HEAT AND

THERMODYNAMICS

Jeremy Tatum
University of Victoria
 University of Victoria
Heat and Thermodynamics

Jeremy Tatum
This open text is disseminated via the Open Education Resource (OER) LibreTexts Project (https://LibreTexts.org) and like the
hundreds of other open texts available within this powerful platform, it is licensed to be freely used, adapted, and distributed.
This book is openly licensed which allows you to make changes, save, and print this book as long as the applicable license is
indicated at the bottom of each page.
Instructors can adopt existing LibreTexts texts or Remix them to quickly build course-specific resources to meet the needs of
their students. Unlike traditional textbooks, LibreTexts’ web based origins allow powerful integration of advanced features and
new technologies to support learning.

The LibreTexts mission is to unite students, faculty and scholars in a cooperative effort to develop an easy-to-use online
platform for the construction, customization, and dissemination of OER content to reduce the burdens of unreasonable
textbook costs to our students and society. The LibreTexts project is a multi-institutional collaborative venture to develop the
next generation of open-access texts to improve postsecondary education at all levels of higher learning by developing an
Open Access Resource environment. The project currently consists of 13 independently operating and interconnected libraries
that are constantly being optimized by students, faculty, and outside experts to supplant conventional paper-based books.
These free textbook alternatives are organized within a central environment that is both vertically (from advance to basic level)
and horizontally (across different fields) integrated.
The LibreTexts libraries are Powered by MindTouch® and are supported by the Department of Education Open Textbook Pilot
Project, the UC Davis Office of the Provost, the UC Davis Library, the California State University Affordable Learning
Solutions Program, and Merlot. This material is based upon work supported by the National Science Foundation under Grant
No. 1246120, 1525057, and 1413739. Unless otherwise noted, LibreTexts content is licensed by CC BY-NC-SA 3.0.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not
necessarily reflect the views of the National Science Foundation nor the US Department of Education.
Have questions or comments? For information about adoptions or adaptions contact info@LibreTexts.org. More information
on our activities can be found via Facebook (https://facebook.com/Libretexts), Twitter (https://twitter.com/libretexts), or our
blog (http://Blog.Libretexts.org).

This text was compiled on 04/29/2021

TABLE OF CONTENTS
Thermodynamics is a subject that has a wide variety of applications, including many in practical and engineering contexts.

1: INTRODUCTORY REMARKS
1.1: INTRODUCTION
1.2: CALORIC, CALORIES, HEAT AND ENERGY
1.3: EXTENSIVE AND INTENSIVE QUANTITIES
1.4: MOLE
1.5: PREPOSITIONS
1.6: APPLICABILITY OF EQUATIONS

2: PARTIAL DERIVATIVES
2.1: INTRODUCTION
2.2: PARTIAL DERIVATIVES
2.3: IMPLICIT DIFFERENTIATION
2.4: PRODUCT OF THREE PARTIAL DERIVATIVES
2.5: SECOND DERIVATIVES AND EXACT DIFFERENTIALS
2.6: EULER'S THEOREM FOR HOMOGENEOUS FUNCTIONS
2.7: UNDETERMINED MULTIPLIERS
2.8: DEE AND DELTA

3: TEMPERATURE
During our studies of heat and thermodynamics, we shall come across a number of simple, easy-tounderstand terms such as entropy,
enthalpy, Gibbs free energy, chemical potential and fugacity, and we expect to have no difficulty with these. There is, however, one
concept that is really quite difficult to grasp, and that is temperature. We shall do our best to understand it in this chapter.

3.2: ZEROTH LAW OF THERMODYNAMICS

3.3: TEMPERATURE SCALES I
3.4: TEMPERATURE SCALES II

4: THERMAL CONDUCTION
4.1: ERROR FUNCTION
4.2: INTRODUCTION
4.3: THERMAL CONDUCTIVITY
4.4: THE HEAT CONDUCTION EQUATION
4.5: A SOLUTION OF THE HEAT CONDUCTION EQUATION

5: THERMODYNAMIC PROCESSES
5.1: SECTION 1-
5.2: SECTION 2-
5.3: SECTION 3-
5.4: SECTION 4-
5.5: SECTION 5-
5.6: SECTION 6-
BACK MATTER
INDEX

6: PROPERTIES OF GASES
6.1: THE IDEAL GAS EQUATION
6.2: REAL GASES
6.3: VAN DER WAALS AND OTHER GASES
6.4: GAS, VAPOUR, LIQUID AND SOLID
6.5: KINETIC THEORY OF GASES- PRESSURE

1 4/29/2021
6.6: COLLISIONS
6.7: DISTRIBUTION OF SPEEDS
6.8: FORCES BETWEEN MOLECULES

7: THE FIRST AND SECOND LAWS OF THERMODYNAMICS

7.1: THE FIRST LAW OF THERMODYNAMICS, AND INTERNAL ENERGY
7.2: WORK
7.3: ENTROPY
7.4: THE SECOND LAW OF THERMODYNAMICS

8: HEAT CAPACITY, AND THE EXPANSION OF GASES

8.1: HEAT CAPACITY
8.2: RATIO OF THE HEAT CAPACITIES OF A GAS
8.3: ISOTHERMAL EXPANSION OF AN IDEAL GAS
8.4: REVERSIBLE ADIABATIC EXPANSION OF AN IDEAL GAS
8.5: THE CLÉMENT-DESORMES EXPERIMENT
8.6: THE SLOPES OF ISOTHERMS AND ADIABATS
8.7: SCALE HEIGHT IN AN ISOTHERMAL ATMOSPHERE
8.8: ADIABATIC LAPSE RATE
8.9: NUMERICAL VALUES OF SPECIFIC AND MOLAR HEAT CAPACITIES
8.10: HEAT CAPACITIES OF SOLIDS

9: ENTHALPY
9.1: ENTHALPY
9.2: CHANGE OF STATE
9.3: LATENT HEAT AND ENTHALPY

10: THE JOULE AND JOULE-THOMSON EXPERIMENTS

10.1: INTRODUCTION
10.2: THE JOULE EXPERIMENT
10.3: THE JOULE-THOMSON EXPERIMENT
10.4: CP MINUS CV
10.5: BLACKBODY RADIATION

11: HEAT ENGINES

11.1: INTRODUCTION
11.2: THE CARNOT CYCLE
11.3: THE STIRLING CYCLE
11.4: THE OTTO CYCLE
11.5: THE DIESEL CYCLE
11.6: THE RANKINE CYCLE (STEAM ENGINE)
11.7: A USEFUL EXERCISE
11.8: HEAT ENGINES AND REFRIGERATORS
11.9: ENTROPY IS A FUNCTION OF STATE

12: FREE ENERGY

12.1: REVIEW OF INTERNAL ENERGY AND ENTHALPY
12.2: FREE ENERGY
12.3: HELMHOLTZ FREE ENERGY
12.4: GIBBS FREE ENERGY
12.5: SUMMARY, THE MAXWELL RELATIONS, AND THE GIBBS-HELMHOLTZ RELATIONS
12.6: THE JOULE AND JOULE-THOMSON COEFFICIENTS
12.7: THE THERMODYNAMIC FUNCTIONS FOR AN IDEAL GAS
12.8: THE THERMODYNAMIC FUNCTIONS FOR OTHER SUBSTANCES
12.9: ABSOLUTE ENTROPY
12.10: CHARGING A BATTERY
12.11: SURFACE ENERGY

2 4/29/2021
12.12: FUGACITY

13: EXPANSION, COMPRESSION AND THE TDS EQUATIONS

13.1: COEFFICIENT OF EXPANSION
13.2: COMPRESSION
13.3: PRESSURE AND TEMPERATURE
13.4: THE TDS EQUATIONS
13.5: EXPANSION, COMPRESSION AND THE TDS EQUATIONS
13.6: YOUNG'S MODULUS
13.7: RIGIDITY MODULUS (SHEAR MODULUS)
13.8: VOLUME, TEMPERATURE AND THE GRÜNEISEN PARAMETER

14: THE CLAUSIUS-CLAPEYRON EQUATION

15: ADIABATIC DEMAGNETIZATION
15.1: INTRODUCTION
15.2: ADIABATIC DECOMPRESSION
15.3: ADIABATIC DEMAGNETIZATION
15.4: ENTROPY AND TEMPERATURE

16: NERNST'S HEAT THEOREM AND THE THIRD LAW OF

THERMODYNAMICS
16.1: NERNST'S HEAT THEOREM
16.2: THE THIRD LAW OF THERMODYNAMICS

17: CHEMICAL THERMODYNAMICS

17.1: EQUILIBRIUM CONSTANT
17.2: HEAT OF REACTION
17.3: THE GIBBS PHASE RULE
17.4: CHEMICAL POTENTIAL
17.5: PARTIAL AND MEAN MOLAR QUANTITIES
17.6: THE GIBBS-DUHEM RELATION
17.7: CHEMICAL POTENTIAL, PRESSURE, FUGACITY
17.8: ENTROPY OF MIXING, AND GIBBS' PARADOX
17.9: BINARY ALLOYS
17.10: TERNARY ALLOYS

18: EXPERIMENTAL MEASUREMENTS

18.1: INTRODUCTION
18.2: THERMAL CONDUCTIVITY
18.3: THE UNIVERSAL GAS CONSTANT
18.4: AVOGADRO'S NUMBER AND BOLTZMANN'S CONSTANT
18.5: SPECIFIC HEAT CAPACITIES OF SOLIDS AND LIQUIDS
18.6: SPECIFIC HEAT CAPACITIES OF GASES
18.7: LATENT HEAT OF FUSION
18.8: COEFFICIENT OF EXPANSION

BACK MATTER
INDEX
GLOSSARY

3 4/29/2021
CHAPTER OVERVIEW
1: INTRODUCTORY REMARKS

1.1: INTRODUCTION
1.2: CALORIC, CALORIES, HEAT AND ENERGY
1.3: EXTENSIVE AND INTENSIVE QUANTITIES
1.4: MOLE
1.5: PREPOSITIONS
1.6: APPLICABILITY OF EQUATIONS

1 4/29/2021
1.1: Introduction
Thermodynamics is a subject that has a wide variety of applications, including many in practical and engineering contexts.
However, by choice, I shall be treating this subject from a very “academic” point of view, in which disembodied forces will
compress ideal gases with frictionless pistons, seemingly far removed from the real engines of steel and gasoline which
engineers must design. This approach may appeal to those with an academic bent – but is it likely to be useful to the aspiring
engineer who lives and works in the real world? Need the practical engineer know and understand all this airy-fairy mumbo-
jumbo? I would just argue this – that the “academic” approach deals with the fundamental physical principles upon which all
practical applications must be built, and that an engineer above all others must thoroughly understand these principles. The
fundamental principles do not cease to apply in the practical world!
I don’t expect to get down to serious thermodynamics in the opening chapter. Instead I shall just discuss a few isolated,
unrelated miscellaneous bits and pieces that I thought worth doing. Furthermore, anyone who opens a book on
thermodynamics will see the symbol ∂ liberally sprinkled over almost every page, so I thought I’d write a short chapter –
Chapter 2 - on partial derivatives. That will not be intended as a formal course in mathematics, but just a brief summary of the
main properties of partial derivatives that you are likely to need. Thus I shan’t get down to serious thermodynamics until
Chapter 3.

Jeremy Tatum 4/1/2021 1.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7207

1.2: Caloric, Calories, Heat and Energy
It has long been understood that heat is a form of energy. But this has not always been so, and indeed it was not generally
accepted until the middle of the nineteenth century. Before then, heat was treated as though it were some sort of “imponderable
(weightless) fluid” known as caloric, which could flow out of one body into another. It is true that as long ago as 1799
Humphrey Davy showed that ice could be melted merely by rubbing two pieces together without the need of any “caloric”,
and indeed this could not be explained by the “caloric” theory. Davy argued – quite correctly – that friction between two
bodies must generate “a motion or vibration of the corpuscles of bodies”, and that the observation of the melting of ice by
rubbing alone showed that “we may reasonably conclude that this motion or vibration is heat”. Likewise at about the same
time Benjamin Thompson, Count Rumford, showed that the boring of cannon continuously produced heat in proportion to the
amount of work done in the boring process, and the amount of heat that could be so produced was apparently inexhaustible.
This again should have sounded the death knell of the caloric theory, and, like Davy, Rumford correctly suggested that heat is
a form of motion.
In spite of this evidence and the arguments of Davy and Rumford, it wasn’t until the middle of the nineteenth century that
caloric theory finally died, and this was a result of the famous experiments of James Prescott Joule to determine the
mechanical equivalent of heat.
There is some question as to whether the name should be pronounced “jool” (to rhyme with fool) or “jowl” (to rhyme with
fowl). Joule was from a beer-brewing family in Manchester, in the North of England. In a north of England accent, “jowl”
would be a preferred pronunciation, while “jool” would come more naturally in the south of England, although most modern
Mancunians, like the rest of us, nowadays say “jool”. The uncertainty in the pronunciation is an old one, and was used by the
brewery (which no longer exists) in Joule’s day as an advertising slogan for the beer. I am indebted to Dr Graham McDonald
of the Joule Laboratory, Salford University, who found the actual advertising slogan for Joule’s Ales:
Do you pronounce it Joule’s to rhyme with Schools, Joule’s to rhyme with Bowls, or Joule’s to rhyme with Scowls?
Whatever you call it, by Joule’s, or Joule’s, or Joule’s. It’s GOOD!
In the nineteenth century (and continuing to today) the metric unit of heat was the calorie (the quantity of heat required to raise
the temperature of a gram of water through one Celsius degree), and the imperial unit was the British Thermal Unit (the
quantity of heat required to raise the temperature of a pound of water through one Fahrenheit degree). What Joule did was to
show that the expenditure of a carefully measured amount of work always produced the same carefully measured amount of
heat. He did this by using falling weights to drive a set of rotating paddles to stir up a quantity of water in a calorimeter, the
motion (kinetic energy) of the water being damped by a system of fixed vanes inside the calorimeter. The amount of energy
expended was determined by the loss of potential energy of the falling weights, and the amount of heat generated was
determined by the rise in temperature of the water. He deduced that the “mechanical equivalent of heat” is 772 foot-pounds per
British thermal unit. That is, 772 foot-pounds of work will raise the temperature of a pound of water through one Fahrenheit
degree. In more familiar metric units, the mechanical equivalent of heat is 4.2 joules per calorie. He wrote: “If my views be
correct,... the temperature of the river Niagara will be raised about one fifth of a degree by its fall of 160 feet.”
(Exercise: Verify this by calculation or by measurement, whichever you find more convenient.)
Once we have accepted that heat is but a form of energy, there should be no further need for separate units, and the joule will
serve for both. That being so, we can interpret Joule’s experiment not so much as determining the “mechanical equivalent of
heat”, but rather as a measurement of the specific heat capacity of water.
In spite of this, the calorie is still (regrettably) used extensively today. Part of the reason for this is that, in measuring heat
capacities, we often drop a hot sample into water and measure the rise in temperature of the water. This tells us rather directly
what the heat capacity of the sample is in calories – i.e. the heat capacity relative to that of water. I suspect, however, that the
calorie remains with us not for scientific reasons, but because old habits die hard. There are several problems associated with
the continued use of the calorie. Roughly, the calorie is the heat required to raise the temperature of a gram of water through 1
Co. For precise work, however, it becomes necessary to state not only the isotopic constitution (and the purity) of the water,
but also through which Celsius degree its temperature is raised. Thus in the past we have defined the calorie as “one hundredth
of the heat required to raise the temperature of a gram of water from 0oC to 100oC”; or again as “the heat required to raise the
temperature of a gram of water from 14.5oC to 15.5oC”. This latter is about 4.184 joules, but there is really no need to know

Jeremy Tatum 3/4/2021 1.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7208

this conversion factor, unless you are specially interested in the specific heat capacity of water (which, by the way is rather
larger than many common substances). (You may have noticed that I have sometimes written oC and sometimes Co, and you
may have wondered which is correct, or whether the degree symbol should be used at all. This will be discussed in Chapter 3.)
The “calories” that nutritionists quote when talking about the calorific value of foods, is actually the kilocalorie and it is
sometimes (but by no means always) written Calorie, with a capital C. How much simpler it would all be if all of us just used
joules!
There is yet another problem associated with the continued use of “calories”. That is that we often come across formulas and
equations in thermodynamics in which a mysterious factor “J” appears. For example, there is a well-known equation CP − CV
= R/J. This relates the specific heat capacities of an ideal gas at constant pressure and volume to the universal gas constant R.
It is supposed to be understood in the equation that CP and CV are to be expressed in calories and R is to be expressed in
joules. The conversion factor between the two units, J, is the mechanical equivalent of heat, or the number of joules in a
calorie. This conversion factor between units will not be used in these notes, and all quantities expressing heat of energy will
be measured in the same units, which will normally be joules. The equation quoted above will be rendered simply as CP − CV
= R. (The letter J, not in italics, will, of course, continue to be used to denote the unit the joule, but not J, in italics, for a
conversion factor.)

Jeremy Tatum 3/4/2021 1.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7208

1.3: Extensive and Intensive Quantities
There is a useful and important distinction in thermodynamics between extensive (or “capacitive”) and intensive quantities.
Extensive quantities are those that depend upon the amount of material. Examples would include the volume, or the heat
capacity of a body. The heat capacity of a body is the amount of heat required to raise its temperature by one degree, and
might be expressed in J Co−1.
Intensive quantities do not depend on the amount of material. Temperature and pressure are examples. Another would be the
specific heat capacity of a substance, which is the amount of heat required to raise unit mass of it through one degree, and it
might be expressed in J kg−1 Co −1. This is what is commonly (though loosely) called “the specific heat”, but we shall use the
correct term: specific heat capacity.
Incidentally, we would all find it much easier to understand each other if we all used the word “specific” in contexts such as
these to mean “per unit mass”.
“Molar” quantities are also intensive quantities. Thus the “molar heat capacity” of a substance is the amount of heat required
to raise the temperature of one mole of the substance through one degree. I shall have to define “mole” in the next section.
Some authors adopt the convention that extensive quantities are written with capital letters, and the corresponding intensive
quantities are written in small letters. Thus C would be the heat capacity of a body in J Co −1 and c would be the specific heat
capacity of a substance in J kg−1 Co −1. This is undeniably a useful distinction and one that many will find helpful. I have a few
difficulties with it. Among these are the following: Some authors (not many) use the opposite convention – small letters for
extensive quantities, capitals for intensive. Some authors make exceptions, using P and T for the intensive quantities pressure
and temperature. Also, how are we to distinguish between extensive, specific and molar quantities? Three different fonts? This
may indeed be a solution – but there is still a problem. For example, we shall become familiar with the equation dU = T dS − P
dV. Here U, S and V are internal energy, entropy and volume. Yet the equation (and many others that we could write) is equally
valid whether we mean extensive, specific or molar internal energy, entropy and volume. How do we deal with that? Write the
equation three times in different fonts?
Because of these difficulties, I am choosing not to use the capital letter, small letter, convention, and I am hoping that the
context will make it clear in any particular situation. This is, I admit, rather a leap of faith, but let’s see how it works out.

Jeremy Tatum 4/29/2021 1.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7209

1.4: Mole
According to our present state of knowledge, a mass of 12 grams of the 12C isotope of carbon contains 6.022 141 99 × 1023
atoms. This number is called Avogadro’s number. A mole of an element is the amount of that element that has the same
number of atoms as there are atoms in 12 grams of 12C, and a mole of a compound is the amount of that compound that has the
same number of molecules as there are atoms in 12 grams of 12C. Likewise a mole of geese is 6.022 141 99 × 1023 geese and a
mole of baseball caps is 6.022 141 99 × 1023 baseball caps.
Why do we define Avogadro’s number in terms of 12 grams of carbon-12? This is a long story involving the history of physics
and chemistry. None of us was born with a complete knowledge of physics and chemistry and it took a long time to reach our
state of knowledge today. We did not always proceed along our path with complete logic, and doubtless, if we did, we might
have defined Avogadro’s number differently. I am not going to go into the history of how we arrived at this particular
definition. Suffice it to say that, if you know that the molecular weight of nitrogen gas (a diatomic molecule) is 28, then 28
grams of nitrogen has Avogadro’s number of molecules in it. Indeed the phrase “molecular weight” is not a happy one; it
would be better to call it the “molar mass”, which is 28 grams.
We might note, however, that when we are doing “SI” calculations, based on MKS units, we shall usually use the kilomole,
which is 6.022 141 99 × 1026 molecules.

Jeremy Tatum 4/29/2021 1.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7210

1.5: Prepositions
Prepositions play an important part in thermodynamics! Heat may be supplied to a system or lost from it. Work may be done
on a gas or by it.
An answer to a question in thermodynamics of “5 joules” is meaningless unless you make it clear and unambiguous whether
the system lost 5 joules of heat or gained 5 joules, or whether the gas did 5 joules of work, or you did 5 joules of work on the
gas. And it is of no avail to say that the answer is “−5 joules” in the vague hope that I might know what you mean by the
minus sign. You must explicitly state in words whether 5 joules was lost or gained – or your reader, or your examiner, will not
understand you (and will give you no marks) or will misunderstand you (and will deduct some marks).
I used to tell students that if they wrote “5 joules” without the necessary preposition, they would get no marks for their answer.
If they used a preposition, but chose the wrong one (the gas lost 5 joules instead of gaining it) I would take a mark off. Be
warned!

Jeremy Tatum 4/8/2021 1.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7211

1.6: Applicability of Equations
There seem to be lots and lots of equations in thermodynamics. Some of them are of very general applicability. For example
the equation dU = dQ + dW, which is known as the First Law of Thermodynamics, tells us that the increase in the internal
energy of a system is equal to the heat supplied to it plus the work done on it, and it is obviously of very general applicability
and is true whatever the nature of the system. An equally well known equation is PV = RT. But this equation, which relates
pressure, molar volume and temperature, applies only to an ideal gas. It doesn’t apply to a vapour (which is a gas that is close
to the temperature at which it will condense), and still less does it apply to a liquid or a solid. Although we often deal in
thermodynamics with a gas held inside a cylinder, thermodynamics is by no means confined to gases, let alone ideal gases.
This section is just an advance warning to be conscious, whenever you see or use an equation, whether the equation is of great
generality or whether it applies only to a particular substance or to some special thermodynamic process or to a narrow set of
circumstances.

Jeremy Tatum 4/15/2021 1.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7212

CHAPTER OVERVIEW
2: PARTIAL DERIVATIVES

2.1: INTRODUCTION
2.2: PARTIAL DERIVATIVES
2.3: IMPLICIT DIFFERENTIATION
2.4: PRODUCT OF THREE PARTIAL DERIVATIVES
2.5: SECOND DERIVATIVES AND EXACT DIFFERENTIALS
2.6: EULER'S THEOREM FOR HOMOGENEOUS FUNCTIONS
2.7: UNDETERMINED MULTIPLIERS
2.8: DEE AND DELTA

1 4/29/2021
2.1: Introduction
Any text on thermodynamics is sure to be liberally sprinkled with partial derivatives on almost every page, so it may be
helpful here to give a brief summary of some of the more useful formulas involving partial derivatives that we are likely to use
in subsequent chapters.

Jeremy Tatum 4/29/2021 2.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7214

2.2: Partial Derivatives
The equation
z = z(x,  y) (2.2.1)

represents a two-dimensional surface in three-dimensional space. The surface intersects the plane y = constant in a plane curve
in which z is a function of x. One can then easily imagine calculating the slope or gradient of this curve in the plane y =
constant. This slope is ( ∂z

∂x
) - the partial derivative of z with respect to x, with y being held constant. For example, if
y

z = y ln x, (2.2.2)

then
∂z y
( ) = , (2.2.3)
∂x x
y

y being treated as though it were a constant, which, in the plane y = constant, it is. In a similar manner the partial derivative of
z with respect to y, with x being held constant, is
∂z
( ) = ln x (2.2.4)
∂y
x

When you have only three variables – as in this example – it is usually obvious which of them is being held constant. Thus ∂z/
∂y can hardly mean anything other than at constant x. For that reason, the subscript is often omitted. In thermodynamics, there
are often more than three variables, and it is usually (I would say always) essential to indicate by a subscript which quantities
are being held constant.
In the matter of pronunciation, various attempts are sometimes made to give a special pronunciation to the symbol ∂. (I have
heard “day”, and “dye”.) My own preference is just to say “partial dz by dy”.
Let us suppose that we have evaluated z at (x , y). Now if you increase x by δx, what will the resulting increase in z be?
Obviously, to first order, it is δx. And if y increases by δy, the increase in z will be δy . And if both x and y increase, the
∂x

∂x
∂z

∂y

corresponding increase in z, to first order, will be

∂z ∂z
δz = δx + δy (2.2.5)
∂x ∂y

No great and difficult mathematical proof is needed to “derive” this; it is just a plain English statement of an obvious truism.
The increase in z is equal to the rate of increase of z with respect to x times the increase in x plus the rate of increase of z with
respect to y times the increase in y.
dy
Likewise if x and y are increasing with time at rates dx

dt
and dt
, the rate of increase of z with respect to time is
dz ∂z dx ∂z dy
= + . (2.2.6)
dt ∂x dt ∂y dt

Jeremy Tatum 4/1/2021 2.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7215

2.3: Implicit Differentiation
Equation 2.2.5 can be used to solve the problem of differentiation of an implicit function. Consider, for example, the unlikely
equation
2 3
ln(xy) = x y (2.3.1)

Calculate the derivative dy/dx. It would be easy if only one could write this in the form y = something; but it is difficult
(impossible as far as I know) to write y explicitly as a function of x. Equation 2.3.1 implicitly relates y to x. How are we going
to calculate dy/dx?
The curve f (x, y) = 0 might be considered as being the intersection of the surface z = f (x, y) with the plane z = 0 . Seen
thus, the derivative dy/dx can be thought of as the limit as δx and δy approach zero of the ratio δy/δx within the plane z = 0 ;
that is, keeping z constant and hence δz equal to zero. Thus equation 2.2.5 gives us that
dy ∂f ∂f
= −( )/( ). (2.3.2)
dx ∂x ∂y

For example, show that, for Rquation 2.3.1,

2 3
dy y(2 x y − 1)
= . (2.3.3)
dx x(1 − 3 x2 y 3 )

Jeremy Tatum 3/18/2021 2.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7216

2.4: Product of Three Partial Derivatives
Suppose x, y and z are related by some equation and that, by suitable algebraic manipulation, we can write any one of the
variables explicitly in terms of the other two. That is, we can write
x = f (y,  z), (2.4.1)

or
y = y(z,  x), (2.4.2)

or
z = z(x,  y). (2.4.3)

Then
∂x ∂x
δx = δy + δz, (2.4.4)
∂y ∂z

∂y ∂y
δy = δz + δx (2.4.5)
∂z ∂x

and
∂z ∂z
δz = δx + δy. (2.4.6)
∂x ∂y

Eliminate δy from Equations 2.4.4 and 2.4.5:

∂x ∂y ∂x ∂x ∂y
δx (1 − ) = δz ( + ), (2.4.7)
∂y ∂x ∂z ∂y ∂z

and δz from Equations 2.4.4 and 2.4.6:

∂x ∂z ∂x ∂x ∂z
δx (1 − ) = δy ( + ). (2.4.8)
∂z ∂x ∂y ∂z ∂y

Since z and x can be varied independently, and x and y can be varied independently, the only way in which Equations 2.4.7 and
2.4.8 can always be true is for all of the expressions in parentheses to be zero. Equating the left-hand parentheses to zero

shows that
∂x ∂y
= 1/ (2.4.9)
∂y ∂x

and
∂x ∂z
= 1/ . (2.4.10)
∂z ∂x

These results may seem to be trivial and “obvious” – and so they are, provided that the same quantity is being kept constant in
the derivatives of both sides of each equation. In thermodynamics we are often dealing with more variables than just x, y and z,
and we must be careful to specify which quantities are being held constant. If, for example, we are dealing with several
∂y
variables, such as u, v, w, x, y, z, it is not in general true that ∂u

∂y
= 1/
∂u
, unless the same variables are being held constant on
both sides of the equation.
Return now to Equation 2.4.7. The left hand parenthesis is zero, and this, together with Equation 2.4.10 , results in the
important relation:
∂x ∂y ∂z
( ) ( ) ( ) = −1. (2.4.11)
∂y ∂z ∂x
z x y

Jeremy Tatum 4/1/2021 2.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7217

2.5: Second Derivatives and Exact Differentials
If z = z(x, y) , we can go through the motions of calculating ∂z

∂x
and ∂z

∂y
, and we can then further calculate the second
2 2 2 2

derivatives ∂

∂x2
z
, ∂

∂y 2
x
, ∂

∂y∂x
z
and ∂

∂y∂x
z
. It will usually be found that the last two, the mixed second derivatives, are equal; that is,
it doesn’t matter in which order we perform the differentiations.

Example 2.5.1
Let z = x sin y . Show that
2 2
∂ z ∂ z
= = cos y. (2.5.1)
∂x∂y ∂y∂x

Solution
We examine in this section what conditions must be satisfied if the mixed derivatives are to be equal.
Figure II.1 depicts z as a “well-behaved” function of x and y. By “well-behaved” in this context I mean that z is
everywhere single-valued (that is, given x and y there is just one value of z), finite and continuous, and that its derivatives
are everywhere continuous (that is, no sudden discontinuities in either the function itself or its slope). “Good behaviour”
in this sense is the sufficient condition that the mixed second derivatives are equal.

Let us calculate the difference δz in the heights of A and C. We can go from A to C via B or via D, and δz is route-
independent. That is, to first order,
(A) (B) (A) (D)
∂z ∂z ∂z ∂z
δz = ( ) δx + ( ) δy = ( ) δy + ( ) δx. (2.5.2)
∂x ∂y ∂y ∂x
y x x y

Here the superscript (A) means “evaluated at A”.

Divide both sides by δx δy:
(B) (A) (D) (A)
∂z ∂z ∂z ∂z
( ) −( ) ( ) −( )
∂y ∂y ∂x ∂x
x x y y
= . (2.5.3)
∂x ∂y

If we now go to the limit as δx and δy approach zero (the equation now becomes exact rather than merely “to first order”),
this becomes:

Jeremy Tatum 4/1/2021 2.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7218

 2 2
∂ z ∂ z
= . (2.5.4)
∂xδy ∂yδx

A further property of a function that is well-behaved in the sense described is that if the differential dz can be written in
the form
dz = A(x,  y)dx + B(x,  y)dy, (2.5.5)

then Equation 2.5.3 implies that,

∂A ∂B
= . (2.5.6)
∂y ∂x

A differential dz is said to be exact if the following conditions are satisfied: The integral of dz between two points is
route-independent, and the integral around a closed path (i.e. you end up where you started) is zero, and if equations 2.5.3
and 2.5.5 are satisfied.
If a differential such as Equation 2.5.4 is exact – i.e., if it is found to satisfy the conditions for exactness – then it should
be possible to integrate it and determine z(x , y). Let us look at an example. Suppose that
dz = (4x − 3y − 1)dx + (−3x + 2y + 4)dy. (2.5.7)

It is readily seen that this is exact. The problem now, therefore, is to find z(x, y).
Let u = ∫ (4x − 3y − 1)dx
So that
2
u = 2x − 3yx − x + g(y). (2.5.8)

Note that we are treating y as constant. The “constant” of integration depends on the value of y – i.e. it is an arbitrary
function of y.
Of course u is not the same as z – unless we can find a particular function g(y) such that u indeed is the same as z.
Now du = ∂u

∂x
+
∂u

∂y
dy ; that is,

dg
du = (4x − 3y − 1)dx + (−3x + ) dy. (2.5.9)
dy

dg
Then du = dz (and u = z plus an arbitrary constant) provided that dy
= 2y + 4 . That is,
2
g(y) = y + 4y + constant. (2.5.10)

Thus
2 2
z = 2x − 3xy + y − x + 4y + constant (2.5.11)

The reader should verify that this satisfies equation 2.5.6. The reader should also try letting
2
ν = −3xy + y + 4y + f (x) (2.5.12)

(where did this come from?) and go through a similar argument to arrive again at equation 2.5.10.

Consider another example

Example 2.5.2
x
dz = 3 ln y dx + dy. (2.5.13)
y

You should immediately find that this differential is not exact, and, to emphasize that, I shall use the symbol đz, the
special symbol đ indicating an inexact differential. However, given an inexact differential đz, it is very often possible to

Jeremy Tatum 4/1/2021 2.5.2 CC-BY-NC https://phys.libretexts.org/@go/page/7218

 find a function H(x , y) such that the differential dw = H(x , y) đz is exact, and dw can then be integrated to find w as a
function of x and y. The function H(x , y) is called an integrating factor. There may be more than one possible integrating
factor; indeed it may be possible to find one simply of the form F(x) or maybe G(y). There are several ways for finding an
integrating factor. We’ll do a simple and straightforward one. Let us try and find an integrating factor for the inexact
differential đz above. Thus, let dw = F(x)dz, so that
xF
dw = 3F ln y dx + dy. (2.5.14)
y

For dw to be exact, we must have

∂ ∂ xF
(3F ln y) = ( ). (2.5.15)
∂y ∂x y

That is,
3F 1 dF
= (F + x ). (2.5.16)
y y dx

Upon integration and simplification we find that

2
F =x , (2.5.17)

or any multiple thereof, is an integrating factor, and therefore

3
x
2
dw = 3 x ln y dx + dy (2.5.18)
y

is an exact differential. The reader should confirm that this is an exact differential, and from there show that
3
w =x ln y + constant (2.5.19)

To anticipate – what has this to do with thermodynamics? To give an example, the state of many simple thermodynamical
systems can be specified by giving the values of three intensive state variables, P, V and T, the pressure, molar volume and
temperature. That is, the state of the system can be represented by a point in PVT space. Often, there will be a known relation
(known as the equation of state) between the variables; for example, if the substance involved is an ideal gas, the variables will
be related by PV = RT, which is the equation of state for an ideal gas; and the point representing the state of the system will
then be represented by a point that is constrained to lie on the two-dimensional surface PV = RT in three-dimensional PVT
space. In that case it will be necessary to specify only two of the three variables. On the other hand, if the equation of state of a
particular substance is unknown, you will have to give the values of all three variables.
Now there are certain quantities that one meets in thermodynamics that are functions of state. Two that come to mind are
entropy S and internal energy U. By function of state it is meant that S and U are uniquely determined by the state (i.e. by P, V
and T). If you know P, V and T, you can calculate S and U or any other function of state. In that case, the differentials dS and
dU are exact differentials.
The internal energy U of a system is defined in such a manner that when you add a quantity dQ of heat to a system and also do
an amount of work dW on the system, the increase dU in the internal energy of the system is given by
dU = dQ + dW . (2.5.20)

Here dU is an exact differential, but dQ and dW are clearly not. You can achieve the same increase in internal energy by any
combination of heat and work, and the heat you add to the system and the work you do on it are clearly not functions of the
state of the system.
Some authors like to use a special symbol, such as đ, to denote an inexact differential (but beware, I have seen this symbol
used to denote an exact differential!). I shall not in general do this, because there are many contexts in which the distinction is
not important, or, if it is, it is obvious from the context whether a given differential is exact or not. If, however, there is some
context in which the distinction is important (and there are many) and in which it may not be obvious which is which, I may,
with advance warning, use a special đ for an inexact differential, and indeed I have already done so earlier in this section.

Jeremy Tatum 4/1/2021 2.5.3 CC-BY-NC https://phys.libretexts.org/@go/page/7218

2.6: Euler's Theorem for Homogeneous Functions
There is a theorem, usually credited to Euler, concerning homogenous functions that we might be making use of.
A homogenous function of degree n of the variables x, y, z is a function in which all terms are of degree n. For example, the
function f (x,  y,  z) = Ax + By + C z + Dx y + Ex z + Gy x + H zx + I zy + Jxyz
3 3 3 2
is a homogenous function of
2 2 2 2

x, y, z, in which all terms are of degree three.

∂f ∂f ∂f
The reader will find it easy to evaluate the partial derivatives ∂x
, 
∂x
, 
∂x
and equally easy (if slightly tedious) to evaluate the
∂f ∂f ∂f
expression x
∂x
+y
∂y
+z
∂z
. Tedious or not, I do urge the reader to do it. You should find that the answer is
3 3 3 2 2 2 2 2 2
3Ax + 3By + 3C z + 3Dx y + 3Ex z + 3F y z + 3Gy x + 3H zx + 3I zy + 3Jxyz.

∂f ∂f ∂f
In other words, x ∂x
+y
∂y
+z
∂z
= 3f . If you do the same thing with a homogenous function of degree 2, you will find that
∂f ∂f ∂f
x
∂x
+y
∂y
+z
∂z
= 2f . And if you do it with a homogenous function of degree 1, such as Ax + By + C z , you will find
∂f ∂f ∂f
that x ∂x
+y
∂y
+z
∂z
=f . In general, for a homogenous function of x, y, z... of degree n, it is always the case that

∂f ∂f ∂f
x +y +z +. . . = nf . (2.6.1)
∂x ∂y ∂z

This is Euler's theorem for homogenous functions.

Jeremy Tatum 4/1/2021 2.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7219

2.7: Undetermined Multipliers
Let ψ(x, y, z) be some function of x, y and z. Then if x, y and z are independent variables, one would ordinarily understand that,
where ψ is a maximum, the derivatives are zero:
∂ψ ∂ψ ∂ψ
= = = 0. (2.7.1)
∂x ∂y ∂z

However, if x, y and z are not completely independent, but are related by some constraining equation such as f(x, y, z) = 0, the
situation is slightly less simple. (In a thermodynamical context, the three variables may be, for example, three “intensive state
variables”, P, V and T, and ψ might be the entropy, which is a function of state. However the intensive state variables may not
be completely independent, since they are related by an “equation of state”, such as PV = RT.)
If we move by infinitesimal displacements dx, dy, dz from a point where ψ is a maximum, the corresponding changes in ψ and
f will both be zero, and therefore both of the following equations must be satisfied.
∂ψ ∂ψ ∂ψ
dψ = dx + dy + dz = 0, (2.7.2)
∂x ∂y ∂z

∂f ∂f ∂f
df = dx + dy + dz = 0. (2.7.3)
∂x ∂y ∂z

Consequently any linear combination of ψ and f, such as Φ = ψ + λf, where λ is an arbitrary constant, also satisfies a similar
equation. The constant λ is sometimes called an “undetermined multiplier” or a “Lagrangian multiplier”, although often some
additional information in an actual problem enables the constant to be identified.
In summary, the conditions that ψ is a maximum (or minimum or saddle point), if x, y and z are related by a functional
constraint f (x, y, z) = 0, are
∂Φ ∂Φ ∂Φ
=0 = 0, = 0, (2.7.4)
∂x ∂y ∂z

where

Φ = ψ + λf . (2.7.5)

Of course, if ψ is a function of many variables x1 , x2 , x3..., and the variables are subjected to several constraints, such as f =
0, g = 0, h = 0, etc., where f, g, h, etc., are functions connecting all or some of the variables, the conditions for ψ to be a
maximum (etc.) are
∂ψ ∂ψ ∂ψ ∂ψ
+λ +μ +ν +. . . = 0,  i = 1,  2,  3 (2.7.6)
∂xi ∂xi ∂xi ∂xi

Jeremy Tatum 3/11/2021 2.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8566

2.8: Dee and Delta
We have discussed the special meanings of the symbols ∂ and đ, but we also need to be clear about the meanings of the more
familiar differential symbols Δ, δ , and d . It is often convenient to use the symbol Δ to indicate an increment (not necessarily a
particularly small increment) in some quantity. We can then use the symbol δ to mean a small increment. We can then say that
if, for example, y = x , and if x were to increase by a small amount δx, the corresponding increment in y would be given
2

approximately by
δy ≅2xδx (2.8.1)

That is,
∂y
≅2x. (2.8.2)
∂x

dy
This doesn’t become exact until we take the limit as \(δx\) and \(δy\) approach zero. We write this limit as dx
and then it is
exactly true that
dy
= 2x. (2.8.3)
dx

There is a valid point of view that would argue that you cannot write dx or dy alone, since both are zero; you can write only
the ratio . It would be wrong, for example, to write
dx

dy

dy = 2x dx, (2.8.4)

or at best it is tantamount to writing 0 = 0. I am not going to contradict that argument, but, at the risk of incurring the wrath of
some readers, I am often going to write equations such as Equation 2.8.4, or, more likely, in a thermodynamical context,
equations such as

dU = T dS − P dV , (2.8.5)

even though you may prefer me to say that, for small increments,

δU ≅T δS − P δV . (2.8.6)

I am going to argue that, in the limit of infinitesimal increments, it is exactly true that dU = T dS − P dV . After all, the
smaller the increments, the closer it becomes to being true, and, in the limit when the increments are infinitesimally small, it is
exactly true, even if it does just mean that zero equals zero. I hope this does not cause too many conceptual problems.

Jeremy Tatum 4/29/2021 2.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8567

CHAPTER OVERVIEW
3: TEMPERATURE
During our studies of heat and thermodynamics, we shall come across a number of simple, easy-tounderstand terms such as entropy,
enthalpy, Gibbs free energy, chemical potential and fugacity, and we expect to have no difficulty with these. There is, however, one
concept that is really quite difficult to grasp, and that is temperature. We shall do our best to understand it in this chapter.

3.2: ZEROTH LAW OF THERMODYNAMICS

Perhaps the simplest concept of temperature is to regard it as a potential function whose gradient determines the direction and rate of
flow of heat. If heat flows from one body to another, the first is at a higher temperature than the second. If there is no net flow of heat
from one body to another, the two bodies are in thermal equilibrium, and their temperatures are equal.

3.3: TEMPERATURE SCALES I

In everyday practice, we use either the Celsius or the Fahrenheit temperature scales, depending on what we are used to, or the fashion
of the day, or what our Government tells us we should be using.

3.4: TEMPERATURE SCALES II

We now know – by definition – the temperatures at the two fixed points on the Celsius and Kelvin scales. But what about
temperatures between the fixed points? We could say that the temperature halfway between the melting point of ice and the boiling
point of water is 50 ºC, or we could divide the temperature between the two fixed points into 100 equal intervals. But: What do we
mean by “halfway” or by “equal intervals” in such a proposal?

1 4/29/2021
3.2: Zeroth Law of Thermodynamics
Perhaps the simplest concept of temperature is to regard it as a potential function whose gradient determines the direction and
rate of flow of heat. If heat flows from one body to another, the first is at a higher temperature than the second. If there is no
net flow of heat from one body to another, the two bodies are in thermal equilibrium, and their temperatures are equal.
We can go further and assert that

If two bodies are separately in thermal equilibrium with a third body, then they are
also in thermal equilibrium with each other.
According to taste, you may regard this as a truism of the utmost triviality or as a fundamental law of the most profound
significance. Those who see it as the latter will refer to it as the Zeroth Law of Thermodynamics (although the "zeroth" does
sound a little like an admission that it was added as an afterthought to the other "real" laws of thermodynamics).
We might imagine that the third body is a thermometer of some sort. In fact it need not even be an accurately calibrated
thermometer. We insert the thermometer into one of our two bodies (we are not thinking particularly of human bodies here),
and it indicates some temperature. Then we insert it into the second body. If it indicates the same temperature as indicated for
the first body, then the Zeroth Law asserts that, if we now place our two bodies into contact with each other, there will be no
net flow of heat from one to the other. There exists some measure which all three bodies have in common and which dictates
that there is no net flow of heat from any one to any other, and the three bodies are in thermal equilibrium. That measure is
what we call their temperature.
To some, this will sound like saying :"if A and C are at the same temperature, and if B and C are at the same temperature, then
A and B are at the same temperature". Others, of philosophical bent, may want to pursue the concept to greater rigor. In any
case, at whatever level of rigor is used, what the Zeroth Law establishes is the existence of some quantity called temperature,
but it doesn't really tell us how to define a temperature scale quantitatively. It is as if we have established the existence of
something called "length" or "mass", but we haven't really specified yet how to measure it or what units to express it in. We
could, for example, discuss the concepts of "length" or of "mass" by describing a test to show whether two lengths, or two
masses, were equal, but without developing any units for expressing such concepts qualitatively. That, I think, is where the
Zeroth Law leaves us.

Jeremy Tatum 4/29/2021 3.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7222

3.3: Temperature Scales I
In everyday practice, we use either the Celsius or the Fahrenheit temperature scales, depending on what we are used to, or the
fashion of the day, or what our Government tells us we should be using. In the Fahrenheit scale, the freezing point of water is
32 oF and the boiling point is 212 oF, so that there are 180 Fo between the two fixed points. In the Celsius scale, the freezing
point of water is 0 oC and the boiling point is 100 oC, so that there are 100 Co between the two fixed points. (When Celsius
originally introduced his scale, he set the temperature of boiling water as 0, and the temperature of melting ice as 100. That
was reversed within a few years!) The Celsius scale was formerly called "the" centigrade scale, but presumably any scale with
100 degrees between two fixed points could be called a centigrade scale, so we now call it (or are supposed to call it) the
Celsius scale.
Conversion is obviously by

F = 1.8C + 32 (3.3.1)

and
F − 32 5
C = = (F − 32). (3.3.2)
1.8 9

Note that "a temperature of so many degrees on the Fahrenheit scale" is written oF and "a temperature of so many degrees on
the Celsius scale" is written oC; whereas "a temperature interval of so many Fahrenheit degrees" is written Fo and "a
temperature interval of so many Celsius degrees" is written Co. In either case, the degrees symbol (o) is mandatory.
In scientific work, we generally use the Kelvin temperature scale. The two fixed points on the Kelvin scale are the absolute
zero of temperature, which is assigned the temperature 0 K, and the triple point of the water-ice-steam system, which is
assigned the temperature 273.16 K. Thus it could reasonably be said that the Kelvin scale is not a centigrade scale, since it
doesn't have 100 degrees between its two fixed points. However, the size of the degree on the Kelvin scale is almost exactly
the same as the size of the Celsius degree, because the absolute zero of temperature is about –273.15 oC and the temperature of
the triple point is about 0.01 oC. The definition of the Kelvin scale, however, does not mention the Celsius scale, and therefore,
although the size of the degrees is about the same on both scales, this is not inherent in the definition. One might speculate
about what might happen in the far distant future if people no longer use the Celsius scale and it is totally forgotten. People
then will wonder what possessed us to divide the Kelvin scale into 273.16 divisions between its two fixed points!
It would not be good enough to define the upper fixed point of the kelvin scale as the temperature of "melting ice", because
this depends on the pressure. The triple point is the temperature at which ice, water and steam are in equilibrium, and it occurs
at a temperature of about 0.01 oC and exactly 273.16 K, and a pressure of about 610.6 Pa.
The Kelvin scale starts at zero at the lowest conceivable temperature. The kelvin (K) is therefore regarded as a unit of
temperature, much as a metre is regarded as a unit of length, or a kilogram as a unit of mass. One therefore does not talk about
a temperature of so many "degrees Kelvin", any more than one would talk about a length of so many "degrees metre" or a
mass of so many "degrees kilogram". When using the Kelvin scale, therefore, we talk simply of a temperature of "280 kelvins"
or "280 K". We do not use the word "degree", nor do we use the symbol o.
In the British Engineering System of units, which is used exclusively in the United States and has never been used in Britain,
the Rankine scale is used. The lower fixed point is the absolute zero of temperature, and it is assigned the temperature 0 R, and
the size of the rankine is equal to the size of the Fahrenheit degree. Melting ice at 0 oC has a temperature of 459.67 R, and the
triple point has a temperature of 459.688 R.
I doubt whether the Réaumur scale has been used anywhere in the last 50 years, but it has probably been used in the last 100.
This had melting ice at 0 o R and steam at 80 oR. I mention this only to point out that if you see a temperature given as so
many oR, you might not know whether the Rankine or Réaumur scale is intended! (Strictly, °R would denote degrees
Réaumur, while R would denote rankines – but can you trust that?)
In these notes, the Kelvin scale will be the scale that is normally used. There may be occasional use of the Celsius scale, but
we shall not use the Fahrenheit, Rankine or Réaumur scales.

Jeremy Tatum 4/22/2021 3.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7223

3.4: Temperature Scales II
We now know – by definition – the temperatures at the two fixed points on the Celsius and Kelvin scales. But what about
temperatures between the fixed points? We could say that the temperature halfway between the melting point of ice and the
boiling point of water is 50 ºC, or we could divide the temperature between the two fixed points into 100 equal intervals. But:
What do we mean by “halfway” or by “equal intervals” in such a proposal? This leaves us rather stumped.
Here is one suggestion.
We could construct a glass capillary tube with a bulb at the bottom containing mercury, which also extends a short way up the
capillary. We could note the length of the mercury column when the tube was immersed in melting ice and call the temperature
0 ºC, and again when it is in boiling water (100 ºC). We could then divide the length of the tube between these two marks into
100 equal intervals of length, and use that to define our temperature scale. But you may ask: How do we know that mercury
expands (relative to glass) uniformly with temperature? Well, it expands uniformly, by definition, with temperature on the
mercury-in-glass temperature scale. Indeed, we can define the temperature in the mercury-in-glass scale by
lt − l0 o
t = 100 ×   C. (3.4.1)
l100 − l0

(I am going to use the symbol T in these notes for temperature in kelvin. Here I am using t for temperature on the Celsius
scale.)
If we place the thermometer (for such it is) in a bowl of warm water, and the length of the mercury column is halfway between
l0 and l100, we could say that the temperature of the water in the bowl is, by definition, 50 ºC on the mercury-in-glass scale.
Now let us repeat the experiment with another type of thermometer, using some different property of matter which is also
known to vary with temperature. We might choose, for example, to use the electrical resistance R of a length of platinum wire;
or the thermoelectric potential difference V that appears when we heat the junction of two different metals; or the pressure P of
some gas when it is heated up but kept at constant volume. We could try immersing each of these thermometers into melting
ice and boiling water and we could interpolate linearly for intermediate temperatures. Thus, using the resistance of the
platinum wire, we could define a platinum resistance temperature scale by
Rf − R0
o
t = 100 ×   C. (3.4.2)
R100 − R0

Or we could define a thermoelectric temperature scale by

Vt − V0
o
t = 100 ×   C. (3.4.3)
V100 − V0

Or we could define a constant volume gas temperature scale by

Pt − P0
o
t = 100 ×   C. (3.4.4)
P100 − P0

But what assurance do we have that all of these temperature scales are the same? What assurance do we have that the
resistance of platinum increases linearly on the temperature scale defined by the mercury-in-glass thermometer? What
assurance do we have that, when we immerse all of these thermometers in the water that registered 50 ºC for the mercury-in-
glass thermometer, they will all register 50 ºC?
The answer is that we have no such assurance.
What we need to do is either choose one particular phenomenon quite arbitrarily to use for our standard temperature scale, or
somehow define an absolute temperature scale which is absolute in the sense that it is defined independently of the properties
of any particular substance. It turns out that it is possible to do the latter, and to define a temperature scale that is absolute and
independent of the properties if any particular substance by means of an idealized theoretical concept called a Carnot Heat
Engine. This imaginary engine uses as its operating medium an equally imaginary substance called an ideal gas, and indeed
the temperature indicated by a constant volume gas thermometer is identical to the absolute temperature defined by a Carnot
engine – provided that the gas used is an ideal gas! The best that can be said for real gases is that, at low pressures, they

Jeremy Tatum 4/29/2021 3.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7224

behave very much like an ideal gas; and indeed if you somehow extrapolate the behaviour or a gas to its behaviour at zero
pressure (when there isn’t any gas at all!), it would behave exactly like a real gas.
Until we have discussed what are meant by a real gas and by a Carnot engine, all this has served to do is to underline what we
said in the Introduction to this chapter – namely that there are a number of relatively easy concepts in thermodynamics, but
temperature is not one of them.
If we do eventually understand what a Carnot engine is and we can construct in our minds a definition of what is meant by an
absolute temperature scale, there will remain the problem of reproducing such a scale in practice. That is the purpose of the
International Temperature Scale 1990 (ITS90). On this scale a number of fixed points, such as
the triple point of hydrogen
the triple point of neon
the triple point of water
the freezing point of zinc
the freezing point of silver
the freezing point of gold
etc.,
are assigned certain values. In the cases of the six points listed, these values are
13.8033
24.5561
273.16
692.677
1234.93
1337.33
kelvin respectively.
A number of standard instruments are to be used in different temperature ranges, with defined interpolation formulas for
temperatures between the fixed points. A complete description of ITS90 would be rather lengthy (see, for example,
http://www.omega.com/techref/intltemp.html), but its purpose is to reproduce as precisely as practically possible the absolute
temperature scale as defined by the Carnot engine.

Jeremy Tatum 4/29/2021 3.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7224

CHAPTER OVERVIEW
4: THERMAL CONDUCTION

4.1: ERROR FUNCTION

4.2: INTRODUCTION
4.3: THERMAL CONDUCTIVITY
4.4: THE HEAT CONDUCTION EQUATION
4.5: A SOLUTION OF THE HEAT CONDUCTION EQUATION

1 4/29/2021
4.1: Error Function
Before we start this chapter, let’s just make sure that we are familiar with the error function erf a. We may need it during this
chapter.
Here is a graph of the gaussian function
1 −x
2

y = −e . (4.1.1)
√π

I have chosen the coefficient 1/√− π so that the area under the curve, from − ∞ to + ∞ is 1. The maximum value, which occurs
−−−
at x = 0, is 1/√−π = 0.5642, and it is easy to show that the half width at half the maximum is √ln 2 = 0.8326. Also of some

interest (though not particularly in this chapter) is the square root of the second moment of area around the y-axis. In a
mechanical context this would be called the radius of gyration. In a statistical context it would be called the standard
–
deviation. Either way, its value is 1/√2 = 0.7071. We shall meet the gaussian function again in Chapter 6.
In the present chapter we shall need to make use of the error function erf a. This is the area under the gaussian curve from x =
-a to x = +a:
+a
1 2
−x
erfa = − ∫ e dx. (4.1.2)
√π −a

The area outside the limits x = ±a, which is the area under the two “tails” of the gaussian function, is sometimes called the
complementary error function:

erfca = 1 − erfa (4.1.3)

It will be clear that erf a goes from 0 to 1 as a goes from 0 to infinity. Note also that
erfc (one standard deviation) = 0.3173
erfc (two standard deviations) = 0.0455.
Here are graphs of erf a (continuous line) and erfc a (dashed line) versus a.

Jeremy Tatum 4/22/2021 4.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7228

Jeremy Tatum 4/22/2021 4.1.2 CC-BY-NC https://phys.libretexts.org/@go/page/7228
4.2: Introduction
While the subject of thermal conduction is an important one, and obviously a proper topic in the theory of heat, it is not really
part of the great logical structure of thermodynamics, not does it require a wide or deep knowledge of thermodynamics to
understand it, at least at an introductory level. In other words, this chapter is more or less a stand-alone chapter. It is not
necessary to understand earlier chapters to understand this one; nor, if your primary interest is in thermodynamics, is it
necessary to understand this chapter before proceeding to later ones. That is – if you wish − you can skip this chapter without
compromising your understanding of any later ones

Jeremy Tatum 3/25/2021 4.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7229

4.3: Thermal Conductivity

Figure IV.1 shows heat flowing at a rate dQ/dt along a bar of cross-sectional area A of material. There is a temperature gradient
along the length of the bar (which is why heat is flowing down it). At a distance x from the end of the bar the temperature is T;
at a distance x + δx it is T + δT. Note that, if heat is flowing in the positive direction as shown, δT must be negative. That is, it
is cooler towards the right hand end of the bar. The temperature gradient dT/dx is negative. Heat flows in the opposite direction
to the temperature gradient.
The ratio of the rate of heat flow per unit area to the negative of the temperature gradient is called the thermal conductivity of
the material:
dQ dT
= −KA . (4.3.1)
dt dx

I am using the symbol K for thermal conductivity. Other symbols often seen are k or λ. Its SI unit is W m−1K−1.
I have defined it in a one-dimensional situation and for an isotropic medium, in which case the heat flow is opposite to the
temperature gradient. One can imagine that, in an anisotropic medium, the rate of heat flow and the temperature gradient may
be different parallel to the different crystallographic axes. In that case the heat flow and the temperature gradient may not be
strictly antiparallel, and the thermal conductivity is a tensor quantity. Such a situation will not concern us in this chapter.
If, in our one-dimensional example, there is no escape of heat from the sides of the bar, then the rate of flow of heat along the
bar must be the same all along the bar, which means that the temperature gradient is uniform along the length of the wire. It
may be easier to imagine no heat loss from the sides than to achieve it in practice. If the bar were situated in a vacuum, there
would be no loss by conduction or convection, and if the bar were very shiny, there would be little loss by radiation.
Order-of-magnitude values of the thermal conductivities of common substances are
Air 0.03 W m−1K−1
Water 0.6
Glass 0.8
Fe 80
Al 240
Cu 400
It is easy to imagine how heat may be conducted along a solid, with the vibrations of the atoms at one end of the solid being
transmitted to the next atoms by one atom nudging the next, and so on. However, it is evident from the table, and in any case is
common knowledge, that some substances (metals) conduct heat much better than others. Indeed, among the metals, there is a
close correlation between the thermal and electrical conductivities (at a given temperature). This suggests that the mechanism
for thermal conductivity in metals is the same as for electrical conductivity. Heat is conducted in a metal primarily by
electrons.
It would be an interesting exercise to find, from the Web or from other references, the thermal and electrical conductivities of a
number of metals. It may be found that thermal conductivities, K, are sometimes quoted in unfamiliar “practical” units, such as
BTU per hour per square foot for a temperature gradient of 1 F° per inch, and converting these to SI units of W m−1K−1 might
be a bit of a challenge. Electrical conductivities, σ, decrease somewhat with rising temperature (so do thermal conductivities,
but rather less so), so it would be important to find them all at the same temperature. Then you could see whether the ratio K/σ
is indeed the same for all metals at a given temperature. This is known as the Wiedemann-Franz Law. First-order theory
(which we do not give here) predicts that

Jeremy Tatum 3/25/2021 4.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7230

 2
K 1 πk −8 −1
= ( ) = 2.44 × 10 WΩ K . (4.3.2)
σT 3 e

Here k is Boltzmann’s constant and e is the electronic charge. This prediction is found to be obeyed well at room temperatures
and higher, but at low temperatures the electrical conductivity increases rapidly with lowering temperature, and the ratio starts
to fall well below the value predicted by equation 4.2.2, approaching zero at 0 K.
The reader may be familiar with the following terms in electricity
Conductivity σ
Conductance G
Resistivity ρ
Resistance R
They are related by G = 1/R, σ = 1/ρ, R = ρl/A, G = σA/l,
where l and A are the length and cross-sectional area of the conductor. The reader probably also knows that resistances add in
series and conductances add in parallel. We may define some analagous quantities related to heat flow. Thus resistivity is the
reciprocal of conductivity, resistance is l/A times resistivity, conductance is A/l times conductivity, and so on. These concepts
may come in useful in the following genre of problems beloved of examiners.
A room has walls of area A1, thickness d1, thermal conductivity K1, a door of area A2, thickness d2, thermal conductivity K2,
and a window of area A3, thickness d3, thermal conductivity K3, The temperature inside is T1 and the temperature outside is T2.
What is the rate of heat loss from the room?
K1 A1 K2 A2 K3 A3
We have three conductances in parallel: d1
, 
d2
, , and d3
, and so we have

dQ K1 A1 K2 A2 K3 A3
=( + + ) (T2 − T1 ). (4.3.3)
dt d1 d2 d3

Of course, the problem need not be exactly like that. Perhaps you are given the rate of heat loss and asked to find the area of
the window. But you get the general idea, and you can probably concoct a few examples yourself. The rate of heat flow is
analogous to the current, and the temperature difference is like the EMF of a battery.

Jeremy Tatum 3/25/2021 4.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7230

4.4: The Heat Conduction Equation
The situation described in Section 4.2 and in figure IV.1 was a steady-state situation, in which the temperature was a function
of x but not of time. We are now going to consider a more general situation in which the temperature may vary in time as well
as in space.

In this case the temperature gradient is written as a partial derivative, \( \frac{\partial T}{\partial x} and is not uniform down
2

the length of the rod. We'll suppose it is \frac{ \deta T}{\partial x} at x and ∂T

∂x
+
∂

∂x
T
2
δx at x + δx.

Consider the heat flow into and out of the portion between x and x + δx. The rate of flow into this portion at x is −KA ∂T

∂x
, and
2 2

the rate of flow out at x + δx is −KA ( ∂T

∂x
+
∂

∂x
T
2
δx) , so that the net flow of heat into that portion is KA ∂

∂x
T
2
. This must
δx

be equal to C ρAδx ∂T

∂t
, where ρ is the density (and hence ρAδx is the mass of the portion), and C is the specific heat capacity.
Therefore
2
∂T ∂ T
Cρ =K . (4.4.1)
2
∂t ∂x

This can be written

2
∂T ∂ T
=D , (4.4.2)
∂t ∂x2

where
K
D = (4.4.3)
Cρ

is the thermal diffusivity (m2 s−1).

Equation 4.3.2 is the heat conduction equation. In three dimensions it is easy to show that it becomes
2
T = D∇ T . (4.4.4)

Jeremy Tatum 4/1/2021 4.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7231

4.5: A Solution of the Heat Conduction Equation
Methods of solving the heat conduction equation are commonly given in courses on partial differential equations. Here we
shall look at a simple one-dimensional example.
A long copper bar is initially at a uniform temperature of 0 oC. At time t = 0, the left hand end of it is heated to 100 oC. Draw
graphs of temperature versus distance x from the hot end of the bar (up to x = 100 cm) at t = 4, 16, 64, 256 and 1024 seconds.
Draw also a graph of temperature versus time at x = 5 cm, up to 1024 seconds. Assume no heat is lost from the sides of the bar.
Data for copper:
K = 400 W m−1 K−1
C = 395 J kg−1 K−1
ρ = 8900 kg m−3
whence
D = 1.137 × 10−4 m2 s−1
The equation to be solved is
2
∂ T ∂T
D = (4.5.1)
2
∂x ∂t

From the form of this equation, it is obvious (once it has been pointed out!) that a solution could be found in which T(x, t) is
solely a function of x2/t, or, for that matter, x/t1/2. Thus, let
1/2
u = x/ t , (4.5.2)

and you will see that equation 4.4.1 reduces to the second order total differential equation
2
d T u dT
D =− . (4.5.3)
2
du 2 du

Let T' = dT/du, and it becomes even easier − a first order equation:
′
dT 1 ′
D =− uT . (4.5.4)
du 2

Upon integration, we obtain

2
u
′
ln T =− + ln A, (4.5.5)
4D

where ln A is an integration constant, to be determined by the initial and boundary conditions. (What are the dimensions of A?)
That is,
′ 2
T = Aexp [−u /(4D)] . (4.5.6)

We have to integrate again, with respect to u:

2
T = A∫ exp [−u /(4D)] du. (4.5.7)

Now, T = 100 oC at x = 0 for any t > 0. That is, T = 100 for u = 0.

And T = 0 oC at t = 0 for any x > 0. That is, T = 0 for u = ∞.
Therefore
0
2
100 − 0 = A ∫ exp [−u /(4D)] du. (4.5.8)
∞

Jeremy Tatum 3/18/2021 4.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7232

 −−
−
The integral is slightly difficult though well known. I'll just state the answer here; it is −√πD . From this, we find that the
integration constant is
−1 1/2
A = −5284 K m s . (4.5.9)

We now have
0
2
100 − T (x,  t) = A ∫ exp [−u /(4D)] du. (4.5.10)
−1/2
xt

The error function erf(r) is defined by

r
2
2
erf(r) = ∫ exp(−s )ds, (4.5.11)
−
√π 0

so that equation 4.4.10 can be written

−−
− x x
T (x,  t) = 100 + A√πDerf ( −
−− ) = 100 [1 − erf ( −− )] .
− (4.5.12)
2 √Dt 2 √Dt

This function is easy to plot provided that your computer will give you the erf function. The solutions are shown in figures
IV.4 and 5.

Jeremy Tatum 3/18/2021 4.5.2 CC-BY-NC https://phys.libretexts.org/@go/page/7232

CHAPTER OVERVIEW
5: THERMODYNAMIC PROCESSES
We shall be considering what happens when we perform certain processes on various systems. The processes will usually entail either
doing work on a system or adding heat to it, or perhaps we shall allow the system to do work on its surroundings, or the system may
lose heat to its surroundings.

Often the system we have in mind will be a gas enclosed by a movable piston inside a cylinder, but it need not be that. The system may
be a solid or a liquid, in which there is little change in volume. Or the system may have several phases, such as gas, liquid and solid.
There may be several components to the system – for example, a mixture of chemicals. Or the system may be a magnetic material, and
we do work on it by putting it in a magnetic field and magnetizing it. Some fundamental thermodynamical laws apply to any
thermodynamical system and are of great generality. Other laws may apply only to certain specific types of system, and we must always
be on our guard to recognize which are general laws applicable to any system, and which are special equations applicable only to
particular systems.

We shall, in our imagination, carry out processes under various ideal conditions. Thus, we may imagine a process to be isothermal
(carried out at constant temperature) or isobaric (constant pressure) or isochoric (constant volume). We may imagine a process in which
no heat is added to or is lost from the system. Such a process is adiabatic.

A process may be quasistatic or nonquasistatic. Let us imagine that we have a box of gas, and we suddenly heat one wall of the box by
pushing that wall up against a source of heat. Not all of the gas will immediately become hotter. At first, the gas near to the heated wall
will start to warm up, while the gas at the far end of the box will scarcely be aware of what has happened. Eventually, heat will permeate
throughout the box, but this may take some time, and the system is not at all in static equilibrium while these changes are taking place.
Likewise, if we have a gas held inside a cylinder by means of a movable piston, and we suddenly move the piston inwards. This will not
result in an immediate change to a higher pressure throughout the gas. At the very most the information about the new position of the
piston can travel through the gas only at the speed of sound. Considerable local turbulence is likely to be caused, and it will be some
time before the gas settles down to its new uniform pressure throughout. Both of these processes are nonquasistatic.

For a process to be quasistatic, the pressure and temperature of the system must differ from those of its surroundings by only an
infinitesimal amount at all times during the process; the process must take place slowly, so that the system passes through an infinite
succession of quasi-equilibrium states. The prefix "quasi" is often translated as "almost"; a more precise meaning is "as it were" or "as if
it were". The reader will conclude that there cannot ever literally be any process that is truly static. This is also true of other processes,
such as isothermal and adiabatic processes. Such processes are limiting theoretical processes. A real process may be intermediate
between the ideal extremes, although it may also be quite close to one of the ideal extremes.

5.1: SECTION 1-
5.2: SECTION 2-
5.3: SECTION 3-
5.4: SECTION 4-
5.5: SECTION 5-
5.6: SECTION 6-
BACK MATTER
INDEX

1 4/29/2021
CHAPTER OVERVIEW
FRONT MATTER

TITLEPAGE
INFOPAGE

1 4/29/2021
 University of Victoria
5: Thermodynamic Processes

Jeremy Tatum
This open text is disseminated via the Open Education Resource (OER) LibreTexts Project (https://LibreTexts.org) and like the
hundreds of other open texts available within this powerful platform, it is licensed to be freely used, adapted, and distributed.
This book is openly licensed which allows you to make changes, save, and print this book as long as the applicable license is
indicated at the bottom of each page.
Instructors can adopt existing LibreTexts texts or Remix them to quickly build course-specific resources to meet the needs of
their students. Unlike traditional textbooks, LibreTexts’ web based origins allow powerful integration of advanced features and
new technologies to support learning.

The LibreTexts mission is to unite students, faculty and scholars in a cooperative effort to develop an easy-to-use online
platform for the construction, customization, and dissemination of OER content to reduce the burdens of unreasonable
textbook costs to our students and society. The LibreTexts project is a multi-institutional collaborative venture to develop the
next generation of open-access texts to improve postsecondary education at all levels of higher learning by developing an
Open Access Resource environment. The project currently consists of 13 independently operating and interconnected libraries
that are constantly being optimized by students, faculty, and outside experts to supplant conventional paper-based books.
These free textbook alternatives are organized within a central environment that is both vertically (from advance to basic level)
and horizontally (across different fields) integrated.
The LibreTexts libraries are Powered by MindTouch® and are supported by the Department of Education Open Textbook Pilot
Project, the UC Davis Office of the Provost, the UC Davis Library, the California State University Affordable Learning
Solutions Program, and Merlot. This material is based upon work supported by the National Science Foundation under Grant
No. 1246120, 1525057, and 1413739. Unless otherwise noted, LibreTexts content is licensed by CC BY-NC-SA 3.0.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not
necessarily reflect the views of the National Science Foundation nor the US Department of Education.
Have questions or comments? For information about adoptions or adaptions contact info@LibreTexts.org. More information
on our activities can be found via Facebook (https://facebook.com/Libretexts), Twitter (https://twitter.com/libretexts), or our
blog (http://Blog.Libretexts.org).

This text was compiled on 04/29/2021

Welcome to the Physics Library. This Living Library is a principal hub of the LibreTexts project, which is a multi-institutional
collaborative venture to develop the next generation of open-access texts to improve postsecondary education at all levels of
higher learning. The LibreTexts approach is highly collaborative where an Open Access textbook environment is under
constant revision by students, faculty, and outside experts to supplant conventional paper-based books.

Campus Bookshelves Bookshelves

Learning Objects

Jeremy Tatum 4/1/2021 1 CC-BY-NC https://phys.libretexts.org/@go/page/7235

Welcome to the Physics Library. This Living Library is a principal hub of the LibreTexts project, which is a multi-institutional
collaborative venture to develop the next generation of open-access texts to improve postsecondary education at all levels of
higher learning. The LibreTexts approach is highly collaborative where an Open Access textbook environment is under
constant revision by students, faculty, and outside experts to supplant conventional paper-based books.

Campus Bookshelves Bookshelves

Learning Objects

Jeremy Tatum 3/18/2021 1 CC-BY-NC https://phys.libretexts.org/@go/page/7236

Welcome to the Physics Library. This Living Library is a principal hub of the LibreTexts project, which is a multi-institutional
collaborative venture to develop the next generation of open-access texts to improve postsecondary education at all levels of
higher learning. The LibreTexts approach is highly collaborative where an Open Access textbook environment is under
constant revision by students, faculty, and outside experts to supplant conventional paper-based books.

Campus Bookshelves Bookshelves

Learning Objects

Jeremy Tatum 4/1/2021 1 CC-BY-NC https://phys.libretexts.org/@go/page/7237

Welcome to the Physics Library. This Living Library is a principal hub of the LibreTexts project, which is a multi-institutional
collaborative venture to develop the next generation of open-access texts to improve postsecondary education at all levels of
higher learning. The LibreTexts approach is highly collaborative where an Open Access textbook environment is under
constant revision by students, faculty, and outside experts to supplant conventional paper-based books.

Campus Bookshelves Bookshelves

Learning Objects

Jeremy Tatum 4/1/2021 1 CC-BY-NC https://phys.libretexts.org/@go/page/7238

Welcome to the Physics Library. This Living Library is a principal hub of the LibreTexts project, which is a multi-institutional
collaborative venture to develop the next generation of open-access texts to improve postsecondary education at all levels of
higher learning. The LibreTexts approach is highly collaborative where an Open Access textbook environment is under
constant revision by students, faculty, and outside experts to supplant conventional paper-based books.

Campus Bookshelves Bookshelves

Learning Objects

Jeremy Tatum 4/8/2021 1 CC-BY-NC https://phys.libretexts.org/@go/page/7239

Welcome to the Physics Library. This Living Library is a principal hub of the LibreTexts project, which is a multi-institutional
collaborative venture to develop the next generation of open-access texts to improve postsecondary education at all levels of
higher learning. The LibreTexts approach is highly collaborative where an Open Access textbook environment is under
constant revision by students, faculty, and outside experts to supplant conventional paper-based books.

Campus Bookshelves Bookshelves

Learning Objects

Jeremy Tatum 4/29/2021 1 CC-BY-NC https://phys.libretexts.org/@go/page/7240

Back Matter

Index

4/29/2021 1 https://phys.libretexts.org/@go/page/27670
Index
A
exact differential M
adiabatic decompression
2.5: Second Derivatives and Exact Differentials Maxwell relations
12.5: Summary, the Maxwell Relations, and the
15.2: Adiabatic Decompression
adiabatic demagnetization
F Gibbs-Helmholtz Relations
Fahrenheit scale
15.3: Adiabatic Demagnetization
3.3: Temperature Scales I R
Réaumur scale
B I 3.3: Temperature Scales I
binary alloys inexact differential radius of gyration
17.9: Binary Alloys
2.5: Second Derivatives and Exact Differentials 4.1: Error Function
Integrating Factors Rankine scale
C 2.5: Second Derivatives and Exact Differentials 3.3: Temperature Scales I
Celsius scale
3.3: Temperature Scales I
J S
Joule coefficient specific heat
D 10.2: The Joule Experiment 7.3: Entropy
Debye temperature specific heat capacity
Joule Experiment
8.10: Heat Capacities of Solids
10.2: The Joule Experiment 7.3: Entropy

E K T
Enthalpy ternary alloy
Kelvin scale
9: Enthalpy
3.3: Temperature Scales I 17.10: Ternary Alloys
entropy
7.3: Entropy
L Z
Equation of state latent heat of freezing zeroth law of thermodynamics
2.5: Second Derivatives and Exact Differentials
9.2: Change of State 3.2: Zeroth Law of Thermodynamics
Error Function Law of Dulong and Petit
4.1: Error Function
8.10: Heat Capacities of Solids
CHAPTER OVERVIEW
6: PROPERTIES OF GASES

6.1: THE IDEAL GAS EQUATION

6.2: REAL GASES
6.3: VAN DER WAALS AND OTHER GASES
6.4: GAS, VAPOUR, LIQUID AND SOLID
6.5: KINETIC THEORY OF GASES- PRESSURE
6.6: COLLISIONS
6.7: DISTRIBUTION OF SPEEDS
6.8: FORCES BETWEEN MOLECULES

1 4/29/2021
6.1: The Ideal Gas Equation
In 1660, the Honorable Robert Boyle, Father of Chemistry and seventh son of the Earl of Cork, and one of the founders of the
Royal Society of London, conducted certain Experiments Physico-Mechanical Touching the Spring of the Air. He held a
quantity of air in the closed arm of a J-shaped glass tube by means of a column of mercury and he measured the volume of the
air as it was subjected to greater and greater pressures. As a result of these experiments he established what is now known as
Boyle's Law:
The pressure of a fixed mass of gas held at constant temperature (i.e. in an isothermal process) is inversely proportional to its
volume.
That is,
P V = constant. (6.1.1)

Later experiments showed that the volume of a fixed mass of gas held at constant pressure increases linearly with temperature.
In particular, most gases have about the same volume coefficient of expansion. At 0oC this is about 0.00366 Co −1 or 1/273 Co
−1
.
If you extrapolate the volume of a fixed mass of gas held at constant pressure to lower and lower temperatures, the
extrapolated volume would fall to zero at −273 oC. This is not directly the basis of our belief that no temperatures are possible
below −273 oC. For one thing, a real gas would liquefy long before that temperature is reached. Nevertheless, for reasons that
will be discussed in a much later chapter, we do believe that this is the absolute zero of temperature. In any case:
The volume of a fixed mass of gas held at constant pressure (i.e. in an isobaric process) is directly proportional to its Kelvin
temperature.
Lastly,
The pressure of a fixed mass of gas held at constant volume (i.e. in an isochoric process) is directly proportional to its Kelvin
temperature.
If P, V and T are all allowed to vary, these three laws become
P V /T = constant (6.1.2)

The value of the constant depends on how much gas there is; in particular, it is proportional to how many moles (hence how
many molecules) of gas there are. That is
P V /T = RN , (6.1.3)

where N is the number of moles and R is a proportionality constant, which is found to be about the same for most gases.
Of course real gases behave only approximately as described, and only provided experiments are performed over modest
ranges of temperature, pressure and volume, and provided the gas is well above the temperature at which it will liquefy.
Nevertheless, provided these conditions are satisfied, most gases do conform quite well to equation 6.1.3 with about the same
proportionality constant for each.
A gas that obeys the equation

P V = N RT (6.1.4)

exactly is called an Ideal Gas, and equation 6.1.4 is called the Equation of State for an Ideal Gas. In this equation, V is the
total volume of the gas, N is the number of moles and R is the Universal Gas Constant. The equation can also be written
P V = RT . (6.1.5)

In this case, V is the molar volume. Some authors use different symbols (such as V, v and Vm) for total, specific and molar
volume. This is probably a good idea, and it is at some risk that I am not going to do this, and I am going to hope that the
context will make it clear which volume I am referring to when I use the simple symbol V in any particular situation. Note
that, while total volume is an extensive quantity, specific and molar volumes are intensive.

Jeremy Tatum 4/15/2021 6.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7242

It is not impossible to go wrong by a factor of 103 when using equation 6.1.5. If you are using CGS units, P will be expressed
in dynes per square cm, V is the volume of a mole (i.e. the volume occupied by 6.0221 × 1023 molecules), and the value of the
universal gas constant is .8.3145×107 erg mole−1 K−1. If you are using SI units, P will be expressed in pascal (N m−2), V will
be the volume of a kilomole (i.e. the volume occupied by 6.0221 x 1026 molecules), and the value of the universal gas constant
is 8.3145 × 103 J kilomole−1 K−1. If you wish to express pressure in Torr, atm. or bars, and energy in calories, you're on your
own.
N NA
You can write equation 6.1.4 (with V = total volume) as P = V
where NA is Avogadro's number, which is 6.0221 ×
RT

NA

1023 molecules per mole, or 6.02221 × 1026 molecules per kilomole. The first term on the right hand side is the total number of
molecules divided by the volume; that is, it is the number of molecules per unit volume, n. In the second term, R/NA is
Boltzmann's constant, k = 1.3807 10−23 × J K−1. Hence the equation of state for an ideal gas can be written
P = nkT . (6.1.6)

Divide both sides of equation 6.1.5 by the molar mass ("molecular weight”) µ. The density ρ of a sample of gas is equal to the
molar mass divided by the molar volume, and hence the equation of state for an ideal gas can also be written
ρRT
P = . (6.1.7)
μ

In summary, equations 6.1.4, 6.1.5, 6.1.6 and 6.1.7 are all commonly-seen equivalent forms of the equation of state for an ideal
gas.
From this point on I shall use V to mean the molar volume, unless stated otherwise, so that I shall use equation 6.1.5 rather
than 6.1.4 for the equation of state for an ideal gas. Note that the molar volume (unlike the total volume) is an intensive state
variable.
In September 2007, the values given for the above-mentioned physical constants on the Website of the National Institute of
Science and Technology (http://physics.nist.gov/cuu/index.html) were:
Molar Gas Constant R = 8314.472 (15) J kmole−1 K−1.
Avogadro Constant NA = 6.022 141 79 (30) × 1026 particles kmole−1.
Boltzmann Constant k = 1.380 6504 (24) × 10−23 J K−1 per particle.
The number in parentheses is the standard uncertainty in the last two figures.
[There is a proposal, likely to become official in 2015, to give defined exact numerical values to Avogadro’s and Boltzmann’s
constants, namely 6.022 14 x 1023 particles mole−1 and 1.380 6 x 10–23 J K−1 per particle. This may at first seem to be
somewhat akin to defining π to be exactly 3, but it is not really like that at all. It is all part of a general shift in defining many
of the units used in physics in terms of fundamental physical quantities (such as the charge on the electron) rather than in terms
of rods or cylinders of platinum held in Paris.]

Jeremy Tatum 4/15/2021 6.1.2 CC-BY-NC https://phys.libretexts.org/@go/page/7242

6.2: Real Gases
How well do real gases conform to the equation of state for an ideal gas? The answer is quite well over a large range of P, V
and T, provided that the temperature is well above the critical temperature. We'll have to see shortly what is meant by the
critical temperature; for the moment we'll say the ideal gas equation is followed quite well provided that the temperature is
well above the temperature at which it can be liquefied merely by compressing it. Air at room temperature obeys the law quite
well. Gases in stellar atmospheres also obey the law well, because there is no danger there of the gas liquefying. (In the cores
of stars, however, where densities are very large, the gases obey a very different equation of state.)
One measure of how well the law is obeyed by real gases is to measure P, V and T, and see how close PV

RT
is to 1. The quantity
RT

PV
is known as the compression factor, and is often given the symbol Z. For most real gases at very high pressures (a few
hundred atmospheres), it is found in fact that Z is rather greater than 1. As the pressure is lowered, Z becomes lower, and then,
alas, it overshoots and is found to be a little less than 1. Then at yet lower pressures Z rises again. The exact shape of the Z : P
curve is different from gas to gas, as is the pressure at which Z is a minimum. Yet, for all gases, as the pressure approaches
zero, PV/T approaches R exactly. For this reason R is sometimes called the Universal Gas Constant as well as the Ideal Gas
Constant. In the limit of very low pressures, all gases behave very closely to the behaviour of an ideal gas. In Section 6.3 we
shall be examining more closely how the compression factor varies with pressure.
Another way to look at how closely real gases obey the ideal gas equation is to plot P versus V for a number of different
temperatures. That is, we draw a set of isotherms. For an ideal gas, these isotherms, PV = constant, are rectangular hyperbolas.
So they are for real gases at high temperatures. At lower temperatures, departures from the ideal gas equation are marked.
Typical isotherms are sketched in figure VI.1. Alas, my limited skills with this infernal computer in front of me allow me only
to sketch these isotherms crudely by hand.

In the PV plane of figure VI.1, you will see several areas marked "gas", "liquid", "vapour", "liquid + vapour". You can follow
the behaviour at a given temperature by starting at the right hand end of each isotherm, and gradually moving to the left – i.e.
increase the pressure and decrease the volume. The hottest isotherm is nearly hyperboloidal. Nothing special happens beyond
the volume decreasing as the pressure is increased, according to Boyle's law. At slightly lower temperatures, a kink develops in
the isotherm, and at the critical temperature the kink develops a local horizontal inflection point. The isotherm for the critical
temperature is the critical isotherm, marked CI on the sketch. Still nothing special happens other than V decreasing as P is
increased, though not now according to Boyle's law.
For temperatures below the critical temperature, we refer to the gas as a vapour. As you decrease the volume, the pressure
gradually increases until you reach the dashed curve. At this point, some of the vapour liquefies, and, as you continue to
decrease the volume, more and more of the vapour liquefies, the pressure remaining constant while it does so. That’s the
horizontal portion of the isotherm. In that region (i.e. outlined by the dashed curve) we have liquid and vapour in equilibrium.
Near the right hand end of the horizontal portion, there is just a small amount of liquid; at the left hand end, most of the
substance is liquid, with only a small amount of vapour left.
After it is all liquid, further increase of pressure barely decreases the volume, because the liquid is hardly at all compressible.
The isotherm is then almost vertical.

Jeremy Tatum 4/29/2021 6.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7243

The temperature of the critical isotherm is the critical temperature. The pressure and molar volume at the horizontal inflection
point of the critical isotherm are the critical pressure and critical molar volume. The horizontal inflection point is the critical
point.

Jeremy Tatum 4/29/2021 6.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7243

6.3: Van der Waals and Other Gases
We have seen that real gases resemble an ideal gas only at low pressures and high temperatures. Various attempts have been
made to find an equation that adequately represents the relation between P, V and T for a real gas – i.e. to find an Equation of
State for a real gas. Some of these attempts have been purely empirical attempts to fit a mathematical formula to real data.
Others are the result of at least an attempt to describe some physical model that would explain the behaviour of real gases. A
sample of some of the simpler equations that have been proposed follows:
van der Waals' equation:
a
(P + ) (V − b) = RT . (6.3.1)
2
V

Berthelot's equation:
a
(R + ) (V − b) = RT . (6.3.2)
2
(T V )

Clausius's equation*:
a
(P + ) (V − b) = RT . (6.3.3)
2
T (V + c)

Dieterici's equation:
a

P (V − b)e ( RT V )
= RT . (6.3.4)

Redlich-Kwong:
RT a 1 1
P = − ( − ). (6.3.5)
1
V −b V V +b
bT 2

Virial equation:
2 3
P V = A + BP + C P + DP +. . . (6.3.6)

*In Clausius’s equation, if we choose c = 3b, we get a fairly good agreement between the critical compression factor of a
Clausius gas and of many real gases. The meaning of “critical compression factor”, and the calculation of its value for a
Clausius gas is described a little later in this section.

There are many others, but by far the best known of these is van der Waals' equation, which I shall
describe at some length.
It is not possible for the voice-box of an English speaker correctly to pronounce the name van der Waals, although the W is pronounced
more like a V than a W, and, to my ear, the v is somewhat intermediate between a v and an f. To hear it correctly pronounced –
especially the vowels − you must ask a native Dutch speaker. The frequent spelling "van der Waal's equation" is merely yet another
symptom of the modern lamentable ignorance of the use of the apostrophe so much regretted by Lynne Truss.
The units in which the constants a and b should be expressed sometimes cause difficulty, and they depend on whether the
symbol V in the equation is intended to mean the specific or molar volume. The following might be helpful.
If V is intended to mean the specific volume, van der Waals’ equation should be written (P + a/V ) (V − b) = RT /μ , 2

where µ is the molar mass (“molecular weight”). In this case the dimensions and SI units of a are M−1 L5 T−2 and Pa m6 kg−2
and the dimensions and SI units of b are M−1 L3 and m3 kg−1
If V is intended to mean the molar volume, van der Waals’ equation should be written in its familiar form
5 −2
(P + a/ V ) (V − b) = RT . In this case the dimensions and SI units of a are ML T
2
mole−2 and Pa m6 kmole−2 and the
3 −1 3 −1
dimensions and SI units of b are L mole and m kmole
The van der Waals constants, referred to molar volume, of H2O and CO2 are approximately:

Jeremy Tatum 3/20/2021 6.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7244

 H2O: a = 5.5 × 105 Pa m6 kmole−2 . b = 3.1 × 10−2 m3 kmole−1
CO2: a = 3.7 × 105 Pa m6 kmole−2 b = 4.3 × 10−2 m3 kmole−1
The van der Waals equation has its origin in at least some attempt to describe a physical model of a real gas. The properties of
an ideal gas can be modelled by supposing that a gas consists of a collection of molecules of zero effective size and no forces
between them, and pressure is the result of collisions with the walls of the containing vessel. In the van der Waals model, there
are supposed to be attractive forces between the molecules. These are known as van der Waals forces and are now understood
to arise because when one molecule approaches another, each induces a dipole moment in the other, and the two induced
dipoles then attract each other. This attractive force reduces the pressure at the walls, the reduction being proportional to the
number of molecules at the walls that are being attracted by the molecules beneath, and to the number of molecules beneath,
which are doing the attracting. Both are inversely proportional to V, so the pressure in the equation of state has to be replaced
by the measured pressure P plus a term that is inversely proportional to V2. Further, the molecules themselves occupy a finite
volume. This is tantamount to saying that, at very close range, there are repulsive forces (now understood to be Coulomb
forces) that are greater than the attractive van der Waals forces. Thus the volume in which the molecules are free to roam has
to be reduced in the van der Waals equation. For more on the forces between molecules, see Section 6.8.
However convincing or otherwise you find these arguments, they are at least an attempt to describe some physics, they do
represent the behaviour of real gases better that the ideal gas equation, and, if nothing else, they give us an opportunity for a
little mathematics practice.
We shall see shortly how it is possible to determine the constants a and b from measurements of the critical parameters. These
constants in turn give us some indication of the strength of the van der Waals forces, and of the size of the molecules.
Van der Waals' equation, equation 6.3.1, can be written
RT a
P = − . (6.3.7)
2
V −b V

A horizontal inflection point occurs where ∂P

∂V
and ∂ P
2
are both zero. That is
∂V

RT 2a
− + =0 (6.3.8)
(V − b)2 V 3

and
2T R 6a
− = 0. (6.3.9)
(V − b)3 V
4

Eliminate RT/a from these to find the critical molar volume of a van der Waals gas:
Vc = 3b. (6.3.10)

Substitute this into equation 6.3.8 or 6.3.9 (or both, as a check on your algebra) to obtain the critical temperature:
8a
Tc = . (6.3.11)
27Rb

Substitute equations 6.3.10 and 6.3.11 into equation 6.3.7 to obtain the critical pressure:
a
Pc = . (6.3.12)
2
27b

From these, we readily obtain

Pc Vc 3
= = 0.375. (6.3.13)
RTc 8

This quantity is often called the critical compression factor or critical compressibility factor, and we shall denote it by the
symbol Zc. For many real gases Zc is about 0.28; thus the van der Waals equation, while useful in discussing the properties of
gases in a qualitative fashion, does not reproduce the observed critical compression factor particularly well.

Jeremy Tatum 3/20/2021 6.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7244

Let us now substitute p = P /P ,  v = V /V ,  t = T /T , and van der Waals' equation, in which the pressure, volume and
c c c

temperature are expressed in terms of their critical values, becomes

2
1 8
(p + 3/ v )(v − ) = t. (6.3.14)
3 3

This can also be written

3 2
3p v − (p + 8t)v + 9v − 3 = 0. (6.3.15)

For volumes less than a third of the critical volume, this equation does not describe the behaviour of a real gas at all well.
Indeed, you can see that p = ∞ when v = 1/3, which means that you have to exert an infinite pressure to compress a van der
Waals gas to a third of its critical volume. You might want to investigate for yourself the behaviour of equations 6.3.14 and 15
for volumes smaller than this. You will find that it goes to infinity at v = 0 and 1/3, and it has a maximum between these two
volumes. But the equation is of physical interest only for v > 1/3 , where the variation of pressure, volume and temperature
bears at least some similarity to the behaviour of real gases, if by no means exact. In figure VI.2, I show the behaviour of a van
der Waals gas for five temperatures – one above the critical temperature, one at the critical temperature, and three below the
critical temperature. The locus of maxima and minima is found by eliminating t between equation 6.3.14 and ∂p/∂v = 0. You
should try this, and show that the locus of the maxima and minima (which I have shown by a blue line in figure VI.2) is given
by
3 2
p = − (6.3.16)
2 3
v v

Don’t confuse the blue curve in this figure (it shows the locus of maxima and minima) with the dashed curve in figure VI.1 (it
shows the boundary between phases.). For the temperatures 0.85, 0.90 and 0.95 I have drawn the constant pressure lines where
liquid and vapour are in equilibrium in the real fluid. These are drawn so that they divide the van ver Waals curve into two
equal areas, above and below. This means that the work done by the real fluid when it changes from liquid to vapour at
constant pressure is equal to the work that would be done by its hypothetical van der Waals equivalent along its wiggly path.
We shall later see that the placement of the horizontal line is a consequence of the fact that the Gibbs function (which we have
not yet met) is constant while the liquid and vapour are in equilibrium. The dashed line of figure VI.1 would correspond on
figure VI.2 to the locus of the ends of the horizontal lines. I have drawn this locus, which outlines the region where liquid and
vapour are in equilibrium, in red in figure VI.2. While the van der Waals equation is only a rough approximation to the
behaviour of real gases, it is nevertheless true that, if pressures, temperatures and molar volumes are expressed in terms of the
critical pressures, temperatures and molar volumes, the actual equations of state of real gases are very similar. Two gases with
the same values of p, v and t are said to be in corresponding states, and the observation that the p : v : t relation is
approximately the same for all gases is called the Law of Corresponding States. We may think of gases as being composed of
particles (molecules) and the only difference between different gases is in the sizes of their molecules (i.e. their different van
der Waals b constants) and their dipole moments or their electrical polarizabilities (i.e. their different van der Waals a
constants). In the dimensionless forms of the equation of state, these van der Waals constants are removed from the equations,
and it is not surprising that all gases then conform to the same equation of state.

Jeremy Tatum 3/20/2021 6.3.3 CC-BY-NC https://phys.libretexts.org/@go/page/7244

I leave it to the reader to show that, for a Berthelot gas, the critical molar volume, temperature and pressure and the critical
−−−− −−−
compression factor are, respectively, 3b, √ 8a

27bR
, 1

b
√
aR

216b
and 0.375, that the equation of state in terms of the dimensionless
variables is
8t 3
p = − , (6.3.17)
2
3v − 1 tv

and that the locus of maxima and minima is

1 4 6
p = ( − ). (6.3.18)
1/2 v 3v − 1
v

These are shown in figure VI.2a. It will be noted that the critical compression factor is the same as (and hence no better than)
for a van der Waals gas.
For a Clausius gas, the critical molar volume, temperature and pressure and the critical compression factor are, respectively, 3b
−−−−−− −−−−−−
+ 2c, √ 8a

27(b+c)R
,
1

(b+c)
√
aR

216(b+c)
and 3b+2c

8(b+c)
.
−−− −−−
If c = 3b. these become 3c, √ 2a

9cR
,
1

c
√
aR

512c
and 9

32
= 0.28125 . I choose c = 3b because that gives a good agreement with the
critical compression factor for many real gases. In dimensionless units, the Clausius equation becomes
32t 48
p = − (6.3.19)
2
9v − 1 t(3v + 1)

The locus of maxima and minima is

1 80 − 144v 16(5 − 9v)

p = − −−− −( ) = (6.3.20)
2 3/2
√ 1 + 3v 1 − 6v − 27v (1 − 9v)(1 + 3v)

These are shown in figure VI.2b

Jeremy Tatum 3/20/2021 6.3.4 CC-BY-NC https://phys.libretexts.org/@go/page/7244

The Clausius equation was hard work. Dieterici’s is a little easier. The critical molar volume, temperature and pressure and the
critical compression factor are, respectively, 2b, a
,
4Rb
a
and 2/e2 = 0.271. Note that the critical compression factor is much
2
4e2 b

closer to that of many real gases. The dimensionless form of the Dieterici equation is
t 2
p = exp(2 − ). (6.3.21)
2v − 1 tv

The locus of maxima and minima is

1 2(v − 1)
p = exp( ). (6.3.22)
2
v 2v − 1

These are shown in figure VI.2c.

The Redlich-Kwong equation of state, like those of van der Waals, Bethelot and Dieterici, has just two parameters (a and b).
All of them are not too bad at temperatures appreciably above the critical temperature, but, close to the critical temperature,
the Redlich-Kwong empirical equation agrees a little better than the van der Waals equation does with what is observed for
real gases. Obtaining the critical constants in terms of the parameters is done by exactly the same method as for the van der
Waals and other equations, but requires perhaps a little more work and patience. The reader might like (or might not like) to
try it. For the critical constants I get
Vc = xb, (6.3.23)

a 2/3

Tc = y ( ) (6.3.24)
bR

and
1/3
2
a R
Pc = z( ) , (6.3.25)
5
b

where
x = 3.847322100 (6.3.26)

y = 0.345039996 (6.3.27)

and

z = 0.029894386. (6.3.28)

The critical compression factor is xz/y , which is exactly 1/3. This is not as close to the compression factor of many real gases
as the Dieterici critical compression factor is.
We can invert these equations to obtain expressions for a and b in terms of the critical temperature and pressure (or
temperature and volume, or pressure and volume). Thus

Jeremy Tatum 3/20/2021 6.3.5 CC-BY-NC https://phys.libretexts.org/@go/page/7244

 2 5/2
R Tc
a =u( ) (6.3.29)
Pc

and
RTc
b =w , (6.3.30)
Pc

where
u = 0.427480233 (6.3.31)

and

w = 0.086640350. (6.3.32)

(You can also do this for the other equations of state, of course.)
In order to reproduce these results, you’ll have to do a little work to see where all the constants come from. It turns out that the
value of the constant x is the positive real root of the equation
3 2
x − 3x − 3x − 1 = 0. (6.3.33)

In the above analysis, I obtained all the constants from a numerical solution of equation 6.3.33, but the solution to this
equation (and all subsequent constants) can also be written in surds. Thus, with
– – −−
, the constants can be calculated from
3 3 3

f = √2 − 1, g = √4 − 1, h = √16 − 1

1 3 − 5g 1 1
2
x = , y =g , z = , u =  and w = f. (6.3.34)
f g+h 9f 3

If we now introduce the dimensionless variables p = P /P , v = V N , t = T /T , and substitute these and equations 6.3.23-
c c c

25 into equation 6.3.6, we obtain the dimensionless Redlich-Kwong equation

lt 1 1 1
p = − ( − ), (6.3.35)
1/2
xv − 1 mt xv xv + 1

where
l = y/z = 11.54196631  and  m = gz = 0.017559994. (6.3.36)

The dimensionless Redlich-Kwong equation is illustrated in figure VI.2d. I have not tried to find an explicit equation for the
locus of maxima, but instead I calculated it numerically, illustrated by the dashed line in figure VI.2d.

Here is a summary of the results for the two-parameter equations of state:

Jeremy Tatum 3/20/2021 6.3.6 CC-BY-NC https://phys.libretexts.org/@go/page/7244

   Pc Vc Tc Zc a b
2 2
a 8a 3 27R Tc RTc
Van der Waals 2
3b
27b 27Rb 8 64Pc 8Pc

−−− −−−− 2 3
1 aR 8a 3 27R Tc RTc
Berthelot √ 3b √
b 216b 27bR 8 64Pe 8Pc

2 2 (6.3.37)
a a RTc 4R Tc
−2
Dieterici 2
2b 2e 2 2
4e b
2 4Rb e Pc e Pc

1/3 2 5/2
2
a R a 2/3 1 R Tc RTc
Redlich-Kwong z( 5
) xb y( ) u( ) w
b bR 3 Pc Pc

  z = 0.0299 x = 3.85 y = 0.345 u = 0.427 w = 0.0866

The reader can try to reproduce these (let me know (jtatum@uvic.ca) if you find any mistakes!) or at least (a useful exercise)
verify their dimensions. We mentioned in Section 6.2 that a useful way of indicating how the behaviour of a real gas differs
from that of an ideal gas is by plotting the compression factor Z = versus pressure. As the pressure approaches zero, the
PV

RT

compression factor approaches 1. This is because the molecules are then so far apart that there are no appreciable forces
(attractive or repulsive) between them. As the pressure is increased from zero, the compression factor generally at first drops a
little below 1, and then rises above 1 as the pressure is increased. It will be interesting to see how the compression factor is
expected to vary with pressure for the various “theoretical” gases that we have been discussing. I’ll do it just for a van der
Waals gas, and I’ll use the dimensionless form of van der Waals’ equation, which was first given as equation 6.3.14:

2
1 8
(p + 3/ v ) (v − ) = t. (6.3.38)
3 3

Pc Vc pv
The compression factor is Z = PV

RT
and the critical compression factor is Z c =
RTc
. From this, we see that Z = t
Zc . For a
3pv
van der Waals gas, Z = , so that Z =
c
3

8
. Unfortunately, in order to plot Z versus p for a given t, we have to be able to
8t

express v in terms of p, which means solving equation 6.3.37, which is a cubic equation in
v [3p v − (p + 8t)v + 9v − 3 = 0] . I have done this numerically, and I show the resulting graphs of Z versus p for several
3 2

temperatures, in figure VI.2e. Notice that at p = 1 and t = 1, (i.e. at the critical point), the compression factor is 0.375. The Z
versus p curves for real gases have the same general shape, but the precise agreement in numerical detail is not quite so good.
Where Z > 1, the pressure is greater than that of an ideal gas, the b (repulsive) part of the van der Waals equation being more
important than the a (attractive) part. Where Z < 1, the pressure is less than that of an ideal gas, the a (attractive) part of the
van der Waals equation being more important than the b (repulsive) part. I haven’t investigated whether the other “theoretical”
equations of state do appreciably better. Why not have a go yourself?!

Figure VI.2e. The compression factor Z = PV / RT versus p (pressure in units of the critical pressure for a van der Waals
gas, for several values of t (temperature in units of the critical temperature.) For a van der Waals gas the compression
factor is greater than 1 for all temperatures greater than t = 27/8 = 3.375. At this temperature, the compression factor is
close to 1 up to p equals approximately 2, and this temperature is known as the Boyle temperature. At the Boyle
temperature, the Z : p curve is flat and close to 1 for a fairly large range of pressures. Thus, at the Boyle temperature,
even a non-ideal gas obeys Boyle’s law fairly closely. For a van der Waals gas, the critical temperature is 8a/(27Rb), so

Jeremy Tatum 3/20/2021 6.3.7 CC-BY-NC https://phys.libretexts.org/@go/page/7244

 the Boyle temperature for a van der Waals is a/(Rb). The reader should calculate this for H2O and CO2, using the values
of the van der Waals constants given in this Chapter. The dot on the t = 1.00 isotherm at p = 1 and Z = 0.375 corresponds
to the critical point. Anyone who feels in need of more mental exercise might like to ask: For what value of p (other than
zero) is Z = 1. For example, can you show that, for t = 1, Z = 1 for p = 152/27 = 5.630?
The last proposed empirical equation of state that we mentioned is the virial equation, equation 6.3.6:
P V = A + BP + C P + DP +…
2
This is sometimes written in the form P V = A + +
3
+ … , but in these
B

V
C
2
D
3
V V

notes we’ll use the form of equation 6.3.6. The coefficients A, B, etc are called the virial coefficients and are functions of
temperature. The first coefficient, A, is just RT. We can also write the virial equation as
′ ′ 2 ′ 3
Z = 1 +B P +C P +D P +… (6.3.39)

We could measure the coefficient B' for a real gas by plotting Z as a function of pressure in a similar manner to figure VI.2e.
The initial slope (
∂Z

∂P
) extrapolated to zero pressure gives the value of B'. At low temperatures B' is negative; at high
T

temperatures B' is positive. At the Boyle temperature B' is zero, and at that temperature the compression factor is unity for a
large range in pressures, and the gas accordingly closely conforms to Boyle’s law. The coefficient C' is small, so the term C'P2
comes into play only at higher pressures. At higher pressures, Z increases, showing that C' is a positive coefficient. The
coefficient D' is smaller still than C'
All the mathematically “well-behaved” equations of state below the critical temperature have a maximum and a minimum –
i.e. the curve shows a “wiggle”. I illustrate this in figure VI.2f. This is the van der Waals isotherm for t = 0.95 in dimensionless
variables. It is the same as one of the curves shown in figure VI.2, drawn to a different scale so as to emphasize the “wiggle”.

Using the little cylinder and piston to the right of the graph, try and imagine what happens to the enclosed liquid or vapour as
you move the piston in and out at constant temperature, moving from a to e and back again on the graph. Start at e. The
cylinder is filled with vapour. Move the piston inwards, going from e to d; the pressure increases and the volume decreases.
Now a real gas doesn’t follow the van der Waals function all the way. At d, something different happens. Actually it is possible
to take a vapour a little way past d towards (but not beyond) n. That would be a supercooled vapour, such as is used in a cloud
chamber. It will condense immediately into a line of liquid droplets as soon as a charged particle flies through the vapour.
However, what usually happens is that some of the vapour starts to condense as liquid, and we move horizontally from d to b.
As we move the piston down at constant temperature, the volume of course decreases, and more and more liquid condenses in
such a manner that the pressure remains constant. In the portion db, we have liquid and vapour existing together in the piston,
in thermodynamic equilibrium. Near to the d end there is only a little liquid; near to the b end it is nearly all liquid, with only a
little vapour left. Beyond b, towards a, the space is completely filled with liquid. We can push and push, increasing the
pressure greatly, but there is very little change in volume, because the liquid is almost (though not quite) incompressible. The
isotherm is very steep there. It is actually possible to take the liquid a little way from b towards (but not beyond) m without
any of it vaporizing. This would be a superheated liquid, such as is used in a bubble chamber. It will vaporize immediately into
a line of bubbles as soon as a charged particle flies through the liquid.

Jeremy Tatum 3/20/2021 6.3.8 CC-BY-NC https://phys.libretexts.org/@go/page/7244

There will be further important material concerning change of state in Chapters 9 and 14. At present, though, I want to ask: At
what pressure does condensation commence? Putting it another way, what is the height of the line bd in figure VI.2f? I have
heard it argued that the height of bd, (the pressure at which condensation occurs) must be such that the area bmc is equal to the
area cnd. I am not sure that I fully understand the arguments leading to this conclusion. After all, a real gas doesn’t conform
exactly to a van der Waals equation or any of the other theoretical/empirical equations that we have discussed. But perhaps it is
not unreasonable to draw bd such that the areas above and below it are equal, and in any case it makes for an interesting (and
challenging) computational exercise. The van der Waals equation, in dimensionless variables, is given as equation 6.3.14. Can
you calculate the pressure such that the area bmc below bd is equal to the area cnd above it? I make it p = 0.812, which is the
height where I have drawn it in the figure. I haven’t done the calculation for the other equations. I leave that to you!

Jeremy Tatum 3/20/2021 6.3.9 CC-BY-NC https://phys.libretexts.org/@go/page/7244

6.4: Gas, Vapour, Liquid and Solid
Our description of the behaviour of a real substance in section 6.2 was incomplete in many ways, not least because it made no
mention of the solid state. At very low temperatures or at very high pressures, most substances will solidify

In figures VI.3 and 4 I have sketched schematically, by hand, the several regions in the PV-plane in which a substance exists in
its several stages. Unlike in Figure VI.1 and VI.2, I have not drawn isotherms. The various lines are intended to represent the
boundaries between phases, and are therefore more akin to the dashed curves in figures VI.1 and VI.2. The one exception is
the critical isotherm, CI, which is indeed the curve that separates gas from vapour or liquid, but which is also, of course, an
isotherm. The difference between figures VI.3 and VI.4 is that figure VI.3 represents a substance that expands when it melts
from solid to liquid, while figure VI.4 represents a substance that contracts when it melts from solid to liquid (that is, the solid
is less dense than the liquid, and will float upon it.) Most substances expand upon melting, but we have to include those
exceptional substances that contract upon melting, because one such substance is one of the most important of all – water.
You can try to understand the figures a little by moving along a horizontal line (isobar) or along a vertical line (isochor) and
noticing where phase changes take place. Can you see, for example, where a solid will change to a vapour without going
through a liquid phase (sublimation)?
You will note, in the figures, the critical isotherm CI, that separates gas from liquid or vapour, and you will note that, at
temperatures above the critical temperature, the only phase possible is gas, and the substance cannot be liquefied merely by
compression. You will note also the critical point CP. You will also see the triple line TL, along which solid, liquid and vapour
co-exist together. What of the region marked O? The substance cannot exist here in solid, liquid or gaseous phase. To that
extent, we see that the van der Waals equation may be a little bit better than we thought it was, because you will remember that
it went up to infinity at a third of the critical volume. All that this means is that by then the molecules are so tightly jammed
together that you simply cannot compress them any further. Although a substance cannot exist in an ordinary solid, liquid or
gas phase in the region marked O, if the matter is degenerate it will be in this region. The electron structure of the atoms
breaks down, so that it then does became possible to jam the atoms closer together. This may mean something to those of you
who are familiar with the concept of degenerate matter. If you have not heard of it, do not worry; you are unlikely to come
across it unless you visit a white dwarf star, or the core of a massive star, or have to take an examination in astrophysics. For
the time being, we shall look the other way and pretend it doesn't exist.
We can get a little more insight by looking at the PT-plane. Figure VI.5 shows a substance that expands on melting, and figure
VI.6 shows a substance (such as water) that expands on freezing. In the PT-plane, the triple point (where solid, liquid and
vapour) are in equilibrium with each other, appears as the triple point, TP. (In PVT-space it is a line, although the critical point
CP remains a genuine point in PVT-space.) The line separating liquid from vapour terminates at the critical point, and the line
is often drawn as though it were somehow left hanging in mid-air, so that one is uncertain whether a given point near the
critical point represents a gas, a vapour or a liquid. But in the PT-plane, the critical isotherm is a vertical line (show as dashed
in the figures), and the liquid/vapour boundary terminates at the critical isotherm, and there is no question what phase is
represented by a point near to the critical point. To the right of the critical isotherm, we have a gas. To the left, we have either a
liquid or a vapour, depending on whether we are above or below the liquid/vapour boundary. As we cross the solid/vapour
boundary, below the critical temperature and below the critical pressure (on Mars!) we have a phase change directly from solid
to vapour or vapour to solid – i.e. sublimation.
(I have often heard that, below the triple point, a solid will "sublime". I think I prefer the verb "to sublimate".)

Jeremy Tatum 4/15/2021 6.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7245

Really to appreciate these diagrams you need to see and to handle a three-dimensional model in 3- space. My skills at making
drawings with my computer are nowhere near good enough yet for me to attempt a three-dimensional drawing, but Mr Charles
Card of the University of Victoria was kind enough to photograph for me a model from the University’s collection, and I
reproduce these below as figures VI 7,8 and 9.

Jeremy Tatum 4/15/2021 6.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7245

I now give some numerical values for the critical temperature and pressure, the compression factor, and the temperature and
pressure of the triple point for H2O and for CO2. These are not intended as definitive values. I looked them up in a number of
sources and I found a surprisingly wide range of the numbers quoted. They are given here merely to give the reader a rough
idea of what the values are for these two substances. The temperature quoted for the triple point of H2O is, of course, exact,
being one of the fixed points of the Kelvin scale. Recall that one atmospheric pressure is about 1.01 × 105 Pa.
The reader might like to see whether these numbers are compatible with the numbers I gave for the van der Waals constants in
Section 6.3. Exact agreement is not to be expected, because the figures I quote are only approximate and are gleaned from a
variety of sources and also, of course, neither gas can be expected to obey van der Waals’ equation exactly. If the numbers
seem to be wildly discrepant, please let me know.
We who live on the surface of Earth are familiar with water in its solid, liquid and vapour forms, and this might suggest that
the conditions on the surface of Earth, the temperature and pressure, must be close to the triple point of water. We see from the
above table that the triple point of water (which is defined to be 273.16 K = 0.01 ºC in the International Temperature Scale), is
indeed near our typical ambient temperatures, but the triple point pressure of water is 611.73 Pa, which is only about 0.006
atm. However, we are near the triple point if the partial pressure of water vapour in the atmosphere is close to 0.006 atm,
which it often is. So we are indeed close to the triple point, which is why we so often see water in its three phases. Incidentally,
the P : T diagram for the water system is a good deal more complicated that the ideal diagram of figure VI.6, particularly in
the “solid” region, since there are apparently many (about 15) different forms, or phases, of water ice.
Some idle thoughts on vapours. There is a question of how to spell “vapour”. In the United States, “vapor” is usual, and in the
United Kingdom “vapour” is usual. “Vaporize” is a bit trickier. The spelling “vaporize” is usual in the United States, but what
to do in the United Kingdom? Is it vapourize, vapourise, vaporize or vaporise? Is there a u or no u? Is it z or s? To answer the
first question: In the United Kingdom, the u, as in the United States, is omitted. Only weak spellers and those who would try to
be “more English than the English” would try to insert a u. As for s or z, either seems to be used in the United Kingdom.
Etymologically, z would be the better choice, so the spelling “vaporize” is perfectly acceptable on both sides of the Atlantic
Ocean.
More idle thoughts on vapours. Is a “vapour” a “gas? What is a “fluid”? And is glass a liquid? Some authors treat “gas” and
“vapour” as though they were quite different things: a gas is not a vapour, and a vapour is not a gas. Others regard a “vapour”
as being a sort of gas – namely a gas whose temperature is below the critical temperature and which can be liquefied by
increasing the pressure. In that case, what do you call a gas that is above the critical temperature? The term permanent gas is
often used. Thus a vapour is a gas below its critical temperature, and a permanent gas is a gas above its critical temperature.
A fluid is something that flows. Thus liquids and gases (including vapours) are fluids. There is, you would imagine, always a
clear distinction between a liquid and a gas. But is the distinction always so clear? I admit that I have never actually seen the
phenomenon that I am about to describe, but it is described so often that I presume someone has seen it! Consider a closed
container with a liquid in equilibrium with its vapour. The liquid and vapour are separated by a sharp, horizontal boundary.

Jeremy Tatum 4/15/2021 6.4.3 CC-BY-NC https://phys.libretexts.org/@go/page/7245

That is to say, the system is on the line separating liquid and vapour in figures VI.5 and 6. This line can be regarded, if you
like, as a graph of boiling point versus temperature, or equally of vapour pressure versus temperature. If you raise the pressure,
the boiling point increases; or if you increase the temperature, the vapour pressure increases. More liquid will enter the vapour
state, and, as the pressure of the vapour increases, so does its density. The liquid, on the other hand, is almost incompressible,
and, because of thermal expansion, its density decreases. As we move up the line separating liquid form vapour in the P:T
plane, the density of the vapour increases and the density of the liquid decreases. Their densities become more and more equal
until, as we approach the critical point, the boundary between liquid and vapour becomes less and less distinct, and less
constrained by gravity to be horizontal, until eventually, at the critical point, the distinction between liquid and vapour blurs
and ultimately disappears. So – what have you got then? It is certainly a fluid, but are you going to call it a gas, a vapour or a
liquid? Since none of these words would seem to have a stronger claim than either of the others, some authors refer to the
substance when a little above and to the right of the critical point in the P:T plane as a supercritical fluid.
There is also the question as to whether glass is a solid or a liquid. A famous radio personality many years ago, on a “Brains
Trust” programme broadcast by the British Broadcasting Corporation, Professor C. E. M. Joad, was famous for his sentence:
“It all depends on what you mean by…” So I suppose the question as to whether glass is a liquid or a solid depends on what
you mean by a liquid or a solid. The moment when I drop a tumbler and it shatters into many viciously sharp fragments is not
a good moment to convince me that glass is a liquid. Those who assert that glass is a liquid say that it has not got a solid
crystalline structure, and that it flows, albeit very slowly. It has a very large viscosity. We are told that windows in ancient
mediaeval cathedrals are thicker at the bottom than at the top, as a result of the viscous liquid flow over the centuries. I don’t
know if any of the many people who have told me that have actually personally measured the thickness of a cathedral window.
At any rate, before you started this chapter, you had a very clear idea in your mind about the differences between a solid, liquid
and a gas. Now that I have painstakingly explained it all, you are completely confused, and are no longer at all sure that you
know the difference.

Jeremy Tatum 4/15/2021 6.4.4 CC-BY-NC https://phys.libretexts.org/@go/page/7245

6.5: Kinetic Theory of Gases- Pressure
There will be more about macroscopic PVT relations for gases when we go further into thermodynamics. In this section, we
deal with microscopic properties, and how pressure and temperature are related to the number density of molecules and their
speed.
We shall consider an ideal gas, containing n molecules per unit volume, each of mass m, held in a cubical box of side l. The
velocity of a particular molecule is to be denoted by c = ui + vj + wk. Here u, v, w are the components of the velocity parallel
to the sides of the box. As ever, I shall use the word velocity to mean "velocity" and the word speed to mean "speed". Thus the
velocity of the molecule is c and its speed is c. We are going to start by calculating the pressure on the walls, assumed to be
caused by the collisions of millions of molecules repeatedly colliding with the walls.
("Why do you keep banging your head against the wall?" "Because it feels so good when I stop.")
Consider the x-motion. Assuming that collisions are elastic, we note that the change of the x-component of momentum when a
molecule bounces off a yz-wall is 2mu. The time taken to cross to the other side of the cube and back again is 2l/u. The number
of collisions that this molecule makes with one yz-wall per unit time is u/(2l). The rate of change of momentum of that
molecule at that wall is therefore 2mu x u/(2l) = mu2/l. The rate of change of the x-component of the momentum at that wall of
all the nl3 molecules in the box is nl
¯
¯¯¯
¯ ¯
¯¯¯
¯ ¯
¯¯¯
¯
3 2
× m u2 /l = nm l u2 . That is, the force on that wall is nml 2
u 2
, and so the pressure on
¯
¯¯¯
¯ ¯
¯¯¯
¯ ¯
¯¯¯
¯ ¯
¯¯¯¯
¯
the wall is nmu
2
. But u
2
=v
2
=w
2
(that’s assuming that the velocities are isotropic and there’s no wind) and
¯
¯¯¯
¯ ¯
¯¯¯
¯ ¯
¯¯¯¯
¯ ¯¯¯
¯ ¯
¯¯¯
¯ 1 ¯¯¯
¯
u
2
+v
2
+w
2
=c
2
(that’s Pythagoras’s theorem), and therefore u 2
=
3
2
c . So the pressure is
1 ¯¯¯
¯
2
1 ¯¯¯
¯
2
P = nm c = ρc . (6.5.1)
3 3

Here ρ is the density = mass ÷ volume = molar mass ÷ molar volume = µ/V, (here V = molar volume) and therefore
1 ¯¯¯
¯
2
PV = μc . (6.5.2)
3

¯¯¯
¯
But μc is of the translational kinetic energy of a mole of gas, and we already know that PV = RT, so that we deduce that
1

3
2 2

the translational kinetic energy of the molecules in a mole of gas is equal to RT . That is to say the mean translational kinetic
3

energy per molecule is kT , where k is Boltzmann's constant (see Section 6.1).

3

Jeremy Tatum 4/15/2021 6.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7246

6.6: Collisions
In this section, we are going to ask: What is the mean time between intermolecular collisions? What is the mean free path
between collisions? How many intermolecular collisions are there per unit volume per unit time? How many collisions with
the walls of a containing vessel are there per unit area per unit time? Since I know little chemistry, I shall assume that
molecules are hard spheres of diameter d. This may not be too bad for monatomic gases such as the rare gases. For others, the
assumption is tantamount to assuming that molecules repel each other when their centres of mass approach within a distance d.
In any case, we shall assume that the collision cross-section is of area πd2.
Notice, from the sketch below, that two equal spheres collide when their centres are separated by their diameter d, and
consequently the collision cross section (shown as a dashed circle) is of area πd2.

In fact in what follows, I’m just going to call the area of the collision cross section σ; in doing that, I don’t even have to
assume that its shape is circular.
In time t, a molecule moving with speed c sweeps out a cylinder of volume σct. If there are n molecules per unit volume, the
number of collisions that that particular molecule will experience in time t would appear to be σctn, which is to say that the
number of collisions it experiences in unit time is σcn. Thus the mean time τ between collisions would appear to be ), τ =
1/(σcn) and the mean free path λ between collisions is ). λ = 1/(σn).
But this isn't quite right, because we have not taken into account the fact that all the molecules in the above-mentioned
cylinder are moving. It is not as though our hero molecule were colliding with a set of stationary molecules. The relevant
speed to use in this analysis is the mean relative speed between molecules, and this is a little greater than the speed c of each.
Let’s see if we can do a little better.
Let’s start by supposing that all of the molecules are moving with speed c. There are two extreme sorts of collision:
The “head on” collision:

For such a collision, the relative speed between the molecules is 2c.
Then there is the sort of collision in which one molecule barely catches up with another one:

In that case the relative speed is zero.

The average relative speed is evidently somewhere between 0 and 2c.
These are extreme cases. The “average” situation is somewhat in between. We may argue that the “average” situation is for the
two molecules to be travelling in perpendicular paths:

Jeremy Tatum 4/1/2021 6.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7247

If we think of this as the “average” situation, then we may argue that the “average” relative speed between two molecules is
– –
√2c . In that case, we may conclude that the mean time between collisions is τ = 1/(√2σcn), and the mean free path is
–
λ = 1/(√2σn) .

This argument may or may not be completely convincing, but it is probably closer to the mark than our previous effort.
Let’s see if we can make a further improvement. As before, we’ll suppose that each molecule is moving with speed c.
Suppose our hero molecule to be moving upward with speed c, and another molecule approaches at an angle θ, as in the sketch
below.

By vector addition of velocities, it will be seen (a little thought will be needed) that the relative speed of approach between the
–
two molecules is √2c(1 + cos θ) . 1/2

Now the fraction of molecules approaching from angles between θ and θ + dθ is sin θdθ . This is because the area of an
1

elemental zone of a sphere of unit radius between θ and θ + dθ is 2πsinθdθ, and the total area of the sphere is 4π - see the
sketch below:

π –
Thus the mean relative speed of all the molecules is ∫ 0
√2c(1 + cos θ)
1/2
×
1

2
sin θdθ , which works out to be 4

3
c .
In this model, then, the mean time between collisions would be τ =
4
1
, and the mean free path would be λ = 4
1
.
σcn σn
3 3

However, we have still assumed that all the molecules are moving at the same speed. I am told (but I have not verified it
myself) that, if you take account of the Maxwell-Boltzmann distribution of speeds (see Section 6.7), the mean relative speed of
−−−
–
collision is ¯
√2c̄ , where ¯
c̄ is the mean speed of the Maxwell-Boltzmann distribution (equal to √
8kT

πm
.) If that is so, then we
obtain τ =
1

¯
√2σc̄ n
and λ =
1

√2σn
.

In any case, since molecules are not hard spheres (they are neither spheres nor hard) and the details of a “collision” depend on
the shape of the molecules and the force law between them, it may not be meaningful to try to obtain an extremely precise
–
formula for the mean free path, but instead settle for τ = and λ =
1
¯¯
bσc n
, and if you wish to take b ≈ √2 , you won’t be far
1

bσn

out.
Of more interest would be to calculate the mean time between collisions for various pressures and temperatures, and ask how
does this compare, for example, with the mean lifetime of an atom in an excited atomic level, or a metastable level. Or to
compare the mean free path between collisions with the mean nearest-neighbour distance between molecules in a gas. I think
under typical familiar conditions, you’ll find that the mean free path is rather longer than the mean nearest-neighbour distance.
Also of interest is the number of collisions per unit volume per unit time. If we suppose that a single molecule experiences
bσ c̄n collisions per unit time, and there are n molecules per unit volume, then the number of collisions per unit volume per
¯

unit time is
Z =
1

2
¯
bσ c̄ n
2
.

Jeremy Tatum 4/1/2021 6.6.2 CC-BY-NC https://phys.libretexts.org/@go/page/7247

The factor of 1

2
is necessary so that we don’t count collisions of A with B and of B with A as two different collisions.
Another useful result is that the number of molecules striking the walls of a containing vessel per unit area per unit time is
1 ¯¯
nc
4

To avoid repetition, I don't derive this here, but you will find a derivation in Chapter 1 Section 1.17 of Stellar Atmospheres,
where I do the derivation with photons rather than with molecules. The only difference is that, in the case of the photons, all
are moving at the same speed c (the speed of light), whereas here we have a distribution of speeds, and we use c, the mean¯¯

speed of the molecules.

Jeremy Tatum 4/1/2021 6.6.3 CC-BY-NC https://phys.libretexts.org/@go/page/7247

6.7: Distribution of Speeds
I am tempted to start by saying "Let f(u)du be the fraction of molecules of which the x-component of their velocities is
between u and u + du." But we can go a little further than this with the realization that this distribution must be symmetric
about u = 0, and therefore, whatever the function is, it must contain only even powers of u. So we can start with:
Let f(u2)du be the fraction of molecules of which the x-component of their velocities is between u and u + du. Then, unless
there is a systematic flow on the x-direction or the x-direction is somehow special, the fraction of molecules with y velocity
components between v and v + dv is f(v2)dv, and the fraction of molecules with z velocity components between w and w + dw
is f(w2)dw. The fraction of molecules in a box du dv dw of velocity space is f(u2)f(v2)f(w2)du dv dw. Since the distribution of
velocity components is independent of direction, this product must be of the form
2 2 2 2
f (u ) f (v ) f (w ) = F (c ) (6.7.1)

or
2 2 2 2 2 2
f (u ) f (v ) f (w ) = F (u +v +w ) (6.7.2)

(Question: Dimensions of f? Of F?)

It is easy to see that this is satisfied by
2 2
2 ±u / cn
f (u ) = Ae , (6.7.3)

where A and cm are constants to be determined. It should also be clear that, of the two possible solutions represented by
equation 6.7.3, we must choose the one with the minus sign.
Since we must have
∞
2
∫ f (u ) du = 1 (6.7.4)
−∞

it follows that
1
A = . (6.7.5)
−
cm √π

−
−
∞
(To see this, you have to know that ∫ .)
2
−ax 1 π
e dx = √
0 2 a

Thus we now have

1 2 2
2 −u / cm
f (u ) = −e (6.7.6)
cm √π

This is the gaussian distribution of a velocity component. We shall shortly find a physical interpretation for the constant cm.
The area under the curve represented by equation 6.7.9 is, of course, unity; the maximum value of ( is) /(1 ). 2 f u cm π
Figure VI.10 illustrates this distribution. In this figure, the unit of speed is cm. The area under the curve is 1. The maximum (at
−−−
u = 0) is 1/√−π = 0.564. Exercise: Show that the FWHM (full width at half maximum) is 2 √ln 2 c = 1.665 c . This gives m m

one physical interpretation of cm; we shall soon give another one, which will explain the use of m as a subscript.

Jeremy Tatum 4/8/2021 6.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8588

The gaussian distribution deals with velocity components. We deal now with speeds. The fraction of molecules having speeds
between c and c + dc is F(c2) times the volume of a spherical shell in velocity space of radii c and c + dc. (Some readers may
recall a similar argument in the Schrödinger equation for the hydrogen atom, in which the probability of the electron's being at
a distance between r and r + dr is the probability density ψψ* times the volume of a spherical shell.
You'll notice that physics becomes easier and easier, because you have seen it all before in different contexts. In the present
context, F is akin to the ψψ* of wave mechanics, and it could be considered to be a "speed density".) Thus the fraction of
molecules having speeds between c and c + dc is
2
2
4c 2
−c / cm
2

Φ (c ) dc = e dc (6.7.7)
3 −
cm √π

∞
I shall leave it to those who are skilled at calculus to show that ∫ Φ (c ) dc = 1 , and also to show that the maximum of this
0
2

distribution occurs for a speed of c = cm and that the maximum value of Φ (c2) is 4/ (c e√− π ). This provides another
m

interpretation of the constant cm. The speed at which the maximum of the distribution occurs is called the mode of the
distribution, or the modal speed – hence the subscript m. Equation 6.7.10 is the Maxwell-Boltzmann distribution of speeds. It is
shown in figure VI.7, in which the unit of speed is cm. The area is 1, and the maximum is 4/(e√− π ) = 0.830.

∞ ∞
The mean speed c̄ is found from ∫ cΦ (c ) dc and the root mean square speed cRMS is found from c
¯
0
2
=∫ c Φ (c ) dc .
2
RMS 0
2 2

If you have not encountered integrals of this type before, you may find that the first of them is easier than the second. If you
can do these integrals, you will find that
−−
2 3
¯¯
c = √
− cm  and cRMS = cm (6.7.8)
√π 2

¯¯¯
¯
The root mean square (RMS) speed, for which I am here using the symbol cRMS, is of course the square root of c
2
. We have
¯¯¯
¯
seen from Section 6.5 that the mean kinetic energy per molecule, 1

2
mc
2
, is equal to 3

2
kT , so now let's bring it all together:

Jeremy Tatum 4/8/2021 6.7.2 CC-BY-NC https://phys.libretexts.org/@go/page/8588

 − −
− −−−− −−−
√π 2 2kT kT
¯ ¯
cm = c̄ = 0.886 c̄ =√ cRMS = 0.816 cRMS = √ = 1.414 √ (6.7.9)
2 3 m m

−−
− −−−− −−−
2 8 8kT kT
¯ √ √ √
c̄ = c
− m = 1.128 cm = cRMS = 0.921 cRMS = = 1.596 (6.7.10)
√π 3π πm m

−
− −−
− −−−− −−−
3 3π 3kT kT
¯ ¯
cRMS = √ cm = 1.225 cm = √ c̄ = 1.085 c̄ =√ = 1.732 √ (6.7.11)
2 8 m m

Gauss:
−−−−− 2
m −
mu
2
f (u ) = √ e 2kT . (6.7.12)
2πkT

Maxwell-Boltzmann:
−
−
3/2 2

2
2 m 2 −
mc

Φ (c ) = √ ( ) c e 2kT
. (6.7.13)
π kT

One last thing occurs to me before we leave this section. Can we calculate the median speed c1/2 of the Maxwell-Boltzmann
distribution? This is the speed such that half of the molecules are moving slower than c1/2, and half are moving faster. It is the
speed that divides the area under the curve in half. If we express speeds in units of cm, we have to find c1/2 such that
c1/2
4 2 1
2 −c
− ∫ c e dc = , (6.7.14)
√π 0
2

or
c1/2 −
2 √π
2 −c
∫ c e dc = = 0.2215567314 (6.7.15)
0
8

That should keep your computer busy for a while. Mine made the answer c1/2 = 1.087 65 cm.

Jeremy Tatum 4/8/2021 6.7.3 CC-BY-NC https://phys.libretexts.org/@go/page/8588

6.8: Forces Between Molecules
We described in a qualitative manner in Section 6.3 the forces between molecules – the long-range attractive van der Waals
forces caused by induced-dipole/induced-dipole interaction, and the shortrange repulsive Coulomb forces as the molecules
approach each other closely, and how these intermolecular forces give rise to deviations from the “Boyle’s Law” expectations
for the equation of state for an ideal gas. Presumably, if we knew the exact equation for the force law as a function of
intermolecular distance, we could in principle calculate the equation of state; conversely, if we knew, through measurement,
the form of the equation of state, we could deduce the form of the intermolecular forces. I have not actually done this myself;
an early reference worthwhile to look up would be Lennard-Jones, Proc. Roy, Soc. A112, 214, (1926).
Qualitatively, the force law for the interaction between molecules would show a repulsive force rapidly falling off with
distance when the molecules are very close (the molecules are “hard”) and a longer-range attractive force at larger distances.
Two of the simpler equations that have been used to describe this are the Lennard-Jones potential:
rc 12 rc 6

V = D [1 + ( ) − 2( ) ] (6.8.1)
r r

and the Morse potential:

2
−(r−re )/a
V = D (1 − e ) (6.8.2)

Each of these goes to V → D as r → ∞, and V = 0 when r = re. The Lennard-Jones potential (but not the Morse potential) goes
to ∞ as r → 0.
These expressions cannot be “derived” in the usual sense; they are merely expressions that are useful for discussion in that
they describe qualitatively the shape of the potential function that you would expect. The Lennard-Jones expression is often
used in discussions of the van der Waals force: if the van der Waals attractive force is due mostly to induced-dipole/induced-
dipole interaction, an r−6 term is about right. The Morse potential is used more often in discussion of the force between atoms
in a bound molecule. If the Morse potential is put into the Schrödinger equation for an anharmonic oscillating diatomic
molecule, it results in a simple solution for the eigenfunctions and eigenvalues, with the energy levels being given as quadratic
(and no higher) in v + .
1

The parameter a in the Morse expression determines how narrow or how broad the minimum is. It is left as an exercise for the
reader to show that the FWHm (full width at half minimum) of the Morse expression is the same as for the Lennard-Jones
potential for
– 1/6 – 1/6
(2 + √2) − (2 − √2)
a = – = 0.177212908 (6.8.3)
ln(3 + √8)

In figure VI.12, I show, as continuous curves, the Morse potentials (in order of increasing width) for a/re = 0.1, 0.1772, 0.3 and
0.4, and the Lennard-Jones potential as a dashed curve. Further comparisons between these two potential functions can be

Jeremy Tatum 3/4/2021 6.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8589

found in T.-C. Lim, Z. Naturforschung, 58a, 615, (2003).

Jeremy Tatum 3/4/2021 6.8.2 CC-BY-NC https://phys.libretexts.org/@go/page/8589

CHAPTER OVERVIEW
7: THE FIRST AND SECOND LAWS OF THERMODYNAMICS

7.1: THE FIRST LAW OF THERMODYNAMICS, AND INTERNAL ENERGY

7.2: WORK
7.3: ENTROPY
7.4: THE SECOND LAW OF THERMODYNAMICS

1 4/29/2021
7.1: The First Law of Thermodynamics, and Internal Energy
The First Law of thermodynamics is:
The increase of the internal energy of a system is equal to the sum of the heat added to the system plus the work done on the
system.
In symbols:
dU = dQ + dW (7.1.1)

You may regard this, according to taste, as any of the following

A fundamental law of nature of the most profound significance;
or A restatement of the law of conservation of energy, which you knew already;
or A recognition that heat is a form of energy.
or A definition of internal energy.
Note that some authors use the symbol E for internal energy. The majority seem to use U, so we shall use U here.
Note also that some authors write the first law as dU = dQ − dW, so you have to be clear what the author means by dW. A
scientist is likely to be interested in what happens to a system when you do work on it, and is likely to define dW as the work
done on the system, in which case dU = dQ + dW. An engineer, in the other hand, is more likely to be asking how much work
can be done by the system, and so will prefer dW to mean the work done by the system, in which case dU = dQ − dW.
The internal energy of a system is made up of many components, any or all of which may be increased when you add heat to
the system or do work on it. If the system is a gas, for example, the internal energy includes the translational, vibrational and
rotational kinetic energies of the molecules. It also includes potential energy terms arising from the forces between the
molecules, and it may also include excitational energy if the atoms are excited to energy levels above the ground state. It may
be difficult to calculate the total internal energy, depending on which forms of energy you take into account. And of course the
potential energy terms are always dependent on what state you define to have zero potential energy. Thus it is really
impossible to define the total internal energy of a system uniquely. What the first law tells us is the increase in internal energy
of a system when heat is added to it and work is done on it.
Note that internal energy is a function of state. This means, for example in the case of a gas, whose state is determined by its
pressure, volume and temperature, that the internal energy is uniquely determined (apart from an arbitrary constant) by P, V
and T – i.e. by the state of the gas. It also means that in going from one state to another (i.e. from one point in PVT space to
another), the change in the internal energy is route-independent. The internal energy may be changed by performance of work
or by addition of heat, or some combination of each, but, whatever combination of work and energy is added, the change in
internal energy depends only upon the initial and final states. This means, mathematically, that dU is an exact differential (see
Chapter 2, Section 2.1). The differentials dQ and dW, however, are not exact differentials.
Note that if work is done on a Body by forces in the Rest of the Universe, and heat is transferred to the Body from the Rest of
the Universe (also known as the Surroundings of the Body), the internal energy of the Body increases by dQ + dW, while the
internal energy of the Rest of the Universe (the Surroundings) decreases by the same amount. Thus the internal energy of the
Universe is constant. This is an equivalent statement of the First Law. It is also sometimes stated as “Energy can neither be
created nor destroyed”.

Jeremy Tatum 4/1/2021 7.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7249

7.2: Work
There are many ways in which you can do work on a system. You may compress a gas; you may magnetize some iron; you
may charge a battery; you may stretch a wire, or twist it; you may stir a beaker of water.
Some of these processes are reversible; others are irreversible or dissipative. The work done in compressing a gas is reversible
if it is quasistatic, and the internal and external pressures differ from each other always by only an infinitesimal amount.
Charging a lead-acid car battery may be almost reversible; charging or discharging a flashlight battery is not, because it has a
high internal resistance, and the chemical reactions are irreversible. Stretching or twisting a wire is reversible as long as you do
not exceed the elastic limit. If you do exceed the elastic limit, it will not return to its original length; that is, it exhibits elastic
hysteresis. When you magnetize a metal sample, you are doing work on it by rotating the little magnetic moments inside the
metal. Is this reversible? To answer this, read about the phenomenon of magnetic hysteresis in Chapter 12, Section 12.6, of
Electricity and Magnetism.
Work that is reversible is sometimes called configuration work. It is also sometimes called PdV work, because that is a
common example. Work that is not reversible is sometimes called dissipative work. Forcing an electric current through a wire
is clearly dissipative.
For much of the time, we shall be considering the work that is done on a system by compressing it. Solids and liquids require
huge pressures to change their volumes significantly, so we shall often be considering a gas. We imagine, for example, that we
have a quantity of gas held in a cylinder by a piston. The work done in compressing it in a reversible process is −PdV. If you
are asking yourself "Is P the pressure that the gas is exerting on the piston, or the pressure that the piston is exerting on the
gas?", remember that we are considering a reversible and quasistatic process, so that the difference between the two is at all
stages infinitesimal. Remember also that in calculus, if x is some scalar quantity, the expression dx doesn't mean vaguely the
"change" in x (an ill-defined word), but it means the increment or increase in x. Thus the symbol dV means the increase in
volume, which is negative if we are doing work on the gas by compressing it. In any case whether you adopt the scientist
convention or the engineer convention (try both) the first law, when applied to the compression or expansion of a gas, becomes
dU = dQ − P dV . (7.2.1)

Jeremy Tatum 3/11/2021 7.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7250

7.3: Entropy
Definition: Entropy Differential
If an infinitesimal quantity of heat dQ is added to a system at temperature T, and if no irreversible work is done on the
system, the increase in entropy dS of the system is defined by
dQ
dS = . (7.3.1)
T

Exercise 7.3.1
What are the SI units of entropy?

Note that, since dQ is supposed to be an infinitesimal quantity of heat, any increase in temperature is also infinitesimal. Note
also that, as with internal energy, we have defined only what is meant by an increase in entropy, so we are not in any position
to state what the entropy of a system is. (Much later, we shall give evidence that the molar entropy of all substances is the
same at the absolute zero of temperature. It may then be convenient to define the zero of the entropy scale as the molar entropy
at the absolute zero of temperature. At present, we have not yet shown that there is an absolute zero of temperature, let alone
of entropy.)
To the question "What is meant by entropy?" a student will often respond with "Entropy is the state of disorder of a system."
What a vague, unquantitative and close to meaningless response that is! What is meant by "disorder"? What could possibly be
meant by a statement such as "The state of disorder of this system is 5 joules per kelvin"? Gosh! I would give nought marks
out of ten for such a response! Now it is true, when we come to the subjects of statistical mechanics, and statistical
thermodynamics and mixing theory, that there is a sense in which the entropy of a system is some sort of measure of the state
of disorder, in the sense that the more disordered or randomly mixed a system is, the higher its entropy, and random processes
do lead to more disorder and to higher entropy. Indeed, this is all connected to the second law of thermodynamics, which we
haven't touched upon yet. But please, at the present stage, entropy is defined as I have stated above, and, for the time being, it
means nothing less and nothing more.
It will have been noted that, in our definition of entropy so far, we specified that no irreversible work be done on the system.
What if some irreversible work is done? Let us suppose that we do work on a gas in two ways. (I choose a gas in this
discussion, because it is easier to imagine compressing a gas with P dV work than it is with a solid or a liquid, because the
compressibility of a solid or a liquid is relatively low. But the same argument applies to any substance.) We compress it with
the piston, but, at the same time, we also stir it with a paddle. In that case, the work done on the gas is more than −P dV .
(Remember that −P dV is positive.) If we didn't compress it at all, but only stirred it, dV would be zero, but we would still
have done work on the gas by stirring. Let's suppose the work done on the gas is
δW = −P dV + δWirr . (7.3.2)

The part δW is the irreversible or dissipative part of the work done on the gas; it is unrecoverable as work, and is
irr

irretrievably converted to heat. You cannot use it to turn the paddle back. Nor can you cool the gas by turning the paddle
backwards.
We can now define the increase of entropy in the irreversible process by
T dS = dQ + dWirr (7.3.3)

that is,
dQ + dWirr
dS = . (7.3.4)
T

In other words, since dW irr is irreversibly converted to heat, it is just as though it were part of the addition of heat.
In summary,

Jeremy Tatum 4/29/2021 7.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7251

 dU = dQ + dW (7.3.5)

and
dU = T dS − P dV (7.3.6)

apply whether there is reversible or irreversible work. But only if there is no irreversible (unrecoverable) work does
dQ = T dS and dW = −P dV . If there is any irreversible work,

dW = −P dV + dWirr (7.3.7)

and
dQ = T dS − dWirr . (7.3.8)

Of course there are other forms of reversible work than P dV work; we just use the expansion of gases as a convenient
example.
Note that P , V , and T are state variables (together, they define the state of the system) and U is a function of state. Thus the
entropy, too, is a function of state. That is to say that the change in entropy as you go from one point in PVT-space to another
point is route-independent. If you return to the same point that you started at (the same state, the same values of P , V and T ),
there is no change in entropy, just as there is no change in internal energy.

Definition: Specific Heat Capacity

The specific heat capacity C of a substance is the quantity of heat required to raise the temperature of unit mass of it by
one degree. We shall return to the subject of heat capacity in Chapter 8. For the present, we just need to know what it
means, in order to do the following exercise concerning entropy.

Example 7.3.1
A litre (mass = 1 kg) of water is heated from 0 oC to 100 oC. What is the increase of entropy? Assume that the specific
heat capacity of water is C = 4184 J kg−1 K−1, that it does not vary significantly through the temperature range of the
question, and that the water does not expand significantly, so that no significant amount of work (reversible or
irreversible) is done.
Solution
The heat required to heat a mass m of a substance through a temperature range dT is mC dT . The entropy gained then is
mC dT

T
. The entropy gained over a finite temperature range is therefore
T2
dT T2
mC ∫ = mC ln( )
Ti T T1

373.15
−1
= 1 × 4184 × ln( ) = 1305 JK .
273.15

Jeremy Tatum 4/29/2021 7.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7251

7.4: The Second Law of Thermodynamics
In a famous lecture entitled The Two Cultures given in 1959, the novelist C. P. Snow commented on a common intellectual
attitude of the day - that true education consisted of familiarity with the humanities, literature, arts, music and classics, and that
scientists were mere uncultured technicians and ignorant specialists who never read any of the great works of literature. He
described how he had often been provoked by such an attitude into asking some of the self-proclaimed intellectuals if they
could describe the Second Law of Thermodynamics – a question to which he invariably received a cold and negative response.
Yet, he said, he was merely asking something of about the scientific equivalent of "Have you read a work of Shakespeare?"
So I suggest that, if you have never read a work of Shakespeare, take a break for a moment from thermodynamics, go and read
A Midsummer Night's Dream, and come back refreshed and ready to complete your well-rounded education by learning the
Second Law of Thermodynamics.
We have defined entropy in such a manner that if a quantity of heat dQ is added reversibly to a system at temperature T, the
increase in the entropy of the system is dS = dQ/T. We also pointed out that if the heat is transferred irreversibly, dS > dQ/T.
Now consider the following situation (figure VII.1).

An isolated system consists of two bodies, A at temperature T1 and B at temperature T2, such that T2 >T1. Heat will eventually
be exchanged between the two bodies, and on the whole more heat will be transferred from B to A than from A to B. That is,
there will be a net transference of heat, dQ, from B to A. Perhaps this heat is transferred by radiation. Each body is sending
forth numerous photons of energy, but there is, on the whole, a net flow of photons from B to A. Or perhaps the two bodies are
in contact, and heat is being transferred by conduction. The vibrations in the hot body are more vigorous than those in the cool
body, so there will be a net transfer of heat from B to A. However, since the emission of photons in the first case, and the
vibrations in the second place, are random, it will be admitted that it is not impossible that at some time more photons may
move from A to B than from B to A. Or, in the case of conduction, most of the atoms in A happen to be moving to the right
while only a few atoms in B are moving to the left in the course of their oscillations. But, while admitting that this is in
principle possible and not outside the laws of physics, it is exceedingly unlikely to happen in practice; indeed so unlikely as
hardly to be taken seriously. Thus, in any natural, spontaneous process, without the intervention of an External Intelligence, it
is almost certain that there will be a net transfer of heat from B to A. And this process, barring the most unlikely set of
circumstances, is irreversible.
The hot body will lose an amount of entropy dQ/T2, while the cool body will gain an amount of entropy dQ/T1, which is
greater than dQ/T2. Thus the entropy of the isolated system as a whole increases by dQ/T1 − dQ/T2.
From this argument, we readily conclude that any natural, spontaneous and irreversible thermodynamical processes taking
place within an isolated system are likely to lead to an increase in entropy of the system. This is perhaps the simplest statement
of the Second Law of Thermodynamics.
I have used the phrase "is likely to", although it will be realised that in practice the possibility that the entropy might decrease
in a natural process is so unlikely as to be virtually unthinkable, even though it could in principle happen without violating any
fundamental laws of physics.
You could regard the Universe as an isolated system. Think of a solid Body sitting somewhere in the Universe. If the Body is
hot, it may spontaneously lose heat to the Rest of the Universe. If it is cold, it may spontaneously absorb heat from the Rest of
the Universe. Either way, during the course of a spontaneous process, the entropy of the Universe increases.
The transference of heat from a hot body to a cooler body, so that both end at the same intermediate temperature, involves, in
effect, the mixing of a set of fast-moving molecules and a set of slow-moving molecules. A similar situation arises if we start

Jeremy Tatum 1/28/2021 7.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7252

with a box having a partition down the middle, and on one side of the partition there is a gas of blue molecules and on the
other there is a gas of red molecules. If we remove the partition, eventually the gases will mix into one homogeneous gas. By
only a slight extension of the idea of entropy discussed in courses in statistical mechanics, this situation can be described as an
increase of entropy – called, in fact, the entropy of mixing. If you saw two photographs, in one of which the blue and red
molecules were separated, and in the other the two colours were thoroughly mixed, you would conclude that the latter
photograph was probably taken later than the former. But only "probably"; it is conceivable, within the laws of physics, that
the velocities of the blue and red molecules separated themselves out without external intervention. This would be allowed
perfectly well within the laws of physics. Indeed, if the velocities of all the molecules in the mixed gases were to be reversed,
the gases would eventually separate into their two components. But this would seem to be so unlikely as never in practice to
happen. The second law says that the entropy of an isolated system is likely (very likely!) to increase with time. Indeed it could
be argued that the increase of entropy is the criterion that defines the direction of the arrow of time. (For more on the arrow of
time, see Section 15.12 of the notes on Electricity and Magnetism of this series. Also read the article on the arrow of time by
Paul Davis, Astronomy & Geophysics (Royal Astronomical Society) 46, 26 (2005). You’ll probably also enjoy H. G. Wells’s
The Time Machine.)
Note that, in the example of our two bodies exchanging heat, one loses entropy while the other gains entropy; but the gain by
the one is greater than the loss from the other, with the result that there is an increase in the entropy of the system as a whole.
The principle of the increase of entropy applies to an isolated system.
In case you have ever wondered (who hasn’t?) how life arose on Earth, you now have a puzzle. Surely the genesis and
subsequent evolution of life on Earth represents an increase in order and complexity, and hence a decrease in the entropy of
mixing. You may conclude from this that the genesis and subsequent evolution of life on Earth requires Divine Intervention, or
Intelligent Design, and that the Second Law of Thermodynamics provides Proof of the Existence of God. Or you may
conclude that Earth is not an isolated thermodynamical system. Your choice.

Jeremy Tatum 1/28/2021 7.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7252

CHAPTER OVERVIEW
8: HEAT CAPACITY, AND THE EXPANSION OF GASES

8.1: HEAT CAPACITY

8.2: RATIO OF THE HEAT CAPACITIES OF A GAS
8.3: ISOTHERMAL EXPANSION OF AN IDEAL GAS
8.4: REVERSIBLE ADIABATIC EXPANSION OF AN IDEAL GAS
8.5: THE CLÉMENT-DESORMES EXPERIMENT
8.6: THE SLOPES OF ISOTHERMS AND ADIABATS
8.7: SCALE HEIGHT IN AN ISOTHERMAL ATMOSPHERE
8.8: ADIABATIC LAPSE RATE
8.9: NUMERICAL VALUES OF SPECIFIC AND MOLAR HEAT CAPACITIES
8.10: HEAT CAPACITIES OF SOLIDS

1 4/29/2021
8.1: Heat Capacity
Definition: The heat capacity of a body is the quantity of heat required to raise its temperature by one degree. Its SI unit is J
K−1.
Definition: The specific heat capacity of a substance is the quantity of heat required to raise the temperature of unit mass of it
by one degree. Its SI unit is J kg−1 K−1.
Definition: The molar heat capacity of a substance is the quantity of heat required to raise the temperature of a molar amount
of it by one degree. (I say "molar amount". In CGS calculations we use the mole – about 6 × 1023 molecules. In SI calculations
we use the kilomole – about 6 × 1026 molecules.) Its SI unit is J kilomole−1 K−1.
Some numerical values of specific and molar heat capacity are given in Section 8.7.
One sometimes hears the expression "the specific heat" of a substance. One presumes that what is meant is the specific heat
capacity.
The above definitions at first glance seem easy to understand – but we need to be careful. Let us imagine again a gas held in a
cylinder by a movable piston. I choose a gas because its volume can change very obviously on application of pressure or by
changing the temperature. The volume of a solid or a liquid will also change, but only by a small and less obvious amount. If
you supply heat to a gas that is allowed to expand at constant pressure, some of the heat that you supply goes to doing external
work, and only a part of it goes towards raising the temperature of the gas. On the other hand, if you keep the volume of the gas
constant, all of the heat you supply goes towards raising the temperature. Consequently, more heat is required to raise the
temperature of the gas by one degree if the gas is allowed to expand at constant pressure than if the gas is held at constant
volume and not allowed to expand. Thus the heat capacity of a gas (or any substance for that matter) is greater if the heat is
supplied at constant pressure than if it is supplied at constant volume. Thus we have to distinguish between the heat capacity
at constant volume CV and the heat capacity at constant pressure CP, and, as we have seen CP > CV.
If the heat is added at constant volume, we have simply that dU = dQ = CVdT.
One other detail that requires some care is this. The specific heat capacity of a substance may well vary with temperature,
even, in principle, over the temperature range of one degree mentioned in our definitions. Therefore, we really have to define
the heat capacity at a given temperature in terms of the heat required to raise the temperature by an infinitesimal amount rather
than through a finite range. Thus it is perhaps easiest to define heat capacity at constant volume in symbols as follows:
∂U
CV = ( ) (8.1.1)
∂T
V

(Warning: Do not assume that CP = (∂U/∂T)P. That isn’t so. The correct expression is given as equation 9.1.13 in Chapter 9 on
Enthalpy.)
As with many equations, this applies equally whether we are dealing with total, specific or molar heat capacity or internal
energy.
If heat is supplied at constant pressure, some of the heat supplied goes into doing external work PdV, and therefore

Cp dT = CV dT + P dV . (8.1.2)

For a mole of an ideal gas at constant pressure, P dV = R dT, and therefore, for an ideal gas,

Cp = Cv + R, (8.1.3)

where, in this equation, CP and CV are the molar heat capacities of an ideal gas.
We shall see in Chapter 10, Section 10.4, if we can develop a more general expression for the difference in the heat capacities
of any substance, not just an ideal gas. But let us continue, for the time being with an ideal gas.
In an ideal gas, there are no forces between the molecules, and hence no potential energy terms involving the intermolecular
distances in the calculation of the internal energy. In other words, the internal energy is independent of the distances between
molecules, and hence the internal energy is independent of the volume of a fixed mass of gas if the temperature (hence kinetic
energy) is kept constant. That is, for an ideal gas,

Jeremy Tatum 2/26/2021 8.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7256

 ∂U
( ) = 0. (8.1.4)
∂V
T

Let us think now of a monatomic gas, such as helium or argon. When we supply heat to (and raise the temperature of) an ideal
monatomic gas, we are increasing the translational kinetic energy of the molecules. If the gas is ideal, so that there are no
intermolecular forces then all of the introduced heat goes into increasing the translational kinetic energy (i.e. the temperature)
of the gas. (Recall that a gas at low pressure is nearly ideal, because then the molecules are so far apart that any intermolecular
forces are negligible.) Recall from Section 6.5 that the translational kinetic energy of the molecules in a mole of gas is RT .3

The molar internal energy, then, of an ideal monatomic gas is

3
U = RT +  constant.  (8.1.5)
2

From equation 8.1.1, therefore, the molar heat capacity at constant volume of an ideal monatomic gas is
3
CV = R. (8.1.6)
2

The molar heat capacities of real monatomic gases when well above their critical temperatures are indeed found to be close to
this.
When we are dealing with polyatomic gases, however, the heat capacities are greater. This is because, when we supply heat,
only some of it goes towards increasing the translational kinetic energy (temperature) of the gas. Some of the heat goes into
increasing the rotational kinetic energy of the molecules. (Wait! Some of you are asking yourselves: "But do not atoms of
helium and argon rotate? Do they not have rotational kinetic energy?" These are very good questions, but I am going to
pretend for the moment that I haven't heard you. Perhaps, before I come to the end of this section, I may listen.)
When two molecules collide head on, there is an interchange of translational kinetic energy between them. But if they have a
glancing collision, there is an exchange of translational and rotational kinetic energies. If millions of molecules are colliding
with each other, there is a constant exchange of translational and rotational kinetic energies. When a dynamic equilibrium has
been established, the kinetic energy will be shared equally between each degree of translational and rotational kinetic energy.
(This is the Principle of Equipartition of Energy.) We know that the translational kinetic energy per mole is RT - that is,
3

RT for each translational degree of freedom ( \frac{1}{2} m \overline{u}^{2}, \frac{1}{2} m \overline{v^{2}}, \frac{1}
1

{2} m \overline{w^{2}}\)). There is an equal amount of kinetic energy of rotation (with an exception to be noted below), so
that the internal energy associated with a mole of a polyatomic gas is 3RT plus a constant, and consequently the molar heat
capacity of an ideal polyatomic gas is
CV = 3R. (8.1.7)

It takes twice the heat to raise the temperature of a mole of a polyatomic gas compared with a monatomic gas.
The exception we mentioned is for linear molecules. These are molecules in which all the atoms are in a straight line. This
necessarily includes, of course, all diatomic molecules (the oxygen and nitrogen in the air that we breathe) as well as some
heavier molecules such as CO2, in which all the molecules (at least in the ground state) are in a straight line. (The molecule
H2O is not linear.) In linear molecules, the moment of inertia about the internuclear axis is negligible, so there are only two
degrees of rotational freedom, corresponding to rotation about two axes perpendicular to each other and to the internuclear
axis. Thus there are five degrees of freedom in all (three of translation and two of rotation) and the kinetic energy associated
with each degree of freedom is RT per mole for a total of RT per mole, so the molar heat capacity is
1

2
5

5
CV = R. (8.1.8)
2

Summary: A monatomic gas has three degrees of translational freedom and none of rotational freedom, and so we would
expect its molar heat capacity to be RT .
3

A diatomic or linear polyatomic gas has three degrees of translational freedom and two of rotational freedom, and so we would
expect its molar heat capacity to be RT .
5

A nonlinear polyatomic gas has three degrees of translational freedom and three of rotational freedom, and so we would expect
its molar heat capacity to be 3R.

Jeremy Tatum 2/26/2021 8.1.2 CC-BY-NC https://phys.libretexts.org/@go/page/7256

How do real gases behave compared with these predictions? The monatomic gases (helium, neon, argon, etc) behave very
well. The diatomic gases quite well, although at room temperature the molar heat capacities of some of them are a little higher
than predicted, while at low temperatures the molar heat capacities drop below what is predicted. Indeed below about 60 K the
molar heat capacity of hydrogen drops to about RT - just as if it had become a monatomic gas or, though still diatomic, the
3

molecules were somehow prevented from rotating. The molar heat capacities of nonlinear polyatomic molecules tend to be
rather higher than predicted.
First let us deal with why the molar heat capacities of polyatomic molecules and some diatomic molecules are a bit higher than
predicted. This is because the molecules may vibrate. When we add heat, some of the heat is used up in increasing the rate of
rotation of the molecules, and some is used up in causing them to vibrate, so it needs a lot of heat to cause a rise in temperature
(translational kinetic energy). The possibility of vibration adds more degrees of freedom, and another RT to the molar heat
1

capacity for each extra degree of vibration. To be strictly correct, the "number of degrees of freedom" in this connection is the
number of squared terms that contribute to the internal energy. Each vibrational mode adds two such terms – a kinetic energy
term and a potential energy term. This means that the predicted molar heat capacity for a nonrigid diatomic molecular gas
would be RT . Polyatomic gases have many vibrational modes and consequently a higher molar heat capacity.
7

So – why is the molar heat capacity of molecular hydrogen not RT at all temperatures? Why is it about RT at room
7

2
5

temperature, as if it were a rigid molecule that could not vibrate? True, at higher temperatures the molar heat capacity does
increase, though it never quite reaches RT before the molecule dissociates. Why does the molar heat capacity decrease at
7

lower temperatures, reaching RT at 60 K, as if it could no longer rotate?

3

Let us ask some further questions, which are related to these. We said earlier that a monatomic gas has no rotational degrees of
freedom. Why not? True, the moment of inertia is very small, but, if we accept the principle of equipartition of energy, should
not each rotational degree of freedom hold as much energy as each translational degree of freedom? Also, we said that a linear
molecule has just two degrees of freedom. It is true that the moment of inertia about the internuclear axis is very small. This is
not the same thing as saying that it cannot rotate about that axis. If all degrees of freedom equally share the internal energy,
then the angular speed about the internuclear axis must be correspondingly large.
Now I could make various excuses about these problems. The fact is, however, that the classical model that I have described
may look good at first, but, when we start asking these awkward questions, it becomes evident that the classical theory really
fails to answer them satisfactorily. In truth, the failure of classical theory to explain the observed values of the molar heat
capacities of gases was one of the several failures of classical theory that helped to give rise to the birth of quantum theory.
Quantum theory in fact accounts spectacularly well and in detail for the specific heat capacities of molecules and how the heat
capacities vary with temperature. This topic is often dealt with on courses on statistical thermodynamics, and I just briefly
mention the explanation here. The solution of Schrödinger's equation for a rigid rotator shows that the rotational energy can
exist with a number of separated discrete values, and the population of these rotational energy levels is governed by
Boltzmann's equation in just the same way as the population of the electronic energy levels in an atom. At temperatures of 60
K, the spacing of the rotational energy levels is large compared with kT, and so the rotational energy levels are unoccupied.
Thus, in that very real sense, the hydrogen molecule does indeed stop rotating at low temperatures. The spacing of the energy
level is inversely proportional to the moment of inertia, and the moment of inertia about the internuclear axis is so small that
the energy of the first rotational energy level about this axis is larger than the dissociation energy of the molecule, so indeed
the molecule cannot rotate about the internuclear axis. Vibrational energy is also quantised, but the spacing of the vibrational
levels is much larger than the spacing of the rotational energy levels, so they are not excited at room temperatures. This has
been only a brief account of why classical mechanics fails and quantum mechanics succeeds in correctly predicting the
observed heat capacities of gases. It is a very interesting subject, and the reader may well want to learn more about it – but that
will have to be elsewhere.

Jeremy Tatum 2/26/2021 8.1.3 CC-BY-NC https://phys.libretexts.org/@go/page/7256

8.2: Ratio of the Heat Capacities of a Gas
The ratio of the heat capacities of a gas at constant pressure and at constant volume plays an important part in many
calculations involving the expansion and contraction of gases. The ratio appears, for one example of many that could be
chosen, in the theoretical expression for the speed of sound in a gas. The higher the ratio CP/CV, the faster the speed of sound.
The ratio is generally given the symbol γ:
Cp
= γ. (8.2.1)
Cv

Apart from any other reason, one reason for its importance is that the ratio is easier to measure precisely than either heat
capacity separately. For example, you could determine it from a measurement of the speed of sound, which is easier than
adding heat to a sample of gas at constant pressure and again at constant volume and measuring the rise in temperature.
We have seen that, for gases that behave as we would like them to behave, the molar heat capacities CV at constant
temperatures for monatomic, diatomic and nonlinear polyatomic gases without molecular vibration are respectively R, R, 3

2
5

and 3R. And since, for an ideal gas, CP = CV + R, (equation 8.1.3), we expect the corresponding values for CP to be R, R 5

2
7

and 4R. and Thus the expected values of γ are 5/3, 7/5 and 4/3.

Jeremy Tatum 4/26/2021 8.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7257

8.3: Isothermal Expansion of an Ideal Gas
An ideal gas obeys the equation of state PV = RT (V = molar volume), so that, if a fixed mass of gas kept at constant
temperature is compressed or allowed to expand, its pressure and volume will vary according to PV = constant. That is,
Boyle's Law. We can calculate the work done by a mole of an ideal gas in a reversible isothermal expansion from volume V1 to
volume V2 as follows.
V2 V2
dV
W =∫ P dV = RT ∫ = RT ln(V2 / V1 ) (8.3.1)
V1 V1 V

Jeremy Tatum 4/8/2021 8.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7258

8.4: Reversible Adiabatic Expansion of an Ideal Gas
An adiabatic process is one in which no heat enters or leaves the system, and hence, for a reversible adiabatic process the first
law takes the form dU = − PdV. But from equation 8.1.1, CV = (∂U/∂T)V. But the internal energy of an ideal gas depends only
on the temperature and is independent of the volume (because there are no intermolecular forces), and so, for an ideal gas, CV
= dU/dT, and so we have dU = CVdT. Thus for a reversible adiabatic process and an ideal gas, CVdT = −PdV. (The minus sign
shows that as V increases, T decreases, as expected.) But for a mole of an ideal gas, PV = RT = (CP − CV)T, or P = (CP −
CV)T/V.
Therefore

Cv dT = − (CP − CV ) T dV /V (8.4.1)

(You may be wondering whether C and V are molar, specific or total quantities. If you look at the equation you'll agree that it
is valid whether the volume and heat capacities are molar, specific or total.)
Separate the variables and write γ for CP/CV:
dT dV
+ (γ − 1) = 0. (8.4.2)
T V

Integrate:
γ−1
TV =  constant . (8.4.3)

This shows how temperature and volume of an ideal gas vary during a reversible adiabatic expansion or compression. If the
gas expands, the temperature goes down. If the gas is compressed, it becomes hot. Of course the pressure varies also, and the
ideal gas conforms to the equation PV/T = constant. On elimination of T we obtain
γ
PV =  constant . (8.4.4)

On elimination of V we obtain
−(γ−1) γ
P T =  constant.  (8.4.5)

In figure VIII.1 I draw, as light curves, five isotherms – i.e. the paths that would be taken by an ideal gas in the PV plane in
isothermal processes at five temperatures. I also show, as a heavier line, an adiabat, PVγ = constant , which I calculated for γ =
5/3. The adiabat is steeper than the isotherms, and the curve shows that, as the gas expands adiabatically, the temperature
drops. If you know the original temperature and the old and new volumes, equation 8.4.3 will enable you to calculate the new
temperature. If you know the original temperature and the old and new pressures, equation 8.4.5 will enable you to calculate
the new temperature. While these purely thermodynamic arguments show that a gas becomes hotter if you compress it, this is
also to be expected at the microscopic level. Thus, if a molecule bounces elastically against a piston that is moving towards it,
it will gain kinetic energy, and it will lose kinetic energy if it bounces off a piston that is moving away from it.

Jeremy Tatum 4/22/2021 8.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7259

Let us calculate the work done by a mole of an ideal gas in a reversible adiabatic expansion from (P1 , V1) to (P2 , V2):
V2

W =∫ P dV . (8.4.6)
V1

For a reversible adiabatic expansion, PVγ = K, and therefore

v2
K −(γ−1) −(γ−1)
−γ
W =K∫ V dV = (V −V ) (8.4.7)
1 2
v1
γ −1

That is,
P1 V1 − P2 V2 R (T1 − T2 )
W = = (8.4.8)
γ −1 γ −1

(Note that T2 < T1 in this adiabatic expansion.)

Compare this with equation 8.3.1 for an isothermal expansion.
Note also that, since R = CP − CV and CP/CV = γ this can also be written
W = CV (T1 − T2 ) (8.4.9)

This is also equal to the heat that would be lost if the gas were to cool from T1 to T2 at constant volume. Think about this! Is it
coincidence, or must it be so?
Here is a useful exercise. In figure VIII.2, a gas goes from (P1, V1) to (P2, V2) via three different reversible routes:
(a) An isobaric expansion followed by an isochoric decrease in pressure;
(b) An isochoric decrease in pressure followed by an isobaric expansion;
(c) An adiabatic expansion.

At each stage, calculate the work done on or by the gas, the heat gained by the gas or lost from the gas, and the increase or
decrease of the internal energy of the gas. This exercise will illustrate that U is a function of state, but Q and W are not. (I
expect the answers to be in algebra; ignore the numbers on the axes – they don’t mean anything in particular.)

Jeremy Tatum 4/22/2021 8.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7259

8.5: The Clément-Desormes Experiment
This is a simple, quick and effective experiment often seen in teaching laboratories for measuring γ for air, or, with some extra
effort, any other gas.
Sometimes this experiment is referred to as the experiment of Clément and Desormes, and sometimes as the experiment of
Clément-Desormes. Apparently Charles-Bernard Desormes was the uncle of Nicolas Clément, and they both worked on the
experiment. Nicolas Desormes later legally changed his name to Nicolas Clément-Desormes. Thus you can refer either to the
experiment of Clément and Desormes or to the experiment of Clément-Desormes!
A bottle of air starts at P1, T1. Pl is a little greater than atmospheric pressure P0. T1 is the ambient room temperature. The
bottle is provided with some device for measuring pressure (for example, a manometer). We'll see that there is no need to
measure temperatures. The stopcock is quickly opened and immediately closed. The pressure at that moment is just
atmospheric pressure, which I'll call P0, and the temperature is T2, which is a little cooler than the original room temperature
T1. The bottle of gas is now allowed slowly to warm up isochorically to its original temperature Tl, by which time the new
pressure P2 is greater than atmospheric pressure P0 but not as large as the original pressure P1. You should sketch these two
stages on a PV diagram.
For the adiabatic process,
−(γ−1) γ −(γ−1) γ
P T =P T . (8.5.1)
1 1 0 2

For the isochoric process,

P0 / T2 = P2 / T1 . (8.5.2)

I'll leave you to do the algebra and eliminate T2/T1 from these equations and hence show that
ln(P1 / P0 )
γ = . (8.5.3)
ln(P1 / P2 )

In the above analysis, we assumed that the gas was ideal and the expansion was adiabatic and reversible. The gas is nearly
ideal if it is a long way above its critical temperature and there are no enormous ranges of P and T. The expansion is adiabatic
if P2 is measured immediately after the stopcock is opened and closed, so that there is no time for heat to enter or leave the
system. It is reversible only if P1 − P0 << P0. If you want to do the experiment yourself right now without getting up from
your comfortable seat, have a look at http://www.univ-lemans.fr/enseignements/physique/02/thermo/clement.html

Jeremy Tatum 4/15/2021 8.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7260

8.6: The Slopes of Isotherms and Adiabats
For an ideal gas in an isothermal process, PV = constant.
In a reversible adiabatic process:
PVγ = constant,
TVγ − 1 = constant,
P1 − γTγ = constant.
From these it is easy to see that the ratios of the adiabatic, isothermal, isobaric and isochoric slopes are as follows:
∂P ∂P ∂V 1 ∂V ∂P γ ∂P
( ) = γ( ) ; ( ) =− ( ) ; ( ) = ( ) . (8.6.1)
∂V ∂V ∂T γ −1 ∂T ∂T γ −1 ∂T
S T S P S V

For example: - isothermal: PV = constant. Take logarithms and differentiate: dP

P
+
dV

V
=0 . Hence ( ∂P

∂V
) =−
P

V
. adiabatic:
T

PVγ = constant. Take logarithms and differentiate: dP

P
+γ
dV

V
=0 . Hence (
∂P

∂V
) = −γ
P

V
. The other two relations can be
S

obtained in a similar manner.

Do these relations hold in general for any equation of state, or are they valid only for an ideal gas? In this section, we shall see
that they are valid in general for any equation of state, and are not restricted to the equation of state for an ideal gas.
Let us imagine that the state of the working substance (be it gas, liquid or solid) starts in PVT space at point A (P, V, TA). We
are going to take it to a new point B (P + δP, V + δV, TB). As I have drawn it in Figure VIII.3, δP is positive, δV is negative,
and TB > TA.

We first suppose that we make this move by a single, adiabatic process. In that case no heat is added to or lost from the
system, and the increase in the internal energy is −PδV.
Alternatively, B can be reached in two stages:
An isochoric path from A to a new point C (P + δP, V, TC), followed by
An isobaric path from C to B.
As I have drawn it in Figure VIII.3, TC > TB > TA .
In the isochoric process, no work is done by or on the system, and the increase in the internal energy is equal to the heat
added to the system, CV (TC − TA).
In the isobaric process, the increase in the internal energy is equal to the work done on the system, −PδV, minus the heat lost
from the system, CP (TC − TB); that is, −CP (TC − TB) − PδV.
Therefore, since the total increase in internal energy is route-independent,
−P δV = CV (Tc − TA ) − CP (TC − TB ) − P δV . (8.6.2)

Jeremy Tatum 3/18/2021 8.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7261

Cancel PδV and write γ for CP/CV, so that
(TC − TA ) = γ (TC − TB ) . (8.6.3)

But TC = TA + (
∂T

∂P
) δP and T B = TC + (
∂T

∂V
) δV .
V P

[Reminder: Here δP means PC − PA (which, in the way in which I have drawn it in figure VIII.3, is positive) and δV means VB
− VC (which, in the way in which I have drawn it in figure VIII.3, is negative).]
Therefore
∂T ∂T
( ) δP = −γ ( ) δV . (8.6.4)
∂P ∂V
V P

Divide both sides by δV and go to the infinitesimal limit, recalling that δP and δV are related through an adiabatic path:
∂T ∂P ∂T
( ) ( ) = −γ ( ) . (8.6.5)
∂P ∂V ∂V
V S P

Therefore
∂P ∂T ∂P
( ) = −γ ( ) ( ) . (8.6.6)
∂V ∂V ∂T
S P V

But ( ∂T

∂V
) (
∂P

∂T
) (
∂V

∂P
) = −1 , so ( ∂T

∂V
) (
∂P

∂T
) = −(
∂P

∂V
) .
P V T P V T

Therefore
∂P ∂P
( ) = γ( ) . (8.6.7)
∂V ∂V
S T

Thus, as for the ideal gas, the slope of the adiabat is γ times the slope of the isotherm, only this time we have made no
assumption about the equation of state.
The other two relations (equations 8.6.1 b,c) can be dealt with as follows.
Equation 8.6.3 can be rearranged to read
TB − TA = −(γ − 1) (TB − TC ) (8.6.8)

But TB = TA + (
∂T

∂V
) δV and T
B = TC + (
∂T

∂V
) δV .
S P

Hence
∂V 1 ∂V
( ) =− ( ) , (8.6.9)
∂T γ −1 ∂T
S P

which is the same as equation 8.6.1 b, but without any assumption about the equation of state.
Note also that
∂P ∂V ∂T
( ) ( ) ( ) = −1. (8.6.10)
∂V ∂T ∂P
T P V

Combine this with equations 8.6.7 and 8.6.9 to obtain

1 ∂P ∂V ∂T
( ) (γ − 1)( ) ( ) = 1. (8.6.11)
γ ∂V ∂T ∂P
S S V

Therefore
∂P γ −1 ∂P ∂V γ −1 ∂P
( ) = ( ) ( ) = ( ) . (8.6.12)
∂T γ ∂V ∂T γ ∂T
V S S S

Jeremy Tatum 3/18/2021 8.6.2 CC-BY-NC https://phys.libretexts.org/@go/page/7261

Therefore
∂P γ ∂P
( ) = ( ) , (8.6.13)
∂T γ −1 ∂T
S V

which is the same as equation 8.6.1 c, but without any assumption about the equation of state.

Jeremy Tatum 3/18/2021 8.6.3 CC-BY-NC https://phys.libretexts.org/@go/page/7261

8.7: Scale Height in an Isothermal Atmosphere
The material in this chapter doubtless has countless applications, most of which I am unaware of, in meteorology. Two simple
topics are easy to mention, namely the scale height in an isothermal atmosphere, dealt with in this section, and the adiabatic
lapse rate dealt with in the next section.
Let us imagine a column of air of cross-sectional area A in an isothermal atmosphere – that is to say the temperature T is
uniform throughout. Consider the equilibrium of the portion of the air between heights z and z + dz. The weight of this portion
is ρgAdz. Let P be the pressure at height z and P + dP be the pressure at height z + dz. (Note that dP is negative.) The net
upward force on the portion dz of the air is −AdP. Therefore dP = − ρgdz. But if we regard air as an ideal gas, it obeys the
equation of state for an ideal gas, equation 6.1.7: P = ρRT/µ where ρ and µ are respectively the density and the “molecular
dρ μg
weight” (molar mass) of the gas. Therefore RT

μ
dρ = −ρgdz , or ρ
=−
RT
dz . Integrate to obtain
−z/H
ρ = ρe (8.7.1)

where H = RT

μg
is the scale height. It is large if the temperature is high, the gas light and the planet’s gravity feeble. It is the
height at which the density is reduced to a fraction 1/e, or 36.8%. of its ground value. What would it be, in kilometres, for an
atmosphere consisting of 80% N2 and 20% O2, at a temperature of 20 ºC, where the gravitational acceleration is 9.8 m s−2?
What fraction is this of the radius of Earth? If you made a model of Earth one metre in diameter (radius = 50 cm), how thick
would be the atmosphere? You’d better look after it - our atmosphere is a very thin skin clinging to the surface!

Jeremy Tatum 3/11/2021 8.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8602

8.8: Adiabatic Lapse Rate
Earth’s atmosphere is not, of course, isothermal. The temperature decreases with height. The temperature lapse rate in an
atmosphere is the rate of decrease of temperature with height; that is to say, it is −dT/dz.
An adiabatic atmosphere is one in which P/ργ does not vary with height. In such an atmosphere, if a lump of air is moved
adiabatically to a higher level, its pressure and density will change so that P/ργ is constant – and will be equal to the ambient
pressure and density at the new height. For such an atmosphere, it is possible to calculate the rate at which temperature
decreases with height – the adiabatic lapse rate. We shall do this calculation, and see how it compares with actual lapse rates.
As in Section 8.7, the condition for hydrostatic equilibrium is
dP = −ρgdz. (8.8.1)

γ
Since we are trying to find a relation between T and z for an adiabatic atmosphere (i.e. one in which P/ρ doesn’t vary with
height), we need to find the adiabatic relations between P and T and between ρ and T.
These are easily found from the adiabatic relation between P and ρ:
γ
P = cρ (8.8.2)

and the ideal gas equation of state:

ρRT
P = . (8.8.3)
μ

Eliminate P:
1/(γ−1)
RT
ρ =( ) . (8.8.4)
cμ

Eliminate ρ:
γ/(γ−1)
R
γ/(γ−1)
P = T , (8.8.5)
γ/(γ−1) 1/(γ−1)
μ c

from which
γ/(γ−1)
γ R
1/(γ−1)
dP = T dT . (8.8.6)
γ − 1 μγ/(γ−1) c1/(γ−1)

Substitute equations (8.8.4) and (8.8.6) into equation (8.8.1), to obtain, after a little algebra, the following equation for the
adiabatic lapse rate:
dT 1 gμ
− = (1 − ) . (8.8.7)
dz γ R

This is independent of temperature.

If you take the mean molar mass for air to be 28.8 kg kmole−1, and g to be 9.8 m s−2 for temperate latitudes, you get for the
adiabatic lapse rate for dry air −9.7 K km−1. The presence of water vapour in humid air reduces the mean value of µ (and
hence the adiabatic lapse rate), and actual lapse rates are usually rather less than the calculated adiabatic lapse rates even for
humid air. (The presence of water vapour also increases slightly the value of γ. This would result in a slightly larger lapse rate,
but the effect is not as great as the reduction in lapse rate caused by the larger value of µ. Try some numbers to convince
yourself of this.) The International Civil Aviation Organization Standard Atmosphere takes the lapse rate in the troposphere
(first 11 km) to be −6.3 K km−1. What happens if the actual lapse rate is faster than the adiabatic lapse rate? If you imagine a
lump of air to be moved adiabatically to a higher level, its pressure and density will change so that P/ργ is constant, and it will
then find itself in a region where its new density is less that the new ambient density. Consequently, it will continue to rise, and
the atmosphere will be convectively unstable, and a storm will ensue. The atmosphere is stable as long as the actual lapse rate

Jeremy Tatum 3/25/2021 8.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8603

is less than the adiabatic lapse rate (which is reduced in humid air) is unstable if the actual lapse rate is greater than the
adiabatic lapse rate.

Jeremy Tatum 3/25/2021 8.8.2 CC-BY-NC https://phys.libretexts.org/@go/page/8603

8.9: Numerical Values of Specific and Molar Heat Capacities
The following table is not intended as a definitive, authoritative table of precise heat capacities. It is intended just to give a
rough idea of the orders of magnitude and the relative magnitudes for a few substances.
For gases, the heat capacities tabulated are at constant pressure. For solids and liquids the difference between Cp and Cv is
much smaller than for gases, because of the much smaller coefficient of expansion. Notice that the molar heat capacities for
gases, when expressed in terms of R, are about what are expected from the theoretical considerations in this chapter. Notice the
relatively large molar heat capacities of organic liquids (the molecules can rotate and can vibrate in many modes), and that, the
more complex the molecule, the larger its molar heat capacity. Notice, however, that, because water has a low molecular
weight (molar mass), water has the largest specific heat capacity of any common liquid or solid. (The specific heat capacities
of gaseous H2 and He are, unsurprisingly, larger still. A kilogram of hydrogen is an enormous number of molecules, so it takes
a lot of heat to warm them all up.) We have not studied the theory of the heat capacities of solids in this chapter, but, when you
do so in a course on solid state physics or on statistical mechanics, you will understand that the expected molar heat capacity
of metals would be about 3R, which is approximately what is shown for the three metals in this table.

Specific Heat Capacity at Constant Pressure Molar Heat Capacity at Constant Pressure

cal g−1 Cº −1 J kg−1 K−1 J kmole−1 K−1 In units of R

Helium (g) 1.25 5250 21000 2.53 R

Argon (g) 0.13 526 21000 2.53 R

H2 (g) 3.44 14400 28800 3.46 R

O2 (g) 0.22 919 29400 3.54 R

N2 (g) 0.25 1040 29100 3.50 R

CO2 (g) 0.20 843 37100 4.46R

H2O (g) 1 4184 75300 9.1 R

C2H5OH (l) 0.58 2430 112000 13.5 R

CCl4 (l) 0.20 852 131000 15.8 R

C6H6 (l) 0.42 1740 136000 16.4 R

Al (s) 0.22 941 25400 3.1 R

Cu (s) 0.092 384 24400 2.9 R

Fe (s) 0.11 450 25100 3.0 R

Jeremy Tatum 4/1/2021 8.9.1 CC-BY-NC https://phys.libretexts.org/@go/page/8604

8.10: Heat Capacities of Solids
I do not deal a great deal with solid state physics in these notes, particularly in this chapter, which has been concerned mostly
with gases. But the inclusion of the heat capacities of three metals in the above table provides an opportunity for a brief
mention of the heat capacities of metals and of other crystalline solids. In a simple model of a crystalline solid, the solid can be
thought of as a regular lattice of atoms held in position near their neighbors by springs, and the atoms have three degrees of
vibrational freedom – in the x, y and z directions. For each of these vibrational modes there are two squared terms (of the form
m v and I ω ) that contribute to the internal energy. The internal energy associated with each of these six terms is
1 2 1 2 1
RT
2 2 2

per mole, which comes to 3RT per mole, and thus you would expect the molar heat capacity to be about 3R – and you can see
from the above table that this is indeed the case. Indeed at room temperature, most metals and simple crystalline solids have a
molar heat capacity of about 3R. (This is sometimes referred to as “Dulong and Petit’s Rule”.) At low temperatures, however,
the molar heat falls below this value, and eventually approaches zero at 0 K. At very low temperatures, the molar heat capacity
varies roughly as the cube of the temperature. As room temperatures are reached, the molar heat capacity asymptotically
approaches the “classical” value of 3R.
The run of molar heat capacity with temperature at low temperatures looks a little like figure VIII.5 for magnesium and figure
VIII.6 for silver bromide. It will be seen that these two curves are the same shape except for a different scale along the
temperature axis – and the same is true for most metals and simple crystalline solids. Indeed we can assign to each solid a
characteristic temperature, known as the Debye temperature, θ , and then, if we express temperature not in kelvin but in
D

units of the Debye temperature for the particular solid, then the curves are indeed the same shape. In other words, the molar
heat capacity of all solids (or at least all solids that behave like this!) is the same function of T/θD. I show this function as
figure VIII.7.
The theory of the heat capacities of solids was investigated by Einstein and by Debye. (Peter Debye – Dutch-American
physicist/chemist. Nobel Chemistry prize 1936.) The Debye temperature is related to the vibrational frequency of the atoms in
their crystalline lattice. Diamond is a very hard substance, with very strong interatomic bonds. Consequently the vibrational
frequencies are very high, and the Debye temperature for diamond is correspondingly high: θD = 1860 K. As a result of this
the heat capacity rises very slowly with increasing temperature, and at room temperature is well below the “classical” value of
3R. Most other solids have weaker bonds and far lower Debye temperatures, and consequently their molar heat capacities have
almost reached the classical Dulong-Petit value of 3R at room temperature. Here are a few Debye temperatures:
Elements Debye temperature

Potassium 100K

Silver bromide 145

Silver 215
Magnesium 290
Copper 315
Iron 420

If it seems that the harder the solid the higher the Debye temperature and the slower the solid is to reach its classical CV of 3R,
this is not coincidence.
I do not derive Debye’s theoretical formula here – it is something to look forward to in courses on solid state physics or
statistical mechanics, but, for interest, the formula (which I used for calculating figures VIII.5-7) is
1/T 4 x
3
x e
CV = 9 T ∫ dx. (8.10.1)
x 2
0 (e − 1)

In this equation CV is in units of R, and T is in units of the Debye temperature.

Jeremy Tatum 3/25/2021 8.10.1 CC-BY-NC https://phys.libretexts.org/@go/page/8605

In case you are wondering what the symbol “x” stands for in equation 8.9.1, it is merely a dummy variable, for the integral in
that expression is a function not of x but of T, the upper limit of the integral.
If you try to reproduce figure VIII.7 yourself by evaluating equation 8.9.1 for a number of different temperatures, you will
soon find that it is a good deal more laborious than may at first be evident.
In my first attempt at doing it, for each of the 400 values of T that I used for plotting Figure VIII.7, I used a 1000 point
Simpson’s Rule integration. Thus I evaluated the integrand 400,000 times, and it took the computer almost half a second.
Later, I found that Gaussian quadrature was much, much more efficient, requiring the calculation of the integrand at only a
very few points.

Jeremy Tatum 3/25/2021 8.10.2 CC-BY-NC https://phys.libretexts.org/@go/page/8605

However, J. Viswanathan of Chennai, India, has since shown me an even better method than the Gaussian quadrature.
He uses the theorem
g(x)
d ′
∫ f (y)dy = f (g(x))g (x)
dx 0

This was a new one on me, but it is very easy to derive and looks almost obvious in hindsight. Applied to our problem, that is,
applied to our equation
1/T 4 x
3
x e
CV = 9 T ∫ dx, (8.10.2)
x 2
0 (e − 1)

it becomes, after a modest amount of work:

dCV 3CV
= − 9T f (1/T ), (8.10.3)
dT T

where
4 x
x e
f (x) = . (8.10.4)
2
(ex − 1)

He evaluates CV at T = 2, using a direct numerical integration of equation 8.9.1 - but this is the only time that he does this! The
answer is 2.9628. Then he moves down by dT at each step and calculates the corresponding dCV by using a fourth order
Runge-Kutta integration on the differential equation 8.9.3. The three methods agree very well, but the Simpson’s Rule method
was by far the most laborious.
Debye’s theory was published in 1912, and they certainly didn’t have electronic computers, or even electronic hand
calculators, in those days. In the 1950s most scientists were using hand-operated mechanical calculators, with electrically-
driven mechanical calculators beginning to come into use towards the end of that decade. I suspect that in 1912 not even hand-
operated mechanical calculators were available, and calculations would have been done using pencil and paper and logarithm
and other tables. One must think of the physical insight and mathematical competence needed to develop the theory of the heat
capacity in the first place, and then the enormous effort needed to calculate the resulting equations.

Jeremy Tatum 3/25/2021 8.10.3 CC-BY-NC https://phys.libretexts.org/@go/page/8605

CHAPTER OVERVIEW
9: ENTHALPY

9.1: ENTHALPY
9.2: CHANGE OF STATE
9.3: LATENT HEAT AND ENTHALPY

1 4/29/2021
9.1: Enthalpy
Enthalpy is sometimes known as "heat content", but "enthalpy" is an interesting and unusual word, so most people like to use
it. Etymologically, the word "entropy" is derived from the Greek, meaning "turning" (I'm not sure why) and "enthalpy" is
derived from the Greek meaning "warming". As for pronunciation, ENtropy is usually stressed on its first syllable, while
enTHALpy is usually stressed on the second. Again, I am not sure why.
Definition: Enthalpy H is defined as

H = U +PV . (9.1.1)

You now know the etymology of enthalpy, you know how to spell it, you know its pronunciation, and you even know its
definition. But you don't yet know what it means. You cannot determine the internal energy of a system to start with (you can
only determine an increase in it), but what on Earth does it mean to add to the (undetermined) internal energy the product of
the pressure and the volume?
Well, let us see how the enthalpy changes if we change the pressure and volume (and hence the internal energy) of a system.
We'll just differentiate Equation 9.1.1.
dH = dU + P dV + V dP (9.1.2)

But dU = T dS − P dV , and so the first law becomes

dH = T dS + V dP (9.1.3)

This helps us to see a little more the meaning of enthalpy. In particular, for a reversible process, T dS = dQ , and so Equations
7.3.2 and 9.1.3 become, respectively,

dU = dQ − P dV (9.1.4)

and

dH = dQ + V dP (9.1.5)

Thus we can say:

The increase of the internal energy of a system is equal to the heat added to it in an
isochoric process,
and

The increase of the enthalpy of a system is equal to the heat added to it in an isobaric
process.
Experiments carried out in open beakers on a laboratory bench are isobaric. Thus the heat generated during a chemical reaction
in an open beaker represents the generation of enthalpy. You will notice that chemists use the symbol H for heat of reaction,
and they are well aware that this means enthalpy. If the reaction were carried out, however, in an autoclave (also known as a
pressure cooker), the heat generated represents the generation of internal energy.
I hope that this now gives some meaning to the concept of enthalpy.
Internal energy U and enthalpy H are both functions of state. From Equation 7.3.2 ( dU = T dS − P dV ) we immediately see
the relations
∂U
( ) =T (9.1.6)
∂S
V

and
∂U
( ) = −P . (9.1.7)
∂V
s

Jeremy Tatum 3/25/2021 9.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7263

From Equation 9.1.3 ( dH = T dS + V dP ) we immediately see the relations
∂H
( ) =T (9.1.8)
∂S
P

and
∂H
( ) =V . (9.1.9)
∂P
S

Also from Equation 7.3.2 (dU = T dS − P dV ) we obtain (since dU is an exact differential)

∂T ∂P
( ) = −( ) , (9.1.10)
∂V ∂S
s V

and from Equation 9.1.3 ( dH = T dS + V dP ) we obtain (since dH is an exact differential)

∂T ∂V
( ) =( ) . (9.1.11)
∂P ∂S
S P

Equations 9.1.10 and 9.1.11 are two of Maxwell's Thermodynamic Relations. (There are two more to come, in a later chapter.)
We also note that, while the heat capacity at constant volume is
∂U
CV = ( ) , (9.1.12)
∂T
V

similarly the heat capacity at constant pressure is

∂H
CP = ( ) . (9.1.13)
∂T
P

Jeremy Tatum 3/25/2021 9.1.2 CC-BY-NC https://phys.libretexts.org/@go/page/7263

9.2: Change of State
According to my dictionary, the word "latent" means "present or existing and capable of development but not manifest".
In a liquid at its freezing point there is present or existing some heat, which is capable of development but is not manifest. That
is, the liquid secretly holds some latent heat. When the liquid freezes, it gives up this latent heat to its surroundings. The heat is
now manifest.
Definition: The latent heat of freezing of a quantity of liquid at its freezing point is the heat given up to its surroundings when
it freezes. Its SI unit is the joule.
Likewise, we define the specific latent heat and the molar latent heat of a liquid at its freezing point as the heat given up when
unit mass, or a molar amount, respectively, freezes. The SI units are J kg−1 and J kilomole−1 respectively.

A distressingly large number of people use the words "latent heat" when they mean "specific latent heat". Thus, when you
read or hear the words "latent heat" you have to be on guard to decide whether this is really what is meant, or whether
"specific latent heat" is intended.

The latent heat of fusion of a solid body at its melting point is the heat required to melt it. This is just equal to the heat given
up when the liquid freezes, so that, numerically, the latent heats of freezing and of fusion (melting) are the same – though
somehow the word "latent" seems less appropriate for freezing, because you are supplying heat to the solid, rather than seeing
latent heat being released by a liquid. If you prefer you could refer to the "latent heat" of fusion simply as the "heat of fusion"
– or as the “enthalpy of fusion”.
Likewise we have a latent heat of condensation of a vapour at its condensation point, and the latent heat of vaporization of a
liquid at its boiling point. These are equal in magnitude. We can also define the specific and molar latent heats of condensation
and vaporization. The term latent heat of transformation will do to cover all four processes. The symbol L (with appropriate
subscripts if need be) can be used for any of the latent heats of transformation.
The specific latent heat of fusion of ice at atmospheric pressure is about 3.36 × 105 J kg−1 or about 80 cal g−1.
The specific latent heat of vaporization of water at atmospheric pressure is about 2.27 × 106 J kg−1 or about 540 cal g−1.

Example 9.2.1
70 g of ice at 0 oC are mixed with 150 g of water at 100 oC. What is the final temperature? (I make it 43º C.)
Solution
We'll reluctantly, for once, work in calories and grams, and of course the specific heat capacity of water is about 1 calorie
per gram per Celsius degree. The heat required to melt the 70 g of ice, and then to raise its temperature from 0 oC to t oC
is 70 × 80 + 70t calories. This heat is supplied by the hot water, which cools from 100 oC to t oC, is 150 % (100 − t)
calories. Equating the two produces t = 43 oC.

Example 9.2.1
Suppose you apply 2.27 × 106 J of energy to a kilogram of water, but, instead of using that energy to vaporize the water,
you use it to raise the water from the ground. How high above the ground could you raise it with this energy? It may
surprise you – it certainly surprised me! If you were to use the energy, not to vaporize the water, and not to raise it above
the ground, but to throw it, how fast, in miles per hour, could you throw it?

For many liquids there is a very rough correlation between molar latent heat of vaporization and boiling point at atmospheric
pressure, the ratio L/T usually being in the range 70,000 to 100,000 J kmole−1 K−1.
One last point before proceeding. Generally it is only crystalline solids (including metals) that have a rather definite melting
point. Amorphous substances such as plastics and glass generally change from solid to liquid over a rather large range of

Jeremy Tatum 4/1/2021 9.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7264

temperature. Indeed is not obvious when to cease calling such a substance a solid and to start calling it a liquid. Some writers
would describe glass as a “liquid” even when it has all the obvious appearances of a solid. See also Section 6.4 for a further
discussion of this. Mixtures, alloys and solutions, too, do not have such a definite melting point as a crystalline solid, and a salt
solution does not have as definite a boiling point (at a given pressure) as a pure liquid does. Thus a salt solution in water at one
atmosphere pressure boils at a little higher temperature than 100 °C. When some of the water boils off, the remaining solution
is a little more concentrated, and so the boiling point becomes a little higher, and so on.

Jeremy Tatum 4/1/2021 9.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7264

9.3: Latent Heat and Enthalpy
Consider a liquid of volume V1 at its boiling point. Suppose a quantity of heat L is supplied, sufficient to vaporize the liquid.
The new volume (of what is now vapour) is V2. If the vapour has expanded against a constant pressure P (e.g. the pressure of
the atmosphere), the work done by it is P(V2 − V1). The increase in the internal energy of the system is the heat supplied to the
system minus the work done by it (this is the engineer's version of the first law of thermodynamics). That is, U2 − U1 = L −
P(V2 − V1), and so
H2 − H1 = L. (9.3.1)

So, during a change of state at constant pressure the increase or decrease of enthalpy is equal to the latent heat of
transformation. This, of course, is just a simple example of our earlier statement, in Section 9.1, that the increase of enthalpy
of a system is equal to the heat supplied to it in an isobaric process.

Jeremy Tatum 3/4/2021 9.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7265

CHAPTER OVERVIEW
10: THE JOULE AND JOULE-THOMSON EXPERIMENTS

10.1: INTRODUCTION
10.2: THE JOULE EXPERIMENT
10.3: THE JOULE-THOMSON EXPERIMENT
10.4: CP MINUS CV
10.5: BLACKBODY RADIATION

1 4/29/2021
10.1: Introduction
Equation 8.4.3, TVγ−1 = constant , tells us how to calculate the drop in temperature if a gas expands adiabatically and
reversibly; it is expanding against an external pressure (e.g., a piston), and, in pushing the piston back, the molecules are doing
external work and are losing kinetic energy. What happens, however, if a gas expands into a vacuum? Suppose that the gas is
held inside a cylinder not by a metal piston but by a thin membrane, and the membrane breaks, so that the molecules rush out
into empty space. This is obviously an irreversible expansion; it is most unlikely that all of the molecules will ever find their
way back to the cylinder. The molecules are doing no external work. If the gas is an ideal gas, there are no intermolecular
forces, so the gas does no internal work. There is nothing to slow down the molecules in their headlong escape from the
cylinder. The temperature will remain unaltered by the expansion. On the other hand, if the gas is not an ideal gas, there will
be van der Waals attractive forces between the molecules, so the molecules will slow down slightly when the gas expands and
there will be a small drop in temperature. But we also recall, from the van der Waals model, that at close intermolecular
distances, the forces between the molecules are predominantly repulsive Coulomb forces, so it is also possible that, if the gas
starts out very dense and it expands irreversibly as we have described, it may initially become slightly warmer as the repulsive
Coulomb forces push the molecules apart and speed them on their way.
The Joule and Joule-Thomson experiments are concerned with these scenari.

Jeremy Tatum 4/1/2021 10.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7270

10.2: The Joule Experiment
In Joule's original experiment, there was a cylinder filled with gas at high pressure connected via a stopcock to a second
cylinder with gas at a low pressure – sufficiently low that, for the purpose of understanding the experiment, we shall assume
the second cylinder to be entirely empty. The two cylinders were immersed in a water bath, and the stopcock was opened so
that gas from the high pressure cylinder flowed into the evacuated cylinder. No heat was supplied to or lost from the system,
nor did the gas do any work, so the internal energy was constant during the expansion. Joule found no temperature fall as a
result of the expansion. This, as we have argued in Section 10.1, is exactly what we would expect for an ideal gas; that is, for
an ideal gas, the temperature is independent of the volume if the internal energy is constant. That is, for an ideal gas,
∂T
( ) = 0. (10.2.1)
∂V
U

Ideal Gas

For a real gas, however, we would expect a small drop in temperature, and (
∂T

∂V
) , which is called the Joule coefficient, is
U

not zero. The heat capacity of the water bath and the cylinders in Joule's original experiment, however, was too large for him
to detect any fall of temperature even with a real gas. More sensitive experiments found that almost all gases cool during a
Joule expansion at all temperatures investigated; the exceptions are helium, at temperatures above about 40 K, and hydrogen,
at temperatures above about 200 K.
∂T
( ) ≠ 0. (10.2.2)
∂V
U

Real Gas

We should be able to derive an expression for the Joule coefficient, given the equation of state, and we should also be able to
show that, if the equation of state is the equation of state for an ideal gas, the Joule coefficient is zero.
Internal energy and enthalpy are both functions of state; that is, they are functions of P, V and T. However, any particular
substance cannot exist at any arbitrary point in PVT-space, but is constrained to be on the two-dimensional surface represented
by its equation of state. Figures VI.7, 8 and 9 of Chapter 6 represent an example of such a surface. In other words, P, V and T
cannot be varied independently; they are connected by an equation of the form f (P , V , T ) = 0 . Thus internal energy and
enthalpy can be described by a function of just two of the state variables P, V and T. In the experiment we are discussing, we
are interested in how temperature varies with volume in an experiment in which the internal energy is constant. We shall
therefore choose U as our state function and V and T as our independent state variables. That is, we shall write U = U (V , T ) ,
so that
∂T ∂U ∂V
( ) ( ) ( ) = −1. (10.2.3)
∂V ∂T ∂U
U V T

Our aim, of course, is to find an expression for the Joule coefficient ( ∂T

∂V
) , for which I shall be using the symbol η.
U

The second of these partial derivatives is C , and therefore

V

∂T 1 ∂U
( ) =− ( ) . (10.2.4)
∂V CV ∂V
U T

Now
dU = T dS − P dV . (10.2.5)

That is,

Jeremy Tatum 4/22/2021 10.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7271

 1
dS = [dU + P dV ] (10.2.6)
T

1 ∂U ∂U
= [( ) dV + ( ) dT + P dV ] . (10.2.7)
T ∂V ∂T
T V

1 ∂U 1 ∂U
= [( ) + P ] dV + ( ) dT . (10.2.8)
T ∂V T ∂T
T V

But we also have

∂S ∂S
dS = ( ) dV + ( ) dT . (10.2.9)
∂V T
∂T V

Therefore
∂S 1 ∂U
( ) = [( ) +P] (10.2.10)
∂V T ∂V
T T

and
∂S 1 ∂U
( ) = ( ) . (10.2.11)
∂T T ∂T
V V

The mixed second derivatives are

2 2
∂ S 1 ∂U 1 ∂ U ∂P
=− [( ) +P]+ [ +( ) ] (10.2.12)
2
∂T ∂V T ∂V T ∂T ∂V ∂T
T V

and
2 2
∂ S 1 ∂ U
= . (10.2.13)
∂V ∂T T ∂V ∂T

But entropy is a function of state and dS is an exact differential, so the mixed second derivatives are equal. Whence, after
simplification:
∂U ∂P
( ) = T( ) −P. (10.2.14)
∂V ∂T
T V

Hence, returning to Equation 10.2.4, we obtain, for the Joule coefficient,

∂T
η =( ) (10.2.15)
∂V
U

1 ∂P
= [P − T ( ) ]. (10.2.16)
CV ∂T
V

Exercise 10.2.1
Show that, for an ideal gas, the Joule coefficient is zero.

Example 10.2.1
Show that, for a van der Waals gas, the Joule coefficient is
∂T a
( ) =− .
2
∂V CV V
U

Hence, for a finite volume change,

Jeremy Tatum 4/22/2021 10.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7271

 a 1 1
T2 = T1 − ( − ).
CV V1 V2

Solution
For example, the volume of a kmole of CO2 at a temperature of 20 °C (293.15 K) and a pressure of 1 atm (1.013 × 105
Pa) is V1 = RT/P = 24.06 m3. (That’s a lot of cubic metres – but then 44 kg of CO2 is a lot of carbon dioxide.). If its
volume were doubled to 48.12 m3 in an irreversible Joule-type expansion, what would be its new temperature? From
Chapter 6, we find a = 3.7 × 105 Pa m6 kmole−1, and from Chapter 8 we find that CP = 37100 J kmole−1 K−1 and
therefore let’s take CV = 28786 J kmole−1 K−1, and so we obtain T2 = 292.88 K = 19.73 °C. This cooling is a result not of
the gas doing external work as in a reversible adiabatic expansion, but of doing work against the internal van der Waals
forces between the molecules. What would be the temperature drop in a reversible adiabatic expansion? The new
γ−1

temperature would be given by T 2 =(

V1

V2
) T1 . Let’s take γ = 37100 ÷ 28786 = 1.29. Then T2 = (1/2)0.29 × 293.15 =
239.77 K = −33.38 °C at which temperature it would easily have sublimated into solid CO2. In this calculation, I used CP
− CV = R and TVγ−1 = constant, which are valid only for an ideal gas. We’ll shortly derive a more general expression for
CP − CV, but the correction for nonideality will obviously be quite small.

Jeremy Tatum 4/22/2021 10.2.3 CC-BY-NC https://phys.libretexts.org/@go/page/7271

10.3: The Joule-Thomson Experiment
The experiment is also known as the Joule-Kelvin experiment. William Thomson was created Lord Kelvin. The experiment is
also known as the porous plug experiment.
In the Joule-Thomson experiment a constant flow of gas was maintained along a tube which was divided into two
compartments separated by a porous plug, such that the pressure and molar volume on the upstream side were P1, V1, and the
pressure and molar volume on the downstream side were P2, V2. Under such circumstances the net work done on a mole of
gas in passing from one compartment to the other is P1V1 − P2V2. (Imagine, for example, that a piston pushes a mole of gas
towards the plug from the upstream side, through a distance x1 ; if A is the crosssectional area of the tube, the work done on
the gas is P1Ax1 = P1V1. Imagine also that the gas on the downstream side pushes a piston away from the plug through a
distance x2. The work done by the gas is P2Ax2 = P2V2. Therefore the net external work done on the gas is P1V1 − P2V2.) If no
heat is supplied to or lost from the system, the increase in internal energy of this gas is just equal to this work done on it:
U2 − U1 = P1 V1 − P2 V2 ,

or
U1 + P1 V1 = U2 + P2 V2 . (10.3.1)

That is, there is no change in enthalpy. Therefore, we want to find ( ∂T

∂P
) , which is the Joule-Thomson coefficient, for which
H

I shall be using the symbol µ.

In the experiment we are discussing, we are interested in how temperature varies with pressure in an experiment in which the
enthalpy is constant. We shall therefore choose H as our state function and P and T as our independent state variables. That is
we shall write H = H(P,T), so that
∂T ∂H ∂P
( ) ( ) ( ) = −1. (10.3.2)
∂P ∂T ∂H
H P T

The second of these partial derivatives is CP, and therefore

∂T 1 ∂H
( ) =− ( ) . (10.3.3)
∂P CP ∂P
H T

Now
dH = T dS + V dP . (10.3.4)

That is,
1 1 ∂H ∂H
dS = [dH − V dP ] = [( ) dP + ( ) dT − V dP ] . (10.3.5)
T T ∂P ∂T
T P

1 ∂H 1 ∂H
dS = [( ) − V ] dP + ( ) dT . (10.3.6)
T ∂P T ∂T
T P

But we also have

∂S ∂S
dS = ( ) dP + ( ) dT . (10.3.7)
∂P ∂T
T P

Therefore
∂S 1 ∂H
( ) = [( ) −V ] (10.3.8)
∂P T ∂P
T T

and

Jeremy Tatum 4/3/2021 10.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7272

 ∂S 1 ∂H
( ) = ( ) . (10.3.9)
∂T T ∂T
P p

The mixed second derivatives are

2 2
∂ S 1 ∂H 1 ∂ H ∂V
=− [( ) −V ]+ [ −( ) ] (10.3.10)
2
∂T ∂P T ∂P T ∂T ∂P ∂T
T P

and
2 2
∂ S 1 ∂ H
= . (10.3.11)
∂P ∂T T ∂P ∂T

But entropy is a function of state and dS is an exact differential, so the mixed second derivatives are equal. Whence, after
simplification:
∂H ∂V
( ) = V −T( ) . (10.3.12)
∂P T
∂T P

Hence, returning to equation 10.3.3, we obtain, for the Joule-Thomson coefficient,

∂T 1 ∂V
μ =( ) = [T ( ) −V ]. (10.3.13)
∂P Cp ∂T
H P

Trivial Exercise: Show that, for an ideal gas, the Joule-Thomson coefficient is zero, and also that, for an ideal gas,
∂H
( ) = 0. (10.3.14)
∂P
T

This is analogous to equation 8.1.4 for an ideal gas, namely ( ∂U

∂V
) =0 .
T

Exercise. Show that, for a van der Waals gas, the Joule-Thomson coefficient is
2 2
∂T V (RT V b − 2a(V − b ) )
( ) =− ⋅ . (10.3.15)
3 2
∂P Cp RT V − 2a(V − b )
H

(Verify the dimensions of this expression.) Hint: It is difficult to calculate (∂V/∂T)P directly, because it is difficult to express V
explicitly as a function of P and T. It is not actually impossible to do it algebraically, because van der Waals' equation is a
cubic equation in V, and a cubic equation does have an algebraic solution. It is easier, however, to calculate (∂V/∂T)P from
(
∂V

∂T
) = −(
∂P

∂T
) /(
∂P

∂V
) , or from ( ∂V

∂T
) = 1/ (
∂T

∂V
) .
P V T P P

Note also that the Joule-Thomson coefficient may be negative or positive; i.e., it may result in cooling or heating. It will result
in heating if you start above a certain temperature called the inversion temperature, and cooling if you start below the
inversion temperature. The Joule-Thomson effect is used in the Linde method for cooling and ultimately liquefying gases. For
most gases, the inversion temperature is higher than room temperature, so that cooling starts immediately. But for hydrogen,
the inversion temperature is about −80 oC, and hydrogen must be cooled below this temperature before the Joule-Thomson
effect can be used to cool it further and to liquefy it. You can see from equation 10.3.14 that the inversion temperature for a
2
2a(V −b)
van der Waals gas is equal to 2
≈
2a

Rb
. Here V is the molar volume.
RV b

Summary:
Joule coefficient
∂T 1 ∂P
η =( ) = [P − T ( ) ] (10.3.16)
∂V CV ∂T
U V

Joule-Thomson coefficient
∂T 1 ∂V
μ =( ) = [T ( ) −V ]. (10.3.17)
∂P CP ∂T
H P

Jeremy Tatum 4/3/2021 10.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7272

10.4: CP Minus CV
In Section 8.1 we pointed out that the heat capacity at constant pressure must be greater than the heat capacity at constant
volume. We also showed that, for an ideal gas, CP = CV + R, where these refer to the molar heat capacities. We said that in
Chapter 10 we would try and develop a more general expression for CP − CV, which was applicable in general and not only for
an ideal gas. Some of the relations that we developed in Sections 10.2 and 10.3 give us the opportunity to try to do that now.
Let us consider an isobaric process and express the internal energy U as a function of V and T. (As we have pointed out, P, V
and T are not independent variables because they are connected through the equation of state, so we may choose any two of
them as independent variables.) Then, if the volume and temperature increase by infinitesimal amounts, the corresponding
increase in the internal energy is given by
∂U ∂U
dU = ( ) dV + ( ) dT . (10.4.1)
∂V T
∂T V

I.e.,
∂U
dU = ( ) dV + CV dT (10.4.2)
∂V
T

Consider how the first law:

dU = dQ + dW . (10.4.3)

In an isobaric process, dQ = C p dT , and in a reversible process, dW = −P dV .

Therefore
∂U
CP dT − P dV = ( ) dV + CV dT (10.4.4)
∂V
T

Divide by dT, recalling that we are considering an isobaric process:

∂V ∂U ∂V
CP − P ( ) =( ) ( ) + CV . (10.4.5)
∂T P
∂V T
∂T P

Hence
∂U ∂V
CP − CV = [P + ( ) ]( ) . (10.4.6)
∂V ∂T
τ P

This is a useful general expression, as long as we know or can determine (∂U/∂V)T. (Note that the extensive quantities can be
total, specific or molar.)
Let us just test this by applying it to an ideal gas to see if it produces the result that it ought to produce. For an ideal gas, the
internal energy at a given temperature is independent of the volume. This is because in an ideal gas there are no intermolecular
forces, so that, as the volume increases and the intermolecular distances increase, there is no change in potential energy; and, if
the temperature is constant, so is the kinetic energy. Thus, for an ideal gas, ( ∂U

∂V
) =0 . The volume of a mole of ideal gas is
T

V = RT/P, so that ( ∂V

∂T
) =
R

P
.
P

Therefore
R
CP − CV = P × = R, (10.4.7)
P

and all is well.

For any substance other than an ideal gas, we shall need to know (
∂U

∂V
) before we can make use of equation 10.4.6. But
T

equation 10.2.12, which we developed in Section 10.2 while analysing the Joule effect, enables us to do just this:

Jeremy Tatum 3/11/2021 10.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7273

 ∂U ∂P
( ) = T( ) −P. (10.4.8)
∂V ∂T
T V

On combining this with equation 10.4.6, we obtain

∂P ∂V
CP − CV = T ( ) ( ) (10.4.9)
∂T ∂T
V P

Depending on the equation of state, it may or may not be easy to evaluate these partial derivatives. For example, for the van
der Waals equation of state (which is a cubic equation in V), it is not easy to evaluate (∂V/∂T)P directly, but one can then make
use of (∂V/∂T)P = (∂V/∂T)P or of (
∂V

∂T
) (
∂T

∂P
) (
∂P

∂V
) = −1 in order to get CP − CV in terms of easily evaluable partial
P V T

derivatives. For example

2
∂P ∂T ∂P ∂P
CP − CV = T ( ) /( )  or  − T ( ) /( ) (10.4.10)
∂T ∂V P
∂T v
∂V T

or several other variants.

Any of equations 10.4.8 or 10.4.9 can be used to calculate CP − CV; it just depends on which of the derivatives, for a particular
equation of state, are easiest to calculate.
The reader will easily be able to show that, for a mole of an ideal gas, this becomes just CP − CV = R. A little more algebra
will be needed to show that, for a mole of a van der Waals gas,
2 2
R P + a/V (RT ) + aP
CP − CV = = R⋅ ≈ R⋅ . (10.4.11)
2 3 2 3 2
1 − 2a(V − b ) / (RT V ) P − a/ V + 2ab/ V (RT ) − aP

In the above analysis, we considered an isobaric process and we chose the internal energy as our function of state and we
started by calculating the increment in U corresponding to increments dV and dT in the volume and temperature. It is tempting
now to go through the same analysis, but this time to consider an isochoric process and to choose the enthalpy as our function
of state. We start by calculating the increment in H corresponding to increments dP and dT in the pressure and temperature:
∂H ∂H
dH = ( ) dP + ( ) dT . (10.4.12)
∂P T
∂T P

I.e.,
∂H
dH = ( ) dP + CP dT . (10.4.13)
∂P
T

Now
H = U +PV , ∴ dH = dU + P dV + V dP = dQ + dW + P dV + V dP . (10.4.14)

Provided that we include in dQ any irreversible work that is being done on the system (irreversible work has the same effect,
as we have seen, as adding heat), so that dW = − PdV, then
dH = V dP + dQ = V dP + CV dT . (10.4.15)

On comparison of equations 10.4.11 and 10.4.12 we obtain

∂H
V dP + CV dT = ( ) dP + CP dT . (10.4.16)
∂P
T

Divide by dT, recalling that we are considering an isochoric process. From this, we obtain an alternative expression for the
difference between the heat capacities:
∂H ∂P
CP − CV = [V − ( ) ]( ) . (10.4.17)
∂P ∂T
T V

Jeremy Tatum 3/11/2021 10.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7273

This is quite analogous to equation 10.4.6. It is left to the reader to show that, for an ideal gas, this reduces to CP−CV=R. This
will be easy if you recall equation 10.3.14, for an ideal gas: ( ∂H

∂P
) =0 .
T

For any substance other than an ideal gas, we shall need to know (
∂H

∂P
) before we can make use of equation 10.4.15. But
T

equation 10.3.12, which we developed in Section 10.3 while analysing the Joule-Thomson effect, enables us to do just this:
∂H ∂V
( ) = V −T( ) . (10.4.18)
∂P ∂T
T P

On combining this with equation 10.4.15, we obtain again equation 10.4.8. We obtained no new result for CP − CV (although
we did obtain the important result 10.4.16 for an ideal gas), but it is satisfying and instructive to have obtained the same result
via internal energy and via enthalpy.
After this, we can hardly resist the temptation to see what happens if we treat P and V as independent variables. Thus, if U =
U(P, V), then increases of dP and dV in the pressure and volume result in an increase dU of the internal energy given by
∂U ∂U
dU = ( ) dP + ( ) dV . (10.4.19)
∂P V
∂V P

But we already know (equation 10.4.1), by choosing the independent variables to be T and V, that
∂U ∂U
dU = ( ) dV + ( ) dT . (10.4.20)
∂V ∂T
T V

And from the equation of state T = T(P, V) , we derive that

∂T ∂T
dT = ( ) dP + ( ) dV . (10.4.21)
∂P ∂V
V P

By elimination of dT from equations 10.4.1 and 10.4.18 we obtain

∂U ∂T ∂U ∂T ∂U
dU = ( ) ( ) dP + ( ) ( ) dV + ( ) dV . (10.4.22)
∂T ∂P ∂T ∂V ∂V
V V V P T

On comparison of equations 10.4.17 and 10.4.19 we deduce the following relations, which are occasionally useful:
∂U ∂U ∂T
( ) =( ) ( ) , (10.4.23)
∂P ∂T ∂P
V V V

(which I hope we already knew!)

and
∂U ∂U ∂T ∂U
( ) =( ) ( ) +( ) . (10.4.24)
∂V P
∂T V
∂V P
∂V T

The first of these is, of course, trivial, and does not require this lengthy derivation. The second is a worthwhile relation, which
we may occasionally find useful.
Summary:
∂U ∂V ∂H ∂P
CP − CV = [P + ( ) ]( ) = [V − ( ) ]( ) (10.4.25)
∂V ∂T ∂P ∂T
T P T V

∂P ∂V
CP − CV = T ( ) ( )  and variants.  (10.4.26)
∂T ∂T
V P

Jeremy Tatum 3/11/2021 10.4.3 CC-BY-NC https://phys.libretexts.org/@go/page/7273

10.5: Blackbody Radiation
Before we forget all the equations in this chapter, let’s use equation 10.2.12 (which we have already used twice – once in the
derivation of the Joule-Thomson coefficient and once in the derivation of CP − CV) in a totally different application:
∂U ∂P
( ) = T( ) −P. (10.5.1)
∂V ∂T
T V

This is a very general thermodynamical relation, and is by no means restricted to Joule’s experiment. Let us apply it to
electromagnetic radiation (rather than molecules) in an enclosure.
You may already have studied the theory of radiation in a cavity and the closely-related theory of blackbody radiation. You
will know that classical electromagnetic theory failed to explain the observed characteristics of blackbody radiation, and that it
was not explained fully until the advent of quantum theory. In the middle of the nineteenth century Kirchhoff argued
theoretically that the energy density inside a cavity was independent of the nature of the walls of the cavity and depended only
on the temperature and the wavelength. Stefan had shown experimentally that the radiation density inside a cavity integrated
over all wavelengths was proportional to the fourth power of the temperature. Later on, Lummer and Pringsheim did some
detailed measurements which showed how the radiation density per until wavelength varied with wavelength and temperature.
It was shown by Rayleigh and Jeans that classical electromagnetic theory failed badly at short wavelengths to explain the
observed distribution of the cavity radiation with wavelength. In 1900 Planck, without quite knowing why, showed that, if he
regarded radiation as being made up of quanta of energy hν, the energy density per unit volume per unit wavelength interval
C1
would be expected to vary as uλ = 5 C /λT )
which agreed very well with the experimental data of Lummer and
λ (e 2 −1)

Pringsheim. You also may know that if you integrate this expression over all wavelengths (not particularly easy), you find that
4
∫ u dλ is proportional to T , thus also agreeing with the observations of Stefan.
λ

However, although quantum theory was necessary to explain the Lummer-Pringsheim measurements of how uλ varies with
temperature, Boltzmann used classical thermodynamical theory to explain Stefan’s T4 law almost immediately after Stefan had
announced his results, and long before the advent of quantum theory. The theory of radiation tells us that the radiation energy
per unit volume u depends only on the temperature (this is Kirchhoff’s radiation law) and that the radiation pressure P is
related to the energy per unit volume by P = u . The derivation of this is very similar to the expression that we derived for
1

the pressure of molecules in a gas. For this situation, equation 10.2.12 becomes
1 du u
u = T − , (10.5.2)
3 dT 3

or
du
4u = T . (10.5.3)
dT

Integration of this (do it!) shows that u ∝ T4, without any need for quantum theory.
This is often written as u ∝ aT4, but beware, here a is not what it generally known as “Stefan’s constant”. See Chapters 1 and 2
(especially Section 1.17) of my Stellar Atmospheres notes for more on this. Stefan’s Law generally refers to the exitance of a
black body surface, M = σT4 , whereas here we are referring to the energy density of radiation in a cavity. The relation between
a and Stefan’s constant σ is a = 4σ/c.
Now suppose that you had some radiation at temperature T in an enclosure (such as The Universe) of volume V. And suppose
that volume were to expand adiabatically, thus diluting the energy density. What would be the new temperature? In what
follows, V means the volume (not the “specific” or “molar” volume) of the enclosure. U is the internal energy of the radiation
inside it, and u is the radiation energy density, such that U = uV, and we shall be making use of P = u and of u = aT . 1

3
4

If the volume were to increase by dV at pressure P, the work done by the radiation would be P dV = udV , and, if we 1

assume that the expansion is adiabatic, this results (by the first law of thermodynamics) in a decrease of the internal energy.
We apply the first law: dU = −PdV. That is

Jeremy Tatum 4/1/2021 10.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7274

 1
d(uV ) = udV + V du = − udV . (10.5.4)
3

dV 3 du
=− . (10.5.5)
V 4 u

Therefore
−3/4 −4/3
V ∝u  or u ∝ V . (10.5.6)

But u ∝ T4 and hence

3
V T  is constant,  (10.5.7)

or the temperature is inversely proportional to the linear dimensions of the enclosure.

Jeremy Tatum 4/1/2021 10.5.2 CC-BY-NC https://phys.libretexts.org/@go/page/7274

CHAPTER OVERVIEW
11: HEAT ENGINES
Topic hierarchy

11.1: INTRODUCTION
11.2: THE CARNOT CYCLE
11.3: THE STIRLING CYCLE
11.4: THE OTTO CYCLE
11.5: THE DIESEL CYCLE
11.6: THE RANKINE CYCLE (STEAM ENGINE)
11.7: A USEFUL EXERCISE
11.8: HEAT ENGINES AND REFRIGERATORS
11.9: ENTROPY IS A FUNCTION OF STATE

1 4/29/2021
11.1: Introduction
In my rarefied, theoretical, academic and unpractical mind, a heat engine consists of a working substance obeying some
idealized equation of state such as that for an ideal gas, held inside a cylinder by a piston, and undergoing, in a closed cycle, a
series of highly idealized processes, such as reversible adiabatic expansions or isothermal compressions. At various stages of
the cycle, the system may be gaining heat from or losing heat to its surroundings; or we may be doing work on the system by
compressing it, or the system may be expanding and doing external work.
The efficiency η of a heat engine is defined as
 net external work done by the engine during a cycle 
η = (11.1.1)
 heat supplied to the engine during a cycle. 

By “net” external work, I mean the work done by the engine during that part of the cycle when it is doing work minus the
work done on the engine during that part of the cycle when work is being done on it. Notice that the word “net” does not
appear in the denominator, which refers only to the heat supplied to the engine during that part of the cycle when it is gaining
heat.
During the compression part of the cycle, the system gives out heat, and only the difference “heat in minus heat out” is
available to do the external work. Thus efficiency can also be calculated from
Qin − Qout
η = . (11.1.2)
Qin

although the definition of efficiency remains as equation 11.1.1.

No heat engine is 100% efficient, and we need to ask what is the most efficient heat engine possible, what are the factors that
limit its efficiency, and what is the greatest possible efficiency? Obviously things like friction in the moving parts of the engine
limit the efficiency, but in my academic mind the engine is built with frictionless bearings and all processes in the cycle of
compressions and expansions are reversible.
During a cycle, a heat engine moves in a clockwise closed path in the PV plane, and, if the processes are reversible, the area
enclosed by this clockwise path is the net external work done by the system. It also moves in a clockwise closed path in the TS
plane, and, if the processes are reversible, the area enclosed by this clockwise path is the net heat supplied to the system. The
two are equal, and when the system returns to its original state, there is no change in the internal energy. That is, internal
energy is a function of state.
Depending upon the nature of the various processes during the cycle, the cycle may carry various names, such as the Carnot,
Stirling, Otto, Diesel or Rankine cycles. Of these, the most important from the theoretical point of view is the Carnot cycle. I
do not know whether anyone has ever built a Carnot heat engine. I do know, however, that no one has ever built an engine
working between a hot heat source and a cold heat sink that is more efficient than a Carnot engine; for, for a given temperature
difference between source and sink, the Carnot engine is the most efficient conceivable. There is another important thing about
the Carnot cycle. In Chapter 3, we struggled to understand that most difficult of all the thermodynamic concepts, namely
temperature, and we wondered if we could define an absolute temperature scale that was independent of the properties of any
particular substance. Consideration of the Carnot cycle enables us to do just that.
Of real heat engines I know very little. I know that one pedal of my car makes the car go faster and the other makes it go
slower – but what is under the hood or bonnet is beyond my ken. Real heat engines may resemble some of the theoretical
engines of academia to a greater or lesser extent. Thus a motor car engine may resemble an Otto cycle, or a steam engine may
resemble a Rankine cycle, or a real Diesel engine may resemble the theoretical Diesel cycle. Engineering students may wonder
whether they need bother with learning about “theoretical” engines that bear little resemblance to the metal and fuel that they
have to work with on a practical basis. I cannot answer that, but there is just one thing I do know about real engines, and that is
that they are subject to and follow all the fundamental laws of thermodynamics that theoretical engines have to follow; and I
suspect that the engineer who designed the engine in my car had a pretty thorough knowledge of the fundamental principles of
thermodynamics.

Jeremy Tatum 4/8/2021 11.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7277

11.2: The Carnot Cycle
I referred above to one of the uses of the theoretical concept known as the Carnot cycle, namely that it enables us to define an
absolute temperature scale. I suggest that, before you read any further, you re-read Section 3.4 of Chapter 3.
Pause while you re-read Section 3.4
As a temporary measure I am going to use the symbol θ to represent the temperature measured on the ideal gas scale. I shall
then define an absolute temperature scale, T, and show that it is identical with the ideal gas temperature scale.
To start with, I shall suppose that the working substance in our Carnot engine is an ideal gas. We shall refer to figure XI.1, in
which ab and cd are isotherms at temperatures θ2 and θ1 respectively (θ2 > θ1), and bc and da are adiabats. Starting at the point
a(P1, V1), a quantity of heat Q2 is supplied to the gas as it expands isothermally from a to b(P2 ,V2) at temperature θ2 on the
ideal gas scale. During this phase, the cylinder is supposed to be uninsulated and placed in a hot bath at temperature θ2. As it
expands isothermally it does external work. Since the working substance is an ideal gas, the internal energy at constant
temperature is independent of volume (there is no internal work against van der Waals forces to be done) so the heat supplied
to the gas is equal to the external work that it does. That is, per mole,
Q2 = Rθ2 ln(V2 / V1 ). (11.2.1)

After the gas has reached b the cylinder is insulated and the gas expands adiabatically and reversibly to c(P3, V3).
It is then placed in a cold bath at temperature θ1, uninsulated, and compressed isothermally to d(P4, V4). During this stage it
gives out a quantity of heat Q1:

Q1 = Rθ1 ln(V3 / V4 ). (11.2.2)

Finally it is insulated again and compressed adiabatically and reversibly to its original state a.
For these four stages we have the equations
lP1 V1 = P2 V2 (11.2.3)

γ γ
P2 V = P3 V (11.2.4)
2 3

P3 V3 = P4 V4 (11.2.5)

γ γ
P1 V = P4 V (11.2.6)
1 4

From these, we readily see that

V2 / V1 = V3 / V4 , (11.2.7)

and therefore

Q2 / Q1 = θ2 / θ1 . (11.2.8)

Jeremy Tatum 3/18/2021 11.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7278

The net heat received is Q2 − Q1, and this is the heat available for doing external work. A quantity of heat must be supplied at
the beginning of each cycle, and so the efficiency of the cycle is
Q2 − Q1 θ2 − θ1
η = = . (11.2.9)
Q2 θ2

Thus the efficiency of the Carnot engine is the fractional temperature difference between source and sink.
We have specified in the above that the working substance is an ideal gas, the temperatures of source and sink being θ1 and θ2
on the ideal gas scale. Let us now not specify what the working substance is, but let us set up a system of 100 Carnot engines
working in tandem, with the sink of one being the source for the next. We’ll have the sink for the coldest engine in a bucket of
melting ice (0 oC) and the source for the hottest engine in a bucket of boiling water (100 oC). They will be working between
isothermals and adiabats on an absolute thermodynamic scale, T, defined such that the net work done by each engine (i.e. the
area of each PV loop) per cycle is the same for each of the engines. This will define the temperature on an absolute scale. It
would take me a while to use the computer to do a decent drawing of 100 isotherms and 2 adiabats, so I’m going to try to
make do with a hand-drawn sketch (figure X1.2) of just five isotherms, two adiabats and four linked Carnot cycles to illustrate
what I am trying to describe.
We suppose that the efficiency of such a Carnot engine depends solely on the temperature of source and sink:
Q1 / Q2 = f (T1 , T2 ) . (11.2.10)

We are making no assumption about the form of this function, which is completely arbitrary. We are free to define it in any
manner that is useful to us in our attempt to define an absolute temperature scale.

Let us consider two adjacent engines, one working between temperatures T1 and T2, and the other working between
temperatures T2 and T3. We have:
lQ1 / Q2 = f (T1 , T2 ) (11.2.11)

Q2 / Q3 = f (T2 , T3 ) (11.2.12)

and for the pair as a whole considered as a single engine,

Q1 / Q3 = f (T1 , T3 ) . (11.2.13)

From these we find that

f (T1 , T3 )
f (T1 , T2 ) = . (11.2.14)
f (T2 , T3 )

This can be only if T3 cancels from the right hand side, so that
ϕ (T1 )
f (T1 , T2 ) = . (11.2.15)
ϕ (T2 )

That is,
Q1 ϕ (T1 )
= . (11.2.16)
Q2 ϕ (T2 )

And since φ is a completely arbitrary function that we can choose at our pleasure to define an absolute scale, we choose

Jeremy Tatum 3/18/2021 11.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7278

 Q1 T1
= . (11.2.17)
Q2 T2

And, with this choice, the absolute thermodynamic temperature scale is identical with the ideal gas temperature scale.
Equation 11.2.17 also implies that entropy in = entropy out. Entropy is conserved around the complete cycle. Entropy is a
function of state.
In Sections 11.3 to 11.5 I give examples of some other cycles. These are largely for reference, and readers who wish to
continue without interruption with the theoretical development of the subject can safely skip these and move on to Sections
11.7 and 11.8.

Jeremy Tatum 3/18/2021 11.2.3 CC-BY-NC https://phys.libretexts.org/@go/page/7278

11.3: The Stirling Cycle
This takes place between two isotherms and two isochors. Note that, provided the working substance is an ideal gas, there is
no change in the internal energy along the isotherms, and that the work done by or on the gas is equal to the heat gained by or
lost from it. No work is done along the isochors. I show the cycle in the PV plane in figure XI.3, and an imaginary schematic
engine in figure XI.4.

The gas is supposed to be held in a cylinder between two pistons. The cylinder is divided into two sections by a porous
partition. One section is kept at a hot temperature T2 and the other is kept at a cold temperature T1.
In stage a, the cold gas is compressed isothermally. The work done on a mole of the gas is RT1 ln(V2/V1); this is converted into
heat, Qa, which is lost from the gas to the cold reservoir.
In stage b, the gas, held at constant volume, is transferred to the hot reservoir. No work is done on or by the gas, but a quantity
of heat Qb = CV(T2 − T1) per mole is supplied to the gas.
In stage c, the hot gas is expanded isothermally to its original volume. The work done by a mole of the gas is RT2 ln(V2/V1); in
order to prevent the gas from cooling down, it has to absorb an equal amount of heat, Qc from the hot reservoir. Note that Qc >
Qa.
In stage d, the gas, held at constant volume, is transferred back to the cold reservoir. No work is done on or by the gas, but the
gas loses a quantity of heat Qd = CV(T2 − T1) to the cold reservoir. Note that Qd = Qb.
Exercise: Show that the efficiency is
R (T2 − T1 ) ln(V2 / V1 )
η = . (11.3.1)
CV (T2 − T1 ) + RT2 ln(V2 / V1 )

Jeremy Tatum 4/1/2021 11.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7279

If the gas is an ideal diatomic gas (to which air is an approximation), then C V =
5

2
R , and then
(T2 − T1 ) ln(V2 / V1 )
η = . (11.3.2)
2.5 (T2 − T1 ) + T2 ln(V2 / V1 )

If helium were used as an ideal gas, the efficiency would be greater, because for helium, C V =
3

2
R .

Jeremy Tatum 4/1/2021 11.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7279

11.4: The Otto Cycle
The Otto cycle (to which the engine under the hood of your car bears some slight resemblance) works between two isochors
and two adiabats (figure XI.5).

The cycle starts at A. From A to B the piston recedes and a valve is open, so that a misture of air and petrol (gasoline) is drawn
in at constant (atmospheric) pressure. The temperature is typically somewhat above ambient temperature because of the
previous operation of the cycle. At B, the valve is closes, and now from B to C a fixed mass of gas is compressed adiabatically,
the temperature being a few hundred K. C is the point of maximum compression. At this point a spark is struck and the
mixture is ignited. In effect heat is added to the system and the temperature goes up instantaneously to perhaps 2000 K at
constant (small) volume. The gas, now having reached D, expands adiabatically to E, doing work, and the temperature drops
somewhat. At E, a (second) valve opens, gas is expelled, the pressure drops to atmospheric, and the temperature drops to its
original value. We are now at F. The piston pushes the remaining gas out, and we end at G. The cycle starts anew.
It is left as an exercise to show:
TB
Net work done by the engine per cycle = C V (TD − TC ) (1 −
TC
) .
1/γ−1)
TB
Volume of stroke = V B − VC = VB [1 − (
TC
) ] .

1/(γ−1)
TD TC
Maximum pressure = P D = PB
TB
(
TB
) .
γ−1
VC TB
Efficiency = 1 − ( VB
) =1−
TC
.

In principle the efficiency could be very large if the temperature at C, at the end of the adiabatic compression, were high. In
practice the temperature at the end of the adiabatic compression is limited (and therefore so is the efficiency) because, if the
temperature were too high, the air-gasoline mixture would ignite spontaneously.

Jeremy Tatum 4/29/2021 11.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7280

11.5: The Diesel Cycle
This difficulty is avoided in the Diesel cycle in that, during the adiabatic compression stage to a high temperature, it is just air
(not an air-fuel mixture) that is compressed. Only then, when the temperature is high, is fuel injected, which then immediately
ignites. The cycle is shown in figure XI.6.
We start at A. A valve opens and the piston moves back, and pure air (no fuel) is sucked into the cylinder. This is followed by
an adiabatic compression from B to C, which can reach a high temperature of 2000 K or so. At C a jet of liquid fuel is forced
at high pressure into the cylinder by a pump that is operated by the engine itself. The fuel immediately ignites. The rate of
injection is held so that the mixture expands at constant pressure until we reach D, at which point the injection of fuel is cut off
and the gas expands adiabatically to E. A valve is then opened so that the pressure drops to atmospheric at F. The piston then
pushes the remainder of the mixture out, and the cycle stars anew.
It is left as an exercise to show:
1−1/γ γ−1
PC PB Tp
Net work done by the engine per cycle = C P [TD − TB (
PB
) ] − CV [TD (
PC TB
) − TB ] .

1/γ
PB
Volume of stroke = V B − VC = VB [1 − (
PC
) ] .
γ−1
P T
B D
CV [TD ( ) −TB ]
P T

Efficiency = 1 − .
C B

1−1/γ
P
C
CP [TD −TB ( ) ]
P
B

Have a look at http://www.univ-lemans.fr/enseignements/physique/02/thermo/diesel.html

Exercise: Assuming γ = 1.4, what are the efficiencies of the Carnot, Otto and Diesel cycles running between 350 K and 2000
K? Assume for the Diesel cycle that the maximum pressure is 30 atmospheres. Assume for the Otto cycle that TC =650 K.

Jeremy Tatum 4/29/2021 11.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7281

11.6: The Rankine Cycle (Steam Engine)
The Titfield Thunderbolt runs on an engine that slightly resembles the Rankine cycle.
The amount of work obtainable from an engine depends on the amount of the working substance and on the temperature.
Internal combustion Otto and Diesel engines work at high temperatures, so they can be small. The steam engine is bulky but
does not require high temperatures. The steam engine has a boiler (which, naturally, boils water into steam) and a condenser
(which, naturally, condenses the steam back again to water).

Steam from the boiler is drawn into a cylinder at constant pressure (A to B), at which point the intake valve is closed and the
remaining expansion (B to C) is adiabatic, taking the temperature down to the temperature of the condenser. The section C to
D corresponds to the condensation of the steam. From D to A the condensed water is transferred to the boiler, and the cycle
starts again.

Jeremy Tatum 4/8/2021 11.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7282

11.7: A Useful Exercise
It would probably not be a useful exercise to try to memorise the details of the several heat engine cycles described in this
chapter. What probably would be a useful exercise is as follows. Note that in each cycle there are four stages, which, in
principle at least (if not always in practice) are well defined and separated one from the next. These stages are described by
one or another of an isotherm, an adiabat, an isochor or an isobar. It would probably be a good idea to ask oneself, for each
stage in each engine, the values of ∆Q, ∆W and ∆U, noting, of course, that in each case, ∆U = ∆Q + ∆W. In each case take care
to note whether heat is added to or lost from the engine , whether the engine does work or whether work is done on it, and
whether the internal energy increases or decreases. By doing this, one could then easily determine how much heat is supplied
to the engine, and how much net work it does during the cycle, and hence determine the efficiency of the engine.
The following may serve as useful guidelines. In these guidelines it is assumed that any work done is reversible, and that
(except for the steam engine or Rankine cycle) the working substance may be treated as if it were an ideal gas.
Along an isotherm, the internal energy of an ideal gas is unchanged. That is to say, ∆U = 0. The work done (per mole of
working substance) will be an expression of the form RT ln(V2/V1), and the heat lost or gained will then be determined by ∆Q
+ ∆W = 0.
Along an adiabat, no heat is gained or lost, so that ∆Q = 0. The expression for the work done per mole will be of the form
R( T1 −T2 ) P1 V1 −P2 V2

γ−1
=
γ−1
where V is the molar volume. Just be sure to understand whether work is done on or by the engine. The
change in the internal energy (be sure to understand whether it is an increase or a decrease) is then given by ∆U = ∆W.
Along an isochor, no work is done. That is, ∆W = 0. The heat lost or gained per mole will be of an expression of the form
CV(T2 − T1), where CV is the molar heat capacity at constant volume. The change in the internal energy (be sure to understand
whether it is an increase or a decrease) is then given by ∆U = ∆Q.
Along an isobar, none of Q, W or U are unchanged. The work done per mole (by or on the engine?) will be an expression of
the form ∆W = P(V2 − V1) = R(T2 − T1).
The heat added to or lost from the engine will be an expression of the form CP(T2 − T1), where CP is the molar heat capacity at
constant pressure. The change in the internal energy (be sure to understand whether it is an increase or a decrease) is then
given by ∆U = ∆Q + ∆W.
It might also be a good idea to try to draw each cycle in the T : S plane (with the intensive variable T on the vertical axes).
Indeed I particularly urge you to do this for the Carnot cycle, which will look particularly simple. Note that, while the area
inside the cycle in the P : V plane is equal to the net work done on the engine during the cycle, the area inside the cycle in the
T : S plane is equal to the net heat supplied to the engine during the cycle.

Jeremy Tatum 3/18/2021 11.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8623

11.8: Heat Engines and Refrigerators

Figure XI.8 illustrates schematically the path taken by the state of a working substance is a generalized heat engine. In the
upper part of the cycle (continuous curve) the working substance is expanding, and the machine is doing work. The work done
by the engine is ∫PdV, or the area under that part of the curve. In the lower part of the cycle (dashed curve) the working
substance is being compressed; work is being done on it. This work is the area under the dashed portion of the cycle. The net
work done by the engine during the cycle is the work done by the engine while it is expanding minus the work done on it
during the compression part of the cycle, and this is the area enclosed by the cycle.
During one part of any heat engine cycle, heat is supplied to the engine, and during other parts, heat is lost from it. As
described in Section 11.1, the efficiency η of a heat engine is defined by
 net external work done by the engine during a cycle
η = (11.8.1)
heat supplied to the engine during a cycle.

Note that the word “net” does not appear in the denominator. The efficiency can also be calculated from
Q in  − Q out 
η = , (11.8.2)
Q in 

though I stress that this is not a definition.

In the Carnot engine, which is the most efficient conceivable engine for given source and sink temperature, the efficiency is
T2 − T1
η = , (11.8.3)
T2

where T2 and T1 are respectively the temperatures of the hot source and cold sink.
If the working substance is taken round a cycle in the PV-plane in the counterclockwise direction, the device is a refrigerator.
In that case the area enclosed by the cycle is equal to the net work that is done on the working substance. If the refrigerator
operates on a reverse Carnot cycle, the working substance takes in (from whatever it is that it is trying to cool) a quantity of
heat Q1 as it expands isothermally from d to c (see figure XI.1, but with the arrows reversed) and expels a (greater) quantity of
heat Q2 as it is compressed isothermally from b to a. This quantity Q2 is expelled into the room – which is why the room gets
warmer when you switch on the fridge. (What – you never noticed?) The refrigerating effect is Q1, since this is the quantity of
heat taken in by the refrigerator from the body that is to be cooled.
The coefficient of performance of a refrigerator is defined by
 refrigerating effect 
(11.8.4)
 net work done on the engine during the cycle. 

By the first law of thermodynamics, the denominator of the expression is Q2 − Q1, and for a reversible Carnot cycle, the
entropy in equals the entropy out, so that Q2/Q1 = T2/T1. Therefore the coefficient of performance for a Carnot refrigeration
cycle can be calculated from

Jeremy Tatum 4/15/2021 11.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8624

 T1
. (11.8.5)
T2 − T1

This, of course, can be much greater than 1 – but no refrigerator working between the same source and sink temperatures can
have a coefficient of performance greater that that of a reversible Carnot refrigerator.
Of course the working substance in a real refrigerator (“fridge”) is not an ideal gas, nor does one follow a Carnot cycle – there
are too many practical difficulties in the way of achieving this ideal dream. As mentioned elsewhere in this course, I am not a
practical man and I am not suited to describing real, practical machines. The fundamental principles described in this section
do, of course, still apply in the real world! In a real refrigerator, the working substance (the refrigerant) is a volatile fluid
which is vaporized in one part of the operation and condensed to a liquid in another part. In industrial refrigerators, the
refrigerant may be ammonia, but this is considered to be too dangerous for domestic use. “Freon”, which was a mixture of
chlorofluorocarbons, such as CCl2F2, was in fashion for a while, but escaping chlorofluorocarbons have been known for some
time to cause breakdown of ozone (O3) in the atmosphere, thus destroying our protection against ultraviolet radiation from the
Sun. The chlorofluorocarbons have been largely replaced by hydrofluorocarbons, such as C2H2F4, which are believed to be
less damaging to the ozone layer. The exact formula or mixture is doubtless a trade secret.
The fluid is forced around a system of tubes by a pump called the compressor. Shortly before the fluid reaches the freezer it is
in liquid form, moving along some rather narrow pipes. It is then forced through a nozzle into a system of wider pipes (the
evaporator) surrounding the freezer, and there it vaporizes, taking heat from the food and from the air in the freezer. A fan
may also distribute the cooled air throughout the rest of the refrigerator. After leaving the freezer, the vapour returns to the
compressor, where it is, of course, compressed (which is why the pump is called the compressor). This produces heat, which is
dissipated into the room as the fluid is forced through a series of pipes and vanes, known as the condenser, at the rear of the
fridge, where the fluid condenses into liquid form again. The cycle then starts anew.
The following summary of Carnot heat engines and refrigerators may be helpful. (But just remember that, while Carnot cycles
are the most efficient engines and refrigerators for given source and sink temperatures, the practical realization of a real engine
or refrigerator may not be identical to this theoretical ideal.)
Notation:
T2 = hotter temperature
T1 = cooler temperature
Q2 = heat gained or lost at T2
Q1 = heat gained or lost at T1
Q1 Q2
ΔS = 0 =
T1 T2

Heat Engine:

ΔU = 0  Net work done by engine  = Q2 − Q1 .

Qin −Q  out   Q2 −Q1 T2 −T1

 Efficiency η = = =
Q  in  Q2 T2

Refrigerator:

Jeremy Tatum 4/15/2021 11.8.2 CC-BY-NC https://phys.libretexts.org/@go/page/8624

 ΔU = 0  Net work done  on refrigerator  = Q2 − Q1

Q  in  Q1 T1
 Coefficient of Performance P = = =
Q  ox t  −Q  in  Q2 −Q1 T2 −T1

Heat Pump:
The principle of a heat pump is the same as that of a refrigerator, except that its purpose is different. The purpose of a
refrigerator is to extract heat from something (e.g. food) and so to make it colder. That the heat so extracted goes into the room
to make the room warmer (at least in principle) is incidental. The important thing is how much heat is extracted from the food,
and that is why it is appropriate to define the coefficient of performance of a refrigerator as the refrigerating effect (i.e. Q1)
divided by the net work done on the refrigerator, per cycle. But with a heat pump, the object is to heat the room by extracting
heat from outside. That the outside may become cooler (at least in principle) is incidental. Thus, for a heat pump, the
appropriate definition of the coefficient of performance is the heating effect (i.e. Q2) divided by the net work done on the
refrigerator, per cycle.

ΔU = 0  Net work done on heat pump  = Q2 − Q1

Qout Q2 T2
Coefficient of Performance P = = =
Qout −Qin Q2 −Q1 T2 −T1

You can see from this equation that, the warmer it is outside (T1), the greater the coefficient of performance. You may
therefore wonder if it is practical to use a heat pump to heat a building in a cold climate, such as the Quebec winter. And, if it
isn’t, can one devise an engine that is simultaneously a refrigerator and a heat pump; that is to say, it extracts heats from (i.e.
cools) the food, and transfers this heat (plus a little bit more because of the work that is done on the refrigerator/heat pump)
into the room in order to heat the room effectively. There’s an answer to that in an article in the Victoria Times-Colonist of
June 11, 2006, which I reproduce, with permission, below.

Jeremy Tatum 4/15/2021 11.8.3 CC-BY-NC https://phys.libretexts.org/@go/page/8624

Air Conditioner
The purpose of a refrigerator (“fridge”) is to pump some heat Q1 from the food (or whatever is to be kept cool). The quantity
Q1 is the “refrigerating effect”. During the operation of the fridge, a somewhat greater quantity Q2 of heat is expelled into the

Jeremy Tatum 4/15/2021 11.8.4 CC-BY-NC https://phys.libretexts.org/@go/page/8624

room, though this should not result in a very noticeable rise in temperature of the room, partly because the room has a large
thermal capacity, and partly because much of this heat will be lost through the windows. The coefficient of performance of the
fridge is the refrigerating effect per cycle, Q1, divided by the net work done on the fridge per cycle, and, for a Carnot cycle it
can be calculated from T1/(T2 − T1).
The purpose of a heat pump is to pump some heat Q1 from outside, and (from the work done on the pump) to pump a larger
quantity Q2 of heat into the room – large enough, indeed to warm the room appreciably, supposing that you don’t keep all the
windows wide open. The coefficient of performance must therefore be defined as Q2 divided by the net work done on the
fridge per cycle. For a Carnot cycle it can be calculated from T2/(T2 − T1).
There is a third possibility, namely an air conditioner. This will incorporate a dehumidifier, but, in our present context we
regard it as a device whose purpose is to pump heat from the room to the outside, rather than from outside to the room. If it is
successful, the room will become cooler than the outside. Thus an air conditioner is more like a refrigerator, in that the
coefficient of performance is the heat Q1 extracted per cycle from the room divided by the net work done on the machine per
cycle. For a Carnot cycle it can be calculated from T1/(T2 − T1).

ΔU = 0  Net work done on air conditioner  = Q2 − Q1 .

Q  in  Q1 T1
 Coefficient of Performance P =
Q  out  −Q  in 
=
Q2 −Q1
=
T2 −T1
.

Those who have read thus far will have an idea that there are things called heat engines, refrigerators, heat pumps and air
conditioners, which are represented by Carnot cycles or similar cycles, with arrows going in different directions, a few
equations with different subscripts, and subtly different definitions of efficiency or coefficient of performance. Since I
prepared these notes I have discovered that there actually exist in the real world, real, solid machines called heat engines,
refrigerators, heat pumps and air conditioners. I have discovered two very nice little pamphlets describing real heat pumps
and real air conditioners, and how you might install them to heat or to cool your home. They are called Heating and Cooling
with a Heat Pump, and Air Conditioning your Home, each about 50 pages. My copies are dated 1996, revised 2004, though I
dare say you might be able to get more recent ones. They are available free from Energy Publications, Office of Energy
Efficiency, Natural Resources Canada, c/o S.J.D.S., 1779 Pink Road, Gatineau, Province of Québec, Canada J9J 3N7. I found
them fascinating.

Jeremy Tatum 4/15/2021 11.8.5 CC-BY-NC https://phys.libretexts.org/@go/page/8624

11.9: Entropy is a Function of State
We have defined temperature on the absolute scale such that the temperature of the source of a reversible Carnot heat engine is
proportional to the heat taken in by the engine during its isothermal expansion at the hot temperature, and the temperature of
the sink is proportional to the heat lost by the engine during its isothermal compression at the cool temperature. No heat is
gained or lost, of course, during the adiabatic phases, and there is no change in internal energy over a complete cycle.
Therefore Q1/Q2 = T1/T2.
Now, any cycle can be represented by an infinite number of infinitesimally narrow Carnot cycles operating in tandem. Thus
∫dQ/T during that part of the cycle in which an engine is losing heat is equal to ∫dQ/T during that part of the cycle in which it is
absorbing heat. Therefore, during the complete cycle, ∫dQ/T is zero. This means that the net change in entropy during a
complete cycle is zero, so that entropy is a function of state. In effect 1/T is an integrating factor which, when it multiplies the
inexact differential đQ, results in the exact differential đQ/T = dS.

Jeremy Tatum 3/25/2021 11.9.1 CC-BY-NC https://phys.libretexts.org/@go/page/8625

CHAPTER OVERVIEW
12: FREE ENERGY

12.1: REVIEW OF INTERNAL ENERGY AND ENTHALPY

12.2: FREE ENERGY
12.3: HELMHOLTZ FREE ENERGY
12.4: GIBBS FREE ENERGY
12.5: SUMMARY, THE MAXWELL RELATIONS, AND THE GIBBS-HELMHOLTZ RELATIONS
12.6: THE JOULE AND JOULE-THOMSON COEFFICIENTS
12.7: THE THERMODYNAMIC FUNCTIONS FOR AN IDEAL GAS
12.8: THE THERMODYNAMIC FUNCTIONS FOR OTHER SUBSTANCES
12.9: ABSOLUTE ENTROPY
12.10: CHARGING A BATTERY
12.11: SURFACE ENERGY
12.12: FUGACITY

1 4/29/2021
12.1: Review of Internal Energy and Enthalpy
We are by now familiar with the equations
dU = TdS − PdV and dH = TdS + VdP,
and with the ideas that the increase in the internal energy is the heat added at constant volume and the increase in enthalpy is
the heat added at constant pressure, and that U is constant in an adiabatic isochoric process and H is constant in an adiabatic
isobaric process. I am now going to examine these equations and statements a bit more critically. In particular I am going to
consider that there may be several types of configuration work involved in addition to just PdV work of compression or
expansion.
The First Law of thermodynamics is dU = dQ + dW.
The work done on a system may comprise an irreversible component dWI (such as stirring with a paddle, or forcing an electric
current through a resistor) plus some reversible components dWR. The irreversible component is dissipated as heat and is
tantamount to adding heat to the system. The heat and the irreversible work contribute to the increase in entropy of the system,
according to dS = (dQ + dWI)/T. Thus we have dQ = TdS − dWI.
The reversible component of the work may consist of work done in compressing the system, −PdV, but there may also be other
kinds of work, such as the work required to create new area, Γdσ , or the work required to twist a rod, τdθ, or the work
required to charge a battery, Edq, or the work required to magnetize a specimen, BdM, and perhaps others. In general the
expression for each of these forms of reversible work is of the form XdY, where X is an intensive state variable and Y is an
extensive state variable. All of these forms of nondissipative work can collectively be called configuration work.
The total work done on the system is therefore of the form

dW = dWI − P dV + ∑ XdY . (12.1.1)

The first law therefore takes the form

dU = dQ + dWI − P dV + ∑ XdY . (12.1.2)

If the system is held at constant volume (e.g. in a pressure cooker or in an autoclave), then no PdV work of expansion or
compression is done. And if no other sort of work is done either (either non-PdV reversible work or irreversible work dWI)
then the increase in internal energy of the system is just equal to the heat added to it.
Enthalpy is defined as H = U + PV, so that dH = dU + PdV + VdP. From this, we obtain

dH = dQ + dWI + V dP + ∑ XdY . (12.1.3)

If heat is added to a system at constant pressure, then the system expands and does external work. However, provided that the
pressure is held constant and if no other sort of work is done either (either non-PdV reversible work or irreversible work dWI)
then the increase in the enthalpy of the system is just equal to the heat added to it.
In summary, the well-known equations dU = TdS − PdV and dH = TdS + VdP are valid for reversible and for irreversible
processes, provided that the only nondissipative work is PdV work; but in general, if there are other types of work being done
(e.g. Γdσ, or τdθ, etc.), the required relations are

dU = T dS − P dV + ∑ XdY (12.1.4)

and

dH = T dS + V dP + ∑ XdY . (12.1.5)

Jeremy Tatum 3/4/2021 12.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7284

12.2: Free Energy
We shall be learning that there are two sorts of free energy.
There is the Helmholtz free energy. Commonly used symbols for this are A (from the German die Arbeit – work) or F.
And there is the Gibbs free energy. Commonly used symbols for this are G − or F!
It is unfortunate that some writers will use simply the term "free energy”, using the symbol F, without specifying which, or
even giving evidence that they are aware of the difference. I have seen the symbol F used about equally often for Helmholtz,
Gibbs or unspecified free energies.
In these notes I shall use the symbol A for the Helmholtz free energy and G for the Gibbs free energy, and I shall avoid the
symbol F.

Jeremy Tatum 4/8/2021 12.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7285

12.3: Helmholtz Free Energy
The Helmholtz free energy A is defined as
A = U − T S. (12.3.1)

As when we first defined enthalpy, this doesn't seem to mean much until we write it in differential form:
dA = dU − T dS − SdT . (12.3.2)

On substitution from equation 12.1.6 (dU = TdS − PdV + ∑XdY), this becomes

dA = −SdT − P dV + ∑ XdY . (12.3.3)

This tells us that in an isothermal process (in which dT = 0), the increase in the Helmholtz function of a system is equal to all
the reversible work (−PdV + ∑XdY) done on it. Conversely, if a machine does any reversible work at constant temperature, the
Helmholtz function decreases, and the decrease in the Helmholtz function is equal (if the temperature is held constant) to the
reversible work (of all types) done by the machine. It is in this sense that the Helmholtz function is called the “free energy”. It
is the energy, so to speak, that is free for the performance of external reversible (i.e. useful) work.

Jeremy Tatum 4/8/2021 12.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7287

12.4: Gibbs Free Energy
The Gibbs free energy G is defined as
G = H −TS (12.4.1)

or, what amounts to the same thing,

G = A+PV . (12.4.2)

As when we first defined enthalpy, this doesn't seem to mean much until we write it in differential form:
dG = dH − T dS − SdT (12.4.3)

or
dG = dA + P dV + V dP . (12.4.4)

Then, either from equations 12.1.5 (dH = TdS + VdP + ∑XdY) and 12.5.3 or from equation 12.4.3 (dA = −SdT − PdV + ∑XdY)
and 12.5.4, we obtain

dG = −SdT + V dP + ∑ XdY (12.4.5)

That is to say that, if the temperature and pressure are constant, the increase in the Gibbs function of a system is equal to the
reversible work (other than PdV work of compression) done on it. Conversely, if the temperature and pressure are held
constant, and a machine is used to do external work (which may include but is not limited to PdV work of expansion), the
Gibbs function decreases by the amount of reversible (i.e.useful) work done by the machine other than the PdV work of
expansion.

Jeremy Tatum 4/22/2021 12.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7288

12.5: Summary, the Maxwell Relations, and the Gibbs-Helmholtz Relations
dU = T dS − P dV + ∑ XdY (12.5.1)

dH = T dS + V dP + ∑ XdY (12.5.2)

dA = −SdT − P dV + ∑ XdY (12.5.3)

dG = −SdT + V dP + ∑ XdY (12.5.4)

If the only reversible work done on or by a system is PdV work of expansion or compression, we have the more familiar forms

dU = T dS − P dV (12.5.5)

dH = T dS + V dP (12.5.6)

dA = −SdT − P dV (12.5.7)

dG = −SdT + V dP (12.5.8)

All four thermodynamic functions are functions of state (and hence their differentials are exact differentials) and therefore
∂U ∂U
( ) =T ( ) = −P (12.5.9)
∂S ∂V
V S

∂H ∂H
( ) =T ( ) =V (12.5.10)
∂S ∂P
P S

∂A ∂A
( ) = −S ( ) = −P (12.5.11)
∂T ∂V
V T

∂G ∂G
( ) = −S ( ) =V (12.5.12)
∂T ∂P
P T

Further, by equating the mixed second derivatives, we obtain the four Maxwell Thermodynamic Relations:
∂T ∂P
( ) = −( ) (12.5.13)
∂V ∂S
S V

∂T ∂V
( ) =( ) (12.5.14)
∂P ∂S
S P

∂S ∂P
( ) =( ) (12.5.15)
∂V ∂T
T V

∂S ∂V
( ) = −( ) (12.5.16)
∂P ∂T
T P

The Gibbs-Helmholtz Relations are trivially found from A = U − TS and together with equations 12.6.11a and 12.6.12a. G = H
− TS They are
∂A
U = A−T( ) (12.5.17)
∂T
V

∂G
H = G−T( ) (12.5.18)
∂T
P

Jeremy Tatum 4/15/2021 12.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7289

12.6: The Joule and Joule-Thomson Coefficients
In Chapter 10, we studied the Joule and Joule-Thomson experiments and we calculated the Joule and Joule-Thomson
coefficients. Now that we are familiar with the Helmholtz and Gibbs functions, and, in particular, with two Maxwell relations
that can be derived from them, we can obtain alternative derivations for these two coefficients. These may be easier than the
derivations we gave in Chapter 10. I am indebted to Dr Greg Trayling for the derivation of the Joule coefficient; the derivation
of the Joule-Thomson coefficient follows a parallel argument.
Let us start with the Joule coefficient. Here we are interested in how the temperature changes with volume in an experiment in
which the internal energy is constant. That is, we want to derive the Joule coefficient, η = (∂T/∂V)U.
Now entropy is a function of state – i.e. of the intensive state variables P, V and T. (V = molar volume.) But the intensive state
variables for a particular substance are related by an equation of state, so we need express the entropy as a function of only two
of P, V or T, and, since we are seeking a relation between V and T, let us choose to express S as a function of V and T, so that
∂S ∂S
dS = ( ) dV + ( ) dT . (12.6.1)
∂V ∂T
T V

Let us look at these three terms in turn.

First, dS. In the Joule experiment, the internal energy of the gas is constant, so that
T dS − P dV = 0. (12.6.2)

That is,
P dV
dS = . (12.6.3)
T

For the first term on the right hand side of equation 12.7.1, we make use of the Maxwell relation, equation 12.6.15, which we
derived from the Helmholtz function:
∂S ∂P
( ) =( ) . (12.6.4)
∂V ∂T
T V

For the second term on the right hand side we obtain

∂S ∂S ∂U ∂U ∂U Cv
( ) =( ) ( ) =( ) /( ) = . (12.6.5)
∂T ∂U ∂T ∂T ∂S T
V v V V V

Thus, equation 12.7.1 becomes

P dV ∂P CV dT
=( ) dV + . (12.6.6)
T ∂T T
V

Multiply through by T, and divide by dV, taking the infinitesimal limit as dV → 0, recalling that we are dealing with an
experiment in which the internal energy is constant, and we arrive at
∂P ∂T
P = T( ) + CV ( ) , (12.6.7)
∂T ∂V
V U

from which we immediately obtain

∂T 1 ∂P
( ) = [P − T ( ) ], (12.6.8)
∂V CV ∂T
U V

quod erat demonstrandum.

Let us now consider the Joule-Thomson coefficient. Here we are interested in how the temperature changes with pressure in an
experiment in which the enthalpy is constant. That is, we want to derive the Joule-Thomson coefficient, µ = (∂T/∂P)H.

Jeremy Tatum 4/29/2021 12.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/8633

Now entropy is a function of state – i.e. of the intensive state variables P, V and T. (V = molar volume.) But the intensive state
variables for a particular substance are related by an equation of state, so we need express the entropy as a function of only two
of P, V or T, and, since we are seeking a relation between P and T, let us choose to express S as a function of P and T, so that
∂S ∂S
dS = ( ) dP + ( ) dT . (12.6.9)
∂P ∂T
T P

Let us look at these three terms in turn.

First, dS. In the Joule-Thomson experiment, the enthalpy of the gas is constant, so that

T dS + V dP = 0. (12.6.10)

That is,
V dP
dS = − . (12.6.11)
T

For the first term on the right hand side of equation 12.7.9, we make use of the Maxwell relation, equation 12.6.16, which we
derived from the Gibbs function:
∂S ∂V
( ) = −( ) . (12.6.12)
∂P ∂T
T P

For the second term on the right hand side we obtain

∂S ∂S ∂H ∂H ∂H CP
( ) =( ) ( ) =( ) /( ) = . (12.6.13)
∂T ∂H ∂T ∂T ∂S T
P P P P P

Thus, equation 12.7.9 becomes

V dP ∂V CP dT
− = −( ) dP + . (12.6.14)
T ∂T T
P

Multiply through by T, and divide by dP, taking the infinitesimal limit as dP → 0, recalling that we are dealing with an
experiment in which the enthalpy is constant, and we arrive at
∂V ∂T
−V = −T ( ) + CP ( ) , (12.6.15)
∂T ∂P
P H

from which we immediately obtain

∂T 1 ∂V
( ) = [T ( ) −V ], (12.6.16)
∂P CP ∂T
H P

quod erat demonstrandum.

Jeremy Tatum 4/29/2021 12.6.2 CC-BY-NC https://phys.libretexts.org/@go/page/8633

12.7: The Thermodynamic Functions for an Ideal Gas
In this section I tabulate the changes in the thermodynamic functions for an ideal gas taken from one state to another.
One mole of an ideal gas going isothermally and reversibly from P1V1T to P2V2T or adiabatically and reversibly from
P1V1T1 to P2V2T2.

Isothermal Adiabatic
P1 V1 −P2 V2 R(T1 −T2 )
Work done by gas RT ln(V2/V1)* γ−1
=
γ−1
= CV (T1 − T2 )

P1 V1 −P2 V2 R(T1 −T2 )

U2 − U1 0 −
γ−1
= −
γ−1
= −CV (T1 − T2 )

Heat absorbed by gas RT ln(V2/V1) 0

S2 − S1 R ln(V2/V1) 0
P1 V1 −P2 V2 R(T1 −T2 )
H2 − H1 0 −
1−1/γ
= −
1−1/γ
= −CP (T1 − T2 )

R(T1 −T2 )
A2 − A1 −RT ln(V2/V1) −
γ−1
− T2 S2 + T1 S1

R(T1 −T2 )
G2 − G1 −RT ln(V2/V1) −
1−1/γ
− T2 S2 + T1 S1

*Note that for isothermal processes on an ideal gas, we can write (V2/V1) = (P1/P2).
A difficulty will be noted in the entries for the increase in the Helmholtz and Gibbs functions for an adiabatic process, in that,
in order to calculate ∆A or ∆G, it is apparently necessary to know S1 and S2, and not merely their difference. For the time being
this is a difficulty to note on one’s shirt-cuff, and perhaps return to it later.

Jeremy Tatum 4/8/2021 12.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8634

12.8: The Thermodynamic Functions for Other Substances
Calculation of the change in the thermodynamic functions of any substance going reversibly from PlV1T1 to P2V2T2.
The first comforting thing to note is that SUHAG are all state functions, and therefore the change in their values is route-
independent.
Entropy.
Entropy is a function of state (i.e. of PVT), but since PVT are related through the equation of state, it is necessary to specify
only two of these quantities. Thus, for example if we express S as a function of T and P, infinitesimal increases in these will
give rise to an infinitesimal increase in S given by
∂S ∂S
dS = ( ) dT + ( ) dP (12.8.1)
∂T ∂P
P T

CP
Now ( ∂S

∂T
) is (for a reversible process) T
(see equation 12.7.5), and ( ∂S

∂P
) is (by a Maxwell relation) equal to −( ∂V

∂T
) .
P T P

If we know CP as a function of temperature, and, if we know the equation of state, we can now calculate
T2 P2
dT ∂V
S2 − S1 = ∫ CP −∫ ( ) dP (12.8.2)
T1 T P1 ∂T
P

This will enable us to calculate the change in entropy of a substance provided that we know how the heat capacity varies with
temperature and provided that we know the equation of state.

For an ideal gas ( ∂V

∂T
) = R/P , and so we obtain, for an ideal gas
P

T2
dT
S2 − S1 = ∫ CP − R ln(P2 / P1 ). (12.8.3)
T1
T

If we want to express the increase of entropy in terms of the change in temperature and volume, and of CV, we can use PV =
RT and CP = CV + R to obtain
T2
dT
S2 − S1 = ∫ CV + R ln(V2 / V1 ) (12.8.4)
T1 T

This agrees with what we had in the previous section for an isothermal expansion.
Here’s another way or arriving at equation 12.9.4. We want to find the change in entropy of a mole of an ideal gas in going
from (P1, V1, T1) to (P2, V2, T2). Since the change in entropy is route-independent, we can choose any simple route for which
the calculation is easy. Let’s go at constant volume from (P1, V1, T1) to (P3, V1, T2) and then at constant temperature from (P3,
V1, T2) to (P2, V2, T2).

T2
To go from (P1, V1, T1) to (P3, V1, T2), the gas has to absorb an amount of heat ∫
T1
CV dT , and so its entropy increases by
T2
∫
T1
C V . To go from (P3, V1, T2) to (P2, V2, T2). The gas does work RT2 ln(V2/V1) without any change in internal energy
dT

(because the internal energy of an ideal gas at constant temperature is independent of its volume), and therefore it absorbs this
amount of heat. Therefore its entropy increases by R ln(V2/V1). Thus we arrive again at equation 12.9.4.

Jeremy Tatum 4/29/2021 12.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8635

Example: If the substance is an ideal monatomic gas, then C P =
5

2
R . From this we calculate
5/2
5 T2 P2 T2 P1
S2 − S1 = R ln( ) − R ln( ) = R ln[ ( ) ]. (12.8.5)
2 T1 P1 T1 P2

Exercise: Go through the same analysis, but starting from S = S(T, V). Show that the result you get for an ideal gas is the same
as above. It will also, of course, necessarily be the same for any substance, though the equality of the expression you get with
equation 12.9.2 may not be immediately apparent.
Exercise: The pressure and volume of an ideal monatomic gas are both doubled. What is the ratio of the new temperature to
the old? What is the increase in the molar entropy?
(I make the answer 2.31 × 104 J kmole−1 K−1.) Now try the same problem with an ideal diatomic gas. (I make the answer 3.46
× 104 J kmole−1 K−1.)
Internal Energy and Enthalpy
These can be calculated if we know how CV and CP vary with temperature, because, by definition, CV = (∂U/∂T)V and CP =
(∂H/∂T)P.
Therefore
T2

U2 − U1 = ∫ CV dT (12.8.6)
T1

and
T2

H2 − H1 = ∫ CP dT . (12.8.7)
T1

Helmholtz and Gibbs Functions

Since A = U − TS, we have
A2 − A1 = U2 − U1 − T2 (S2 − S1 ) − S1 (T2 − T1 ) . (12.8.8)

In the special case of an ideal gas, we obtain

T2 T2
CV dT
A2 − A1 = ∫ CV dT − T2 ∫ − RT2 ln(V2 / V1 ) − S1 (T2 − T1 ) . (12.8.9)
T1 T1
T

Since G = H − TS, we have

G2 − G1 = H2 − H1 − T2 (S2 − S1 ) − S1 (T2 − T1 ) (12.8.10)

In the special case of an ideal gas, we obtain

T2 T2
CP dT
G2 − G1 = ∫ Cp dT − T2 ∫ − RT2 ln(P1 / P2 ) − S1 (T2 − T1 ) . (12.8.11)
T1 T1
T

There is, however, a serious difficulty with equations 12.9.9 and 12.9.11, in that, in order to calculate the change in the
Helmholtz and Gibbs functions, we need to know the initial absolute entropy S1.

Jeremy Tatum 4/29/2021 12.8.2 CC-BY-NC https://phys.libretexts.org/@go/page/8635

12.9: Absolute Entropy
We can, of course, calculate the molar entropy of a substance at some temperature provided that we define the entropy at a
temperature of absolute zero to be zero. By way of example, assuming that the molar entropy of hydrogen at 0 K is zero,
calculate the absolute entropy of a kmole of H2 gas at a temperature of 25oC (298.15 K) and a pressure of one atmosphere. We
can do this in five stages, as follows. You will find it helpful to sketch these stages on a drawing similar to figure VI.5.
1. Heat the solid hydrogen from 0 K to 13.95 K at a pressure of 7173 Pa. (That’s the triple point.) The increase in entropy is
∫CP d(lnT). Assuming that we know CP as a function of temperature in this range, that comes to 2080 J K−1 kmole−1.
2. Liquefy it at the same temperature and pressure. The molar latent heat of fusion is 117000 J kmole−1. Increase in entropy =
117000/13.95 = 8400 J K−1 kmole−1.
3. Vaporize it at the same temperature and pressure. The molar latent heat of vaporization is 911000 kmole−1. Increase in
entropy = 911000/13.95 = 65300 J K−1 kmole−1.
4. Increase temperature to 298.15 K at constant pressure. See equation 12.9.3. The increase in entropy is ∫CP d(lnT). Assuming
that we know CP as a function of temperature in this range, that comes to 70000 J K−1 kmole−1.
5. Increase pressure to 1 atmos = 1.013 × 105 Pa at constant temperature. See equation 12.9.4, from which we see that there is
a decrease of entropy equal to R ln(P2/P1) = 8314ln(1.103 × 105 / 7173) = 22000 J K−1 kmole−1.
Hence, taking the entropy to be zero at 0 K, the required entropy is 124000 J K−1 kmole−1.
Now that we have calculated the absolute entropy at a given temperature and pressure, we can calculate the increase in the
Helmholtz and Gibbs functions from equations 12.9.9 and 12.9.11. But this leaves us in a rather uncomfortable position. After
all, all we have done in this example is to calculate the increase in entropy as we took the sample up to 25 oC and 1
atmosphere – we haven’t really calculated the absolute entropy. The entropy appearing in equations 12.9.9 and 12.9.11 is
surely the absolute entropy, and we cannot calculate this unless we know the entropy at T = 0 K. This slight puzzle will remain
with us until Chapter 16, when we meet Nernst’s Heat Theorem and the Third Law of Thermodynamics.
Many of the examples of thermodynamical calculations have hitherto involved PdV work in a system in which the working
substance has been an ideal gas. Let us now look at two entirely different situations, both involving non-PdV work. Let us look
at charging a battery, and creating new surface by distorting a spherical drop of liquid.

Jeremy Tatum 4/8/2021 12.9.1 CC-BY-NC https://phys.libretexts.org/@go/page/8636

12.10: Charging a Battery
The concept of “non-PdV work” sometimes causes difficulty, so am going to illustrate it in this section by using the charging
of a battery as an example, and in the next section by a discussion of surface tension. This section will also give us an
opportunity of using a Gibbs-Helmholtz relation.
Suppose that we force a charge q into an electric cell whose electromotive force (EMF) is E, at constant temperature and
pressure. What is the increase in the Gibbs function of the cell? And what is the increase in its enthalpy?
The answer to the first question is easy. It is just qE. The increase in the enthalpy is given by
ΔH = ΔG + T ΔS

and, by a Maxwell relation (equation 12.6.12a), this is

∂G
ΔH = ΔG − T Δ( ) , (12.10.1)
∂T
P

which is one of the Gibbs-Helmholtz relations. But since ∆G = qE, this becomes
∂E ∂E
ΔH = qE − T q ( ) = q [E − T ( ) ]. (12.10.2)
∂T ∂T
P P

Thus we can calculate the increase in enthalpy from a measurement of how the EMF of the cell changes with temperature.

Jeremy Tatum 3/25/2021 12.10.1 CC-BY-NC https://phys.libretexts.org/@go/page/8637

12.11: Surface Energy
For a second example of non-PdV work we shall consider the phenomenon of “surface tension”.
It is well known that a liquid tends to contract to a shape that minimizes its surface area. In the absence of other forces, this
means that it will become spherical. The effect is often conveniently described in terms of “surface tension”. We describe the
tendency of a surface to contract by drawing an imaginary line in the surface, and we say that the surface to one side of the line
pulls the surface of the other, and we call the force per unit length perpendicular to the line the surface tension. It is expressed
in dynes per cm or newtons per metre. In this section I shall use the following symbols:
Surface tension: Γ
Area: σ
However, from the point of view of thermodynamics, it is easier to think of surface energy. How much work is needed to
increase the surface area? And how is this related to what we have described as “surface tension”? It may be noted in passing
that energy per unit area (J m−2) is dimensionally similar to force per unit length (N m−1).
A non-spherical blob of liquid will, under the action of surface tension, contract into a spherical blob – i.e. a blob of least
surface area for a given volume. It should not come as a surprise to learn that, at least in principle, as the blob adjusts (in an
adiabatic process) to its spherical shape of least surface area, it becomes warmer. Molecules near the surface have a high
potential energy. As many of them fall beneath the surface as the surface area is decreased, this potential energy is converted to
kinetic energy. Conversely, if a spherical drop is distorted from its spherical shape, it becomes cooler.
We have already pointed out that the surface tension can be regarded as the work required to create new area. Increasing the
area will result in a fall in temperature, so, if the temperature is kept constant, some heat must be absorbed from the
surroundings, and hence the increase in the internal energy is a little more than the surface tension. It may at first seem
surprising that doing work on a liquid, in order to create new surface, results in a fall of temperature, but the work is being
used not to increase the kinetic energy of the molecules, but rather to increase their potential energy by pulling them to the
surface.
One way in which we can imagine work being done on a liquid to increase its surface area is simply to imagine distorting a
spherical drop into a nonspherical shape. Another way, which might lend itself more easily to the sort of thermodynamical
analysis we are accustomed to in discussing gases, is to imagine a film of soapy water held in a wire frame, constructed of a
fixed U-shaped portion A (see figure XII.1), and a bridge B which we can move in and out, allowing us to do work on the
liquid by pulling it to the right, or the liquid to do work by pulling the bridge to the left. We could even refer to these two parts
as the “cylinder” A and the “piston” B. A difference between this picture and that of a gas inside a real cylinder is that when
we pull the “piston” out, we are doing work on the liquid. Nevertheless, as explained above, the temperature of the liquid then
drops. If we allow the film to contract and to pull the “piston” to the left, the temperature will rise.

If the width of the “cylinder” is a, the surface tension force with which the liquid is pulling on the “piston” is 2aΓ, where Γ is
the surface tension. The factor 2 arises because there are two surfaces, above and below. If we pull the piston to the right
through a distance dx, the work we do on the liquid is 2aΓdx. If we do this adiabatically (quickly), the liquid cools. If we do it
isothermally (slowly), the liquid has to absorb some heat from its surroundings.

Jeremy Tatum 4/1/2021 12.11.1 CC-BY-NC https://phys.libretexts.org/@go/page/8638

Let us now take the liquid around a Carnot cycle, as shown in figure XII.2. Notice that, as we move the “piston” to the right,
provided that the temperature remains constant the surface tension force between the “piston” and the liquid does not change;
thus the isotherms are horizontal lines, with the warmer isotherms lying lower than the cooler isotherms.
Let us start by moving the piston to the right, isothermally at a temperature T1, through a distance ∆x, being the portion AB of
figure XII.2. The work done on the liquid is 2aΓ1 ∆x, where Γ1 is the surface tension at temperature T1. In order that the
process should be isothermal, the liquid has to absorb an amount of heat Q1 from its surroundings. The internal energy
increases by 2aΓ1∆x + Q1.
Now expand the liquid further, but this time adiabatically, from B to C. Work is being done on the liquid, but no heat is being
absorbed. The temperature drops to T2. The new surface tension is Γ2, which is greater than Γ1, because surface tension
generally decreases at warmer temperatures.
Now allow the liquid to contract isothermally at temperature T2, from C to D. The liquid does an amount of work 2aΓ2∆x, and
it must lose an amount of heat Q2 (which, as we shall see, is less than Q1) to its surroundings. The internal energy decreases by
2aΓ2∆x + Q2.
Finally, return the liquid to its original state A along the adiabatic path DA. As many molecules on the surface fall back
beneath the surface, the temperature rises to its original value T1. Work is being done by the liquid; the work done by the liquid
along DA is equal to the work done on it along BC.
The net work done by the liquid around the complete cycle is 2a(Γ2 − Γ1)∆x and the net heat absorbed by the liquid around the
cycle is Q1 − Q2. Since there is no change in the internal energy around the cycle (because U is a function of state), these two
are equal. Also, there is no change in entropy around the cycle (because S is a function of state), and therefore Q2/T2 = Q1/T1.
(This justifies our earlier assertion that Q2 < Q1.)
From these two equations we obtain
Q1
(T1 − T2 ) = 2a (Γ2 − Γ1 ) Δx. (12.11.1)
T1

Go to the infinitesimal limit and drop the subscripts, and this becomes
dΓ
Q = −T × 2aΔx. (12.11.2)
dT

The right hand side is a positive quantity, because is negative. We have seen that, in order to create new surface
dΓ

dT

isothermally, heat must be absorbed. What equation 12.12.2 says is that the heat absorbed to create the new area ∆σ = 2a∆x
created is equal to Q = −T × Δσ .
dΓ

dT

Now the work required to create the new area is Γ ×∆σ.

Thus the increase in internal energy when new area dσ is created at constant temperature is

Jeremy Tatum 4/1/2021 12.11.2 CC-BY-NC https://phys.libretexts.org/@go/page/8638

 dΓ
ΔU = (Γ − T ) Δσ. (12.11.3)
dT

This will remind you of equation 12.11.1, ΔH = ΔG − T Δ(

∂G

∂T
) , for the increase in enthalpy of a battery when we add
P

charge to it at constant pressure. This time we are adding new area to a liquid at constant volume.
Here is another way at arriving at the same result: It will remind you of the way in which, in this Chapter, we derived the
expression for the Joule coefficient.
The increase in internal energy and Helmholtz functions of a system when we add heat to it and do work on it is given by the
familiar equations

dU = T dS − P dV + ∑ XdY (12.11.4)

and

dA = −SdT − P dV + ∑ XdY . (12.11.5)

We are most familiar with them when the term ∑XdY is zero, but in this case we are dealing with a liquid at constant volume,
and the one XdY term is Γdσ, so that the equations become
dU = T dS + Γdσ (12.11.6)

and
dA = −SdT + Γdσ. (12.11.7)

Divide equation 12.12.6 by dσ at constant temperature:

∂U ∂S
( ) = T( ) + Γ. (12.11.8)
∂σ ∂σ
T T

From equation 12.12.7 obtain a Maxwell relation:

∂S ∂Γ
( ) = −( ) , (12.11.9)
∂σ ∂T
T σ

except that Γ is in any case independent of σ, so the right hand term is actually a total derivative, dΓ/dT.
Substitute this into equation 12.12.8 and we have the same result as in our previous argument:
∂U dΓ
( ) = Γ−T . (12.11.10)
∂σ dT
T

In summary, the increase in internal energy in creating dσ of new surface at constant temperature is the sum of the work
required, Γdσ, and the heat absorbed, −T dσ .
dΓ

dT

Here’s yet another way of getting there! It will remind you of the way in which we derived the expression for the Joule
coefficient in Chapter 10. In general the internal energy of a drop of liquid depends on its volume, temperature and surface
area:

U = U (V , T , σ). (12.11.11)

However, let us ignore the very small change in energy resulting from the very small amount of PdV work that the drop would
do if it expands a tiny bit as a result of temperature increase. We shall be concerned only with internal energy as a function of
temperature and of surface tension (which may vary with temperature.) Thus, we’ll assume

U = U (T , σ). (12.11.12)

For infinitesimal increases in temperature and surface tension, the corresponding increase in the internal energy is
∂U ∂U
dU = ( ) dT + ( ) dσ. (12.11.13)
∂T ∂σ
σ T

Jeremy Tatum 4/1/2021 12.11.3 CC-BY-NC https://phys.libretexts.org/@go/page/8638

The internal energy could increase by the addition of heat to the drop, dQ, plus work done on it, dW. The former is TdS, and
the latter is +Γdσ. Thus
dU = T dS + Γdσ. (12.11.14)

From these we obtain

1 ∂U ∂U
dS = [( ) dT + {( ) − Γ} dσ] . (12.11.15)
T ∂T ∂σ
σ T

Since entropy is a function of state, dS is an exact differential, and therefore

1 ∂ ∂U ∂ 1 ∂U Γ
( ) = [ ( ) − ]. (12.11.16)
T ∂σ ∂T ∂T T ∂σ T
σ T

2 2
1 ∂ U 1 ∂U 1 ∂ U Γ 1 ∂Γ
=− ( ) + + − ( ) . (12.11.17)
2 2
T ∂σ∂T T ∂σ T ∂T ∂σ T T ∂T
T σ

Therefore
∂U ∂Γ
( ) = Γ−T( ) . (12.11.18)
∂σ ∂T
T σ

Again, we point out that Γ cannot in any case depend on σ, so that last derivative is really a total derivative, so that
∂U dΓ
( ) = Γ−T (12.11.19)
∂σ dT
T

Surface tension generally decreases with temperature, so this equation shows that the increase of internal energy at constant
temperature per unit new area is a little greater than the surface tension, as expected.
Can we calculate the fall in temperature if new area is created adiabatically and reversibly (i.e. isentropically)? Yes, because
equation 12.12.15 (with dS = 0) tells us that then
∂U ∂U
( ) dT = − [( ) − Γ] dσ. (12.11.20)
∂T ∂σ
σ T

On making use of equation 12.12.19, we obtain

∂U ∂Γ
( ) dT = T ( ) dσ. (12.11.21)
∂T ∂T
σ σ

We are assuming that the volume is constant so that ( ∂U

∂T
) = CV , and therefore the increase in temperature with area is
0

T ∂Γ T dΓ
dT = ( ) dσ = dσ. (12.11.22)
CV ∂T CV dT
σ

Since dΓ

dT
is generally negative, this means that the temperature falls as the area is increased, as expected. In this equation, if
dσ means the increase in area of a sample, in m2, then CV means the heat capacity of that sample, in J K−1.
Measurement of the surface tension of a liquid is very sensitive to how clean the surface is, but, for the record, the following
figures for the surface tension of clean water in contact with air are taken from the Website
www.engineeringtoolbox.com/watersurface-tension-d_597.html

Temperature - t - (oC) Surface Tension in contact with air - Γ - (N/m)

0 0.0756

5 0.0749

10 0.0742

20 0.0728

Jeremy Tatum 4/1/2021 12.11.4 CC-BY-NC https://phys.libretexts.org/@go/page/8638

30 0.0712

40 0.0696

50 0.0679

60 0.0662

70 0.0644

80 0.0626

90 0.0608

100 0.0589

Exercise: A drop of water 1 mm in diameter at 45 °C is broken up into two equal droplets, each half the volume of the original
drop. Calculate the change in temperature, and say whether it is cooler or warmer.

Jeremy Tatum 4/1/2021 12.11.5 CC-BY-NC https://phys.libretexts.org/@go/page/8638

12.12: Fugacity
Problem: The pressure of a mole of an ideal gas is increased isothermally from P0 to P. What is the increase G − G0 in its
Gibbs free energy?

Solution: By integration of equation 12.6.12b, ( ∂G

∂P
) =V , or by use of dG = −SdT + V dP , we have
T

G − G0 = ∫ V dP . (12.12.1)
P0

For a mole of an ideal gas, V = RT/P, and hence

G − G0 = RT ln(P / P0 ), (12.12.2)

which agrees with equation 12.9.11.

Equation 12.13.1 enables us to calculate the change in the Gibbs free energy of a substance while its pressure is increased at
constant temperature. Equation 12.13.2 gives the result for a mole of an ideal gas. If the substance is not an ideal gas, then we
need to know the equation of state, V = V(P , T) in order to integrate equation 12.13.1. For example, the equation of state for a
van der Waals gas is (P + a/V ) (V − b) = RT , where V is the molar volume, or P V − (bP + RT )V + aV − ab = 0 .
2 3 2

Integrating equation 12.13.1 with this van der Waals equation of state may appear formidable. I am grateful to Dr J.
Visvanathan of Chennai, India, for pointing out that it is not necessary. Instead one can calculate the change in the Helmholtz
V
function, which, at constant temperature, is given by A − A = − ∫ P dV , which is easy, and then use
0
V0

G−G = A −A +PV −P V
0 0 . I am also indebted to Dr Justin Albert for pointing out that this amounts to integrating
0 0

P
∫
P0
V dP by parts, even if you had never heard of the Helmholtz function!
The fugacity f of a substance is defined in such a manner that, if the molar Gibbs free energy increases from G0 to G, the ratio
of the new fugacity to the initial fugacity, f/f0, is given by
G − G0 = RT ln(f / f0 ) (12.12.3)

.
In other words, for a real substance, we can use all (or at least most!) of the equations that we know for an ideal gas as long as
we substitute fugacity for pressure.
That is,
G − G0
f / f0 = exp( ). (12.12.4)
RT

As for internal energy, only the difference between the Gibbs free energies of two states can be defined; likewise, only the
ratio of the fugacities of two states is defined.
Combining equations 12.13.4 and 12.13.1 we obtain
P
1
ln(f / f0 ) = ∫ V dP , (12.12.5)
RT P0

which should enable us to find the relation between pressure and fugacity if we know the equation of state.
We note also that at very low pressures, a real gas behaves more and more like an ideal gas, and we can define the fugacity in
units of pressure (pascal) in such a manner that, in the limit, as the pressure approaches zero, the fugacity equals the pressure.
Indeed, we can then define the ratio of the fugacity to the pressure as the activity coefficient, which has the value unity at zero
pressure.
Problem: Show that for a substance having the equation of state P(V − b) = RT (V = molar volume), as the pressure increases
from P0 to P, the ratio of the final to initial fugacities is

Jeremy Tatum 3/25/2021 12.12.1 CC-BY-NC https://phys.libretexts.org/@go/page/8639

 b (P − P0 )
ln(f / f0 ) = ln(P / P0 ) + . (12.12.6)
RT

That is,
b (P − P0 )
ln f − ln f0 = ln P − ln P0 + . (12.12.7)
RT

Now suppose that P0 is very small, and in the limit, as P0 → 0, f0 → P0. We now find that the fugacity at temperature T and
pressure P is given by
bP
ln f = ln P + . (12.12.8)
RT

This can be written

f bP
= exp( ). (12.12.9)
P RT

The ratio f/P is called the activity coefficient. You can see that f ≈ P if P is small, or if b is small, as expected.

Jeremy Tatum 3/25/2021 12.12.2 CC-BY-NC https://phys.libretexts.org/@go/page/8639

CHAPTER OVERVIEW
13: EXPANSION, COMPRESSION AND THE TDS EQUATIONS
Topic hierarchy

13.1: COEFFICIENT OF EXPANSION

13.2: COMPRESSION
13.3: PRESSURE AND TEMPERATURE
13.4: THE TDS EQUATIONS
13.5: EXPANSION, COMPRESSION AND THE TDS EQUATIONS
13.6: YOUNG'S MODULUS
13.7: RIGIDITY MODULUS (SHEAR MODULUS)
13.8: VOLUME, TEMPERATURE AND THE GRÜNEISEN PARAMETER

1 4/29/2021
13.1: Coefficient of Expansion
Notation: In an ideal world, I’d use α, β, γ respectively for the coefficients of linear, area and volume expansion. Unfortunately
we need γ for the ratio of heat capacities. Many people use β for volume expansion, so I’ll follow that. What, then, to use for
area expansion? I’ll use b, so we now have α, b, β, which is very clumsy. However, we shall rarely need b, so maybe we can
survive.
Coefficient of linear expansion: α
Coefficient of area expansion: b
Coefficient of volume expansion: β
For small ranges of temperature, the increases in length, area and volume with temperature can be represented by
l2 = l1 [1 + α̂ (T2 − T1 )] (13.1.1)

^
A2 = A1 [1 + b (T2 − T1 )] (13.1.2)

and

V2 = V1 [1 + β̂ (T2 − T1 )] (13.1.3)

Here α b and β are the approximate coefficients of linear, area and volume expansion respectively over the temperature range
^, ^ ˆ

T1 to T2. For all three, the units are degree−1 – that is Cº−1 or K−1.
For anisotropic crystals, the coefficient may be different in different directions, but for isotropic materials we can write
2 2 2 ~
A2 = l =l [1 + α
^ (T2 − T1 )] = A1 [1 + 2 α (T2 − T1 ) + …] (13.1.4)
2 1

3 3 3 ~
V2 = l =l [1 + α
^ (T2 − T1 )] = V1 [1 + 3 α (T2 − T1 ) + …] (13.1.5)
2 1

Thus for small expansions, ^b ≈ 2α

~
and βˆ ≈ 3α
^.

Equations 13.1.1, 2 and 3 define the approximate coefficients over a finite temperature range. The coefficients at a particular
temperature are defined in terms of the derivatives, i.e.
1 ∂l
α = ( ) , (13.1.6)
l ∂T
P

1 ∂A
b = ( ) (13.1.7)
A ∂T
P

1 ∂V
β = ( ) . (13.1.8)
V ∂T
P

The relations b = 2α and β = 3α are exact.

We specify “at constant pressure” because obviously we don’t want, in our definition, to prevent the material from expanding
by increasing the pressure on it when we heat it.
For solids, the coefficient of linear expansion is usually the appropriate parameter; for liquids and gases the volume coefficient
is usually appropriate. For most familiar common metals the coefficient of linear expansion is of order 10−5 K−1. Alloys such
as the nickel-steel alloy, “invar”, used in clock construction, may have much smaller coefficients. Ordinary glass has a
coefficient only a little less than that of metals; pyrex and fused quartz have a much smaller expansion – hence their use in
telescope mirrors. For liquids and gases it is usually the volume coefficient that is quoted. The volume coefficient of mercury
is about 0.00018 K−1. Water actually contracts between 0 and 4 oC, and expands above that temperature. The volume
coefficient of air at 0 oC is 0.0037 K−1.
At room temperatures and above, the coefficient of linear expansion of metals doesn’t vary a huge amount with temperature,
but at low temperatures the coefficient of expansion varies much more rapidly with temperature – and so does the specific heat

Jeremy Tatum 3/25/2021 13.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7291

capacity (see Section 8.10). Indeed, for a given metal, the variation of expansion coefficient and the specific heat capacity vary
with temperature in a rather similar manner, so that, for a given metal, the ratio α/CP is constant over a large temperature
range.
Exercise: A square metal plate has a circular hole of area 300 cm2 in the middle of it. If the coefficient of linear expansion is 2
× 10−5 Cº−1, calculate the area of the hole when the temperature of the plate is raised through 100 degrees.
Exercise: Show that the coefficient of volume expansion of an ideal gas is 1/T. Compare this with the numerical value for air
given above.
Although classical thermodynamics does not deal with detailed microscopic processes, it is of interest to ask why a solid
material expands upon heating. Let us imagine a crystalline solid to be made up of atoms connected to each other by little
springs, and each spring is governed by Hooke’s Law, and consequently each atom is vibrating in a parabolic potential well
and is moving in simple harmonic motion. If we increase the temperature, we increase the amplitude of the vibrations, but we
do not change the mean positions of the atoms. Consequently, in such a model, we would not expect any expansion upon
heating. However, the real potential is not parabolic, but is shaped, at least qualitatively, something like the Lennard-Jones or
Morse potentials mentioned in Chapter 6, Section 6.8. If the material is heated, the amplitude of the vibrations increases, and,
because of the higher-order terms in the potential, which give the potential its asymmetric anharmonic shape, the mean
separation of the atoms does indeed increase, and so we have expansion. Thus the expansion upon heating of a solid material
is a consequence of the anharmonicity of the atomic vibrations and the asymmetry of the potential in which they are moving.
In the next two exercises, I shall be thinking of the expansion of a metal rod as the temperature is increased, and the pressure
will be assumed to be constant at all times. Thus I am going to assume that pressure is not a variable in the discussion, and I
shall define the coefficient of linear expansion as α = 1

l
dl

dT
rather than the more general 1

l
(
∂l

∂T
) . A small point that I make
P

at this stage is this: Suppose that the length of a metal rod increases linearly with temperature, so that that this does not dl

dT

mean that the coefficient of expansion is independent of temperature. And if α is independent of temperature, l does not
increase linearly with temperature. The next two exercises will illustrate that, and will also illustrate how the exact coefficient
l2 −l1
α =
1

l
dl

dT
is related to what I have called (for want of a better term) the “approximate” coefficient α̂ = 1

l1 T2 −T1
.

Exercise. Suppose that the length of a metal rod increases with temperature according to l = l (1 + α T ) where l0 is the 0 0

length at 0 K, and α0 is the coefficient at 0 K. This means that and lα are independent of temperature, and each is equal to
dl

dT

l0α0. Show that the coefficient at temperature T is given by

α0
α = . (13.1.9)
1 + α0 T

Show that α̂ , the approximate coefficient over the temperature range T1 to T2, is equal to the exact coefficient α evaluated at T
= T1.
Exercise. Suppose that the coefficient α is independent of temperature. Show that the length of the rod increases with
α(T −T )

temperature according to l = l , where l0 is the length at 0 K. Show also that α .

i 1
aT e −1
0e
^ =
T2 −T1

By this time, it may have occurred to the reader that what we have called α ) , for all its usefulness in the equation
l = l [1 + α̂ (T − T )] , is not “the” coefficient of expansion at temperature T1, nor is it the mean coefficient in the
2 1 2 1

T2
temperature range T1 to T2. The mean coefficient in this range must be defined by α (T − T ) = ∫ αdT . So now, one more
¯¯
¯
2 1 T1

exercise:
Exercise. Suppose that the length of a metal rod increases with temperature according to l = l0 (1 + α0 T ) , where l0 is the
length at 0 K, and α0 is the coefficient at 0 K. Show that
1 1 + α0 T2
¯¯
¯
α = ln( ). (13.1.10)
(T2 − T1 ) 1 + α0 T1

Summary

Jeremy Tatum 3/25/2021 13.1.2 CC-BY-NC https://phys.libretexts.org/@go/page/7291

 dl
 If   is constant   If α is constant 
dT

αT
l = l0 (1 + α0 T ) l = l0 e
α0
α = α = α0
1+α0 T
(13.1.11)
α(T −T )
i
e −1
α̂ = α1 α̂ =
T2 −T1

1 1+α0 T2
¯¯ ¯¯
ᾱ = ln( ) ᾱ =α
( T2 −T1 ) 1+α0 T1

Of course, you may feel that this distinction between α, α , α

^ and α is splitting hairs. Let us discover for ourselves how much
0
¯¯
¯

they differ, by putting in some numbers. Let us suppose that α0 = 1.7 × 10−5 K−1 and that l0 = 1 m. Then, assuming that T1 =
280 K (6.85 °C) and T2 = 380 K (106.85 °C), we obtain
dl
 If   is constant   If α is constant 
dt

l1 = 1.004760m 1.004771m

−5 −1 −5 −1
α(280K) = 1.691946 × 10 K α(280K) = 1.700000 × 0 K

−5 −1 −5 −1
α̂ = 1.691946 × 10 K ^ = 1.701446 × 10
α K

¯¯
¯ −5 −1 ¯¯
¯ −5 −1
α = 1.690516 × 10 K α = 1.700000 × 0 K

In general, if the length at T1 is l1, the length l2 at T2 will be given by

T2

l2 = l1 exp( ∫ αdT ). (13.1.12)

T1

α0
In the case where dl/dT is constant, so that α = 1+α0 T
, this becomes

1 + α0 T2
2
l2 = l1 ( ) = l1 (1 + α0 (T2 − T1 ) − α T1 (T2 − T1 ) + …) . (13.1.13)
0
1 + α0 T1

In the case where α is constant, so it becomes

1 2 2
l2 = l1 exp(α (T2 − T1 )) = l1 (1 + α (T2 − T1 ) + α (T2 − T1 ) + …) (13.1.14)
2

Thus to the first order of small quantities, all varieties of α are equal.
Coefficient of Expansion as a Tensor Quantity. In Chapter 4, I briefly mentioned that, in the case of an anistropic crystal, the
coefficient of thermal conduction is a tensor quantity. The same is true, for an anisotropic crystal, of the coefficient of
expansion. Thus, if, during an physics examination, you were asked to give examples of tensor quantities, you could give these
as examples – though a small risk might be involved if your teacher had not thought of these as tensors! The coefficient of
expansion of an anisotropic crystal may vary in different directions. (In Iceland Spar – calcium carbonate – in one direction the
coefficient is actually negative.) If you cut an anisotropic crystal in the form of a cube, whose edges are not parallel to the
crystallographic axis, the sample, upon heating, will not only expand in volume, but it will change in shape to become a non-
rectangular parallelepiped. However, it is possible to cut the crystal in the form of a cube such that, upon heating, the sample
expands to a rectangular parallelepiped. The edges of the cube (and the resulting parallelepiped) are then parallel to the
principal axes of expansion, and the coefficients in these directions are the principal coefficients of expansion. These
directions will be parallel to the crystallographic axes if the crystal has one of more axes of symmetry (but obviously not
otherwise)

Jeremy Tatum 3/25/2021 13.1.3 CC-BY-NC https://phys.libretexts.org/@go/page/7291

13.2: Compression
The way in which the volume of a material decreases with pressure at constant temperature is described by the isothermal
compressibility, κ:
1 ∂V
κ =− ( ) . (13.2.1)
V ∂P
T

Note the necessary minus sign.

Later, we shall need to distinguish between “isothermal compressibility” and “adiabatic compressibility”, and we shall need a
subscript to the symbol κ in order to distinguish between the two. For the time being, however, κ with no subscript will be
taken to mean the isothermal compressibility.
The reciprocal of κ is called the isothermal bulk modulus, sometimes (understandably) called the isothermal incompressibility.
Question: What are the SI units for compressibility and bulk modulus?
Exercise: Show that the isothermal compressibility of an ideal gas is 1/P.
Exercise: What is the bulk modulus of air at atmospheric pressure?

Jeremy Tatum 3/25/2021 13.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7292

13.3: Pressure and Temperature
The way in which the pressure of a material increases with temperature at constant volume is described by ( ∂P

∂T
) .
V

Exercise: By making use of equation 2.4.11, show that

∂P β
( ) = . (13.3.1)
∂T κ
V

Exercise: By making use of equation 10.4.8, show that

2 2
TV β 9T V α
CP − CV = = . (13.3.2)
κ κ

Thus we can determine CP − CV from measurements of the expansion coefficient and the isothermal compressibility without
knowing the equation of state. We have already shown that the expansion coefficient of an ideal gas is 1/T, and the isothermal
compressibility of an ideal gas is 1/P. Note that, for an ideal gas, β = 1/T and κ = 1/P, so that equation 13.3.2 reduces to R.
Note that, in equation 13.3.2, κ is the isothermal compressibility. CP and CV may denote the molar heat capacities (in which
case V is the molar volume); or they may denote the specific heat capacities (in which case V is the specific volume or
reciprocal of density); or they may denote the total heat capacities (in which case V is the total volume).
Recall that the physical reason that CP is greater than CV is that when a substance is heated and expands at constant pressure, it
does work, whereas if held at constant volume it does no work. In the case of an ideal gas expanding reversibly, the work done
is all external work. A real gas, or a van der Waals gas, on expanding also does internal work against the intermolecular forces.
Therefore CP is greater than CV by more than R − but only a little more, because the intermolecular (van der Waals) forces are
not very large. In Chapter 10 we developed an explicit expression for CP − CV for a van der Waals gas (equation 10.4.10).
When a solid is heated, it expands very little compared with a gas, and hence does very little external work. The intermolecular
forces, however, are quite large, and hence an expanding solid does quite a lot of internal work. Thus for a gas, most of the
work of expansion is external; for a solid, most of the work of expansion is internal.
Here are order-of-magnitude figures for copper at room temperature (for exact figures, we would have to specify the exact
temperature).
Specific heat capacity at constant pressure = 384 J K−1 kg−1
Molar mass (“atomic weight”) = 63.5 kg kmole−1
Molar heat capacity at constant pressure = 24400 J K−1 kmole−1 = 2.93 R.
Density = 8960 kg m−3
Molar volume = 7.09 × 10−3 m3 kmole−1
Coefficient of linear expansion = 1.67 × 10−5 K−1
Coefficient of volume expansion = 5.00 × 10−5 K−1
Isothermal bulk modulus = 1.40 × 1011 Pa
Isothermal compressibility = 7.14 × 10−12 Pa−1
Equation 13.3.2 will give us, at a temperature of 20 °C = 293.15 K, CP − CV (molar) = 728 J K−1 kmole−1 = 0.09R. CP − CV
(specific) = 11 J K−1 kg−1 This is only about 3 percent of CP.
Equation 13.3.2 raises an interesting problem concerning water. It will be understood that the reason why CP for an ideal gas is
greater than CV is as follows. When heat is added to an ideal gas at constant volume, all of the heat goes into raising the
temperature. When heat is added at constant pressure, however, some of the heat goes into doing external work. Hence CP >
CV. That argument is correct. However...
Water at 2 ºC (or indeed at any temperature in the range between 0 ºC and 4 ºC) contracts upon heating (i.e. β is negative), so
that, if we add heat at constant pressure, work is done on the water by its surroundings, and hence (we might argue, though

Jeremy Tatum 3/18/2021 13.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7293

erroneously), for water at 2 ºC, CP < CV. Equation 13.3.2, however, shows that CP ≥ CV regardless of the sign of β. (The
equality applies where β = 0, which occurs at 4 ºC.) Thus we have a paradox.
In fact, equation 13.3.2 is correct, and, at 2 ºC, . CP > CV. The explanation is as follows. It is true that, when heat is added to an
ideal gas at constant volume, all of the heat goes into raising the temperature – but this is true only for an ideal gas in which
the internal energy is all kinetic. But for real substances, including water, the correct statement (which is really just the first
law of thermodynamics) is that when heat is added to a substance at constant volume, all of the heat goes into raising the
internal energy, and, for a nonideal substance the internal energy is partly kinetic and partly potential. When we add heat
isobarically to water at 2 ºC, more of this heat goes into increasing the potential energy than if we add heat isochorically, and
hence CP is still greater than CV. A very clear account of this problem, from both the thermodynamical and statistical
mechanical points of view, is to be found in a paper by McDougall and Feistel, Deep-Sea Research I 50, 1523 (2003).
(You may remember a similar apparent paradox in connection with surface tension of a liquid. When we do work adiabatically
and reversibly to create new surface, the temperature drops. So doing work on a system or adding heat to it doesn’t necessarily
result in a rise in temperature. It does result in an increase of internal energy, which include potential energy.)

Jeremy Tatum 3/18/2021 13.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7293

13.4: The TdS Equations
The three TdS equations have been known to generations of students as the “tedious equations” − though they are not at all
tedious to a true lover of thermodynamics, because, among other things, they enable us to calculate the change of entropy
during various reversible processes in terms of either dV and dT, or dP and dT, or dV and dP, and even in terms of directly
measurable quantities such as the coefficient of expansion and the bulk modulus.
i.) We can express entropy in terms of any two of PVT. Let us first express entropy as a function of V and T
∂S ∂S
dS = ( ) dV + ( ) dT . (13.4.1)
∂V T
∂T V

Therefore
∂S ∂S
T dS = T ( ) dV + T ( ) dT . (13.4.2)
∂V ∂T
T V

From a Maxwell relation (equation 12.6.15), (

∂S

∂V
) =(
∂P

∂T
) . Also, in a constant volume process, TdS = dU so that
T V

T(
∂S

∂T
) =(
∂U

∂T
) = CV .
V V

Therefore T dS = T ( ∂P

∂T
) dV + CV dT .
V

This is the first of the TdS equations.

ii.) This time, let us express entropy as a function of P and T
∂S ∂S
dS = ( ) dP + ( ) dT . (13.4.3)
∂P ∂T
T P

Therefore
∂S ∂S
T dS = T ( ) dP + T ( ) dT . (13.4.4)
∂P ∂T
T P

From a Maxwell relation (equation 12.6.16), (

∂S

∂P
) = −(
∂V

∂T
) . Also, in a constant pressure process, TdS = dH so that
T P

T(
∂S

∂T
) =(
∂H

∂T
) = CP .
P P

Therefore
∂V
T dS = −T ( ) dP + CP dT . (13.4.5)
∂T
P

This is the second of the TdS equations.

iii.) If we express entropy as a function of P and V (recall that we can choose to express a function of state as a function of any
two of P, V or T) we have
∂S ∂S
dS = ( ) dP + ( ) dV . (13.4.6)
∂P ∂V
V P

Therefore
∂S ∂S
T dS = T ( ) dP + T ( ) dV . (13.4.7)
∂P ∂V
V P

In a constant volume process, TdS = CVdT, so that T ( ∂S

∂P
) = CV (
∂T

∂P
) .
V V

And in a constant pressure process, TdS = CPdT, so that

Jeremy Tatum 4/1/2021 13.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7294

 ∂S ∂T
T( ) = CP ( ) . (13.4.8)
∂V ∂V
p P

Therefore
∂T ∂T
T dS = CV ( ) dP + CP ( ) dV . (13.4.9)
∂P ∂V
V P

This is the third of the TdS equations.

In summary, then, these are the three TdS equations:
∂P
T dS = T ( ) dV + CV dT (13.4.10)
∂T
V

∂V
T dS = −T ( ) dP + CP dT (13.4.11)
∂T
P

∂T ∂T
T dS = CV ( ) dP + CP ( ) dV (13.4.12)
∂P ∂V
V P

These equations can be used, for example, to calculate, by integration, the change of entropy between one state and another,
provided that the equation of state is known in order that we can evaluate the partial derivatives.

Jeremy Tatum 4/1/2021 13.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7294

13.5: Expansion, Compression and the TdS Equations
It will be recalled, from equations 13.3.1 and 13.1.8, that
∂P β ∂V
( ) =  and  ( ) = βV . (13.5.1)
∂T κ ∂T
V P

That is,
∂T κ ∂T 1
( ) =  and  ( ) = . (13.5.2)
∂P β ∂V βV
V P

With these, the TdS equations become

β CV
dS = dV + dT (13.5.3)
κ T

Cp
dS = −βV dP + dT (13.5.4)
T

and
CV κ CP
dS = dP + dV (13.5.5)
Tβ T βV

These equations can be used, for example, to calculate, by integration, the change of entropy between one state and another,
provided that β, κ and the heat capacities are known as functions of temperature and pressure or specific volume. You don’t
even have to know the equation of state.
They won’t tell us anything about an ideal gas that we don’t already know, but let’s just apply them to an ideal gas in any case,
just to see if we have made any mistakes so far. For an ideal gas, as we saw in Sections 13.1 and 13.2, β = 1/T and κ = 1/P. The
first two TdS equations become
T dS = P dV + CV dT (13.5.6)

and
T dS = −V dP + CP dT . (13.5.7)

That is to say,
T dS = P dV + dU (13.5.8)

and
T dS = −V dP + dH (13.5.9)

so all is well with the world so far. The third equation becomes
dP dV
dS = CV + CP . (13.5.10)
P V

For a reversible adiabatic process, dS = 0, so what do you get if you integrate equation 13.5.10 for a reversible adiabatic
process for an ideal gas? This should complete your happiness – though there is more to come.
If a material (be it solid, liquid or gas) is compressed reversibly and adiabatically (i.e. dS = 0), equation 13.5.3 will tell you
how the temperature changes with volume:
∂T βT
( ) =− . (13.5.11)
∂V κCV
S

If it is the pressure, rather than the volume, that is changed reversibly and adiabatically, equation 13.5.4 will tell you how the
temperature changes with pressure:

Jeremy Tatum 4/15/2021 13.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7295

 ∂T βV T
( ) =+ . (13.5.12)
∂P CP
S

In equation 13.5.11, κ is the isothermal compressibility, defined in equation 13.2.1 as κ = − 1

V
(
∂V

∂P
) . To emphasize that this
T

is the isothermal compressibility, I’ll add a subscript: κiso. There is also a need to define an adiabatic compressibility,
κ ad  = −
1

V
(
∂V

∂P
) . (Note - I used the word adiabatic, but I used the subscript S to the partial derivative. Are the words
S

adiabatic and isentropic synonymous?) This is going to be less that the isothermal compressibility, because, if you try to
compress a material adiabatically it will become hot and therefore not be as readily compressible as if the compression were
CV κ  is o  CP
isothermal. Now refer to equation 13.5.5, dS = dP + dV . Divide both sides by dP and go to the infinitesimal
Tβ T βV

limit, recalling that in a reversible adiabatic process S is constant, and this equation then gives us CV κiso = −CP ⋅
1

V
(
∂V

∂P
) .
S

But − V
1
(
∂V

∂P
) is the adiabatic compressibility, and CP/CV = γ, CP CV so we arrive at
S

κiso
= γ, (13.5.13)
κad

where γ is the ratio of the isobaric and isochoric heat capacities. In particular, recall that, for an ideal gas, κiso = 1/P. Hence, for
an ideal gas, κad = 1/(γP).
2
T Vβ
In equation 13.3.2, we deduced the relation C − C = P. In equation 13.5.13, we have deduced an expression for the
V
κis o

ratio of the isothermal to adiabatic compressibilities, the isothermal compressibility being greater. Combining these now with γ
= CP/CV, we can now deduce an expression for the difference between the isothermal and adiabatic compressibilities, namely:
2
TV β
κ iso  − κ ad  = . (13.5.14)
CP

Bad
In terms of bulk modulus B, which is the reciprocal of compressibility, equations 13.5.13 and 13.5.14 are, of course, Bis o
=γ
2
T Vβ
and 1

Bis o
−
1

Bad
=
CP
.

Comparison of equations 13.3.2 and 13.5 14 shows that

CP κiso
= . (13.5.15)
CV κad

−−−−−−
Sir Isaac Newton in his Principia correctly deduced that the speed of sound in a gas is equal to √1/(ρκ), where ρ is the
density, and without making any distinction between κiso and κad. The measured speed was faster than predicted from his
theory, and Newton tried, not completely successfully, to account for the difference. I haven’t gone into the history, but there is
a story – probably apocryphal – that, in order to secure agreement between observation and theory, he “fudged his lab” and
“adjusted” his experimental results a little. But the trouble was not with the experimental results. If you take for κ the
isothermal value, namely 1/P for an ideal gas (to which air approximates quite well over the small pressure changes involved),
−−−−
the theory gives √P /ρ for the sound speed. In fact, however, the compressions and rarefactions in a sound wave are so rapid
−− −−−
that they are, in effect, adiabatic, so that it is the adiabatic compressibility κad that should be used, giving √γP /ρ as the
theoretical expression, which agrees well with the observed speed.

Jeremy Tatum 4/15/2021 13.5.2 CC-BY-NC https://phys.libretexts.org/@go/page/7295

13.6: Young's Modulus
This Section is under revision.

Jeremy Tatum 4/15/2021 13.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7296

13.7: Rigidity Modulus (Shear Modulus)
When we are discussing the bulk modulus of a material we are usually thinking in terms of applying pressure and noting the
compression, so the adiabatic bulk modulus is usually greater than the isothermal bulk modulus. We could in principle also
imagine a situation in which we are moving a material into a vacuum, thus decreasing the external pressure, and then
measuring the resulting expansion. In that case we would find that the adiabatic bulk modulus is less than the isothermal bulk
modulus – but that is a rather artificial situation. In Section 13.5 we derived (see equation 13.5.14) the usual relation for
compression:
2
1 1 β T
− = , (13.7.1)
Biso Bad ρCP

in which β is the volume coefficient of expansion, and CP is the specific heat capacity at constant pressure. (Compare this with
equations 13.6.12 and 13.6.20.)
We now must ask ourselves what is the difference between the adiabatic and isothermal rigidity moduli (also known as shear
modulus). If you are unfamiliar with the rigidity modulus, see my Classical Mechanics notes, Chapter 20, Section 20.3.
The rigidity modulus involves no change in volume or length, and hence there is no difference between the adiabatic and
isothermal rigidity moduli.

Jeremy Tatum 4/15/2021 13.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8647

13.8: Volume, Temperature and the Grüneisen Parameter
If you compress a material adiabatically and reversibly (i.e. isentropically) its temperature goes up. The amount by which it
goes up can be represented by the partial derivative ( ∂T

∂V
) . Here, V could mean the total volume , the specific volume or the
S

molar volume, according to context, and you would have to specify your units accordingly. The derivative is negative, because
the temperature goes up as the volume is decreased.

[Compare this with the definition of the volume coefficient of expansion β =

1

V
(
∂V

∂T
) , which is positive. Think about the
P

difference.]

A dimensionless version which also expresses the variation of temperature with volume would be V

T
(
∂T

∂V
) =(
∂ ln T

∂ ln V
) , and
S S

here there is no need to specify whether V means total, specific or molar. The derivative could also be written as −( ∂ ln T

∂ ln ρ
) ,
S

where ρ is the density. The positive value, −(

∂ ln T

∂ ln V
) = +(
∂ ln T

∂ ln ρ
) is called the Grüneisen parameter. We have already
S S

used the symbols G, g, Γ and γ for various things in these notes, so I am stuck for a suitable symbol. Sometimes non-italic
symbols are used for dimensionless parameters, such as R for Reynolds number in aerodynamics. Let’s try Gr for the
Grüneisen parameter.
For an ideal gas, the relation between volume and temperature in a reversible adiabatic expansion is TVγ − 1 = constant, and
therefore the Grüneisen parameter for an ideal gas is γ − 1.
In thinking about volume and temperature changes, we often have some sort of a gas (ideal or otherwise) in mind. However,
geophysicists have to deal with very large pressures in the interior of the Earth, where volume and temperature changes of
solids under pressure are not negligible, and geophysicists often make use of the Grüneisen parameter for solid materials.
For a bit of practice in deriving relationships between some of the quantities described in this chapter, see if you can show that
β β
Gr = = (13.8.1)
ρCV κiso ρCP κad

and
γ = 1 + Gr βT . (13.8.2)

If ρ in these questions stands for density (mass per unit volume), what, precisely, are CV and CP? Total, specific or molar? Or
does it not matter? What do these equations become in the case of an ideal gas?

Jeremy Tatum 4/8/2021 13.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8648

14: The Clausius-Clapeyron Equation
Before starting this chapter, it would probably be a good idea to re-read Sections 9.2 and 9.3 of Chapter 9.
The Clausius-Clapeyron equation relates the latent heat (heat of transformation) of vaporization or condensation to the rate of
change of vapour pressure with temperature. Or, in the case of a solid-liquid transformation, it relates the latent heat of fusion
or solidification to the rate of change of melting point with pressure.
Let us imagine a vapour in equilibrium with its liquid held in a cylinder by a piston, at a constant temperature – namely the
temperature at which the liquid and vapour are in equilibrium − that is to say, the boiling (or condensation) point for that
pressure. We imagine the piston to be pulled out, at constant temperature; liquid evaporates and the pressure remains constant.
If the piston is pushed in, vapour condenses, at constant temperature and pressure. During this process the pressure and
temperature remain constant, so the Gibbs free energy of the system is constant.
Let G1 be the specific Gibbs free energy for the liquid
and G2 be the specific Gibbs free energy for the vapour.
Suppose that a mass dm of the liquid vaporizes, so that the Gibbs free energy for the liquid decreases by G1dm and the Gibbs
free energy for the vapour increases by G2dm. But the Gibbs free energy for the system is constant. This therefore shows that,
when we have a liquid in equilibrium with its vapour (i.e. at its boiling point) the specific Gibbs free energies of liquid and
vapour are equal. (The same is true, of course, for the molar Gibbs free energies.) That is:
H1 − T S1 = H2 − T S2 (14.1)

or
T (S2 − S1 ) = H2 − H1 , (14.2)

in which the enthalpy and entropy are specific. The left hand side is the specific latent heat of vaporization, and we already
knew from Chapter 9 that this was equal to the difference in the specific enthalpies of liquid and vapour.
The equality of the specific Gibbs free energies of liquid and vapour can also be written
U1 − T S1 + P V1 = U2 − T S2 + P V2 , (14.3)

or
T (S2 − S1 ) = (U2 − U1 ) + P (V2 − V1 ) . (14.4)

This shows that the latent heat of vaporization goes into two things: To increase the internal energy upon vaporization
(especially the increase of potential energy as the molecules are pulled apart from each other) and the PdV work done against
the external pressure as the volume increases. Thus we could divide the latent heat into an internal latent heat and an external
latent heat.
In the foregoing, we imagined that some liquid vaporized as we withdrew the piston. Now let us imagine that we cause some
liquid to vaporize as we add some heat at constant volume. The specific Gibbs free energies of liquid and vapour both
increase, but they increase by the same amount because, as we have seen, when a liquid and its vapour are in equilibrium at the
boiling point, their specific Gibbs free energies are equal. Thus

−S1 dT + V1 dP = −S2 dT + V2 dP , (14.5)

or
dP S2 − S1
= . (14.6)
dT V2 − V1

The left hand side is the rate of increase of vapour pressure with temperature, while S2 − S1 is equal to L/T, where L is the
specific latent heat of vaporization. Thus we arrive at the Clausius-Clapeyron equation:
dP L
= . (14.7)
dT T (V2 − VL )

Jeremy Tatum 4/8/2021 14.1 CC-BY-NC https://phys.libretexts.org/@go/page/7304

Example: At 100 oC the rate of increase of vapour pressure of steam is 27.1 mm Hg per Celsius degree, and a gram of steam
occupies 1674 cm3. What is the specific latent heat of vaporization?
Answer: L = T (V 2 − V1 )
dP

dT
.
T = 373.15K. V2 − V1 = 1.673 m kg
3 −1
.
dP

dT
= 1.36 × 10
4
× 9.81 × 2.71 × 10
−2
= 3.616 PaK
−1
.
Hence
L = 2.26 × 10 Jkg
6 −1
.
–––––––––––––––––––
––––––––––––––––––––
–

The same argument can be used to relate the rate of change of melting point with pressure of a solid with its latent heat of
fusion. The Clausius-Clapeyron equation then takes the form
dT T (V2 − V1 )
= . (14.8)
dP L

For most substances, the specific volume of the liquid (V2) is greater than the specific volume of the solid (V1); but for H2O, Bi
and Ga, V2 < V1 and dT/dP is negative.
Example: For the ice-water system,
5 −1
L = 3.36 × 10 Jkg

−3 3 −1
V2 = 10 m kg

−3 3 −1
V1 = 1.091 × 10 m kg

T = 273.15K

Hence
dT −8 −1
= −7.4 × 10 KPa
dP

That’s about −7.4 × 10−3 kelvins per atmosphere

Solids, Liquids, Gases, Entropy and the Gibbs Function.
Of the three phases, solid, liquid and vapour, solid is the most ordered (has the least entropy) and vapour is the most disordered
(has the most entropy). Now equation 12.6.12a tells us that (∂G/∂T) = −S. This means that, at a given pressure, the Gibbs
function of the vapour decreases rapidly with increasing temperature, whereas the Gibbs function of a solid decreases
relatively slowly. Schematically the Gibbs function of the three phases for H2O at atmospheric pressure looks something like
this:

Below 0 ºC, the Gibbs function is lowest for the solid, and that is the stable phase. Between 0 ºC and 100 ºC, the Gibbs
function is lowest for the liquid, and that is the stable phase. Above 100 ºC, the Gibbs function is lowest for the vapour, and
that is the stable phase. At 0 ºC, the molar Gibbs function of solid and liquid are equal; the two phases there are in equilibrium.
At 100 ºC, the molar Gibbs function of gas and liquid are equal; the two phases there are in equilibrium.
The slopes and intercepts of these lines vary not only from substance to substance, but also, for a given substance, with
pressure. The Maxwell relation 12.6.16, ( ∂S

∂P
) = −(
∂V

∂T
) , tells us that the manner in which entropy changes with pressure
T P

Jeremy Tatum 4/8/2021 14.2 CC-BY-NC https://phys.libretexts.org/@go/page/7304

is related to the expansion coefficient. For most substances (water between 0 ºC and 4 ºC is an exception), the coefficient of
expansion is positive, so this tells us that entropy decreases with increasing pressure, and increases with decreasing pressure.
The change in entropy with pressure is greatest for the vapour, so that, at lower pressures the slope of the vapour line in the
graph of Gibbs function with temperature will be much steeper, and the situation will look like this:

At temperatures below A, the Gibbs function is lowest for the solid, and that is the stable phase. At temperature above A, the
Gibbs function is lowest for the vapour, and that is the stable phase. At the pressure represented in the above diagram, the
liquid is never the stable phase. The substance sublimates from solid to vapour as the temperature is raised.
At the pressure corresponding to the triple point line (remind yourself by looking at figures VI.3, VI.4 and VI.8), the diagram
looks like:

At the triple point (A) the molar Gibbs functions of all three phases are equal, and all three phases are in equilibrium. As you
increase the temperature from below A to above A, the substance sublimates directly from solid to vapour, as can also be seen
from figure VI.5.

Jeremy Tatum 4/8/2021 14.3 CC-BY-NC https://phys.libretexts.org/@go/page/7304

CHAPTER OVERVIEW
15: ADIABATIC DEMAGNETIZATION

15.1: INTRODUCTION
15.2: ADIABATIC DECOMPRESSION
15.3: ADIABATIC DEMAGNETIZATION
15.4: ENTROPY AND TEMPERATURE

1 4/29/2021
15.1: Introduction
One way to cool a gas is as follows. First compress it isothermally. This means compress it in a vessel that isn’t insulated, and
wait for the gas to lose any heat that is generated so that it returns to room temperature. Then insulate the vessel and allow the
gas to expand adiabatically. We could call this cooling by adiabatic decompression.
You can cool a rubber band as follows. First stretch it isothermally. That means, stretch it slowly, so that it has lots of time to
lose any heat that is generated. Then, suddenly destretch it, and before it has time to gain any heat from its surrounding,
measure its temperature by immediately holding it up to your lips. You will find that it has cooled by adiabatic de-stretching.
(If you stretch the band quickly (i.e. adiabatically) and immediately hold it up to your lips, you will find that it is hot. BUT ...
before you try that experiment, close your eyes tightly. You don’t want the stretched elastic band to break and hit you in the
eye. Believe me, you do not want that to happen.)
The method of adiabatic demagnetization has been used to obtain extremely low temperatures. A sample of a paramagnetic
salt (such as cerium magnesium nitrate), already cooled to low temperatures by other means, is magnetized isothermally. The
sample is often suspended in an atmosphere of helium, which can conduct away any heat that is produced, and hence keeps the
process isothermal. It is then insulated (by pumping out the helium) and suddenly and adiabatically demagnetized. This
process of isothermal magnetization followed by adiabatic demagnetization can be repeated over and over again. Temperatures
close to 0 K have been reached in this manner. You could actually reach a temperature of absolute zero if you did this an
infinite number of times – but not for any fewer.
In the analysis that follows, I shall have to assume that you are familiar with the concepts of B, H, magnetic moment and
magnetization from electricity and magnetism.
In brief, the magnetic dipole moment pm of a sample is the maximum torque it experiences in unit field B. That is, the torque is
given by τ = pm × B. The magnetization M of a specimen is defined by B = µH = µ0 (H + M). The magnetization is also equal
to the magnetic moment per unit volume.
Now consider the following.
If the tension in an elastic string is F, the work done on the string when its length is increased by dx is F dx.
If the pressure of a gas is P, the work done on the gas when its volume is increased by dV is −P dV.
And the work done per unit volume on an isotropic sample in increasing its magnetization from M to M + dM in a magnetic
field B is BdM. (I am assuming here that the sample is isotropic and that the magnetic moment and the magnetic field are in
the same direction, and hence I am no longer using boldface to indicate vector quantities.)
Note that, in all of these examples, the work done is the product of an intensive state variable (P, F, B) and the differential of an
extensive state variable (dV, dx, dM).
If we add heat to a magnetizable sample, and do work per unit volume on it by putting it in a magnetic field B and thereby
increasing its magnetization by dM, then, provided there is no change in volume, the increase in its internal energy per unit
volume is given by
dU = T dS + BdM , (15.1.1)

In this magnetic context, we can define state functions H, A and G per unit volume by
H = U − BM (15.1.2)

A = U −TS (15.1.3)

G = H − T S = A − BM (15.1.4)

In differential form, these become

dH = T dS − M dB (15.1.5)

dA = −SdT + BdM (15.1.6)

dG = −SdT − M dB (15.1.7)

Jeremy Tatum 4/8/2021 15.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7305

Here M is the dipole moment per unit volume, in N m T−1 m−3, which is the same as the magnetization, in A m−1. (Other
equivalent units for magnetization would be Pa T−1 or T m H−1, but I recommend N m T−1 m−3 as being the most readily
understandable in the present context.)
In Section 15.2 I am going to derive an expression for the lowering of the temperature in an adiabatic decompression, (∂T/
∂P)S. And then, in Section 15.3, I am going to derive an expression, by exactly the same argument, step-by-step, for the
lowering of the temperature in an adiabatic demagnetization, (∂T/∂B)S.

Jeremy Tatum 4/8/2021 15.1.2 CC-BY-NC https://phys.libretexts.org/@go/page/7305

15.2: Adiabatic Decompression
We are going to calculate an expression for (∂T /∂P ) . The expression will be positive, since T and P increase together. We
S

shall consider the entropy as a function of temperature and pressure, and, with the variables

we shall start with the cyclic relation

∂S ∂T ∂P
( ) ( ) ( ) = −1. (15.2.1)
∂T ∂P ∂S
P S T

The middle term is the one we want. Let’s find expressions for the first and third partial derivatives in terms of things that we
can measure.
In a reversible process dS = dQ/T , and, in an isobaric process, dQ = C P dT . Therefore

∂S Cp
( ) = . (15.2.2)
∂T T
p

Also, we have a Maxwell relation (Equation 12.6.16). ( ∂S

∂P
) = −(
∂V

∂T
) . Thus Equation 15.2.1 becomes
T P

∂T T ∂V
( ) = ( ) . (15.2.3)
∂P CP ∂T
S P

Check the dimensions of this. Note also that CP can be total, specific or molar,
provided that V is correspondingly total, specific or molar. (∂T/∂P)S is, of course,
intensive.
If the gas is an ideal gas, the equation of state is P V = RT , so that
∂V R V
( ) = = . (15.2.4)
∂T P T
P

Equation 15.2.3 therefore becomes

∂T V
( ) = . (15.2.5)
∂P CP
S

Jeremy Tatum 4/22/2021 15.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7306

15.3: Adiabatic Demagnetization
We are now going to do the same argument for adiabatic demagnetization.
We are going to calculate an expression for (∂T /∂B) . The expression will be positive, since T and B increase together. We
S

shall consider the entropy as a function of temperature and magnetic field, and, with the variables

we shall start with the cyclic relation

∂S ∂T ∂B
( ) ( ) ( ) = −1. (15.3.1)
∂T B
∂B S
∂S T

The middle term is the one we want. Let’s find expressions for the first and third partial derivatives in terms of things that we
can measure.
In a reversible process dS = dQ/T , and, in a constant magnetic field, dQ = C dT . Here I am taking S to mean the entropy
B

per unit volume, and CB is the heat capacity per unit volume (i.e. the heat required to raise the temperature of unit volume by
one degree) in a constant magnetic field.
CB
Thus we have ( ∂S

∂T
) =
T
.
B

The Maxwell relation corresponding to ( ∂S

∂P
) = −(
∂V

∂T
) is ( ∂S

∂B
) =(
∂M

∂T
) . Thus Equation 15.3.1 becomes
T P T B

∂T T ∂M
( ) =− ( ) (15.3.2)
∂B CB ∂T
S B

.
Now for a paramagnetic material, the magnetization, for a given field, is proportional to B and it falls off inversely as the
temperature (that’s the equation of state). That is, M = aB/T. and therefore ( ∂M

∂T
) =−
aB
2
=−
M

T
. Equation 15.3.2 therefore
B T

becomes
∂T M
( ) = . (15.3.3)
∂B CB
s

You should check the dimensions of this equation.

The cooling effect is particularly effective at low temperatures, when C is small. B

Jeremy Tatum 4/15/2021 15.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7307

15.4: Entropy and Temperature
Cooling by adiabatic demagnetization involves successive isothermal magnetizations followed by adiabatic demagnetizations,
and this suggests that some insight into the process might be obtained by following it on an entropy : temperature (S : T)
diagram.

In figure XV.1 I draw schematically with a thin curve the variation of entropy of the specimen with temperature in the absence
of a magnetizing field, and, with a thick curve, the (lesser) entropy of the more ordered state in the presence of a magnetizing
field. The process a represents an isothermal magnetization, and the process b is the following adiabatic (isentropic)
demagnetization, and it is readily seen how this results in a lowering of the temperature.
By the time when we reach the point A, an isothermal magnetization is represented by the process c, and the following
adiabatic demagnetization is the process d, which takes us down to the absolute zero of temperature.
We shall find out in the next Chapter, however, that there is a fundamental flaw in this last argument, and that getting down to
absolute zero isn’t going to be quite so easy.

Jeremy Tatum 4/8/2021 15.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7308

CHAPTER OVERVIEW
16: NERNST'S HEAT THEOREM AND THE THIRD LAW OF
THERMODYNAMICS
Topic hierarchy

16.1: NERNST'S HEAT THEOREM

16.2: THE THIRD LAW OF THERMODYNAMICS

1 4/29/2021
16.1: Nernst's Heat Theorem
At the beginning of the twentieth century, Walther Nernst (Nobel Prize in Chemistry 1920) had investigated heat capacities
and heats of reaction at progressively lower temperatures. As a result of his studies, he enunciated an important principle that
initially was restricted to the behaviour of reactions involving solids and liquids but which is now believed to apply to all
processes and substances.
The subject of chemical thermodynamics is dealt with more fully in Chapter 17, but for the present we shall note that some
chemical reactions require an input of heat to initiate them; other chemical reactions generate heat. The former are known as
endothermic reactions; the latter are exothermic reactions. If the reaction takes place at constant pressure (i.e. on an open
laboratory bench) the heat gained or lost is an increase or decrease in enthalpy H. The heat of reaction is usually given as ∆H,
being positive for an endothermic reaction (in which the system gains heat) and negative for an exothermic reaction. It should
be noted that spontaneous reactions are by no means always exothermic; some spontaneous reactions result in the absorption
of heat from their surroundings and in a corresponding increase of enthalpy.
Nernst had noticed that, at progressively lower temperatures, the change in enthalpy and the change in the Gibbs function
during a chemical reaction become more and more equal. And (as we shall see, what amounts to the same thing) the rate of
change of the Gibbs function with temperature becomes less and less as the temperature is lowered. That this amounts to the
same thing is evident from the Gibbs-Helmholtz relation
∂(ΔG)
ΔH = ΔG − T ( ) . (16.1.1)
∂T
P

What Nernst proposed was that, in the limit, as the temperature approaches zero, the changes in the enthalpy and Gibbs
function are equal – or, what amounts to the same thing, the temperature rate of change of the Gibbs function at constant
pressure approaches zero at zero temperature. And since
∂(ΔG)
( ) = −ΔS, (16.1.2)
∂T
P

this implies that chemical reactions at a temperature of absolute zero take place with no change of entropy. This is Nernst’s
Heat Theorem.
Planck later extended this to suppose that, not only does ∆G → ∆H, but that, as T → 0, the enthalpy and the Gibbs function of
the system approach each other asymptotically in such a manner that, in the limit, as T → 0, G → H and (∂G/∂T)P →0.
This has a number of consequences. For example, until now, we had defined only what is meant by a change in entropy. In
particular, in order to state what the entropy of a system is at some temperature, we would need to know what the entropy is at
a temperature of zero kelvin. In Sections 12.8 and 12.9 we attempted to calculate the change in the Helmholtz and Gibbs
functions as a system was changed from one state to another. We found that the right hand sides of equations 12.9.9 and
12.9.11 for calculating the changes in these functions contained the entropy. We later went on to show how we could calculate
the difference in entropy in some state to that at zero temperature, but there was still a matter of an arbitrary constant, namely –
what is the entropy at zero temperature? We now have the answer, resulting from the observed behaviour of (∂G/∂T)P [= −S] as
the temperature approaches zero – namely that the arbitrary constant is no longer arbitrary, and the entropy approaches zero as
the temperature approaches zero.
Another consequence is

Jeremy Tatum 4/15/2021 16.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7312

16.2: The Third Law of Thermodynamics
Nernst’s heat theorem and Planck’s extension of it, while originally derived from observing the behaviour of chemical
reactions in solids and liquids, is now believed to apply quite generally to any processes, and, in view of that, it is time to
reconsider our description of adiabatic demagnetization. We see immediately that figure XV.1 needs to be redrawn to reflect
the fact that the entropy of the substance approaches zero whether or not it is situated in a magnetic field. The revised drawing
is shown as figure XVI.1, in which I have drawn three consecutive magnetization-demagnetization operations, and it will be
readily seen that we shall never reach a temperature of exactly zero in a finite number of operations.
The same applies to any operation in which we attempt to lower the temperature by a series of isothermal constraints that
decrease the entropy followed by adiabatic relaxations – whether we are compressing a gas isothermally and then
decompressing it adiabatically, or stretching a rubber band isothermally and loosening it adiabatically. In all cases, owing to
the convergence of the two entropy curves at zero temperature, we are led to conclude:
It is impossible to reduce the temperature of a material body to the absolute zero of temperature in a finite number of
operations.
This is the Third Law of Thermodynamics, and it is an inevitable consequence of Planck’s extension of Nernst’s Heat Theorem.

This is usually taken to mean that it is impossible ever to reduce the temperature of anything to absolute zero. From a practical
point of view, that may be true, though that is not strictly what the third law says. It says that it is impossible to do it in a finite
number of operations. I cannot help but think of a bouncing ball (see Classical Mechanics Chapter V), in which the ball
bounces an infinite number of times before finally coming to rest after a finite time. After every bounce, there are still an
infinite number of bounces yet to come, yet it is all over in a finite time. Now, perhaps some reader of these notes one day will
devise a method of performing an infinite number of isothermal stress/adiabatic relaxation operations in a finite time, and so
attain absolute zero.
The third law also talks about a finite number of operations – by which I take it is meant operations such as an entropy-
reducing constraint followed by an adiabatic relaxation. I am not sure to what extent this applies to processes such as laser
cooling. In such experiments a laser beam is directed opposite to an atomic beam. The laser frequency is exactly equal to the
frequency need to excite the atoms to their lowest excited level, and so it stops the atoms in their tracks. As the atoms slow
down, the required frequency can be changed to allow for the Doppler effect. Such experiments have reduced the temperature
to a fraction of a nanokelvin. These experiments do not seem to be of the sort of experiment we had in mind when developing
the third law of thermodynamics. We might well ask ourselves if it is conceptually possible or impossible to reduce the speeds
of a collection of atoms to zero for a finite period of time. We might argue that it is conceptually possible – but then we may
remember that atoms attract each other (van der Waals forces), so, if all the atoms are instantaneously at rest, they will not
remain so. Of course if we had an ideal gas (such as a real gas extrapolated to zero pressure!) such that there are no forces
between the molecules, the concept of zero temperature implies that all the atoms are stationary – i.e. each has a definite
position and zero momentum. This is, according the Heisenberg’s uncertainty principle, inconceivable. So I leave it open as a
subject for lunchtime conversations exactly how strictly the third law prevents us from ever attaining the absolute zero of
temperature.

Jeremy Tatum 4/15/2021 16.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7313

Exercise. If the kinetic temperature of a set of hydrogen atoms is reduced to a tenth of a nanokelvin, what is the root-mean-
square speed of the atoms?

Jeremy Tatum 4/15/2021 16.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7313

CHAPTER OVERVIEW
17: CHEMICAL THERMODYNAMICS

17.1: EQUILIBRIUM CONSTANT

17.2: HEAT OF REACTION
17.3: THE GIBBS PHASE RULE
17.4: CHEMICAL POTENTIAL
17.5: PARTIAL AND MEAN MOLAR QUANTITIES
17.6: THE GIBBS-DUHEM RELATION
17.7: CHEMICAL POTENTIAL, PRESSURE, FUGACITY
17.8: ENTROPY OF MIXING, AND GIBBS' PARADOX
17.9: BINARY ALLOYS
17.10: TERNARY ALLOYS

1 4/29/2021
17.1: Equilibrium Constant
There are many types of chemical reaction, but to focus our attention we shall consider a reaction involving two reactants A
and B which, when mixed, form two resultants C and D. The reaction will proceed at a certain rate (fast or slow), and the rate
at which the reaction proceeds is part of the subject of chemical kinetics, which is outside the scope of this chapter, and to
some extent, though by no means entirely, outside the scope of this writer! We shall not, therefore, be concerned with how fast
the reaction proceeds, but with what the final state is, and whether the reaction needs some heat to get it going, or whether it
proceeds spontaneously and generates heat as it does so.
We shall suppose that the reaction is reversible. That is, that either

A+B → C +D (17.1.1)

or

C +D → A+B (17.1.2)

is possible.
That is
A + B ↔ C + D. (17.1.3)

The end result is a dynamic equilibrium in which the rates of forward and backward reaction are the same, and there is an
equilibrium amount of A, of B, of C and of D. The question is: How much of A? Of B? Of C? Of D?
Let us suppose that in the equilibrium mixture there are NA moles of A, NB of B, NC of C and ND of D. If we make the
reasonable assumption that the rate of the forward reaction is proportional to NANB and the rate of the backward reaction is
proportional to NCND, then, when equilibrium has been achieved and these two rates are equal, we have
NA NB
=  "constant".  (17.1.4)
NC ND

The “constant”, which is called the equilibrium constant for the reaction, is constant only for a particular temperature; in
general it is a function of temperature.
A simpler type of reaction is the dissociation-recombination equilibrium of a diatomic molecule:
AB ↔ A + B. (17.1.5)

The dissociation equilibrium constant is then

NA NB
. (17.1.6)
NAB

This “constant” is a function of the temperature and the dissociation energy of the molecule.
A similar consideration obtains for the ionization of an atom:
+ −
A ↔ A +e . (17.1.7)

In this situation,
N+ N−
, (17.1.8)
N0

the ionization equilibrium constant, is a function of the temperature and the ionization energy. The equilibrium constants can
be determined either experimentally or they can be computed from the partition functions of statistical mechanics. Some
details of how to calculate the dissociation and ionization constants and how to use them to calculate the numbers of atoms,
ions and molecules of various species in a hot gas are discussed in Stellar Atmospheres, Chapter 8, as well as in papers by the
writer in Publ. Dom. Astrophys. Obs., XIII (1) (1966) and by A. J. Sauval and the writer in Astrophys. J. Supp., 56, 193 (1984).

Jeremy Tatum 4/22/2021 17.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7319

17.2: Heat of Reaction
In some reactions, heat is produced by the reaction, and such reactions are called exothermic. If no heat is allowed to escape
from the system, the system will become hot. In other reactions, heat has to be supplied to cause the reaction. Such reactions
are endothermic.
The heat of reaction is the heat required to effect the reaction, or the heat produced by the reaction – some authors use one
definition, others use the other. Here we shall define the heat of reaction as the heat required to effect the reaction, so that it is
positive for endothermic reactions and negative for exothermic reactions. (In your own writing, make sure that your meaning
is unambiguous – don’t assume that there is some “convention” that everyone uses.) If the reaction is carried out at constant
pressure (i.e. on an open laboratory bench), the heat required to effect the reaction is the increase of enthalpy of the system. In
other words, ∆H is positive for an endothermic reaction. If the reaction produces heat, the enthalpy decreases and ∆H is
negative. Heats of reaction are generally quoted as molar quantities at a specific temperature (often 25 oC) and pressure (often
one atmosphere). The usual convention is to write
A + B → C ∆H = x J mole−1
One can make it yet clearer by specifying the temperature and pressure at which the enthalpy of reaction is determined, and
whether the reactants are solid (s), liquid (l) or gas (g).
If the reaction is carried out at constant volume (in a closed vessel), the heat required to effect the reaction is the increase of
the internal energy, ∆U. In either case, in our convention (which seems to be the most common one) ∆H or ∆U is positive for
an endothermic reaction and negative for an exothermic reaction.
The heat of reaction at constant pressure (∆H) is generally a little larger than at constant volume (∆U), though if all reactants
are liquid or solid the difference is very small indeed and often negligible within the precision to which measurements are
made.

Jeremy Tatum 4/29/2021 17.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7320

17.3: The Gibbs Phase Rule
Up to this point the thermodynamical systems that we have been considering have consisted of just a single component and,
for the most part, just one phase, but we are now going to discuss systems consisting of more than one phase and more than
one component. The Gibbs Phase Law provides a relation between the number of phases, the number of components and the
number of degrees of freedom. But Whoa, there! We have been using several technical terms here: Phase, Component,
Degrees of Freedom. We need to describe what these mean.
The state of a system consisting of a single component in a single phase (for example a single gas – not a mixture of different
gases) can be described by three intensive state variables, P, V and T. (Here V is the molar volume – i.e. the reciprocal of the
density in moles per unit volume – and is an intensive variable.) That is, the state of the system is described by a point in three-
dimensional PVT space. However, the intensive state variables are connected by an equation of state f(P, V, T) = 0, so that the
system is constrained to be on the two-dimensional surface described by this equation. Thus, because of the constraint, only
two intensive state variables suffice to describe the state of the system. Just two of the intensive state variables can be
independently varied. The system has two degrees of freedom.
Definition. A phase is a chemically homogeneous volume, solid, liquid or gas, with a boundary separating it from other
phases.
Definition. The number of intensive state variables that can be varied independently without changing the number of phases in
a system is called the number of degrees of freedom of the system.
These are easy. Defining the number of components in a system needs a bit of care. I give a definition, but what the definition
means can, I hope, be made a little clearer by giving a few examples.
Definition. The number of components in a system is the least number of constituents that are necessary to describe the
composition of each phase.
Let us look at a few examples to try and grasp what this means.
First, let us consider an aqueous solution of the chlorides and bromides of sodium and potassium co-existing with the
crystalline solids NaCl, KCl, NaBr, KBr, illustrated schematically in figure XVII.1.

There are five phases – four solid and one liquid – but how many components? There are six elements: H, O, Na, K, Cl, B –
but the quantities of each cannot be varied independently. There are two constraints: n(H) = 2n(O), and n(Na) + n(K) = n(Cl) +
n(Br). That is, if we know the number of hydrogen atoms, then the number of oxygen atoms is known. And if we know the
number of any three of Na, K, Cl or Br, then the fourth is known. Thus the number of constituents that that can be
independently varied is four. The number of components is four.
Or again, consider an aqueous solution of a moles of H2SO4 in b moles of water. There is just one phase. There are three
elements: H, O and S. These may be distributed among several species, such as H2O, H2SO4, H3O+, OH−, SO4−−, but that
doesn’t matter. There is just one constraint, namely that
2(a + b)n(H) = an(S) + (4a + b)n(O) .
That is, if we know the number of any two of H, O or S, we also know the number of the third. The number of components is
two.
Or again, consider the reversible reaction

Jeremy Tatum 4/29/2021 17.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7321

 CaCO3 (s) ↔ CaO (s) + CO2 (g) .
If the system is in equilibrium, and we know the numbers of any two of these three molecules, the number of the third is
determined by the equilibrium constant. Thus the number of components is two.
In each of these three examples, it was easy to state the number of phases and slightly more difficult to determine the number
of components. We now need to ask ourselves what is the number of degrees of freedom. This is what the Gibbs phase law is
going to tell us.
If there are C components in a system, the composition of a particular phase is fully described if we know the mole fraction of
C − 1 of the components, since the sum of the mole fractions of all the components must be 1. This is so for each of the P
phases, so that there are in all P(C − 1) mole fractions to be specified, as well as any two of the intensive state variables P, V
and T. Thus there are P(C − 1) + 2 intensive state variables to be specified. (The mole fraction of each component is an
intensive state variable.) But not all of these can be independently varied, because the molar Gibbs functions of each
component are the same in all phases. (To understand this important statement, re-read this argument in Chapter 14 on the
Clausius-Clapeyron equation.) For each of the C components there are P − 1 equations asserting the equality of the specific
Gibbs functions in all the phases. Thus the number of intensive state variables that can be varied independently without
changing the number of phases – i.e. the number of degrees of freedom, F − is P(C − 1) + 2 − C(P − 1), or

F = C − P + 2. (17.3.1)

This is the Gibbs Phase Rule.

In our example of the sodium and potassium salts, in which there were C = 4 components distributed through P = 5 phases,
there is just one degree of freedom. No more than one intensive state variable can be changed without changing the number of
phases.
In our example of sulphuric acid, there was one phase and two components, and hence three degrees of freedom.
In the calcium carbonate system, there were three phases and two components, and hence just one degree of freedom.
If we have a pure gas, there is one phase and one component, and hence two degrees of freedom. (We can vary any two of P, V
or T independently.)
If we have a liquid and its vapour in equilibrium, there are two phases and one component, and hence F = 1. We can vary P or
T, but not both independently if the system is to remain in equilibrium. If we increase T, the pressure of the vapour that
remains in equilibrium with its liquid increases. The system is constrained to lie on a line in PVT space.
If we have a liquid, solid and gas co-existing in equilibrium, there are three phases and one component and hence no degrees
of freedom. The system exists at a single point in PVT space, namely the triple point.
I have often been struck by the similarity of the Gibbs phase rule to the topological relation between the number of faces F,
edges E and vertices V of a solid polyhedron (with no topological holes through it). This relation is F = E − V + 2. E.g.
E V F

Tetrahedron: 6 4 4
(17.3.2)
Cube: 12 8 6

Octahedron: 12 6 8

As far as I know there is no conceivable connection between this and the Gibbs phase rule, and I don’t even find it useful as a
mnemonic. I think we just have to put it down as one of life’s little curiosities.
Since writing this section, I have added some additional material on binary and ternary alloys, which provide additional
examples of the Gibbs phase rule. I have added these at the end of the chapter, as sections 17.9 and 17.10.

Jeremy Tatum 4/29/2021 17.3.2 CC-BY-NC https://phys.libretexts.org/@go/page/7321

17.4: Chemical Potential
It is a truth universally acknowledged that, if we add some heat reversibly to a closed thermodynamic system at constant
volume, its internal energy will increase by (
∂U

∂S
) dS ; or, if we allow it to expand without adding heat, its internal energy
V

will increase by (
∂U

∂V
) dV . (In most cases the derivative (
∂U

∂V
) is negative, so that an increase in volume results in a
S S

decrease of internal energy.) If we do both, the increase in internal energy will be

∂U ∂U
dU = ( ) dS + ( ) dV . (17.4.1)
∂S ∂V
V S

By application of the first and second laws of thermodynamics, we find that this can be written
dU = T dS − P dV . (17.4.2)

Likewise, it is a truism that, if we add some heat reversibly to a closed thermodynamic system at constant pressure, its
enthalpy will increase by (
∂H

∂S
) dS ; or if we increase the pressure on it without adding heat, its enthalpy will increase by
P

(
∂H

∂P
) dP . If we do both, the increase in internal energy will be
S

∂H ∂H
dH = ( ) dS + ( ) dP . (17.4.3)
∂S ∂P
P S

By application of the first and second laws of thermodynamics, we find that this can be written
dH = T dS + V dP . (17.4.4)

Likewise, it is a truism that, if we increase the temperature of a closed thermodynamic system at constant volume, its
Helmholtz function will increase by (
∂A

∂T
) dT ; or, if we allow it to expand at constant temperature, its Helmholtz function
V

will increase by ( ∂A

∂V
) dV . (In most cases both of the derivatives are negative, so that an increase in temperature at constant
T

volume, or of volume at constant temperature, results in a decrease in the Helmholtz function.) If we do both, the increase in
the Helmholtz function will be
∂A ∂A
dA = ( ) dT + ( ) dV . (17.4.5)
∂T ∂V
V T

By application of the first and second laws of thermodynamics, we find that this can be written
dA = −SdT − P dV . (17.4.6)

Likewise, it is a truism that, if we increase the temperature of a closed thermodynamic system at constant pressure, its Gibbs
function will increase by (
∂G

∂T
) dT . (In most cases the derivative |9 \left(\frac{\partial G}{\partial T}\right)_{P}\) is
P

negative, so that an increase in temperature at constant pressure results in a decrease in the Gibbs function.) If we increase the
pressure on it at constant temperature, its Gibbs function will increase by (
∂G

∂P
) dP . If we do both, the increase in Gibbs
T

function will be
∂G ∂G
dG = ( ) dT + ( ) dP . (17.4.7)
∂T ∂P
P T

By application of the first and second laws of thermodynamics, we find that this can be written
dG = −SdT + V dP . (17.4.8)

So much, we are already familiar with. However, we can increase any of these thermodynamical functions of a system without
adding any heat to it or doing any work on it – merely by adding more matter. You will notice that, in the above statements, I
referred to a “closed” thermodynamical system. By a “closed” system, I mean one in which no matter is lost or gained by the
system. But, if the system is not closed, adding additional matter to the system obviously increases the (total)

Jeremy Tatum 3/25/2021 17.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7322

thermodynamical functions. For example, consider a system consisting of several components. Suppose that we add dNi moles
of component i to the system at constant temperature and pressure, by how much would the Gibbs function of the system
increase?
We might at first make the obvious reply: “dNi times the molar Gibbs function of component i”. This might be true if the
component were entirely inert and did not interact in any way with the other components in the system. But it is possible that
the added component might well interact with other components. It might, for example, shift the equilibrium position of a
reversible reaction A + B ↔ C + D. The best we can do, then, is to say merely that the increase in the (total) Gibbs function of
the system would be ( ∂G

∂Ni
) dNi . Here, Nj refers to the number of moles of any component other than i.
T ,P ,Nj

In a similar manner, if dNi moles of component were added at constant volume without adding any heat, the increase in the
internal energy of the system would be (
∂U

∂Ni
) dNi . Or if dNi moles of component were added at constant pressure
V ,S,Nj

without adding any heat, the increase in the enthalpy of the system would be ( ∂H

∂Ni
) dNi . Or if dNi moles of component
P ,S,Nj

were added at constant temperature and volume, the increase in the Helmholtz function of the system would be
(
∂A

∂Ni
) dNi . If we added a little bit more of all components at constant temperature and volume, the increase in the
T ,V ,Nj

Helmholtz function would be ∑ ( ∂A

∂Ni
) dNi , where the sum is over all components.
T ,V ,Nj

Thus, if the system is not closed, and we have the possibility of adding or subtracting portions of one or more of the
components, the formulas for the increases in the thermodynamic functions become
∂U ∂U ∂U
dU = ( ) dS + ( ) dV + ∑ ( ) dNi , (17.4.9)
∂S V ,Ni
∂V S,Ni
∂Ni V ,S,Nj

∂H ∂H ∂H
dH = ( ) dS + ( ) dP + ∑ )S,P ,N dNi , (17.4.10)
j
∂S P ,Ni
∂P S,Ni
∂Ni

∂A ∂A ∂A
dA = ( ) dT + ( ) dV + ∑ ( ) dNi , (17.4.11)
∂T ∂V ∂Ni
V ,Ni T ,Ni T ,V ,Nj

∂G ∂G ∂G
dG = ( ) dT + ( ) dP + ∑ ( ) dNi . (17.4.12)
∂T ∂P ∂Ni
P ,Ni T ,Ni T ,P ,Nj

The quantity (
∂U

∂Ni
) is the same as (
∂H

∂Ni
) or as (
∂A

∂Ni
) or as (
∂G

∂Ni
) , and it is called the chemical
V ,S,Nj P ,S,Nj T ,V ,Nj T ,P ,Nj

potential of species i, and is usually given the symbol µi. Its SI units are J kmole−1. (We shall later refer to it as the “partial
molar Gibbs function” of species i − but that is jumping slightly ahead.) If we make use of the symbol µi, and the other things
we know from application of the first and second laws, we can write equations 17.4.9 to 17.4.12 as

dU = T dS − P dV + ∑ μi dNi (17.4.13)

dH = T dS + V dP + ∑ μi dNi , (17.4.14)

dA = −SdT − P dV + ∑ μi dNi (17.4.15)

and

dG = −SdT + V dP + ∑ μi dNi (17.4.16)

It will be clear that

Jeremy Tatum 3/25/2021 17.4.2 CC-BY-NC https://phys.libretexts.org/@go/page/7322

 ∂U ∂U ∂U
( ) = T; ( ) = −P ; ( ) = μi ;
∂S ∂V ∂Ni
V ,Ni S,Ni V ,S,Nj

∂H ∂H ∂H
( ) = T; ( ) =V ; ( ) = μi ;
∂S ∂P ∂Ni
P ,Ni S,Ni P ,S,Nj
(17.4.17)
∂A ∂A ∂A
( ) = −S; ( ) = −P ; ( ) = μi ;
∂T ∂V ∂Ni
V ,Ni T ,Ni V ,T ,Nj

∂G ∂G ∂G
( ) = −S; ( ) =V ; ( ) = μi .
∂T ∂P ∂Ni
P ,Ni T ,Ni P ,T ,Nj

Since the four thermodynamical functions are functions of state, their differentials are exact and their mixed second partial
derivatives are equal. Consequently we have the following twelve Maxwell relations:
∂μ ∂μ
∂T ∂P ∂T i ∂P i
( ) = −( ) ; ( ) = +( ) ; ( ) = −( ) ;
∂V ∂S ∂Ni ∂S ∂Ni ∂V
S,Ni V ,Ni S,V ,Nj V ,Ni S,V ,Nj S,Ni

∂T ∂V ∂T ∂μi ∂V ∂μi
( ) = +( ) ; ( ) = +( ) ; ( ) = +( ) ;
∂P ∂S ∂Ni ∂S ∂Ni ∂P
S,Ni P ,Ni S,P ,Nj P ,Ni S,P ,Nj S,Ni
(17.4.18)
∂μ ∂μ
∂S ∂P ∂S i ∂P i
( ) = +( ) ; ( ) = −( ) ; ( ) = −( ) ;
∂V ∂T ∂Ni ∂T ∂Ni ∂V
T ,Ni V ,Ni T ,V ,Nj V ,Ni T ,V ,Nj T ,Ni

∂S ∂V ∂S ∂μi ∂V ∂μi
( ) = −( ) ; ( ) = −( ) ; ( ) = +( ) .
∂P ∂T ∂Ni ∂T ∂Ni ∂P
T ,Ni P ,Ni T ,P ,Nj P ,Ni T ,P ,Nj T ,Ni

Refer to equations 17.4.13 to 17.4.16, and we understand that:

If we add dN1 moles of species 1, dN2 moles of species 2, dN3 moles of species 3, etc., in a insulated constant-volume vessel
(dS and dV both zero), the increase in the internal energy is

dU = ∑ μi dNi . (17.4.19)

If we do the same in an insulated vessel at constant pressure (for example, open to the atmosphere, but in a time sufficiently
short so that no significant heat escapes from the system, and dS and dP are both zero), the increase in the enthalpy is

dH = ∑ μi dNi . (17.4.20)

If we do the same in a closed vessel (e.g. an autoclave or a pressure cooker, so that dV = 0) in a constant temperature water-
bath (dT = 0), the increase in the Helmholtz function is

dA = ∑ μi dNi . (17.4.21)

If we do the same at constant pressure (e.g. in an open vessel on a laboratory bench, so that dP = 0) and kept at constant
temperature (e.g. if the vessel is thin-walled and in a constant-temperature water bath, so that dT = 0), the increase in the Gibbs
free energy is

dG = ∑ μi dNi . (17.4.22)

We have called the symbol µi the chemical potential of component i – but in what sense is it a “potential”? Consider two
phases, α and β, in contact. The Gibbs functions of the two phases are Gα and Gβ respectively, and the chemical potential of
species i is µiα in α and µiβ in β. Now transfer dNi moles of i from α to β. The increase in the Gibbs function of the system is
µiβdNi − µiαdNi. But for a system of two phases to be in chemical equilibrium, the increase in the Gibbs function must be zero.
In other words, the condition for chemical equilibrium between the two phases is that µiβ = µiα for all species, just as the
condition for thermal equilibrium is that Tα = Tβ, and the condition for mechanical equilibrium is that Pα = Pβ.
Students of classical mechanics may see an analogy between equation 17.4.44 and the principle of Virtual Work. One way of
finding the condition of static equilibrium in a mechanical system is to imagine the system to undergo an infinitesimal change
in its geometry, and then to calculate the total work done by all the forces as they are displaced by the infinitesimal geometrical
alteration. If the system were initially in equilibrium, then the work done by the forces, which is an expression of the form
∑Fidxi, is zero, and this gives us the condition for mechanical equilibrium. Likewise, if a system is in chemical equilibrium,
and we make infinitesimal changes dNi, at constant temperature and pressure, in the chemical composition, the corresponding
change in the Gibbs function of the system, ∑µidNi, is zero. At chemical equilibrium, the Gibbs function is a minimum with
respect to changes in the chemical composition.

Jeremy Tatum 3/25/2021 17.4.3 CC-BY-NC https://phys.libretexts.org/@go/page/7322

17.5: Partial and Mean Molar Quantities
Consider a single phase with several components. Suppose there are Ni moles of component i, so that the total number of
moles of all species is

N = ∑ Ni . (17.5.1)

The mole fraction of species i is

Ni
ni = , (17.5.2)
N

and of course ∑ni = 1.

Let V be the volume of the phase. What will be the increase in volume of the phase if you add dNi moles of component i at
constant temperature and pressure? The answer, of course, is
∂V
dV = ( ) dNi (17.5.3)
∂Ni
T ,P ,Nj

If you increase the number of moles of all species at constant temperature and pressure, the increase in volume will be
∂V
dV = ∑ ( ) dNi . (17.5.4)
∂Ni
T ,P ,Nj

The quantity ( ∂V

∂Ni
) is called the partial molar volume of species i:
T ,P ,Nj

∂V
vi = ( ) (17.5.5)
∂Ni
T ,P ,Nj

Let us suppose that the volume of a phase is just proportional to the number of moles of all species in the phase. It might be
thought that this is always the case. It would indeed be the case if the phase contained merely a mixture of ideal gases.
However, to give an example of a non-ideal case: If ethanol C2H5OH is mixed with water H2O, the volume of the mixture is
less than the sum of the separate volumes of water and ethanol. This is because each molecule has an electric dipole moment,
and, when mixed, the molecules attract each other and pack together more closely that in the separate liquids. However, let us
go back to the ideal, linear case.
In that case, if a volume V contains N moles (of all species) and you add Ni moles of species i at constant temperature and
pressure, the ratio of the new volume to the old is given by
V + dV N + dNi
= , (17.5.6)
V N

and hence
dV dNi
= , (17.5.7)
V N

or
∂V V
( ) = vi = . (17.5.8)
∂Ni N
P ,T ,Nj

Example. (You’ll need to think long and carefully about the next two paragraphs fully to appreciate what are meant by molar
volume and partial molar volume. You’ll need to understand them before you can understand more difficult things, such as
partial molar Gibbs function.)
A volume of 6 m3 contains 1 mole of A, 2 moles of B and 3 moles of C. Thus the molar volumes (not the same thing as the
partial molar volumes) of A, B and C are respectively 6, 3 and 2 m3.

Jeremy Tatum 2/25/2021 17.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7323

Assume that the mixing is ideal. In that case, equation 17.5.8 tells us that the partial molar volume of each is the total volume
divided by the total number of moles. That is, the partial molar volume of each is 1 m3. You could imagine that, before the
component were mixed (or if you were to reverse the arrow of time and un-mix the mixture), we had 1 mole of A occupying 1
m3, 2 moles of B occupying 2 m3 and 3 moles of C occupying 3 m3, the molar volume of each being 1 m3.
The mean molar volume per component is
V
¯¯
v̄ = . (17.5.9)
N

If the components are ideal, each component has the same partial molar volume, and hence the mean molar volume is equal to
the partial molar volume of each – but this would not necessarily be the case for nonideal mixing.
The total volume of a phase, whether formed by ideal or nonideal mixing, is

V = ∑ Ni vi . (17.5.10)

If you divide each side of this equation by N, you arrive at

¯
¯¯
v = ∑ ni vi . (17.5.11)

Note that the partial molar volume of a component is not just the volume occupied by the component divided by the number of
moles. I.e. the partial molar volume is not the same thing as the molar volume. In our ideal example, the molar volume of the
three components would be, respectively, 6, 3 and 2 m3.
Another way of looking at it: In the mixture, Ni moles of species i occupies the entire volume V, as indeed does every
component, and its molar volume is V/Ni. The pressure of the mixture is P. Now remove all but species i from the mixture and
then compress it so that its pressure is still P, it perforce must be compressed to a smaller volume, and the volume of a mole
now is its partial molar volume.
Let Φ be any extensive quantity (such as S, V, U, H, A, G).
Establish the following notation:
Φ = total extensive quantity for the phase;
φi = partial molar quantity for component i;
φ = mean molar quantity per component.
The partial molar quantity φi for component i is defined as
∂Φ
ϕi = ( ) . (17.5.12)
∂Ni
P ,T ,Nj∗i

The total value of Φ is given by

Φ = ∑ Ni ϕi , (17.5.13)

and the mean value per component is

¯¯
¯
ϕ = ∑ ni ϕi . (17.5.14)

If the extensive quantity Φ that we are considering is the Gibbs function G, then equation 17.5.12 becomes
∂G
gi = ( ) (17.5.15)
∂Ni
P ,T ,Nj+i

Then we see, by comparison with equation 17.4.28 that the chemical potential µi of component i is nothing other than its
partial molar Gibbs function.
Note that this is not just the Gibbs function per mole of the component, any more than the partial molar volume is the same as
the molar volume.

Jeremy Tatum 2/25/2021 17.5.2 CC-BY-NC https://phys.libretexts.org/@go/page/7323

Recall (Chapter 14 on the Clausius-Clapeyron equation) that, when we had just a single component distributed in two phases
(e.g. a liquid in equilibrium with its vapour), we said that the condition for thermodynamic equilibrium between the two
phases was that the specific or molar Gibbs functions of the liquid and vapour are equal. In Section 17.5 of this chapter, when
we are dealing with several components distributed between two phases, the condition for chemical equilibrium is that the
chemical potential µi of component i is the same in the two phases. Now we see that the chemical potential is synonymous
with the partial molar Gibbs function, so that the condition for chemical equilibrium between two phases is that the partial
molar Gibbs function of each component is the same in each phase. Of course, if there is just one component, the partial molar
Gibbs function is just the same as the molar Gibbs function.
Although pressure is an intensive rather than an extensive quantity, and we cannot talk of “molar pressure” or “partial molar
pressure”, opportunity can be taken here to define the partial pressure of a component in a mixture. The partial pressure of a
component is merely the contribution to the total pressure made by that component, so that the total pressure is merely

P = ∑ pi , (17.5.16)

where pi is the partial pressure of the ith component,

Dalton’s Law of Partial Pressures states that for a mixture of ideal gases, the partial pressure of component j is proportional
mole fraction of component j. That is, for a mixture of ideal gases,
pj pj Nj Nj
= = = = nj . (17.5.17)
P ∑ pi N ∑ Ni

That is,
pj = nj P . (17.5.18)

Jeremy Tatum 2/25/2021 17.5.3 CC-BY-NC https://phys.libretexts.org/@go/page/7323

17.6: The Gibbs-Duhem Relation
In a mixture of several components kept at constant temperature and pressure, the chemical potential µi of a particular
component (which, under conditions of constant T and P, is also its partial molar Gibbs function, gi) depends on how many
moles of each species i are present. The Gibbs-Duhem relation tells us how the chemical potentials of the various components
vary with composition. Thus:
We have seen that, if we keep the pressure and temperature constant, and we increase the number of moles of the components
by N1, N2, N3, the increase in the Gibbs function is

dG = ∑ μi dNi . (17.6.1)

We also pointed out in section 17.5 that, provided the temperature and pressure are constant, the chemical potential µi is just
the partial molar Gibbs function, gi, so that the total Gibbs function is

G = ∑ gi Ni = ∑ μi Ni , (17.6.2)

the sum being taken over all components. On differentiation of equation 17.7.2 we obtain

dG = ∑ μi dNi + ∑ Ni dμi . (17.6.3)

Thus for any process that takes place at constant temperature and pressure, comparison of equations 17.6.1 and 17.6.3 shows
that

∑ Ni dμi = 0, (17.6.4)

which is the Gibbs-Duhem relation. It tells you how the chemical potentials change with the chemical composition of a phase.

Jeremy Tatum 4/22/2021 17.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7324

17.7: Chemical Potential, Pressure, Fugacity
Equation 12.9.11 told us how to calculate the change in the Gibbs function of a mole of an ideal gas going from one state to
another. For N moles it would be
P2
ΔG = N ∫ CP dT − N T2 ∫ CP d(ln T ) + N RT2 ln( ) − N S (T2 − T1 ) , (17.7.1)
P1

where CP and S are molar, and G is total.

Since we know now how to calculate the absolute entropy and also know that the entropy at T = 0 is zero, this can be written

G(T , P ) = N (RT ln P +  constant ) (17.7.2)

The “constant” here depends on the temperature, but is not a function of the pressure, being in fact the value of the molar
Gibbs function extrapolated to the limit of zero pressure. Sometimes it is convenient to write Equation 17.7.2 in the form

G = N RT (ln P + ϕ) (17.7.3)

where φ is a function of temperature.

If we have a mixture of several components, the total Gibbs function is

G(T , P ) = ∑ Ni (RT ln pi +  constant ) (17.7.4)

We can now write this in terms of the partial molar Gibbs function of the component i – that is to say, the chemical potential of
the component i, which is given by μ = (∂G/∂ N )
i i , and the partial pressure of component i. Thus we obtain
P ,T ,Nj≠1

0
μi = μ (T ) + RT ln pi (17.7.5)
i

and
μi = RT (ln pi + ϕi ) (17.7.6)

Here I have written the “constant” as 0 µi0 (T), or as RTφi. The constant µi0 (T) is the value of the chemical potential at
temperature T extrapolated to the limit of zero pressure. If the system consists of a mixture of ideal gases, the partial pressure
of the ith component is related to the total pressure simply by Dalton’s law of partial pressures:
pi = ni P , (17.7.7)

where ni is the mole fraction of the ith component. In that case, equation 17.7.4 becomes
0
μi = μ (T ) + RT ln ni + RT ln P . (17.7.8)
i

and equation 17.7.5 becomes

μi = RT (ln ni + ln P + ϕi ) . (17.7.9)

However, in a common deviation from ideality, volumes in a mixture are not simply additive, and we write equation 17.7.4 in
the form
0
μi = μ (T ) + RT ln fi , (17.7.10)
i

or equation 17.7.5 in the form

μi = RT (ln fi + ϕi ) . (17.7.11)

where fi is the fugacity of component i.

Jeremy Tatum 4/15/2021 17.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8667

17.8: Entropy of Mixing, and Gibbs' Paradox
In Chapter 7, we defined the increase of entropy of a system by supposing that an infinitesimal quantity dQ of heat is added to
it at temperature T, and that no irreversible work is done on the system. We then asserted that the increase of entropy of the
system is dS = dQ/T. If some irreversible work is done, this has to be added to the dQ.
We also pointed out that, in an isolated system any spontaneous transfer of heat from one part to another part was likely (very
likely!) to be from a hot region to a cooler region, and that this was likely (very likely!) to result in an increase of entropy of
the closed system − indeed of the Universe. We considered a box divided into two parts, with a hot gas in one and a cooler gas
in the other, and we discussed what was likely (very likely!) to happen if the wall between the parts were to be removed. We
considered also the situation in which the wall were to separate two gases consisting or red molecules and blue molecules. The
two situations seem to be very similar. A flow of heat is not the flow of an “imponderable fluid” called “caloric”. Rather it is
the mixing of two groups of molecules with initially different characteristics (“fast” and “slow”, or “hot” and “cold”). In either
case there is likely (very likely!) to be a spontaneous mixing, or increasing randomness, or increasing disorder or increasing
entropy. Seen thus, entropy is seen as a measure of the degree of disorder. In this section we are going to calculate the increase
on entropy when two different sorts of molecules become mixed, without any reference to the flow of heat. This concept of
entropy as a measure of disorder will become increasingly apparent if you undertake a study of statistical mechanics.
Consider a box containing two gases, separated by a partition. The pressure and temperature are the same in both
compartments. The left hand compartment contains N1 moles of gas 1, and the right hand compartment contains N2 moles of
gas 2. The Gibbs function for the system is
G = RT [ N1 (ln P + ϕ1 ) + N2 (ln P + ϕ2 )] . (17.8.1)

Now remove the partition, and wait until the gases become completely mixed, with no change in pressure or temperature. The
partial molar Gibbs function of gas 1 is
μ1 = RT (ln p1 + ϕ1 ) (17.8.2)

and the partial molar Gibbs function of gas 2 is

μ2 = RT (ln p2 + ϕ2 ) . (17.8.3)

Here the pi are the partial pressures of the two gases, given by and p1 = n1P, p2 = n2P where the ni are the mole fractions.
The total Gibbs function is now N1µ1 + N2µ2, or
G = RT [ N1 (ln n1 + ln P + ϕ1 ) + N2 (ln n2 + ln P + ϕ2 )] . (17.8.4)

The new Gibbs function minus the original Gibbs function is therefore
ΔG = RT (N1 ln n1 + N2 ln n2 ) = N RT (n1 ln n1 + n2 ln n2 ) . (17.8.5)

This represents a decrease in the Gibbs function, because the mole fractions are less than 1.
∂(ΔG)
The new entropy minus the original entropy is ΔS = −[ ∂T
] , which is
P

ΔS = −N R (n1 ln n1 + n2 ln n2 ) . (17.8.6)

This is positive, because the mole fractions are less than 1.

Similar expressions will be obtained for the increase in entropy if we mix several gases.
Here’s maybe an easier way of looking at the same thing. (Remember that, in what follows, the mixing is presumed to be ideal
and the temperature and pressure are constant throughout.)
Here is the box separated by a partition:

Jeremy Tatum 4/29/2021 17.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8668

Concentrate your attention entirely upon the left hand gas. Remove the partition. In the first nanosecond, the left hand gas
expands to increase its volume by dV, its internal energy remaining unchanged (dU = 0). The entropy of the left hand gas
therefore increases according to dS = P dV

T
=N R 1 . By the time it has expanded to fill the whole box, its entropy has
dV

increased by ln( / ). RN1 ln(V/V1). Likewise, the entropy of the right hand gas, in expanding from volume V2 to V, has
increased by RN2 ln(V/V2). Thus the entropy of the system has increased by R[ N1 ln(V/V1) ln(V/V2)], and this is equal to RN[
n1 ln(1/n1) ln(1/n2)] = − NR[n1 ln n1 + n2 ln n2].
Where there are just two gases, n2 = 1 − n1, so we can conveniently plot a graph of the increase in the entropy versus mole
fraction of gas 1, and we see, unsurprisingly, that the entropy of mixing is greatest when n = n = , when ∆S = NR ln 2 =
1 2
1

0.6931NR.

What is n1 if ΔS = 1

2
NR ? (I make it n1 = 0.199 710 or, of course, 0.800 290.)
We initially introduced the idea of entropy in Chapter 7 by saying that if a quantity of heat dQ is added to a system at
temperature T, the entropy increases by dS = dQ/T. We later modified this by pointing out that if, in addition to adding heat,
we did some irreversible work on the system, that irreversible work was in any case degraded to heat, so that the increase in
entropy was then dS = (dQ + dWirr)/T. We now see that the simple act of mixing two or more gases at constant temperature
results in an increase in entropy. The same applies to mixing any substances, not just gases, although the formula −NR[n1 ln n1
+ n2 ln n2] applies of course just to ideal gases. We alluded to this in Chapter 7, but we have now placed it on a quantitative
basis. As time progresses, two separate gases placed together will spontaneously and probably (very probably!) irreversibly
mix, and the entropy will increase. It is most unlikely that a mixture of two gases will spontaneously separate and thus
decrease the entropy.
Gibbs’ Paradox arises when the two gases are identical. The above analysis does nothing to distinguish between the mixing of
two different gases and the mixing of two identical gases. If you have two identical gases at the same temperature and pressure
in the two compartments, nothing changes when the partition is removed – so there should be no change in the entropy. Within
the confines of classical thermodynamics, this remains a paradox – which is resolved in the study of statistical mechanics.
Now consider a reversible chemical reaction of the form Reactants ↔ Products − and it doesn’t matter which we choose to call
the “reactants” and which the “products”. Let us suppose that the Gibbs function of a mixture consisting entirely of “reactants”
and no “products” is less than the Gibbs function of a mixture consisting entirely of “products”. The Gibbs function of a
mixture of reactants and products will be less than the Gibbs function of either reactants alone or products alone. Indeed, as we
go from reactants alone to products alone, the Gibbs function will look something like this:

Jeremy Tatum 4/29/2021 17.8.2 CC-BY-NC https://phys.libretexts.org/@go/page/8668

The left hand side shows the Gibbs function of the reactants alone. The right hand side shows the Gibbs function for the
products alone. The equilibrium situation occurs where the Gibbs function is a minimum.
If the Gibbs function of the reactants were greater than that of the products, the graph would look something like:

Jeremy Tatum 4/29/2021 17.8.3 CC-BY-NC https://phys.libretexts.org/@go/page/8668

17.9: Binary Alloys
(This section is a little out of order, and might be better read after Section 17.3.)
If two metals are melted together, and subsequently cooled and solidified, interesting phenomena occur. In this section we look
at the way tin and lead mix. I do this in an entirely schematic and idealized way. The details are bit more complicated (and
interesting!) than I present them here. For the detailed description and more exact numbers, the reader can refer to the
specialized literature, such as Constitution of Binary Alloys by M. Hansen and K. Anderko and its subsequent Second
Supplement by F. A. Shunk. In my simplified description I am assuming that when tin and lead are melted, the two liquids are
completely miscible, but, when the liquid is cooled, the two metals crystallize out separately. The phenomena are illustrated
schematically in the figure below, which is a graph of melting point versus composition of the alloy at a given constant
pressure (one atmosphere).

The melting point of pure Pb is 327 ºC

The melting point of pure Sn is 232 ºC
In studying the diagram, let us start at the upper end of the dashed line, where the temperature is 350 ºC and we are dealing
with a mixture of 70 percent Pb and 30 percent Sn (by mole – that is to say, by relative numbers of atoms, not by relative
mass). If you review the definitions of phase, component, and degrees of freedom, and the Gibbs phase rule, from Section
17.3, you will agree that there is just one phase and one component (there’s no need to tell me the percentage of Sn if you have
already told me the percentage of Pb), and that you can vary two intensive state variables (e.g. temperature and pressure)
without changing the number of phases.
Now, keeping the composition and pressure constant, let us move down the isopleth (i.e down the dashed line of constant
composition). Even after the temperature is lower than 327 ºC, the mixture doesn’t solidify. Nothing happens until the
temperature is about 289 ºC. Below that temperature, crystals of Pb start to solidify. The full curve represents the melting
point, or solidification point, of Pb as a function of the composition of the liquid. Of course, as some Pb solidifies, the
composition of the liquid changes to one of a lesser percentage of Pb, and the composition of the liquid moves down the
melting point curve. As long as the liquid is at a temperature and composition indicated along this curve, there is only one
remaining degree of freedom (pressure). You cannot change both the temperature and the composition without changing the
number of phases. As the temperature is lowered further and further, more Pb solidifies, and the composition of the liquid
moves further and further along the curve to the left, until it reaches the eutectic point at a temperature of 183 ºC and a
composition of 26% Pb. Below that temperature, both Pb and Sn crystallize out.
If we had started with a composition of less than 26% Pb, Sn would have started to crystallize out as soon as we had reached
the left hand curve, and the composition of the liquid would move along that curve to the right until it had reached the eutectic
point.

Jeremy Tatum 4/29/2021 17.9.1 CC-BY-NC https://phys.libretexts.org/@go/page/8669

Below, we show similar (highly idealized and schematic) eutectic curves for Pb-Bi and for Bi-Sn. (For more precise
descriptions, and more exact numbers, see the literature, such as the references cited above). The data for these three alloys
are:
For the pure metals:
Melting point
Pb 327 ºC
Bi 271
Sn 232
Sn-Pb Eutectic 183 ºC 26% Pb
Pb-Bi Eutectic 125 ºC 56% Bi
Bi-Sn Eutectic 139 ºC 57% Sn

Jeremy Tatum 4/29/2021 17.9.2 CC-BY-NC https://phys.libretexts.org/@go/page/8669

17.10: Ternary Alloys
In this section we look at what happens with an alloy of three metals, and we shall use as an example Pb-Bi-Sn. Our
description is merely illustrative of the principles; for more exact details, see the specialized literature.
To illustrate the phase equilibria of an alloy of these three metals, I have pasted the eutectic diagrams of the previous section to
the faces of a triangular prism, as shown below. The vertical ordinate is the temperature.

Jeremy Tatum 4/1/2021 17.10.1 CC-BY-NC https://phys.libretexts.org/@go/page/8670

On each of the three faces only two of the metals are present. The situation where all three metals are present on comparable
quantities would be illustrated by a surface inside the prism, but creating this inner surface is unfortunately beyond my skills.
Anywhere above the surface outlined by the curves on each face is completely liquid. Below it one or other of the constituent
metals solidifies. The surface goes down to a deep well, terminating in a eutectic temperature well below the 125 ºC of the Pb-
Bi eutectic.
In lieu of building a nice three-dimensional model, the next best thing might be to take a horizontal slice through the prism at
constant temperature. If I do that at, say, 200ºC, the ternary phase diagram might look something like this:

You can imagine what happens as you gradually lower the temperature. First a bit of Pb solidifies out. Then a bit of Bi. Lastly
a bit of Sn. You have to try and imagine what this ternary diagram would look like as you lower the temperature. Eventually
the solidification parts spread out from the corners of the triangle, and meet at a single eutectic point where there are no
degrees of freedom. Below that temperature, all is solid, whatever the composition.

Jeremy Tatum 4/1/2021 17.10.2 CC-BY-NC https://phys.libretexts.org/@go/page/8670

CHAPTER OVERVIEW
18: EXPERIMENTAL MEASUREMENTS
Topic hierarchy

18.1: INTRODUCTION
18.2: THERMAL CONDUCTIVITY
18.3: THE UNIVERSAL GAS CONSTANT
18.4: AVOGADRO'S NUMBER AND BOLTZMANN'S CONSTANT
18.5: SPECIFIC HEAT CAPACITIES OF SOLIDS AND LIQUIDS
18.6: SPECIFIC HEAT CAPACITIES OF GASES
18.7: LATENT HEAT OF FUSION
18.8: COEFFICIENT OF EXPANSION

1 4/29/2021
18.1: Introduction
Most of these notes on heat and thermodynamics have been largely theoretical, and almost no attention has been given to
laboratory measurements of the various quantities discussed. This is not because experiment is any less important that theory.
Rather it is more the consequence of my own interests and personal lack of expertise in experiment. Indeed, laboratory physics
equipment has a tendency to disintegrate as soon as I approach it. However, in this chapter we shall endeavour to describe,
however inadequately, some of the early classical experimental measurements.
I am under the impression that today, in order to measure any physical quantity, you purchase some expensive equipment,
attach one end of it to the thing to be measured, and the other end to a computer, and one instantaneously obtains a digital
readout of the quantity in question, without necessarily having any idea how the equipment works. And I, certainly, have little
idea how much of modern high technology works. Consequently I shall restrict this chapter to brief descriptions of some of the
earlier classical historical determinations of thermal quantities, many of which were performed during the nineteenth century
or the early twentieth century.
Of all the many experimental determinations of physical quantities in various branches in physics, accurate determinations in
the laboratory of thermal quantities are among the most difficult classical measurements of all. It would be easy to dismiss the
various early experiments that I shall describe in this chapter as quaint, crude and of no modern interest. Far from it. Some of
these experiments were extremely difficult to carry out accurately, and it is astonishing how accurate many of the early
measurements were, as a result of the careful design, attention to detail and allowance for heat losses. The early experimenters
deserve our great admiration and our gratitude for the important fundamental contributions they made to our understanding of
physical science.

Jeremy Tatum 4/29/2021 18.1.1 CC-BY-NC https://phys.libretexts.org/@go/page/7326

18.2: Thermal Conductivity
18.2.1 Solid Good Conductors (Metals)
The difference between the thermal conductivities of metals and non-metals is so large that different experimental approaches
are needed for the two classes of solids, and in this subsection we deal with metals.
I remember as long ago as when I was in high school one of the experiments we had to do was to measure the conductivity of
a metal rod, which, as far as I remember, was about a foot (30 cm) long and maybe two centimetres in diameter. The
experiment was called Searle's Rod, or Searle's Bar, after an experimenter in the early years of the twentieth century. The
experiment was simple in principle, but very difficult in practice to do accurately, and we were always advised to avoid doing
heat experiments in our matriculation examinations. Heat was supplied by an electrical coil wrapped around one end the rod,
and the rate of supply of heat was determined from the current through the wire and the potential difference across it,
measured with an ammeter and voltmeter respectively. Heat was collected at the other end of the rod by means of a stream of
water flowing through a helical tube wrapped around that end of the rod. Thermometers at the beginning and end of this
helical tube measured the rise in temperature of the water. Hence one could determine the rate of flow of heat out of the cool
end of the rod. If no heat were lost from side of the rod, the rate of flow of heat into the rod (determined electrically) should
equal the rate of flow out of it (measured by the rise in temperature of the stream of water through the helical tube). The
difference between the two would be a measure of how much heat was lost from the side of the rod. The rod was supposedly
well lagged with cotton wool to keep the heat loss small. (I am talking of a high-school experiment here. One could improve
on this in a more advanced laboratory by having the rod in a vacuum – so there is no loss of heat by conduction or convection,
and highly polished to reduce heat loss by radiation). The temperature gradient along the length of the tube was determined by
drilling pits at two points along the rod, filling these with mercury (for good thermal contact), and sticking mercury-in-glass
thermometers into these little pools of mercury. One can easily imagine how difficult such an experiment was! At any rate,
there was by then enough information to determine the thermal conductivity, for one knew the temperature gradient, from the
thermometers stuck into the little mercury pools, and one knew the rate of flow of heat into and out of the rod, and of course
one knew the cross-sectional area of the rood. In a more advanced laboratory today, rather than sticking mercury-in-glass
thermometers into two mercuryfilled holes, one could measure the temperature at several points along the length of the rod by
means of thermocouples or thermistors welded into the rod. If there were no heat losses along the length of the rod, the
temperature gradient would be uniform along the rod. In practice, the thermistors would show a nonuniform temperature
gradient, and from this one could calculate and allow for the heat loss along the rod. Likewise the temperatures at the inflow
and outflow ends of the little helical tube could be measured with some tiny modern device, and all of these electrical
connections today would be connected to a computer, which would immediately do all the necessary calculations, including
correction of heat loss, and the thermal conductivity would be instantly displayed!
Lees developed the details of the equipment so that much smaller specimens could be used – e.g. a rod just a few cm in length
and a few mm in diameter – so that he could enclose it in a Dewar flask and make measurements down to the temperature of
liquid air. Three small coils of varnished copper wire were wound round the rod. (By varnished copper wire I mean copper
wire whose surface was painted with a layer of varnish of sufficient thickness to insulate the coils electrically but sufficiently
thin that good thermal contact with the rod was made.) One of these coils was wound round the upper end of the rod, and
supplied heat at that end. The other two coils could be slid up and down to any desired positions on the rod, and they served as
resistance thermometers. That is, the local temperature of the rod could be measured by measuring the resistance if the coils.
This set-up provided in principle what was necessary to determine the thermal conductivity of the rod, for the rate of input of
heat to the rod was determined by the current in the uppermost coil, and the temperature gradient down the rod was measured
with the two movable thermometer coils.
An interesting method that has been used (using a rod of roughly the same dimensions as in Searle's Rod experiment – that is
to say, about a foot (30 cm) long and one or two cm in diameter − is to pass an electrical current along length of the rod, thus
heating it. However, the two ends of the rod are kept at the same temperature (T1) by keeping them in constant-temperature
baths. The temperature of the rod is greatest (T2) at its mid-point. It can be shown, by a solution of the heat conduction
equation, that
2
V
σ therm  = σ elect  . (18.2.1)
8 (T2 − T1 )

Jeremy Tatum 3/25/2021 18.2.1 CC-BY-NC https://phys.libretexts.org/@go/page/7327

Here, V is the electrical potential difference, in volts, across the ends of the rod, σtherm is the thermal conductivity in W m−1
K−1, and σelect is the electrical conductivity, in S m−1 . I haven't derived that equation here (if I can, I may do so later!), but at
least you can (I hope!) show that it is dimensionally correct. This method has been used to measure the ratio of the thermal to
the electrical conductivity down to temperatures of a few kelvin, as well as at high temperatures.
18.2.2 Solid Poor Conductors (Non-metals)
The most obvious modification that has to be made for the measurement of the thermal conductivity of a poor conductor is in
the shape of the sample to be measured. Instead of a long, thin rod, one needs a thin disc. In Lees' Discs experiment, the disc-
shaped sample is clamped between two copper discs, one of which is heated with an electrical coil. The temperatures of the
two copper discs are measured with thermocouples. This gives enough information, in principle, for the determination of the
thermal conductivity, but, as in all thermal experiments, there are numerous refinements both for minimizing heat losses, and
for allowing for what heat losses remain.
18.2.3 Liquids and Gases
Several methods have been used. Here I mention one straightforward method that has been used for both liquids and gases (i.e.
fluids). The fluid is held in a long cylinder of radius b. A wire, of radius a, down the axis of the cylinder is heated electrically.
The temperature T2 of the wire can be measured by measuring its resistance, and the temperature T1 the wall of the cylinder
can be measured with a thermocouple. The rate of flow of heat Q ˙
through the fluid is equal to the rate at which electrical
2
energy is supplied to the wire - I R. Anyone who has been able to work out the electric field between two coaxial cylinders in
an elementary electricity course (see the Electricity and Magnetism section of these Notes) will be able to work out the
relevant equations, but here goes, anyway.
Consider an elemental cylindrical shell, radii r, r + dr. Its area is 2πrl, where l is the length of the cylinder. If the temperature
gradient there is dT/dr (which is negative), the rate of flow of heat, Q˙
(which is known, as explained above) is given by
dT
˙
Q = −2πrlσ . (18.2.2)
dr

Integrate this from r = a, T = T2 to r = b, T = T1, and we get

˙
Q b
σ = ln( ). (18.2.3)
2πl (T2 − T1 ) a

This assumes a very long cylinder, and ignores end effects. End effects can be kept small by using a long, thin tube, and can be
allowed for by experimenting with tubes of several lengths.

Jeremy Tatum 3/25/2021 18.2.2 CC-BY-NC https://phys.libretexts.org/@go/page/7327

18.3: The Universal Gas Constant
If you had an ideal gas, all you would have to do is to measure its pressure, its temperature, and the volume occupied by a
mole, for then PV = RT. (Measuring P and T is relatively easy. Measuring the volume occupied by a mole is less so.) In real
life, however, we have to make measurements on real gases. What has to be done is to measure the product PV (at a given
temperature) at progressively lower and lower pressures, and extrapolate the value of PV/T to the limit of zero pressure. (See
notes in Chapter 6 on the compression factor.)

Jeremy Tatum 4/8/2021 18.3.1 CC-BY-NC https://phys.libretexts.org/@go/page/7328

18.4: Avogadro's Number and Boltzmann's Constant
Avogadro's number is best determined by electrolytic deposition. That is, you have to measure the quantity of electricity
(current times time) that will deposit a mole of a monovalent element from an electrolytic solution on to an electrode. This
quantity of electricity is generally called a faraday, and is about 96,484 coulombs, and is the product of the electronic charge
and Avogadro's number.
Boltzmann's constant is given by k = R/NA.
[It is likely that, in 2015, Avogadro’s Number and Boltzmann’s constant will be given defined values. See Section 6.1 of
Chapter 6.]

Jeremy Tatum 4/22/2021 18.4.1 CC-BY-NC https://phys.libretexts.org/@go/page/7329

18.5: Specific Heat Capacities of Solids and Liquids
In elementary instructional methods often used at high school, the method of mixtures is generally used. For example, to
measure the specific heat capacity of copper, one would need a calorimeter (a small cup) made of copper, and of known mass.
Pour a measured mass of boiling water (100 °C) into this. The temperature of the copper rises from room temperature, t1 °C, to
a final temperature, t2 °C, while the temperature of the copper falls from 100 °C to t2 °C. The specific heat capacity of the
copper is then given by mCu CCu (t2 − t1) = mH2O (100 − t2). Since the specific heat capacity of water is, by definition, 1 cal g−1
C°−1 (at least to the precision expected at this level of experimentation), the specific heat capacity of the copper is determined.
To determine the specific heat capacity of another liquid, you could pour a measured mass of the hot liquid into the calorimeter
(whose heat capacity is now known), and measure the fall in temperature of the liquid and the rise in temperature of the
calorimeter, and hence deduce the specific heat capacity of the liquid by means of a similar equation to the above.
To determine the specific heat capacity of another metal, for example, iron, one can warm an iron specimen (of measured
mass) to 100 °C, and then drop it into the copper calorimeter, which contains water at room temperature, t1 °C, and then
measure the final temperature t2 °C to which the iron cools down and the copper and water heat up. Then mCu CCu (t2 − t1) +
mH2O CH2O (t2 − t1) = mFe CFe (100 − t2).
In all such experiments, precautions must be taken to minimize heat losses, and to allow for such heat losses as remain. Most
of us will remember such experiments from our schooldays, and will remember how difficult it was to get reliable results, and
will be aware that there are much more accurate methods available. Furthermore, the method of mixtures measures the relative
values of the specific heat capacities of the materials being mixed, rather than their absolute specific heat capacities. This is all
right if we accept that the specific heat capacity of water is unity by definition, but, as soon as it is recognized that heat is a
form of energy, we want to be able to measure heat capacities in joules rather than in calories, and the method of mixtures does
not do this.
It must not be supposed, however, that the method of mixtures is confined to the schoolroom, and is never used in professional
research laboratories. It has been found particularly useful in the measurement of heat capacities at high temperatures. While
the details of such experiments are much more sophisticated than as described above, the principle of the method of mixtures
still remains.
Nevertheless it remains true that the method of mixtures is really a method of comparing the specific heat capacities of
different materials, or of comparing the specific heat capacity of a substance with that of water. The first reasonably accurate
direct determination of the amount of energy needed to raise the temperature of a measured mass of water through a measured
temperature rise was Joule’s famous experiment. In Joule's experiment, water was warmed by stirring it with paddles, which
were operated by a set of falling weights, and the amount of work done by these falling weights could be accurately calculated
in units of work (which, to Joule, were foot-pounds, but which today, we would calculate in joules.) To Joule, the object of the
experiment was to demonstrate that a given amount of work always produced the same amount of heat, and hence to determine
what he called the mechanical equivalent of heat. Today, we recognize the experiment as a direct measurement, in units of
mechanical work, of the specific heat capacity of water, no longer defined to be 1 calorie per gram per degree, but measured to
be 4184 joules per kilogram per kelvin. We can look back today at Joule's experiment in amazement – amazement not only at
how difficult it must have been and what great experimental skills it entailed, but amazement, too, at how accurate a result he
obtained. He wrote: "After reducing the result to the capacity for heat for a pound of water, it appeared that for each degree of
heat evolved by the friction of water, a mechanical power equal to that which can raise a weight of 890 lbs to the height of one
foot had been expended." Bearing in mind that his "degree of heat" would have been a Fahrenheit degree, this is equivalent to
4790 joules per kilogram per Celsius degree. In addition to his famous paddle-wheel experiment, Joule performed two other
experiments - of a quite different nature – to determine the "mechanical equivalent of heat", and he took, as the average of the
three experiments, a figure of 817 pounds, which, in modern units, would be equivalent to 4398 joules per kilogram per
Celsius degree – only five percent greater than the modern value.
Of course much more accurate measurements of the energy required to raise the temperature of a solid or a liquid can be made
by heating the sample electrically – that is, in the case of a liquid, immersing a heating coil in the liquid, or, in the case of a
solid, wrapping a heating coil around the solid. Admittedly, this does not have the direct frontal approach of heating the
sample by mechanical work, but at least the heat input (I2R) can be accurately measured. Of course, as in all thermal

Jeremy Tatum 4/8/2021 18.5.1 CC-BY-NC https://phys.libretexts.org/@go/page/7330

measurements, precautions must be taken to minimize heat losses, and to allow for what heat losses remain, and these
considerations must go into the detailed design of the experiment and its procedures. One technique is to surround the
calorimeter by an outer vessel, which, by means of suitably-designed thermostats, is kept always at the same temperature as
the calorimeter itself, thus (at least in principle) avoiding heat losses from the calorimeter altogether.
Quite precise measurements of the specific heat capacities of solids and liquids (relative to that of water) can be made with the
Bunsen Ice Calorimeter, which is described in Section 18.7.

Jeremy Tatum 4/8/2021 18.5.2 CC-BY-NC https://phys.libretexts.org/@go/page/7330

18.6: Specific Heat Capacities of Gases
We have to consider the measurement of the specific heat capacity at constant pressure and at constant volume.
The most famous of the early experiments to measure directly the specific heat capacity of a gas at constant pressure were
Regnault's experiments of around 1860. Gas from a large storage cylinder was passed at constant pressure (measured with a
manometer) through a series of helical copper tubes. The first helix was immersed in a constant high-temperature bath, which,
of course, warmed the gas up. The warm gas then continued its flow through a smaller helix, which was immersed in a small
copper calorimeter filled with cold water. The gas, of course, cooled down, and the water in the calorimeter warmed up. The
fall in temperature of the gas and the rise in temperature of the water were measured, and hence the specific heat capacity of
the gas at constant pressure was calculated. While the principle of the experiment was simple and straightforward, its actual
practical execution required an experimental skill of the very highest order.
The most famous of the early experiments to measure directly the specific heat capacity of a gas at constant volume is Joly's
differential steam calorimeter of around 1890. Two equal hollow copper spheres were suspended from the arms of a balance.
One of the spheres was filled with the gas under investigation; the other was evacuated (or at least as far as the vacuum
technology of the day could achieve). The two spheres were surrounded by a chamber into which steam could be pumped. I'm
not very good at art, but I'll try to indicate very schematically, in figure XVIII.1, what I am trying to describe.

Steam was pumped into the chamber, and some of it condensed on the two spheres. Naturally, more steam condensed on the
sphere that held the gas, and the mass of extra condensate was measured by adding weights to the other scale pan. The mass of
extra condensate times its specific latent heat of condensation was equated to the heat required to raise the temperature of the
gas inside the filled sphere from its initial room temperature to 100 °C. It was a brilliant experiment requiring superb
experimental skill.

Jeremy Tatum 4/8/2021 18.6.1 CC-BY-NC https://phys.libretexts.org/@go/page/7331

18.7: Latent Heat of Fusion
The most straightforward method for measuring the specific latent heat L of ice is to drop a lump of
Ice of mass m and specific latent heat L at its melting point T0 into a
Calorimeter of mass MC and specific heat capacity CC and initial (warm) temperature T2,
which contains
a mass MW of Water of specific heat capacity CW at the same warm temperature T2.
After the ice has melted everything comes to a final (cool) temperature T1. Then
m [L + CW (T2 − T0 )] = (MC CC + Mw Cw ) (T2 − T1 ) . (18.7.1)

If the temperatures in this equation are supposed to be in degrees Celsius, so that T0 = 0, and if masses are in grams and heat in
calories, so that C2 = 1, this equation becomes
m (L + T2 ) = (MC CC + MW ) (T2 − T1 ) . (18.7.2)

For good results, heat losses must be minimized and allowed for, and precautions must be taken to minimize and allow for any
water initially clinging to the lump of ice.
Quite precise measurements of the latent heat of fusion of ice can be made with the Bunsen Ice Calorimeter, an apparatus that
can also be used to measure specific heat capacities of other substances. My limited artistic skills with the computer do not
allow me to draw all the minute details of the practical construction of a Bunsen ice calorimeter that makes it a precision
instrument, but may, perhaps, suffice to show the general principles, in figure XVIII.2. A test-tube T, is fitted with an outer
glass sleeve S, the lower end of which leads to a manometer M. The portion of the sleeve above the level B is filled with air-
free pure water at its freezing point. The manometer from level A to level B is filled with mercury. The entire apparatus is
generally enclosed in a large ice-box, so that the entire apparatus is at 0°C. Some ice is formed outside the bottom of the test-
tube, at I. In order to measure the specific latent heat of fusion of ice, a measured quantity of hot water is poured into test-tube.
This water, in cooling down to 0°C, gives up a known amount of heat to the ice, some of which melts. So – how do you know
how much ice has melted? Water ice contracts on melting into liquid. Consequently the level B moves up and the level A
(which can be in a quite narrow capillary tube) goes down, so the reduction in volume (and hence the mass of ice melted) can
be determined quite accurately. Thus the latent heat of fusion of ice can be determined. Once the equipment has been
calibrated (i.e. when we know how much movement of level A corresponds to how much transfer of heat), the calorimeter can
be used to measure specific heat capacities of other substances, simply by dropping a known mass of the substance at a known
temperature into the test-tube, and measuring the movement of the level A. It will be understood, I think, that, in using the
apparatus to measure the specific latent heat of ice, it is necessary to know the densities of ice and of water precisely. To use it
for measuring the specific heat capacities of other substances, it is not necessary to know this, or even to know the specific
latent heat of fusion of ice. You do have to know the specific heat capacity of water – which is not much of a burden,
especially if you are content to express heat in calories!

Jeremy Tatum 3/4/2021 18.7.1 CC-BY-NC https://phys.libretexts.org/@go/page/8678

18.8: Coefficient of Expansion
If a specimen can be obtained in the form of a long rod, the simplest and most direct method is merely to rest the rod
horizontally on some support, immersed in a water bath by which means the temperature can be varied. Two scratches, one at
each end of the rod, can be observed with a pair of measuring microscopes held on a support at constant temperature. The
measuring microscopes can either be fixed and fitted with a fine scale in the eyepiece of each, or they may be movable by
means of a fine precision screw (96 turns to the inch). The movement of the microscopes can be measured either by means of a
wheel fitted with a vernier that turns the precision screw, or by attaching a corner reflector to each moving microscope, and
reflecting a laser beam off the reflector and counting the number of half wavelengths traversed by the microscope.
If the specimen cannot be obtained in the form of a long rod, but can be obtained in the form of a thin, flat plate with parallel
faces, another method can be used. A hole might be cut in the flat specimen, and the specimen can be rested on top of a flat
glass plate. A second flat glass plate rests on the upper face of the specimen. The arrangement can be illuminated with an
extended monochromatic light source, to create a system of interference fringes. When the temperature is raised, the specimen
expands and the distance between the glass plates increases by an amount that can be measured by measuring the movement of
the interference fringes. Some materials may not be easily obtainable either in the form of a long rod or a thin plate, but
perhaps they can be obtained in the form of a small cube. The specimen is placed side-by-side with a similar cube of quartz,
whose expansion coefficient is very small, the two resting on the horizontal surface of a polished shiny metal or glass block.
On top of the two specimens rests a thin, flat glass plate. A narrow beam of light, preferably from a laser, is directed from
above to the arrangement, and two reflections are observed, one from the thin glass plate that rests on top of the specimen and
its quartz companion, the other from the upper surface of the block on which the specimens are resting. When the specimen
and the quartz are warmed, the specimen expands more than the quartz does, and so the upper thin glass plate tilts, and the
reflection from it is deflected. The displacement of one reflected beam from the other can be measured with a microscope, and
hence the tilt of the upper glass plate can be calculated, and hence the excess of expansion of the specimen over that of the
quartz can be determined. The experiment gives the difference in expansion coefficient between the specimen and the quartz.
The latter is very small, and its exact value need not be known with great precision in order to obtain the absolute coefficient
of expansion of the specimen.
For nonvolatile liquids, a weight thermometer can be used. This is a glass (or, better, fused quartz) bulb fitted with a narrow
capillary tube as shown in figure XVIII.3.

The bulb (whose weight empty is known) is completely filled (including the capillary to the very tip) with the liquid, and
weighed, so that the weight, hence mass, of the liquid is known. The temperature is increased, so some liquid escapes, and the
bulb is weighed again. Thus we know the weight of the liquid held by the bulb at two temperatures. If we assume that the
volume is constant (the bulb being made of fused quartz) this enables us to calculate the coefficient of expansion of the liquid.
Of course, the bulb does expand a little, so what we have determined is the difference between the volume expansions of the
liquid and the quartz. If we know the volume expansion of the quartz (which need not be known to high precision, since it is
small), we can then determine the absolute coefficient of expansion of the liquid.
In another method for measuring the coefficient of expansion of liquids, the liquid is contained in a U-tube, the two arms of
which are maintained at different temperatures, as shown in figure XVIII.4. The upper ends of the two arms of the U-tube are
connected to vertical tubes containing mercury for controlling and measuring the pressure. The apparatus is maintained so that
the volumes of the liquids in the two arms of the U-tube are equal – but because the two arms are at different temperatures,
their densities (hence specific volumes) are different, so a little extra mercury is needed to balance the hot arm against the cold
arm. Thus it is possible to determine the difference in densities at the two temperatures, and hence to determine the volume

Jeremy Tatum 4/29/2021 18.8.1 CC-BY-NC https://phys.libretexts.org/@go/page/8679

coefficient of expansion. The figure shows the principle of the method; some practical refinements are needed in the actual
equipment.

Jeremy Tatum 4/29/2021 18.8.2 CC-BY-NC https://phys.libretexts.org/@go/page/8679

Index
A
exact differential M
adiabatic decompression
2.5: Second Derivatives and Exact Differentials Maxwell relations
12.5: Summary, the Maxwell Relations, and the
15.2: Adiabatic Decompression
adiabatic demagnetization
F Gibbs-Helmholtz Relations
Fahrenheit scale
15.3: Adiabatic Demagnetization
3.3: Temperature Scales I R
Réaumur scale
B I 3.3: Temperature Scales I
binary alloys inexact differential radius of gyration
17.9: Binary Alloys
2.5: Second Derivatives and Exact Differentials 4.1: Error Function
Integrating Factors Rankine scale
C 2.5: Second Derivatives and Exact Differentials 3.3: Temperature Scales I
Celsius scale
3.3: Temperature Scales I
J S
Joule coefficient specific heat
D 10.2: The Joule Experiment 7.3: Entropy
Debye temperature specific heat capacity
Joule Experiment
8.10: Heat Capacities of Solids
10.2: The Joule Experiment 7.3: Entropy

E K T
Enthalpy ternary alloy
Kelvin scale
9: Enthalpy
3.3: Temperature Scales I 17.10: Ternary Alloys
entropy
7.3: Entropy
L Z
Equation of state latent heat of freezing zeroth law of thermodynamics
2.5: Second Derivatives and Exact Differentials
9.2: Change of State 3.2: Zeroth Law of Thermodynamics
Error Function Law of Dulong and Petit
4.1: Error Function
8.10: Heat Capacities of Solids
Glossary
Sample Word 1 | Sample Definition 1