derived in section 6.3 as:
(dG t )T , P 0 ..(6.36b)
TU
As already explained, the above equation implies that if a closed system undergoes a process of change
while being under thermal and mechanical equilibrium, for all incremental changes associated with the
compositions of each species, the total Gibbs free energy of the system would decrease. At complete
equilibrium the equality sign holds; or, in other words, the Gibbs free energy of the system corresponds
to the minimum value possible under the constraints of constant (and uniform) temperature and
JN

pressure. Since the criterion makes no assumptions as to the nature of the system in terms of the
number of species or phases, or if reactions take place between the species, it may also be applied to
determine a specific criterion for a reactive system under equilibrium.
As has been explained in the opening a paragraph of this chapter, at the initial state of a
reaction, when the reactants are brought together a state of non-equilibrium ensues as reactants begin
ll

undergoing progressive transformation to products. However, a state of equilibrium must finally attain
when the rates of forward and backward reactions equalize. Under such a condition, no further change
A

in the composition of the residual reactants or products formed occurs. However, if we consider this
particular state, we may conclude that while in a macroscopic sense the system is in a state of static
equilibrium, in the microscopic sense there is dynamic equilibrium as reactants convert to products and
vice versa. Thus the system is subject to minute fluctuations of concentrations of each species.

11 | P a g e
However, by the necessity of maintenance of the dynamic equilibrium the system always returns to the
state of stable thermodynamic equilibrium. In a macroscopic sense then the system remains under the
under equilibrium state described by eqn. 6.36b. It follows that in a reactive system at the state of
chemical equilibrium the Gibbs free energy is minimum subject to the conditions of thermal and
mechanical equilibrium.

ld
The above considerations hold regardless of the number of reactants or the reactions occurring
in the system. Since the reaction co-ordinate is the single parameter that relates the compositions of all

or
the species, the variation of the total Gibbs free energy of the system as a function of the reaction co-
ordinate may be shown schematically as in fig. 8.3; here e is the value of the reaction co-ordinate at

equilibrium.

W
TU
Fig. 8.3 Variation of system Gibbs free energy with equilibrium conversion
JN

8.5 The Equilibrium Constant of Reactions

Since chemical composition of a reactive system undergoes change during a reaction, one may use the
eqn. 6.41 for total differential of the Gibbs free energy change (for a single phase system):
d (nG ) = (nV )dP (nS )dT + dn
i i ..(6.41)
ll

For simplicity considering a single reaction occurring in a closed system one can rewrite the last
equation using eqn. 8.3:
A

d (nG ) = (nV )dP (nS )dT + d

i i ..(8.40)

( nG )
It follows that: i i = ..(8.41)
T , P

12 | P a g e
On further applying the general condition of thermodynamic equilibrium given by eqn. 6.36b it follows
that:
( nG ) G t
=
0 ..(8.42)
T , P T , P

ld
Hence by eqn. 8.41 and 8.42:

i i =0 ..(8.43)

Since the reactive system is usually a mixture one may use the eqn. 6.123:

or
i dG
d= =i RTd ln fi ; at constant T ..(6.123)
Integration of this equation at constant T from the standard state of species i to the reaction pressure:
fi
i,T = Gi,T + RT ln

W
o
..(8.44)
fi o

The ratio fi /fi o is called the activity ai of species i in the reaction mixture, i.e.:

fi
ai = ..(8.45)
fi o
TU
Thus, the preceding equation becomes: =
i ,T Gi0,T + RT ln ai ..(8.46)

Using eqns. 8.46 and 8.44 in eqn. 8.43 to eliminate i gives:

(G i
o
i,T + RT ln ai )= 0
..(8.47)
On further re-organization we have:
JN

G + RT ln ( ai ) i =
o
i i ,T 0

ln ( ai ) i =
iGio,T ..(8.48)
RT
Where, signifies the product over all species i. Alternately:

i Gi,To
( i )
i
ll

a = exp ..(8.49)
RT

( )
i
( ai )
i
A

=
fi / fi 0 =
KT ..(8.50)

On comparing eqns. 8.49 and 8.50 it follows: KT = exp i i,T

Go
..(8.51)
RT

13 | P a g e
The parameter K T is defined as the equilibrium constant for the reaction at a given temperature. Since
the standard Gibbs free energy of pure species, Gi,To , depends only on temperature, the equilibrium

constant K T is also a function of temperature alone. On the other hand, by eqn. 8.50 K T is a function of
fi , which is in turn a function of composition, temperature and pressure. Thus, it follows that since

ld
temperature fixes the equilibrium constant, any variation in the pressure of the reaction must lead to a
change of equilibrium composition subject to the constraint of K T remaining constant. Equation (8.51)
may also be written as:

or
iGio,T =
RT ln KT = GT0 ..(8.52)

GT0
ln KT = ..(8.53)
RT

W
Taking a differential of eqn. 8.53:

d ln KT
=
(
d GT0 RT ) ..(8.53)
dT dT
Now using eqn. 8.18:
TU
d ln KT H T0
= ..(8.54)
dT RT 2
On further use of eqn. 8.13:
T

d ln KT H 0 + T0 (A) + (B)T + (C )T + ...

o 2

=
dT RT 2
JN

Lastly, upon integration one obtains the following expression:

T
H 0o + (A) + (B)T + (C )T 2 + ... dT
= KT ln KT0
T0
ln ..(8.55)
RT 2
Where, KT0 is the reaction equilibrium constant at a temperature T0 .

If H T0 , is assumed independent of T (i.e. H avg , over a given range of temperature (T2 T1 ) , a simpler
ll

relationship follows from eqn. 8.54:

A

K H avg 1 1
0

ln T 2 =
..(8.55)
KT 1 R T2 T1

The above equation suggests that a plot of ln KT vs. 1/ T is expected to approximate a straight line. It

also makes possible the estimation of the equilibrium constant at a temperature given its values at

14 | P a g e
another temperature. However, eqn. 8.55 provides a more rigorous expression of the equilibrium
constant as a function of temperature.
Equation 8.54 gives an important clue to the variation of the equilibrium constant depending on
the heat effect of the reaction. Thus, if the reaction is exothermic, i.e., H T0 < 0, the equilibrium

ld
constant decreases with increasing temperature. On the other hand, if the reaction is endothermic, i.e.,
H T0 > 0, equilibrium constant increases with increasing temperature. As we shall see in the following
section, the equilibrium conversion also follows the same pattern.

or
--------------------------------------------------------------------------------------------------------------------------
Example 8.4
Consider again the reaction: C 2 H 4 (g) + H 2 O(g) C 2 H 5 OH(g). If an equimolar mixture of

W
ethylene and water vapor is fed to a reactor which is maintained at 500 K and 40 bar
determine the equilibrium constant, assuming that the reaction mixture behaves like an ideal
gas. Assume the following ideal gas specific heat data: C p ig = a + bT + cT2 + dT3 + eT2
(J/mol); T(K).
TU
Species a bx103 cx106 dx109 ex10-5
C2H4 20.691 205.346 99.793 18.825 -
H2O 4.196 154.565 81.076 16.813 -
C 2 H 5 OH 28.850 12.055 - - 1.006
JN

(Click for solution)

--------------------------------------------------------------------------------------------------------------------------
8.6 Reactions involving gaseous species
We now consider eqn. 8.50 that represents a relation that connects equilibrium composition with the
ll

equilibrium constant for a reaction. The activities ai in eqn. 8.50 contains the standard state fugacity of

each species which as described in section 8.1 is chosen as that of pure species at 1 bar pressure.
A

The assumption of such a standard state is necessarily arbitrary, and any other standard state may be
chosen. But the specific assumption of 1 bar pressure is convenient from the point of calculations.
Obviously the value of the state Gibbs free energy Gio of the species needs to correspond to that at the

15 | P a g e
standard state fugacity. In the development that follows we first consider the case of reactions where
all the species are gaseous; the case of liquids and solids as reactants are considered following that.
For a gas the standard state is the ideal-gas state of pure i at a pressure of 1 bar. Since a
gaseous species at such a pressure is considered to be in an ideal gas state its fugacity is equal to its
pressure; hence at the standard state assumed at the present, f i o = 1 bar for each species of a gas-phase

ld
reaction. Thus, the activity and hence eqn. 8.50 may be re-written as follows:
ai = fi /f i o = fi ..(8.56)

or
( f )
i
K= i ..(8.57)

For the use of eqn. 8.57, the fugacity fi must be specified in bar [or (atm)] because each fi is

W
implicitly divided by f i o 1 bar [or 1(atm)]. It follows that the equilibrium constant K T is dimensionless.
This is true also for the case of liquid and/or solid reactive species, though, as is shown later, the
standard state fugacity is not necessarily 1 bar, since for condensed phases the fugacity and pressure
need not be identical at low pressures.
By eqn. 6.129, for gaseous species, fi = i yi P. Thus eqn. 8.57 may be rewritten as:
TU
( )
i
KT = i yi P ..(8.58)

On further expanding the above equation:

i
KT = { ( ) }{ ( y ) }{ ( P ) }
i
i
i i
..(8.59)
JN

Or:
K T = K K y P ..(8.60)

Where:

K = { ( ) } i
i
..(8.61)
ll

Ky = { ( y ) } i
i
..(8.62)
A

i
{ ( P)
i
} =P i
=P ..(8.63)

An alternate from of eqn. 8.60 is:

K K y = K T P ..(8.64)

16 | P a g e
Both the terms K and K y contain the mole fraction yi of each species. As given by eqn. 8.29 or 8.39,

all the mole fractions may be expressed as a function of the reaction co-ordinate of the reaction(s).
Hence, for a reaction under equilibrium at a given temperature and pressure the only unknown in eqn.
8.64 is the equilibrium reaction co-ordinate e . An appropriate model for the fugacity coefficient

ld
(based on an EOS: virial, cubic, etc.) may be assumed depending on the pressure, and eqn. 8.64 may
then solved using suitable algorithms to yield the equilibrium mole fractions of each species. A
relatively simple equation ensues in the event the reaction gas mixture is assumed to be ideal; whence

or
i = 1. Thus, eqn. 8.64 simplifies to:

K y = K T P ..(8.65)

W
Or:
( yi ) i =

P K ..(8.66)

Yet another simplified version of eqn. 8.64 results on assuming ideal solution behavior for which (by
eqn.7.84): i = i . Thus:

{ ( ) }
TU
i
K = i ..(8.67)

This simplification renders the parameter K independent of composition. Once again a suitable model

for fugacity coefficient (using an EOS) may be used for computing each i and eqn. 8.64 solved for the

equilibrium conversion.
JN

--------------------------------------------------------------------------------------------------------------------------
Example 8.5
Consider the reaction: C 2 H 4 (g) + H 2 O(g) C 2 H 5 OH(g). If an equimolar mixture of ethylene and
water vapor is fed to a reactor which is maintained at 500 K and 40 bar determine the degree of
conversion, assuming that the reaction mixture behaves like an ideal gas. Assume the following ideal
ll

gas specific heat data: C p ig = a + bT + cT2 + dT3 + eT2 (J/mol); T(K).

Species a bx103 cx106 dx109 ex10-5
A

C2H4 20.691 205.346 99.793 18.825 -

H2O 4.196 154.565 81.076 16.813 -
C 2 H 5 OH 28.850 12.055 - - 1.006

17 | P a g e
(Click for solution)
--------------------------------------------------------------------------------------------------------------------------

8.7 Reaction equilibria for simultaneous reactions

While we have so far presented reaction equlibria for single reactions, the more common situation that

ld
obtains in industrial practice is that of multiple, simultaneous reactions. Usually this occurs due to the
presence of side reactions that take place in addition to the main, desired reaction. This leads to the

or
formation of unwanted side products, necessitating additional investments in the form of purification
processes to achieve the required purity of the product(s). An example of such simultaneously
occurring reaction is:
CH 4 ( g ) + 2O2 ( g ) CO2 ( g ) + 2 H 2O( g )

W
CH 4 ( g ) + H 2O( g ) CO( g ) + 3H 2 ( g )
Clearly the challenge in such cases is to determine the reaction conditions (of temperature, pressure
and feed composition) that maximize the conversion of the reactants to the desired product(s).
Essentially there are two methods to solve for the reaction equilibria in such systems.
TU
Method 1: Use of reaction-co-ordinates for each reaction
This is an extension of the method already presented in the last section for single reactions. Consider,
for generality, a system containing i chemical species, participating in j independent parallel reactions,
each defined by a reaction equilibrium constant K j and a reaction co-ordinate j . One can then write a
JN

set of j equations of the type 8.64 as follows:

j
( K ) j ( K y ) j = K T , j P ..(8.68)

Where, j and yi are given by eqns. 8.37 and 8.39 respectively (as follows):

j = i, j ..(8.37)
ll

nio + i, j j

And, yi = j
A

(i = 1,2,....,N) ..(8.39)
no + j j
j

Therefore there are j unknown reaction co-ordinates which may be obtained by solving simultaneously
j equations of the type 8.68.

18 | P a g e
--------------------------------------------------------------------------------------------------------------------------

ld
Example 8. 6
The following two independent reactions occur in the steam cracking of methane at 1000 K and 1 bar:
CH 4 (g) + H 2 O(g) CO(g) + 3H 2 (g); and CO(g) + H 2 O(g) CO 2 (g) + H 2 (g). Assuming ideal gas

or
behaviour determine the equilibrium composition of the gas leaving the reactor if an equimolar mixture
of CH 4 and H 2 O is fed to the reactor, and that at 1000K, the equilibrium constants for the two
reactions are 30 and 1.5 respectively.

W
(Click for solution)
--------------------------------------------------------------------------------------------------------------------------
Method 2: Use of Lagrangian Undetermined Multipliers
This method utilizes the well-known Lagrangian method of undetermined multipliers typically
employed for optimizing an objective function subject to a set of constraints. As outlined in section 8.3
TU
at the point of equilibrium in a reactive system, the total Gibbs free energy of the system is a
minimum. Further, during the reaction process while the total number of moles may not be conserved,
the total mass of each atomic species remains constant. Thus, in mathematical terms, the multi-reaction
equilibria problem amounts to minimizing the total Gibbs free energy of the system subject to the
constraint of conservation of total atomic masses in the system. The great advantage that this approach
JN

offers over the previous method is that one does not need to explicitly determine the set of independent
chemical reactions that may be occurring in the system.
We formulate below the set of equations that need to be solved to obtain the composition of the
system at equilibrium. Let there be N chemical (reactive) species and p (corresponding) elements in a
system; further, n i = initial no of moles of species i; ik = number of atoms of kth element in the ith
ll

chemical species; k = total number of atomic masses of kth element as available in the initial feed

composition.
A

n
i
i ik = k ; ( k = 1, 2 , p ) ..(8.69)

Or: ni ik k =
0; (=
k 1, 2 , p ) ..(8.70)
i

19 | P a g e
Use of p number of Lagrangian multipliers (one for each element present in the system) give:

k ni ik k =
0; (=
k 1, 2 , p ) ..(8.71)
i
These equations are summed over p, giving:

ld
n
p
k
i
i ik - k = 0

..(8.72)

Let Gt be the total Gibbs free energy of the system. Thus, incorporating p equations of the type 8.72

or
one can write the total Lagrangian L for the system as follows:

G t + k ni ik k
L= ..(8.73)
p i
It may be noted that in eqn. 8.73, L always equals Gt as the second term on the RHS is identically zero.

W
Therefore, minimum values of both Land G t occur when the partial derivatives of L with respect to all
the n i and k are zero.

F G t
Thus: = + k = 0; =
(i 1, 2, , N ) ..(8.74)

ik
i T , P ,n ji i T , P ,n ji
n n
TU
k

However, the first term on the RHS is the chemical potential of each reactive species in the system;
thus eqn. 8.74 may be written as:
i + k ik = ( i 1, 2,, N )
0; = ..(8.75)
k

But by eqn. 8.44:

JN

i,T = Gi,To + RTln fi / f i o ( ) ..(8.44)

Once again, we consider, for illustration, the case of gaseous reactions for which the standard state
pressure for each species is 1 bar, whence, fi 0 = 1bar.

i,T = Gi,To + RTln fi ( ) ..(8.76)

ll

i ,T =
G of ,T i
+ RT ln ( y P )
i i ..(8.77)
A

In the above equation Gi,To may be equated to G ofi ,T , the latter being the standard Gibbs free energy of

formation of the i species (at temperature T). In arriving at this relation, the standard Gibbs free
energy of formation of the elements comprising the ith species are arbitrarily set to zero (for
convenience of calculations). Thus combining eqns. 8.75 and 8.77 one obtains:
20 | P a g e
 ( )
G ofi ,T + RT ln yii P + k ik =
k
( i 1, 2,, N ) ;
0; = ..(8.78)

In eqn. 8.78, the reaction pressure P needs to be specified in bar (as fi 0 = 1bar ). Also, if the ith species

is an element, the corresponding G 0fi ,T =

0

ld
Further taking the partial derivative of the Lagrangian L (of eqn. 8.73) (L k ) ni ,nk with respect to

each of the p undetermined multipliers, an additional set of p equations of type 8.70 obtains. Thus there
are a total of ( N + p ) equations which may be solved simultaneously to obtain the complete set of

or
equilibrium mole fractions of N species.
--------------------------------------------------------------------------------------------------------------------------
Example 8.7

W
The gas n-pentane (1) is known to isomerise into neo-pentane (2) and iso-pentane (3) according to the
P2 ; P2
following reaction scheme: P1 P3 ; P3
P1 . 3 moles of pure n-pentane is fed into a
o
reactor at 400 K and 0.5 atm. Compute the number of moles of each species present at equilibrium.
o
Species G 0f at 400 K (Cal/mol)
TU
P1 9600

P2 8900

P3 8200
JN

--------------------------------------------------------------------------------------------------------------------------
8.8 Reactions involving Liquids and Solids
In many instances of industrially important reactions, the reactants are not only gaseous but are also
liquids and / or solids. Such reactions are usually heterogeneous in nature as reactants may exist in
separate phases. Some examples include:
ll

Removal of CO 2 from synthesis gas by aqueous solution of potassium carbonate

Removal of H 2 S by ethanolamine or sodium hydroxide
A

Air oxidation of aldehydes to acids

Oxidation of cyclohexane to adipic acid
Chlorination of benzene

21 | P a g e
 Decomposition of CaCO 3 to CaO and CO 2
In all such instances some species need to dissolve and then diffuse into another phase during the
process of reaction. Such reactions therefore require not only reaction equilibria considerations, but
that of phase equilibria as well. For simplicity, however we consider here only reaction equilibria of
instances where liquid or solid reactive species are involved. The thermodynamic treatment presented

ld
below may easily be extended to describe any heterogeneous reaction. The basic relation for the
equilibrium constant remains the starting point. By eqn. 8.50 we have:

KT = ( ai )
i

or
..(8.50)
On expanding (by eqn. 6.171):

ai = fi (T , P, xi ) / fi 0
..(8.79)

W
As already mentioned in section 8.1 above, for solids and liquids the usual standard state is the pure
solid or liquid at 1 bar [or 1(atm)] and at the temperature (T) of the system. However, unlike in the
case of gaseous species, the value of f i o for such a state cannot be 1 bar (or 1 atm), and eqn.(8.50)
cannot be reduced to the form simple form of eqn. 8.57.
TU
Liquid-phase reactants
On rewriting eqn. 8.79:

fi (T , P, xi ) = xi i fi
JN

Thus:
ai = xi i fi (T , P ) / f i 0 (T ,1bar ) ..(8.80)
By eqn. 6.115:
RTdlnf i = Vi dP
Thus on integrating:
ll

f i (T , P ) P Vi
0
fi (T ,1bar )
d ln fi =
1 RT
dP
..(8.81)
A

As we have already seen in section 6.10, the liquid phase properties, such as molar volume, are weakly
dependent on pressure; hence their variation with respect to pressure may be, for most practical
situations, considered negligible. Thus, if one considers that in the last equation the molar volume V i is
constant over the range 1 P bar, one obtains:

22 | P a g e
 f V ( P 1)
ln i0 = i
fi RT ..(8.82)

V ( P 1)
fi / fi o = exp i
RT ..(8.83)

ld
Thus, using eqn. 8.53 in 8.50:

i ( xi i ) i ( f i / fi o )
i
T = ( ai )

K=
..(8.84)

or
P -1
Or: ( fi / fi o ) =
i
exp iVi
RT
Thus:
( P -1) iVi

W
KT = ( xi i ) i exp


RT ..(8.85)
Except for very high pressure the exponential term on the right side of the above equation:
( P -1) iVi << RT .
TU
Thus one may approximate:
( P -1) iVi
exp 1.0
RT
Whence:

KT ( xi i )
i
JN

..(8.86)
To use the above equation for prediction of the equilibrium composition one needs to employ a
suitable activity coefficient model. The activity coefficient models described in section 6.18 are based
on physical interactions between molecules. Hence, their use is not expected to be adequately
representative of a situation where molecules are subject to specific chemical forces and are chemically
transformed due to formation of intermediate reactive complexes. While it is possible to write models
ll

for physical interactions it is generally not feasible to establish quantitative relations that describe the
microscopic-level interactions between reactive molecular species. Discussions on approaches to solve
A

such problems may be found elsewhere (J.M. Prausnitz, R.N. Lichtenthaler and E.G. Azevedo,
Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed., Prentice Hall, 1998). Nevertheless, for

23 | P a g e
the purpose of illustrating an approximate solution, one may simplify eqn. 8.86 by assuming ideal
solution behavior, wherein i = 1.0. Hence:

K = ( xi )
i
..(8.87)
However, since reactive solutions can never be ideal, one way to overcome the difficulty is by defining

ld
a reaction equilibrium constant based on molar concentration (say in moles/m3), rather than in terms of
mole fractions. Thus:

K c = ( Ci )
i
..(8.88)

or
Where, C i = molar concentration of each species.
It is generally difficult to predict the equilibrium constant K C , and one needs to use experimentally
determine values of such constants in order to predict equilibrium compositions.

W
--------------------------------------------------------------------------------------------------------------------------
Example 8.8
Consider the liquid phase reaction: A(l) + B(l) C(l) + D(l). At 50oC, the equilibrium constant is
0.09. Initial number of moles, n A,0 = 1 mole; n B,0 = 1 mol Find the equilibrium conversion. Assume
TU
ideal solution behaviour.
(Click for solution)
--------------------------------------------------------------------------------------------------------------------------
Solid-phase reactants
Consider a solid reactive species now, for which one again starts from eqn. 8.80:
JN

ai = fi (T , P, xi ) / fi 0
..(8.80)
Thus as for a liquid reactant one has
( P 1)Vi
ai = ( xi i ) ( f i f i 0 ) = ai = ( xi i ) exp
RT
As it is for liquid species, V i for solids is also small and remains practically constant with pressure,
ll

thus:
( P -1) iVi
A

exp 1.0
RT
In addition, the solid species is typically pure as any dissolved gas or liquid (for a multi-phase
reaction) is negligible in amount.

24 | P a g e
Thus xi ~ 1.0, i =
1.

( P 1)Vi
Therefore, for solids ai = ( xi i ) exp 1.0 ..(8.89)
RT
--------------------------------------------------------------------------------------------------------------------------

ld
Example 8.9
Consider the following reaction: A(s) + B(g) C(s) + D(g). Determine the equilibrium fraction of B
which reacts at 500oC if equal number of moles of A and B are introduced into the reactor initially.

or
The equilibrium constant for the reaction at 500oC is 2.0.
Assignment- Chapter 8

W
TU
JN
ll
A

25 | P a g e
Chapter 9: Other Topics in Phase Equilibria
This chapter deals with relations that derive in cases of equilibrium between combinations of
two co-existing phases other than vapour and liquid, i.e., liquid-liquid, solid-liquid, and solid-
vapour. Each of these phase equilibria may be employed to overcome difficulties encountered
in purification processes that exploit the difference in the volatilities of the components of a
mixture, i.e., by vapour-liquid equilibria. As with the case of vapour-liquid equilibria, the

ld
objective is to derive relations that connect the compositions of the two co-existing phases as
functions of temperature and pressure.

or
9.1 Liquid-liquid Equilibria (LLE)
9.1.1 LLE Phase Diagrams
Unlike gases which are miscible in all proportions at low pressures, liquid solutions (binary

W
or higher order) often display partial immiscibility at least over certain range of temperature,
and composition. If one attempts to form a solution within that certain composition range the
system splits spontaneously into two liquid phases each comprising a solution of different
composition. Thus, in such situations the equilibrium state of the system is two phases of a
TU
fixed composition corresponding to a temperature. The compositions of two such phases,
however, change with temperature. This typical phase behavior of such binary liquid-liquid
systems is depicted in fig. 9.1a. The closed curve represents the region where the system
exists
JN
ll
A

Fig. 9.1 Phase diagrams for a binary liquid system showing partial immiscibility
in two phases, while outside it the state is a homogenous single liquid phase. Take for
example, the point P (or Q). This point defines a state where a homogeneous liquid solution
exists at a temperature TP (or TQ ) with a composition x1P (or x1Q ). However, if one tries to

form a solution of composition x1Y (at a temperature TY ) the system automatically splits into

two liquid phases (I and II) with compositions given by x1I and x1II . The straight line MYN
represents a tie-line, one that connects the compositions of the two liquid phases that co-exist

ld
in equilibrium. Any attempt to form a solution at TY with a composition corresponding to any

point within MYN always results in two phases of fixed compositions given by x1I and x1II .

or
The total mass (or moles) of the original solution is distributed between the two phases
accordingly. Note, however, that the compositions at the end tie lines (horizontal and parallel
to MYN) change as the temperature changes. Indeed as one approaches either point U or L,

W
the tie line reduces to a point, and beyond either point the system exists as a single
homogeneous solution. The temperatures corresponding to points U and L, i.e.,
TUCST and TLCST are called the Upper Consolute Solution Temperature (UCST) and Lower
Consolute Solution Temperature (LCST) respectively. They define the limit of miscibility of
TU
the components of the binary solution. Not all liquid-liquid systems, however, depict the
behaviour described by fig. 9.1a. Other variants of phase behaviour are shown in figs. 9.1b
9.1c. In each case there is at least one consolute temperature.

9.1.2 Phase Stability Criteria

JN

The passage of a liquid solution from a state of a single, homogeneous phase to a biphasic
LLE state occurs when a single phase is no longer thermodynamically stable. It is instructive
to derive the conditions under which this may take place. Let us consider a binary liquid
mixture for example for which the molar Gibbs free energy of mixing (Gmix ) is plotted as

function of mole fraction ( x1 ) of component 1 for two temperatures T1 and T2 (fig. 9.2). At
ll

temperature T1 the value of Gmix is always less than zero and also goes through a minimum.
A

This signifies that the mixture forms a single liquid phase at all compositions at the given
temperature. The mathematical description of this condition is given by the two following
equations:
Gmix
= 0 ..(9.1)
x1 T , P
And:
2Gmix
2
>0 ..(9.2)
x1 T , P
Alternately:
Gmix
= 0 ..(9.3)
x1 T , P

ld
And:
2 Gmix
2 >0 ..(9.4)
x1 T , P

or
However, at certain other temperature T2 the value of Gmix , while being always less than

W
zero, passes through a region where it is concave to the composition axis. Consider a solution
with a composition given by the point A. At this composition the Gibbs free energy of the
mixture is given by:
G=
A
mix x1AG1 + x2AG2 + Gmix
A
..(9.5)
TU
JN
ll
A

Fig. 9.2 Molar Gibbs free energy mixing vs. mole fraction x1 for a binary solution

Let the points B and C denote the points of at which a tangent BC is drawn on the Gmix

curve for the isotherm T2 , and D the point vertically below A on the straight line BC

(which is a tangent to the Gmix curve at points B and C). If the solution at A splits into two
phases with compositions characterized by x1B and x1C then its molar Gibbs free energy is

given by:
G=
B
mix x1AG1 + x2AG2 + Gmix
D
..(9.6)

Clearly, since Gmix

D
< Gmix
A
, (or Gmix
D
< Gmix
A
) a phase splitting is thermodynamically favoured

over a single phase solution of composition x1A . Point D represents the lowest Gibbs free

ld
energy that the mixture with an overall composition x1A can have at the temperature T2 . In

other words, at the temperature T2 a solution with an overall composition x1A exists in two

or
phases characterized by compositions x1B and x1C . If nB and nC represent the quantities of

solution in each phase then the following equation connects x1A to x1B and x1C :

( nB x1B + nC x1C ) ( nB + nC )
x1A = ..(9.7)

W
The above considerations allow one to formulate a phase separation criteria for liquid
solutions. Essentially, for phase splitting to occur, the Gmix curve must in part be concave to

the composition axis. Thus in general mathematical terms a homogeneous liquid solution
becomes unstable if:
TU
( G
2
mix x 2 )
T ,P
< 0; (where, x xi ) ..(9.8)

It follows that the alternate criterion for instability of a single phase liquid mixture is:

( G
2
mix x 2 )
T ,P
<0 ..(9.9)

It may be evident that at T2 the above conditions hold over the composition range
JN

x1B < x1 < x1C . Conversely, a homogeneous liquid phase obtains for compositions over the

ranges 0 < x1 < x1B and x1C < x1 < 1. In these ranges, therefore, the following mathematical
condition applies:

( G
2
x 2 ) >0 ..(9.10)
ll

mix T ,P

Or:

( G x 2 )
A

2
mix >0 ..(9.11)
T ,P

Referring again to fig. 9.2 it may be evident that at points X and Y the following
mathematical condition holds:

( G
2
mix x 2 )
T ,P
=0 ..(9.12)
 A series of Gmix curves at other temperatures (say between T1 and T2 ) may be drawn, each
showing different ranges of unstable compositions. All such curves may be more concisely
expressed by the T x plot in fig. 9.2. There is an absolute temperature T C above which the
mixture is stable at all compositions since the condition described by eqn. 9.11 applies at all
compositions. The binodal curve in fig. 9.2 represents the boundary between the single phase
region and the two phase regions. Within the two phase region the spinodal curve represents

ld
(
the locus of points at which 2 Gmix x 2 )T ,P
=0. As may be evident from the T x plot in

fig. 9.2, it is the boundary between the unstable ( 2 Gmix x 2 ) < 0 and metastable

or
T ,P

( 2 G x 2 ) > 0 regions.
mix T ,P
The above mathematical conditions may be used to explain the phase behaviour

W
depicted in fig. 9.1. Since the binary mixture displays both UCST and LCST, it follows that
at T > TUCST and T < TLCST the mixture forms single phase at all compositions. Thus for such

(
temperatures the Gmix curve is described by eqn. 9.11, i.e., 2 Gmix x 2 )T ,P

compositions. Conversely for TLCST < T < TUCST the Gmix curve displays a form corresponding
> 0 for all
TU
to that for temperature T2 in fig. 9.2. That is, for TLCST < T < TUCST there is partial immiscibility
of the mixture. The mathematical conditions for the existence of the consolute temperatures
may then be surmised by the following relations.
For the existence of UCST:
JN

2G
= 2 0= for some value of x1 at T TUCST
x1
and
2G
> 0 for all values of x1 at T > TUCST
x12
ll

For the existence of UCST:

2G
= 2 0= for some value of x1 at T TLCST
A

x1
and
2G
> 0 for all values of x1 at T < TLCST
x12

The foregoing mathematical relations are more conveniently expressed in terms of the
excess molar Gibbs free energy function as follows:
 =
G E
Gmix Gmix
id
..(9.13)

Or:
G E =Gmix x1G1 + x2G2 + RT ( x1 ln x1 + x2 ln x2 ) ..(9.14)

Thus:
Gmix =G E + x1G1 + x2G2 + RT ( x1 ln x1 + x2 ln x2 ) ..(9.15)

ld
On applying the criterion for instability given by eqn. 9.8, an equivalent criterion obtains as
follows:
2G E 1 1
+ + <0

or
2 RT
x T , P x1 x2

2G E RT
2
+ <0 ..(9.16)
x T , P x1 x2

W
For an ideal solution G E = 0; hence the condition in eqn. 9.16 can never hold. Thus, and ideal
solution can never display phase-splitting behaviour.
Consider now the case of a binary mixture described by the excess Gibbs free energy
function: G E = x1 x2 . Thus:
TU
2G E
2
= 2 ..(9.17)
x T , P
On applying the condition given by eqn. 9.16 one obtains the following relation:
RT
2 + <0 ..(9.18)
JN

x1 x2
Or:
RT
2 > ..(9.19)
x1 x2

The maximum value of the term x1 x2 = 0.25; hence the values of that satisfies the
ll

inequality 9.19 is given by:

2RT ..(9.20)
A

An equivalent criterion of instability may be derived using activity coefficient as a

parameter in place of the excess Gibbs free energy function. For simplicity we consider a
binary solution. Thus we have:
GE
= x1 ln 1 + x2 ln 2 ..(9.21)
RT
 On differentiating eqn. 9.21 one obtains:
d ( G E / RT ) d ln 1 d ln 2
= ln 1 ln 2 + x1 + x2 ..(9.22)
dx1 dx1 dx1
By Gibbs-Duhem relation:
d ln 1 d ln 2
x1 + x2 0
= ..(9.23)
dx1 dx1

ld
Hence:
d ( G E / RT )
= ln 1 ln 2 ..(9.24)
dx1

or
d 2 ( G E / RT ) d ln 1 d ln 2
= ..(9.25)
dx12 dx1 dx1

W
Or, multiplying both sides by x2 :

d 2 ( G E / RT ) d ln 1 d ln 2
x2 2
(1 x1 )
= x2 ..(9.26)
dx1 dx1 dx1
Applying the Gibbs-Duhem relation again and upon simplification eqn. 9.26 reduces to:
d 2 ( G E / RT )
TU
1 d ln 1
2
= ..(9.27)
dx 1 x2 dx1
For stability of a solution using eqn. 9.16 we have:
2G E RT
2
+ >0 ..(9.28)
x T , P x1 x2
JN

Upon substituting eqn. 9.27 in 9.28 one obtains after due algebraic simplification:
d ln( x1 1 )
>0 ..(9.28)
dx1
The above relation provides a convenient starting point for determining if a certain activity
coefficient model may predict phase instability. As an illustration we apply it to Wilson eqn.
ll

for a binary (see table 6.2) wherein:

12 21
A

ln =
1 ln( x1 + x2 12 ) + x2 ..(9.29)
x1 + 12 x2 21 x1 + x2
On using eqn. 9.29 in 9.28 and after algebraic simplification one obtains:
d ln( x1 1 ) 2
x2 12 221
= + ..(9.30)
dx1 x1 ( x1 + x2 12 ) 2 ( x2 + x1 21 ) 2
 It may be seen that the RHS of the above equation is always > 0, hence eqn. 9.28 is satisfied.
Hence it follows that the Wilson equations cannot be used to predict instability of a liquid
solution.

9.1.3 Solving the liquid-liquid phase equilibria problem

We now present the equations needed to obtain the phase compositions for a biphasic liquid-

ld
liquid equilibria condition. Consider two liquid phases I and II , in equilibrium with each
other and each containing N species. One starts with the basic phase equilibria criterion:
=fi I f=
i ;
II
(i 1, 2,...N ) ..(9.31)

or
In practice the majority of liquid-liquid extraction operations are carried out at low to
moderate pressures. Thus the fugacities may be expressed in terms of activity coefficients.
On expanding each term in the above equation we get:
fi I = ( xi i f i ) I

fi II = ( xi i fi ) II
W
Where, xi = mole fraction of ith species in a phase
..(9.32)

..(9.33)
TU
Since both the phases are at the same temperature and pressure at equilibrium, the pure
component fugacities are equal: fi I = fi II
Thus it follows that:
( xi i ) I = ( xi i ) II

xi I i I x=
= i i ; (i 1, 2,...N ) ..(9.34)
JN

II II

Further by mole balance:

x i
I
=1 ..(9.35)

x II
=1 ..(9.36)
ll

Equation 9.34 represents the generalized LLE criterion applicable to systems with any
A

number of components distributed between two co-existing partially miscible liquid phases.
Additionally, a suitable activity coefficient model needs to be assumed for solving the LLE
problem. The phase compositions are derived by solving a set of N equations of the type 9.34
(one for each component), along with two constraining equations of the type 9.35 and 9.36.
------------------------------------------------------------------------------------------------------------
Example 9.1
 Use the van Laar activity coefficient expression to predict the compositions of co-existing
liquid phases (I and II) comprised of two partially miscible liquids (1) and (2) at 50oC and 4
bar. At these conditions the van Laar equations are given by:
A12 A21
ln 1
= = ;ln 2 = ; A12 2.5;
= A21 3.5
A x A x
[1 + 12 1 ]2 [1 + 21 2 ]2
A21 x2 A12 x1
(Click for solution)

ld
-------------------------------------------------------------------------------------------------------------
9.1.4 Ternary liquid-liquid equilibria

or
As may be evident from fig. 9.1a, under two-phase conditions, one of the liquid phases
usually contains one of the components in a more concentrated form. This feature of liquid-
liquid equilibria is exploited in the process industry, typically in the form of ternary liquid-

W
liquid extraction. In many instances of practical interest a binary (or higher) liquid mixture
cannot be adequately concentrated in one of the components by vapour-liquid separation
processes; this, either because of inadequate difference in volatility of the components, or
because there is an azeotrope formation. In such instances an alternate approach to purifying
a species to the required extent from other components could be through the addition of a
TU
third liquid (solvent) that is partially miscible with the original solution. In such a case a
second liquid phase results, where one of the components in the original mixture may become
preferentially concentrated due to its higher affinity for the added solvent. As with binary
systems one can solve this ternary LLE problem by application of eqns. 9.34 9.36. Such
phase equilibria data (either experimental or computed) are typically plotted as a triangular
JN

diagram. A very simple example of such a diagram is shown in fig. 9.3. More complex
versions of such diagrams may be found elsewhere (see S.I. Sandler, Chemical, Biochemical
and Engineering Thermodynamics, 4th Edition, Wiley India, 2006).
ll
A

Fig. 9.3 Triangular diagram for ternary liquid-liquid phase equilibria

In this figure A and B comprise the original binary mixture to which a third, partially
miscible solvent C is added to preferentially extract A. Each apex of the triangle represents
100% mole (or mass) fraction of the species indicated at the apex. Each side of the triangle
represents a mixture of two species that are indicated at the two ends of the side. The
concentration of each species decreases linearly with distance along the perpendicular line

ld
that may be drawn from the apex to the opposite side of the triangle. The mass or mole
fractions of each species corresponding to a point within the triangle is found by first drawing
a line through the point parallel to the side opposite to the apex for that species. The point of

or
intersection of the line with the appropriate side of the triangle provides the concentration
fraction of the species. For example, for the point J the constituent compositions are shown as
x AJ , xBJ and xCJ respectively on the three sides of the triangle.

W
The dome shaped region DPE corresponds to the two phase region in the triangular
diagram, while any point outside it represents either a binary or a ternary homogeneous
mixture. The lines FG, HI, JK are typical tie lines across the two-phase region. The two end
points of the line correspond to the two phases that co-exist in equilibrium. For example, the
TU
end point on the left side (here, F) of the dome DPE corresponds to the solvent-rich extract
phase where A is preferentially concentrated, while the point G represents the raffinate
phase that is rich in B. Thus any point within the dome represents a biphasic system with
the end points of the tie line through the point corresponding to the extract and raffinate
phases. As may also be observed, the tie lines become shorter in length as one approaches the
JN

top part of the dome, until it reduces to a point at P. This point is termed the plait point
and it signifies the condition at which the compositions of the two liquid phases in
equilibrium become identical, and indeed transform to a single phase.

9.2 Solid-Liquid Equilibria

ll

The prediction of solubility of a solid in a liquid is important for design of separation

processes that utilize either preferential dissolution or crystallization as a route to purification
A

of a species. The presence of a dissolved solid in a liquid also changes the freezing (as well as
boiling) point of the latter, and its estimation is important for design of heat exchange
equipments such as crystallizers (and evaporators). A wide variety of phase behaviour has
been observed in systems which comprise mixtures of solids, mixture of a solid and a liquid
and mixtures comprised of a liquid and more than a single solid (see R. T. DeHoff,
Thermodynamics in Materials Science, chaps. 9 and 10, McGraw-Hill, New York, 1993).
This section is devoted to consideration of thermodynamic relations that allow computation
of the solubility of solids in liquids as a function of temperature. Such relations may also be
extended to estimate changes in freezing or boiling point of a liquid in presence of dissolved
solids.
The phase equilibria relations developed here are limited to the simplest case of

ld
systems comprising a solid (1) and a liquid (2) (solvent). As with all other phase equilibria,
one starts with the basis of equality of the species fugacity in the solid (s) and liquid (l)
phases respectively.

or
fi L = fi S ..(9.37)

xi iL f i L = zi iS f i S

W
..(9.38)

Where, xi and zi are, respectively, the mole fractions of species i in the liquid and solid
solutions. Equivalently,

xi iL = zi iS i ..(9.39)
TU
Where, i f i S (T , P ) / f i L (T , P ) ..(9.40)

The right side of this equation, defines i as the ratio of fugacities at the temperature T and
pressure P of the system. Note that for each pure species at its (say, normal) melting point
JN

Tm ,i we have: fi S (Tm,i , P) = fi L (Tm ,i , P) ..(9.41)

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

Thus, in general one needs to derive a suitable expression for i defined by eqn. 9.40.
This may be done by through estimation of the enthalpy and entropy of fusion at the system
temperature T, with values of the enthalpy and entropy at the melting point at the system
ll

pressure as datum. However, since the properties of solids are not dependent on pressure in a
significant way, for convenience of computation one can assume the reference point as the
A

normal melting point Tm ,i . Accordingly, we may write for either the solid or the solvent

species:
Tm T
=
H CP dT + HTm + CP dT ..(9.42)
fus L fus S
T
T Tm
 T
Or: H =
H + (C CPL ) dT ..(9.43)
fus fus S
T Tm P
Tm

In the same manner one may write:

S fus
=
S fus
T
+
(C S
P - CPL )
dT ..(9.44)
T Tm
Tm
T

At temperature Tm , GTfus 0 , thus:

=

ld
m

H Tfus
m
Tm STfus
m
0
= ..(9.45)

H Tfus

or
Or: S fus
Tm =m ..(9.46)
Tm
Thus, using eqn. 9.46 in 9.44 we have:

H Tfus T
(C S
CPL )

W
S = + ..(9.47)
fus P
T
m
dT
Tm Tm
T

Further, at temperature T, we have:

GTfus =H Tfus T STfus ..(9.48)
Therefore, substituting eqns. 9.43 and 9.47 in 9.48 and re-arranging we have:
TU
H Tfus T
+ (C C
T
(C S
CPL )
G T
fus
=
H fus
Tm T m

Tm
S
P
L
P ) dT T P

T
dT ..(9.49)
Tm Tm

If we further make the simplifying assumption that (C S

P CPL ) is constant over the

temperature range Tm to T one obtains:

JN

T T
GTfus =
H Tfus 1 + ( CP CP ) (T Tm ) (CP CP )T ln ..(9.50)
S L S L
m
Tm Tm

S fS

Now, dGT = RT d ln f
ll

L fL

Or:
A

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

fS T T
RT ln H Tfum 1 s
= + ( CP CP ) (T Tm ) (CP CP )T ln ..(9.52)
S L S L
L
f Tm Tm
 Therefore using eqn. 9.52 in eqn. 9.40, we have:
T T
RT ln =
H Tfum 1 s + ( CP CP ) (T Tm ) (CP CP )T ln ..(9.53)
S L S L

Tm Tm
Upon suitable rearrangement:
H Tfus 1 1 (CPS CPL ) Tm T
ln
= m
+ 1 ln ..(9.54)
R T Tm R T Tm

ld
Now by eqn. 9.39:

xi iL = zi iS i ..(9.39)

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

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

W
respectively and then eqn. 9.39 may be solved to yield the phase compositions. We illustrate
the typical methodology of such calculations for a binary system.
Let us assume a solid (1) is to be solubilized in a liquid (2). The starting point is eqn.
9.39. However, in applying it we make the simplifying assumption that the liquid does not
TU
dissolve in the solid phase, and so the latter phase is a pure component solid. This is a
reasonable approximation for many system of practical interest in the chemical industry. The
liquid phase, however, will be a binary with the solid transforming to a liquid phase and
then mixing with the liquid solvent. Thus, one may rewrite the phase equilibrium as:
f1S (T , P ) = f1L (T , P, x1 ) ..(9.55)
JN

Now f1L (T , P, x1 ) = x1 1 (T , P, x1 ) f1L ..(9.56)

Thus from the last two equations, the equilibrium solubility of the solid ( x1 ) in the liquid is
given by:
f1S 1
x1 = ( L )( ) ..(9.57)
ll

f1 1 (T , P, x1 )
Since activity coefficients are functions of temperature and composition only:
A

f S 1
x1 = 1L ..(9.58)
f1 1 (T , x1 )
It follows that if the temperature at which the solubility is desired is the melting point Tm of

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

Therefore, we have:
 1
x1 = ..(9.59)
1 (Tm , x1 )
However, for any temperature T (other than Tm ) eqn. 9.58 needs to be employed. Thus, we
have:
fS
ln 1 ln 1L
ln x1 = ..(9.60)
f1

ld
Or using eqns. 9.40 and 9.52 (or 9.53) we have:
H Tfus 1 1 (CPS CPL ) Tm T
ln 1
ln x1 = m
+ 1 ln ..(9.61)

or
R T Tm R T Tm

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

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

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

H Tfus 1 1
TU
ln 1
ln x1 = m
..(9.62)
R T Tm
------------------------------------------------------------------------------------------------------------
Example 9.2
Estimate solubility of a solid A in a liquid B at 300oK, using (i) ideal solution assumption, (ii)
regular solution model for liquid phase. The following data are available:
JN

( cal / cc ) ; B 7.5 ( cal / cc ) .

1/2 1/2
=VAL 100
= cm3 / mol ; VBL 125
= cm3 / mol ; A 9.5= Heat of

fusion for A: 17.5 kJ/mol. Melting point for A = 350oK.

(Click for solution)
------------------------------------------------------------------------------------------------------------
ll

For the purpose of further illustration of the results that typically obtain from application of
the foregoing phase equilibria relations, we consider two limiting or special cases for a binary
A

system assuming that eqn. 9.62 applies. It may be noted, however, that this assumption is not
essential to the considerations below. Using eqn. 9.54 we write:
H Tfus 1 1
=ln m
..(9.63)
R T Tm

Or, for any of the species:

 H i fus T Tm ,i
i = exp ..(9.64)
RTm ,i T

The limiting cases are:

1. Both phases form ideal solutions, i.e., iL = 1 and iS = 1 for all temperature and

ld
compositions.
2. The liquid phase behaves as an ideal solution ( iL = 1 ), and all species in the solid

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

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

x2 = z2 2
Further:
1
x1 + x2 = W ..(9.66)

..(9.67)
TU
1
z1 + z2 = ..(9.68)
Using eqns. 9.57 9.60, one obtains:
1 (1 2 )
x1 = ..(9.69)
1 2
1 2
z1 =
JN

..(9.70)
1 2
If one uses the simplified form for , then by eqn. 9.64:

H1fus T Tm ,1
1 = exp ..(9.71)
RTm ,1 T
ll

H 2fus T Tm ,2
2 = exp ..(9.72)
RTm ,2 T
A

On putting x=
1 z=
1 1 it follows that 1 = 1; thus, by eqn. 9.71, T = Tm ,1. In the same manner
if we put x=
2 z=
2 1 then 2 = 1; and T = Tm ,2 . On choosing a series of temperature T
between Tm ,1 and Tm ,2 and for each case solving for xi and zi one obtains a T x z plot that

is shown schematically in fig. 9.4. The upper curve in fig. 9.4 is the freezing curve and the
lower curve is the melting curve. Any T x point within the lens-shaped region between the
two curves corresponds to a two phase situation, while points outside the region depict a
single phase. If a point is above the upper curve the state of the system is a homogeneous
binary liquid mixture, whereas if the point lies beneath the melting curve, the state of the
system is binary solid mixture. Any straight line parallel to the composition axis depicts a tie
line (for example XY), the compositions at the two ends signifying the liquid and solid phase

ld
compositions. This behaviour is analogous to that of Raoults law T x y plot shown in
fig.7.7.

or
W
TU
Figure 9.4 Case I, ideal liquid and solid solutions
JN

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

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

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

immiscibility for all species in the solid state ( zi is = 1 ). Thus the governing phase equilibria

relations are:
x1 = 1 ..(9.73)

x2 = 2 ..(9.74)
Since, 1 and 2 are functions of temperature alone (by eqns.9.71 and 9.72), it follows that

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

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

fig. 9.5. However, if eqns. 9.64 and 9.65 apply simultaneously, then we have: 1 + 2 =
1 and

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

ld
or
W
TU
Fig. 9.5 T-x-z plot for Case II

H1fus T Tm ,1 H 2fus T Tm ,2
exp + exp =1 (9.75)
RTm ,1 T RTm ,2 T
JN

On solving the above equation one obtains the Eutectic temperature Te shown in fig. 9.5. This
particular phase diagram is to be interpreted as follows. When eqn. 9.73 applies one has:

H fus T Tm1
x1 = exp 1 ..(9.76)
ll

RTm1 T
A

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

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

diagram a liquid solution can be in equilibrium with only pure species 1 as a solid phase. This
is represented by region I on fig. 9.5 where liquid solutions with compositions x 1 given by
line YE are in equilibrium with pure solid 1. Consider the system to be initially at a state
defined by point M. If one cools the system it eventually reaches the point N on the T x1
plot, at which condition only the component 1 just begins to freeze. This happens till one
reaches the eutectic temperature Te at which point component 2 begins to co-precipitate. In
the same manner region II obtains by solving eqn. 9.74, i.e.,

H 2sl T Tm2
exp 1 x1 ..(9.77)

ld
x2
RTm2 T

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

or
eutectic composition. Equation 9.77 therefore applies where a liquid solution is in
equilibrium with pure species 2 as a solid phase. This is represented by region II on fig. 9.5
where liquid solutions with compositions x1 given by line XE are in equilibrium with pure

W
solid 2. Consider an initial system state defined by point P which is a binary solution. On
cooling, the system eventually reaches the point Q where only pure solid 2 begins to
crystallize. On continued cooling, the eutectic temperature is reached, whereupon component
1 also precipitates simultaneously.
TU
Finally, if one commences cooling from the point R with a solution composition given
by x1 = x1e , no precipitation of any solid occurs till the temperature Te (or point E) is
reached; at which state, as mentioned above, eqn. 9.75 holds. This is a state where two solids
are in equilibrium with the solution. Further cooling below Te results in co-precipitation of
JN

both species, and the system enters a state where two immiscible solids are present.
---------------------------------------------------------------------------------------------------------------
Example 9.3
Compute the eutectic composition and temperature for a mixture of two substances A and B
using the following data:
ll

Property A B
Normal T m (oK) 180 181
6600 9075
A

H fus (J/mol)

(Click for solution)

---------------------------------------------------------------------------------------------------------------
9.3 Solid-Vapour Equilibrium
This section deals with the solubility of solids in gases. The phase equilibria consideration
here are relatively straightforward as the solubility of gases in solids is usually negligible.
Therefore, the solid phase may be considered as pure; all non-ideality in the system therefore,
derives from that of the gas phase. In the following analysis we, thus, consider that a pure
solid phase (1) co-exists in equilibrium with a vapour-phase mixture of the gas (or solvent, 2)
and the solute (i.e., 1). The phase equilibria relation for the component 1 may be written as:

ld
f1S = f1V ..(9.78)
A relation of the form of eqn. 6.119 used for pure liquid phase fugacity calculation may be
employed to compute the fugacity of a pure solid at the temperature T and pressure P of the

or
system.
Vi l ( P Pi sat )
fi (T , P ) = isat Pi sat exp[ ] ..(6.119)
RT
For a pure solid the corresponding equation is:

f =P
1
S
1
sat sat
1
V1S ( P P1sat )
exp
RT

W ..(9.79)

Here V1S is the molar volume of the solid, and P1sat is the vapour-phase saturation pressure for
TU
the solid solute at the given temperature. Now for the vapour phase fugacity of the solute on
has:
f1V = y11 P ..(9.80)

Thus using eqns. 9.78 9.80 one obtains:

JN

V1S ( P P1sat )
P
1
sat sat
1 exp = y11 P ..(9.81)
RT

The above equation may be written in a compressed form as follows:

P sat
y1 = 1 E ..(9.82)
P
ll

Where,
V1S ( P P1sat )
A

1sat
E= exp ..(9.83)
1 RT

The term E is called the enhancement factor. The vapour pressure P1sat of solids is typically

very low (~ 10-5 10-3 bar), hence 1sat 1.0. Further if the system pressure is also low or as

P P1sat , 1 1. Thus, on inspecting eqn. 9.82, it is evident that under such a condition
E = 1. The corresponding solubility y1 is termed the ideal solubility. At high pressures (with

P >> P1sat ) , the exponential Poynting factor tends to be greater than one, and the values of 1
are lower than unity. As a result, the factor E typically is higher than one. The enhancement
factor, therefore, provides a measure of the degree to which pressure augments the solubility
of the solid in the gas phase.
------------------------------------------------------------------------------------------------------------

ld
Example 9.4
A certain solid A has a vapour pressure of 0.01 bar at 3000K. Compute its solubility at the
same temperature in a gas B at a pressure of 1.0bar. The molar volume of the solid is

or
125cc/mol.
(Click for solution)
------------------------------------------------------------------------------------------------------------
Assignment- Chapter 9

W
TU
JN
ll
A