Notes

LECTURE NOTES ON
THERMODYNAMICS
Joseph M. Powers
Department of Aerospace and Mechanical Engineering

University of Notre Dame
Notre Dame, Indiana 46556-5637
USA
updated
03 July 2024, 3:40pm
2
CC BY-NC-ND. 03 July 2024, J. M. Powers.

Contents
Preface 9
1 Introduction 11
1.1 Some semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2 Historical milestones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 Philosophy of science note . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4 Some practical applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5 Example to illustrate homework solution style . . . . . . . . . . . . . . . . . 24
1.6 Thermodynamic system and control volume . . . . . . . . . . . . . . . . . . 29
1.7 Macroscopic versus microscopic . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.8 Properties and state of a substance . . . . . . . . . . . . . . . . . . . . . . . 33
1.9 Processes and cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.10 Fundamental variables and units . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.11 Zeroth law of thermodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.12 Secondary variables and units . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2 Properties of a pure substance 43

2.1 The pure substance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.2 Vapor-liquid-solid phase equilibrium . . . . . . . . . . . . . . . . . . . . . . . 43
2.3 Independent properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.4 Thermal equations of state . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4.1 Ideal gas law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.4.2 Non-ideal thermal equations of state . . . . . . . . . . . . . . . . . . 62
2.4.2.1 van der Waals . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4.2.2 Redlich-Kwong . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.4.3 Compressibility factor . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.4.4 Tabular thermal equations of state . . . . . . . . . . . . . . . . . . . 65
2.4.4.1 Saturated liquid-vapor water, temperature tables . . . . . . 66
2.4.4.2 Saturated liquid-vapor water, pressure tables . . . . . . . . 68
2.4.4.3 Superheated water tables . . . . . . . . . . . . . . . . . . . 71
2.4.4.4 Compressed liquid water tables . . . . . . . . . . . . . . . . 71
2.4.4.5 Saturated water, solid-vapor . . . . . . . . . . . . . . . . . . 71
3
4 CONTENTS
2.4.4.6 Tables for other materials . . . . . . . . . . . . . . . . . . . 73

2.4.4.7 Linear interpolation of tabular data . . . . . . . . . . . . . . 73
2.4.4.7.1 Single interpolation . . . . . . . . . . . . . . . . . . 74
2.4.4.7.2 Double interpolation . . . . . . . . . . . . . . . . . 76
3 The first law of thermodynamics 81

3.1 Mathematical preliminaries: exact differentials . . . . . . . . . . . . . . . . . 81
3.1.1 Partial derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.1.2 Total derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.2 Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.2.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.2.2 Work for a simple compressible substance . . . . . . . . . . . . . . . 94
3.2.3 Other forms of work . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.3 Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.4 Representations of the first law . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.4.1 Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.4.2 Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.5 Specific internal energy for general materials . . . . . . . . . . . . . . . . . . 122
3.6 Specific enthalpy for general materials . . . . . . . . . . . . . . . . . . . . . 123
3.7 Specific heat capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.8 Caloric equations of state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.8.1 Ideal gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.8.1.1 Calorically perfect . . . . . . . . . . . . . . . . . . . . . . . 128
3.8.1.2 Calorically imperfect . . . . . . . . . . . . . . . . . . . . . . 131
3.8.2 Liquids and solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.8.3 General materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.9 Time-dependency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
3.10 Final comments on conservation . . . . . . . . . . . . . . . . . . . . . . . . . 155
4 First law analysis for a control volume 157

4.1 Detailed derivations of control volume equations . . . . . . . . . . . . . . . . 158
4.1.1 Relevant mathematics . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.1.1.1 Fundamental theorem of calculus . . . . . . . . . . . . . . . 159
4.1.1.2 Divergence theorem . . . . . . . . . . . . . . . . . . . . . . 159
4.1.1.3 Leibniz’s rule . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.1.1.4 General transport theorem . . . . . . . . . . . . . . . . . . . 165
4.1.1.5 Reynolds transport theorem . . . . . . . . . . . . . . . . . . 166
4.1.1.6 Fixed (control) volumes . . . . . . . . . . . . . . . . . . . . 167
4.1.2 Conservation axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.1.2.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.1.2.2 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.1.2.2.1 Total energy term . . . . . . . . . . . . . . . . . . 179

CONTENTS 5
4.1.2.2.2 Work term . . . . . . . . . . . . . . . . . . . . . . 180

4.1.2.2.3 Heat transfer term . . . . . . . . . . . . . . . . . . 181
4.1.2.2.4 The first law of thermodynamics . . . . . . . . . . 181
4.2 Mass conservation in brief . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
4.3 Energy conservation in brief . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
4.4 Some devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
4.4.1 Throttling device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
4.4.2 Nozzles and diffusers . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4.4.3 Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
4.4.4 Pumps and compressors . . . . . . . . . . . . . . . . . . . . . . . . . 198
4.4.5 Heat exchanger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
4.5 Introduction to the Rankine cycle . . . . . . . . . . . . . . . . . . . . . . . . 203
4.6 Preview: equations of continuum mechanics . . . . . . . . . . . . . . . . . . 209
4.6.1 Full set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
4.6.2 Static solids equations . . . . . . . . . . . . . . . . . . . . . . . . . . 211
4.6.3 Incompressible fluid mechanics equations . . . . . . . . . . . . . . . . 212
4.6.4 Compressible fluid mechanics equations . . . . . . . . . . . . . . . . . 212
4.6.5 Heat transfer in incompressible static solids . . . . . . . . . . . . . . 213
5 The second law of thermodynamics 217

5.1 Statements of the second law . . . . . . . . . . . . . . . . . . . . . . . . . . 218
5.1.1 Entropy-based statement . . . . . . . . . . . . . . . . . . . . . . . . . 219
5.1.2 Clausius statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
5.1.3 Kelvin-Planck statement . . . . . . . . . . . . . . . . . . . . . . . . . 220
5.1.4 Carathéodory statement . . . . . . . . . . . . . . . . . . . . . . . . . 221
5.1.5 Judeo-Christian statement . . . . . . . . . . . . . . . . . . . . . . . . 223
5.1.6 Historian-inspired statement . . . . . . . . . . . . . . . . . . . . . . . 224
5.1.7 Literature-inspired statement . . . . . . . . . . . . . . . . . . . . . . 224
5.1.8 Food service-inspired statement . . . . . . . . . . . . . . . . . . . . . 224
5.2 Reversible and irreversible processes . . . . . . . . . . . . . . . . . . . . . . . 224
5.3 Analysis of Carnot heat engines . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.4 The absolute temperature scale . . . . . . . . . . . . . . . . . . . . . . . . . 229
5.5 Analysis of Carnot refrigerators and heat pumps . . . . . . . . . . . . . . . . 231
5.6 Rejected thermal energy on a national scale . . . . . . . . . . . . . . . . . . 233
6 Entropy 237
6.1 Theoretical development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
6.2 Second law in terms of entropy . . . . . . . . . . . . . . . . . . . . . . . . . 243
6.3 The Gibbs equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
6.4 Entropy for ideal gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
6.4.1 Calorically perfect . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
6.4.2 Calorically imperfect . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

6 CONTENTS
6.5 Entropy for an incompressible solid or liquid . . . . . . . . . . . . . . . . . . 258

6.6 Iso- curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
6.6.1 Isochores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
6.6.2 Isobars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
6.6.3 Isentropes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
6.7 Isentropic relations for an ideal gas . . . . . . . . . . . . . . . . . . . . . . . 261
6.7.1 Calorically perfect . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
6.7.2 Calorically imperfect . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
6.8 Two cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
6.8.1 Carnot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
6.8.2 Otto . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
6.9 Entropy of thermo-mechanical mixing . . . . . . . . . . . . . . . . . . . . . . 291
6.10 Probabilistic approach to entropy . . . . . . . . . . . . . . . . . . . . . . . . 298
6.11 Summary statement of thermodynamics . . . . . . . . . . . . . . . . . . . . 309
7 Second law analysis for a control volume 311

7.1 Irreversible entropy production . . . . . . . . . . . . . . . . . . . . . . . . . 311
7.2 Bernoulli principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
7.2.1 Incompressible limit . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
7.2.2 Calorically perfect ideal gas limit . . . . . . . . . . . . . . . . . . . . 320
7.2.3 Torricelli’s formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
7.3 Component efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
8 Cycles 329
8.1 Rankine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
8.2 Brayton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
8.3 Refrigeration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
9 Mathematical foundations 357

9.1 Maxwell relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
9.2 Functions of two independent variables . . . . . . . . . . . . . . . . . . . . . 359
9.3 Legendre transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
9.4 Specific heat capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
9.5 The first law and coordinate transformations . . . . . . . . . . . . . . . . . . 369
9.6 The van der Waals gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
9.7 The Redlich-Kwong gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
9.8 Adiabatic sound speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
9.9 Introduction to compressible flow . . . . . . . . . . . . . . . . . . . . . . . . 387
9.9.1 Acoustics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
9.9.2 Compressible Bernoulli principle . . . . . . . . . . . . . . . . . . . . . 391
9.9.3 Steady flow with area change . . . . . . . . . . . . . . . . . . . . . . 395
Appendix 399

CONTENTS 7
Bibliography 419
Index 428

8 CONTENTS

Preface
These are lecture notes for AME 20231, Thermodynamics, a sophomore-level undergraduate
course taught in the Department of Aerospace and Mechanical Engineering at the University
of Notre Dame. The objective of the course is to survey practical and theoretical problems
in classical thermodynamics. The emphasis is on the axiomatic approach to equilibrium
thermodynamics and its application to practical devices. However, there are some brief
appeals to the non-equilibrium and molecular underpinnings of thermodynamics.
The notes draw often on the text specified for the course, Borgnakke and Sonntag’s
(BS) Fundamentals of Thermodynamics, Ninth Edition, John Wiley, Hoboken, NJ, 2017,
especially Chapters 1-7, 9, 10, 12, and 15. In general the nomenclature of BS is used, and
much of the notes follow a similar structure as the text. Some example problems have been
directly taken from BS and other texts; specific citations are given where appropriate.
These notes emphasize problem-solving and rigorous development of the underlying clas-
sical theory; the student should call on other reference materials to fill some of the gaps.
It should also be remembered that practice is essential to the learning process; the student
would do well to apply the techniques presented here by working many problems.
An unusual aspect of the notes is the inclusion of accompanying text and figures related
to the history of classical thermodynamics. Where possible, citations of original work are
provided; most citations are linked directly to the original source material via click-able
text. Additionally, an extensive annotated bibliography is provided. This is an attempt
to fill gaps in most modern undergraduate texts, which at best provide small guidance to
original thermodynamics scholarship.
The notes and course data can be found at https://www3.nd.edu/∼powers/ame.20231.
At this stage, anyone is free to make copies for their own use. I would be happy to hear
about suggestions for improvement. Thanks to many of the students, past and present,
whose diligence in identifying dozens of errors has been useful.
Joseph M. Powers
powers@nd.edu
https://www3.nd.edu/∼powers
Notre Dame, Indiana; USA
CC BY: $ = 03 July 2024
\
The content of this book is licensed under Creative Commons Attribution-Noncommercial-No Derivative Works 3.0.
9
10 CONTENTS

Chapter 1
Introduction
Read BS, Chapter 1
1.1 Some semantics

We introduce here classical thermodynamics. The word “thermo-dynamic,” used first by
Thomson (later Lord Kelvin),1 has Greek origin, and is translated2 as the combination of
• θǫ́ρµη, therme: heat, and
• δύναµις, dynamis: power.
An image of Thomson and his 1849 first use of the word is given in Fig. 1.1.
Figure 1.1: William Thomson (Lord Kelvin) (1824-1907), Ulster-born Scottish scientist;
image from https://mathshistory.st-andrews.ac.uk/Biographies/Thomson.html and
image giving the first use of “thermo-dynamic” extracted from his 1849 work.
The modifier “classical” is used to connote a description in which quantum mechanical

effects, the molecular nature of matter, and the statistical nature of molecular behavior
1
W. Thomson (later Kelvin), 1849, “An account of Carnot’s theory of the motive power of heat; with
numerical results deduced from Regnault’s experiments on steam,” Transactions of the Royal Society of
Edinburgh, 16:541-574. See also C. W. Smith, 1977, “William Thomson and the creation of thermodynamics:
1840-1855,” Archive for History of Exact Sciences, 16(3): 231-288.
2
All Greek spellings and etymologies are drawn from the Oxford English Dictionary, 2nd edition, 1989.
11
12 CHAPTER 1. INTRODUCTION
are not considered in any detail. These effects will not be completely ignored; however,
they will be lumped into simple averaged models that are valid on the macroscale. As an
example, for ordinary gases, our classical thermodynamics will be valid for systems whose
characteristic length scale is larger than the mean free path between molecular collisions.
For air at atmospheric density, this about 0.1 µm = 10−7 m).
Additionally, “classical” also connotes a description in which the effects of finite time-
dependency are ignored. In this sense, thermodynamics resembles the field P of statics from
2 2
Newtonian mechanics. Recall Newton’s second law of motion, m d x/dt = F, where m
is the mass,
P x is the position vector, t is time, and F is the force vector. In the statics limit
where F = 0, inertial effects are ignored, as is time-dependency. Now, a Newtonian would
consider dynamics to imply motion, and so would consider thermodynamics to imply the
time-dependent motion of heat. So a Newtonian would be more inclined to call the subject
of these notes “thermostatics.” However, if we return to the earlier Greek translation of
dynamics as power, we are actually truer to the classical connotation of thermodynamics.
For the fundamental interplay of thermodynamics is that between so-called thermal energy
(as might be thought of when considering heat) and mechanical energy (as might be thought
of when considering power, a work rate). More formally, we will take the definition
• thermodynamics: the science that deals with heat and work and those properties of
matter that relate to heat and work.
One of the main goals of these notes will be to formalize the relationship between heat, work,
and energy.
We close this section by noting that the concept of energy has evolved through time, but
has ancient origins. The word itself had its first recorded use by Aristotle.3 His portrait,
along with an image of the relevant section of an 1818 translation of his treatise, is depicted in
,
Figs. 1.2. In the Greek, the word ǫνǫ́ργǫια, “energeia,” connotes activity or operation. While
the word was known to Aristotle, its modern usage was not; it was the English polymath
Thomas Young who first used the word “energy,” consistent with any sort of modern usage,
in this case kinetic energy.4 A portrait of Young and an image of his text defining energy,
in actuality kinetic energy, in modern terms are shown in Fig. 1.3.
3
Aristotle, ∼335 BC, The Rhetoric, Poetic, and Nicomachean Ethics, Book III, Ch. XI, English transla-
tion by T. Taylor, 1818, Black, London, see pp. 242-243.
4
T. Young, 1807, Lectures on Natural Philosophy, William Savage, London, p. 52.

1.1. SOME SEMANTICS 13
Figure 1.2: Aristotle (384 BC-322 BC), Greek philosopher who gives the first recorded use
of the word “energy” and whose method of logic permeates classical thermodynamics; im-
age from https://mathshistory.st-andrews.ac.uk/Biographies/Aristotle.html and
an image of Aristotle’s usage of the word “energy” from his Nicomachean Ethics.
Figure 1.3: Thomas Young (1773-1829), English natural philosopher; image from
https://en.wikipedia.org/wiki/Thomas Young (scientist), and a reproduction of his
more modern 1807 definition of (kinetic) energy.

Finally, though she did not use the word “energy,” the notion of what is now known as
kinetic energy being related to the square of velocity was first advanced by du Châtelet,5
pictured in Fig. 1.4.
Figure 1.4: Gabrielle Émilie Le Tonnelier de Breteuil, marquise du Châtelet (1706-1749),

French physicist; image from https://en.wikipedia.org/wiki/Emilie du Chatelet.
1.2 Historical milestones

Thermodynamics has a long history; unfortunately, it was not blessed with the crispness of
development that mechanics realized with Newton. In fact, its growth is filled with false
steps, errors, and debate that continues to this day. Truesdell6 and Müller7,8 summarize the
development in their idiosyncratic histories. Some of the milestones of its development are
given here:
• first century AD: Hero of Alexandria documents many early thermal engines.
• 1593: Galileo develops a water thermometer.
• 1650: Otto von Guericke designs and builds the first vacuum pump.
• 1662: Robert Boyle develops his law for isothermal ideal gases.
5
É. du Châtelet, 1740, Institutions de Physique, Chez Prault, Paris.
6
C. Truesdell, 1980, The Tragicomical History of Thermodynamics, 1822-1854, Springer, New York.
7
I. Müller, 2007, A History of Thermodynamics: the Doctrine of Energy and Entropy, Springer, Berlin.
8
I. Müller and W. H. Müller, 2009, Fundamentals of Thermodynamics and Applications with Historical
Annotations and Many Citations from Avogadro to Zermelo, Springer, Berlin.

1.2. HISTORICAL MILESTONES 15
• 1679: Denis Papin develops his steam digester, forerunner to the steam engine.
• 1687: Isaac Newton publishes PhilosophiæNaturalis Principia Mathematica.
• 1698: Thomas Savery patents an early steam engine.
• 1710: Thomas Newcomen creates a more practical steam engine.
• 1760s: Joseph Black develops calorimetry.
• 1780s: James Watt improves the steam engine.
• 1798: Benjamin Thompson (Count Rumford) considers the mechanical equivalent of

heat from cannon boring experiments.
• 1824: Nicolas Lèonard Sadi Carnot discusses idealized heat engines.
• 1840: Germain Henri Hess considers an early version of the first law of thermodynamics
for work-free chemical reactions.
• 1840s: Julius Robert von Mayer relates heat and work.
• 1840s: James Prescott Joule relates heat and work.
• 1847: Hermann von Helmholtz publishes his theory of energy conservation.
• 1848: William Thomson (Lord Kelvin) postulates an absolute zero of temperature.
• 1850: Rudolf Julius Emanuel Clausius formalizes the second law of thermodynamics.
• 1865: Clausius introduces the concept of entropy.
• 1871: James Clerk Maxwell develops the Maxwell relations.
• 1870s: Josiah Willard Gibbs further formalizes mathematical thermodynamics.
• 1870s: Maxwell and Ludwig Boltzmann develop statistical thermodynamics.
• 1889: Gibbs develops statistical mechanics, giving underlying foundations for classical
and statistical thermodynamics.
• 1909: Constantin Carathéodory publishes a rigorous axiomatic formulation of the laws

of thermodynamics.
• 1931: Lars Onsager advances non-equilibrium thermodynamics with his reciprocity

relations.
Much development of the twentieth century was recognized by the Nobel prize, including
that of Nobel laureates

• Jacobus Henricus van’t Hoff (1901),
• Johannes van der Waals (1910),
• Heike Kamerlingh Onnes (1913),
• Max Planck (1918),
• Walther Nernst (1920),
• Albert Einstein (1921),
• Erwin Schrödinger (1933),
• Enrico Fermi (1938),
• Percy Bridgman (1946),
• William Giauque (1949),
• Lars Onsager (1968),
• Ilya Prigogine (1977), and
• Kenneth Wilson (1982).
Strictly speaking, Newton did not consider the subject matter of thermodynamics. How-
ever, much of his work is concerned with energy. His theories are appropriate only for
mechanical energy. The notion that thermal energy existed and that it could be equivalent
to mechanical energy was not part of Newtonian mechanics. However, temperature was
known to Newton, as was Boyle’s law. When he tried to apply his theories to problems
of thermodynamics, such as calculation of the speed of sound in air, they notably failed.
This will be discussed at the conclusion of these notes in Ch. 9.8. The reason for the failure
required consideration of the yet-to-be-developed second law of thermodynamics.
1.3 Philosophy of science note

As with science in general, thermodynamics is based on empirical observation. Moreover, it
is important that those observations be repeatable. A few postulates, also known as axioms,
will serve as the foundation of our science. Following Occam’s razor,9 we shall seek as few
axioms as possible to describe this behavior. We will supplement these axioms with some
necessary definitions to describe nature. Then we shall use our reason to deduce from the
axioms and definitions certain theorems of engineering relevance.
This approach, that has its foundations in Aristotelian methods, is not unlike the ap-
proach taken by Euclid to geometry, Aquinas to theology, or Newton to mechanics. A

1.4. SOME PRACTICAL APPLICATIONS 17
Figure 1.5: Euclid of Alexandria (∼ 325 BC- ∼265 BC), Greek mathematician whose rational
exposition of geometry formed a model for how to present classical thermodynamics; image
from https://mathshistory.st-andrews.ac.uk/Biographies/Euclid.html.
depiction of Euclid is given in Fig. 1.5. Consider for example that Euclid defined certain en-
tities such as points, lines, and planes, then adopted certain axioms such as parallel lines do
not meet at infinity, and went on to prove a variety of theorems. Classical thermodynamics
follows the same approach. Concepts such as system and process are defined, and axioms,
known as the laws of thermodynamics, are proposed in such a way that the minimum amount
of theory is able to explain the maximum amount of data.
Now, in some sense science can never be formally proved; it can only be disproved.
We retain our axioms as long as they are useful. When faced with empirical facts that
unambiguously contradict our axioms, we are required to throw away our axioms and develop
new ones. For example, in physics, the Michelson-Morley experiment forced Einstein to
abandon the axioms of Euclid, Newton, and Clausius for his theory of general relativity.
It turns out that we can still use these axioms, as long as we are considering problems in
which the speed of our reference frame is far less than the speed of light. In an example
from biology that is the topic of a popular science book,10 it was noted that it was once
believed that all swans were white. This working hypothesis was perfectly acceptable until
1697, when a black swan was discovered in Australia. Thus, the “theory” (though it is not
a highly profound theory) that all swans were white was unambiguously discredited. It will
be briefly seen in Ch. 6.10 that non-classical thermodynamics actually has a deep relation
to probability and statistics and information, a topic that transcends thermodynamics.
1.4 Some practical applications

It turns out that the classical approach to thermodynamics has had success in guiding the
engineering of devices. People have been building mechanical devices based on thermal
9
from William of Occam, (c. 1288-c. 1348) English Franciscan friar and philosopher.
10
N. N. Taleb, 2007, The Black Swan: The Impact of the Highly Improbable, Random House, New York.

energy inputs for centuries, without the benefit of a cleanly enunciated theory. Famously,
Hero of Alexandria, perhaps the first recognized thermal engineer, documented a variety of
devices. These include an early steam engine11 known as the æolipile, pumps, and a device
to use fire to open doors. Hero and a nineteenth century rendition of his steam engine are
shown in Fig. 1.6. While Hero’s contributions are a matter of some speculation inspired by
Figure 1.6: Hero of Alexandria (10-70 AD), Greek engineer and mathematician who devised
some early ways to convert thermal energy into mechanical energy, and his æolipile; images
from https://en.wikipedia.org/wiki/Hero of Alexandria.
ancient artistry, the much later works of Denis Papin (1647-1712) are more certain. Papin
invented the so-called steam digester, that anticipated both the pressure cooker and the
steam engine. The device used steam power to lift a weight. Depictions of Papin and his
device are found in Fig. 1.7. Significant improvements were led by James Watt (1736-1819)
of Scotland. An image of Watt and one of his engines is shown in Fig. 1.8.
11
P. Keyser, 1990, “A new look at Heron’s steam engine,” Archive for History of Exact Sciences, 44(2):
107-124.

Figure 1.7: French-born inventor Denis Papin (1647-1712) and his steam digester; images
from https://en.wikipedia.org/wiki/Denis Papin.

a) b)
Figure 1.8: a) Scottish engineer James Watt (1736-1819); image from

https://en.wikipedia.org/wiki/James Watt, b) Sketch of one of Watt’s steam en-
gines; image from W. J. M. Rankine, 1859, A Manual of the Steam Engine and Other Prime
Movers, First Edition, Griffin, London.

These engines were adopted for transportation. In 1807, the American engineer Robert
Fulton (1765-1815) was the first to use steam power in a commercial nautical vessel, the
Clermont, that was powered by a Boulton and Watt steam engine. Soon after, in 1811 in
Scotland, the first European commercial steam vessel, the Comet, embarked. We have a
sketch of the Comet and its steam power plant in Fig. 1.9. On land, steam power soon
Figure 1.9: Sketch of the Comet and its steam engine; image from W. J. M. Rankine, 1859,
A Manual of the Steam Engine and Other Prime Movers, First Edition, Griffin, London.
enabled efficient rail transportation. A famous early steam locomotive was the English
engineer Robert Stephenson’s (1803-1859) Rocket, sketched in Fig. 1.10.
Figure 1.10: Sketch of the Rocket; image from W. J. M. Rankine, 1859, A Manual of the
Steam Engine and Other Prime Movers, First Edition, Griffin, London.

The effect of steam power, a contribution driven by engineers, on the development of

the world remains remarkable. It is what is commonly known as a disruptive technology as
its widespread adoption displaced other well-established technologies. While it is difficult to
quantify historical pronouncements, it is likely that the effect on the world was even more
profound than the introduction of networked computers in the late twentieth century. In
short, steam power was the linchpin for the industrial revolution. Steam power replaced
animal power as a prime mover throughout much of the world and, where implemented,
enabled rapid development of broad economic segments: mining, manufacturing, land and
sea transportation, among others. Large scale population movements ensued as opportunities
in urban manufacturing centers made industrial work more appealing than agricultural work.
Certainly, changes precipitated by the advent of steam power were contributing factors in
widespread social unrest in the nineteenth century, ranging from labor strife to war between
nation states.
Some more modern devices, listed here, are relevant to thermodynamics:
• simple steam power plant,
• fuel cells,
• vapor-compression refrigeration cycle,
• air separation plant,
• the gas turbine, and
• the chemical rocket engine.
As an example, the main power plant of the University of Notre Dame, depicted in Fig. 1.11,
is based on a steam power cycle that will be a topic of study in this course. Additionally,
one might consider the following topics to have thermodynamic relevance:
• gasoline and Diesel engines,
• the weather,
• cooking,
• heating, ventilation, air conditioning, and refrigeration (HVAC&R), or
• materials processing (metals, polymers, etc.).

Figure 1.11: University of Notre Dame Power Plant; image from Matt Cashore, University
of Notre Dame.

1.5 Example to illustrate homework solution style

Proper technical communication is important for engineering. Here is an example of how
one might construct a homework solution. We take an example involving mechanical energy
from introductory physics:
Example 1.1
A mass of m = 1 kg is initially at rest and is dropped from a height of y = yo = 10 m above the
ground, where gravitational acceleration g = 9.81 m/s2 . Neglect drag forces. Find the time to reach
the ground, the kinetic energy as a function of time, and the potential energy as a function of time.
Plot key results.
The scenario is sketched in Fig. 1.12. The principle governing the motion of the body is Newton’s
m = 1 kg
g = 9.81 m/s2
yo = 10 m
y=0m
Figure 1.12: Sketch of a configuration for particle motion in a gravitational field.
second law, embodied in a second order differential equation. The only force is the gravitational force
acting in the negative y direction. This gives the equation
d2 y
m = −mg. (1.1)
dt2
Note the mass m cancels here, giving
d2 y
= −g. (1.2)
dt2
Integrate once to get
dy
= −gt + C1 , (1.3)
dt
where C1 is a constant. Integrate a second time to get
1
y(t) = − gt2 + C1 t + C2 . (1.4)
2
We need two initial conditions for this second order ordinary differential equation. At time t = 0, we
know from the problem statement that
dy
y(0) = yo , (0) = 0. (1.5)
dt

1.5. EXAMPLE TO ILLUSTRATE HOMEWORK SOLUTION STYLE 25
Applying the first initial condition, we get

1
yo = − g(0)2 + C1 (0) + C2 = C2 . (1.6)
2
Thus, we have
1
y(t) = − gt2 + C1 t + yo . (1.7)
2
Apply the second initial condition to get
0 = −g(0) + C1 = C1 . (1.8)
Thus, we have
1
y(t) = − gt2 + yo . (1.9)
2
For the velocity, we get
dy
= −gt. (1.10)
dt
When the mass reaches the ground, y = 0. Solving for the time when y = 0, we get
1
0 = − gt2 + yo , (1.11)
2
1 2
gt = yo , (1.12)
2 r
2yo
t = ± . (1.13)
g
We are considering t going forward, so we take the positive root, giving
r
2yo
t = , (1.14)
g
s
2(10 m)
= , (1.15)
9.81 sm2
= 1.43 s. (1.16)
The kinetic energy, KE, is

2
1 dy
KE = m , (1.17)
2 dt
1
= mg 2 t2 . (1.18)
2
The gravitational potential energy, P E, is
PE = mgy, (1.19)
1
= mgyo − mg 2 t2 . (1.20)
2
Note that
KE + P E = mgyo , (1.21)
that is a constant. Thus, mechanical energy is conserved here! By conserved, we mean it does not
change.

Numerically, we have for y(t), KE(t), and P E(t),
1 m
y(t) = − 9.81 2 t2 + (10 m), (1.22)
2 s
m 2
= − 4.904 2 t + (10 m), (1.23)
s
1 m 2
KE(t) = (1 kg) − 9.81 2 t , (1.24)
2
s
kg m2 2
= 48.12 t , (1.25)
s4

kg m2 J
= 48.12 t2 , (1.26)
s4 kg m2 /s2

J
= 48.12 2 t2 , (1.27)
s
m 1 m 2
P E(t) = (1 kg) 9.81 2 (10 m) − (1 kg) 9.81 2 t2 , (1.28)
s 2 s

J
= (98.1 J) − 48.12 2 t2 . (1.29)
s
The position as a function of time is plotted in Fig. 1.13. The kinetic, potential, and total mechanical
y (m)
10
t (s)
0.4 0.8 1.2
Figure 1.13: Position versus time for a particle accelerating in a gravitational field with no
drag force.
energies as functions of time are plotted in Fig. 1.14. One can tell by inspection that as potential energy
decreases, kinetic energy increases just as much, rendering the total mechanical energy to be constant.
If we include drag forces, the total mechanical energy is not constant; in fact, it dissipates with
time. We will omit the details, but if we include a simple drag force proportional to the particle velocity,
we get the equations
d2 y dy dy
m 2
= −b − mg, y(0) = yo , (t = 0) = 0. (1.30)
dt dt dt
Skipping the details of calculation, if we take b = 0.1 N s/m, and all other parameters as before, we

1.5. EXAMPLE TO ILLUSTRATE HOMEWORK SOLUTION STYLE 27
energy (J)
100 total mechanical energy
po
80 ten
tia
le
ne
r gy
60
40
gy
e ner
20 tic
kine
t (s)
0.4 0.8 1.2
Figure 1.14: KE(t), P E(t), and total mechanical energy for a particle accelerating in a
gravitational field with no drag force.
find

gm2 −bt mg
y(t) = yo + 2 1 − exp − t, (1.31)
b m b

−t 1
= (991 m) − (981 m) exp − 98.1 t, (1.32)
10 s s

−t −t 1
KE(t) + P E(t) = (14533.5 J) + (4811.8 J) exp − (19247.2 J) exp − 962.361 t.
20 s 10 s s
(1.33)
The kinetic, potential, and total mechanical energies as functions of time are plotted in Fig. 1.15.
energy (J)
100 total mechanical ener
gy
p ot
ent
80 ia l
en
erg
y
60
40
rg y
ene
20 etic
kin
t (s)
0.4 0.8 1.2
Figure 1.15: KE(t), P E(t), and total mechanical energy for a particle accelerating in a
gravitational field in the presence of a drag force.

When drag forces are included, we begin with the same amount of total mechanical energy and
potential energy. At the end of the calculation, we have the same amount of potential energy (zero), but
less kinetic energy and less total mechanical energy. Where did this energy go? In fact, it is transformed
into another form of energy, thermal energy, that is not accounted for in Newtonian mechanics. When
we properly account for thermal energy, we will again impose a conservation of total energy, one of the
main topics of this course, to be considered at the outset of Chapter 3.
We close this section with an image of Newton in Fig. 1.16, who began to study issues
related to thermodynamics, and whose scientific methods imbue its development.
Figure 1.16: English genius Sir Isaac Newton (1643-1727), in a

1702 portrait by Sir Godfrey Kneller, whose classical mechanics
broadly influenced the development of thermodynamics; image from
https://commons.wikimedia.org/wiki/File:Sir Isaac Newton by Sir Godfrey Kneller, Bt.jpg.

1.6. THERMODYNAMIC SYSTEM AND CONTROL VOLUME 29
1.6 Thermodynamic system and control volume

We take the following definitions:
• Thermodynamic system: a quantity of fixed mass under investigation,
• Surroundings: everything external to the system,
• System boundary: interface separating system and surroundings, and
• Universe: combination of system and surroundings.
The system, surroundings, and system boundary for a universe are shown for a potato-
shaped system in Fig. 1.17. We allow two important interactions between the system and
work out
surroundings
system
heat in system boundary
universe
Figure 1.17: Sketch of a universe composed of a system, its surroundings, and the system
boundary.
its surroundings:
• heat can cross into the system (our potato can get hot), and
• work can cross out of the system (our potato can expand).
Now, the system boundaries can change, for example the potato might expand on heating,
but we can still distinguish the system and the surroundings. We now define an
• isolated system: a system that is not influenced by its surroundings.
Note that a potato with thick and inelastic skin will be isolated. We distinguish the system,
that has constant mass, but possible variable volume, from the

• Control Volume: fixed volume over which mass can pass in and out of its boundary.
The control volume is bounded by the
• Control Surface: boundary of the control volume.
The mass within a control volume may or may not be constant. If there is fluid flow in
and out there may or may not be accumulation of mass within the control volume. We will
mainly study cases in which there is no accumulation, but this need not be the case. A
sketch contrasting scenarios in which a fluid is compressed in which the system approach
would be used against those where the control volume approach would be used is shown in
Fig. 1.18. In summary,
system approach control volume approach
control
flow in flow out
volume
system
piston-cylinder compressor
Figure 1.18: Comparison of system (fixed mass) and control volume (fixed volume) ap-
proaches in thermodynamics for two common scenarios: piston-cylinder compression (left)
and compression in a flow device whose details are not shown (right).
• system → fixed mass, closed, and

• control volume → potentially variable mass, open.
1.7 Macroscopic versus microscopic

In principle, we could solve for the forces acting on every molecule and use Newton’s laws
to determine the behavior of systems. This is difficult for even modestly sized systems.
• If we had a volume of 1 m3 of gas at atmospheric pressure and temperature, we would

find that it was composed of 2.4 × 1025 molecules.
• We would need six equations of motion for each molecule, three for x, y, z, position,
and three for u, v, w, velocity. This would require then a total of 1.4 × 1026 differential
equations to solve simultaneously.

1.7. MACROSCOPIC VERSUS MICROSCOPIC 31
• Even with our largest computers, this is impossible today. Note most desktop com-
puters only can store roughly 109 bytes of data in Random Access Memory (RAM).
• We can however model the average behavior of the molecules statistically.
• We can also use simple empirical relations that can be formally proved to capture the
statistical nature of the flow. This will be our approach.
• classical thermodynamics will treat macroscopic effects only and ignore individual
molecular effects. For example molecules bouncing off a wall exchange momentum
with the wall and induce pressure. We could use Newtonian mechanics for each par-
ticle collision to calculate the net force on the wall. Instead our approach amounts to
considering the average over space and time of the net effect of millions of collisions on
a wall.
We will in fact assume that matter can be modeled as a
• Continuum: the limit in which discrete changes from molecule to molecule can be
ignored and distances and times over which we are concerned are much larger than
those of the molecular scale. This will enable the use of calculus in our continuum
thermodynamics.
The continuum theory can break down in important applications where the length and time
scales are of comparable magnitude to molecular time scales. Important applications where
the continuum assumption breaks down include
• rarefied gas dynamics of the outer atmosphere (relevant for low orbit space vehicles),
and
• nano-scale heat transfer (relevant in cooling of computer chips).
To get some idea of the scales involved, we note that for air at atmospheric pressure and
temperature that the time and distance between molecular collisions provide the limits of
the continuum. Under these conditions, we find for air
• length > 0.1 µm, and
• time > 0.1 ns,
will be sufficient to use the continuum assumption. For denser gases, these cutoff scales are
smaller. For lighter gases, these cutoff scales are larger. Details of collision theory can be
found in advanced texts such as that of Vincenti and Kruger.12 They show for air that the
mean free path λ is well modeled by the following equation:
M
λ= √ . (1.34)
2πNρd2
12
W. G. Vincenti and C. H. Kruger, 1965, Introduction to Physical Gas Dynamics, John Wiley, New York,
pp. 12-26.

Here, M is the molecular mass, N is Avogadro’s number, and d is the molecular diameter.
Example 1.2
Find the variation of mean free path with density for air.
We turn to Vincenti and Kruger for numerical parameter values, that are seen to be M =
28.9 kg/kmole, N = 6.02252 × 1023 molecule/mole, d = 3.7 × 10−10 m. Thus,

kg 1 kmole
28.9 kmole 1000 mole
λ = √ 2, (1.35)
2π 6.02252 × 1023 molecule
mole ρ (3.7 × 10−10 m)
kg
7.8895 × 10−8 molecule m2
= . (1.36)
ρ
Note that the unit “molecule” is not really a dimension, but really is literally a “unit,” that may well
be thought of as dimensionless. Thus, we can safely say
kg
7.8895 × 10−8 m2
λ = . (1.37)
ρ
A plot of the variation of mean free path λ as a function of ρ is given in Fig. 1.19. Vincenti and Kruger
10-4
10-5
10-6
10-7
10-8
10-3 10-2 10-1 100 101
Figure 1.19: Mean free path length, λ, as a function of density, ρ, for air.
go on to consider an atmosphere with density of ρ = 1.288 kg/m3 . For this density
kg
7.8895 × 10−8 m2
λ = kg
, (1.38)
1.288 m3
−8
= 6.125 × 10 m, (1.39)
= 6.125 × 10−2 µm. (1.40)
Vincenti and Kruger also show the mean molecular speed under these conditions is roughly c = 500 m/s,
so the mean time between collisions, τ , is
λ 6.125 × 10−8 m
τ∼ = = 1.225 × 10−10 s. (1.41)
c 500 ms

1.8. PROPERTIES AND STATE OF A SUBSTANCE 33
1.8 Properties and state of a substance

We define
• Phase: a quantity of matter that is homogeneous throughout, and
• Phase Boundaries: interfaces between different phases.
An example of a single phase is ice. Another single phase is liquid water. A glass of ice
water is a two-phase mixture with the phase boundaries at the edge of each ice cube.
We next define (circularly)
• State: condition described by observable macroscopic properties, and
• Property: quantity that only depends on the state of the system and is independent
of the history of the system.
Examples of properties include temperature and pressure. Two states are equivalent if they
have the same properties. So if state 1 is defined by temperature T1 and pressure P1 , and
state 2 is defined is by temperature T2 and P2 , state 1 is equivalent to state 2 iff (that is, if
and only if) T1 = T2 and P1 = P2 .
There are two important classes of properties we consider in thermodynamics:
• Extensive Property: a property that depends on the mass (or the extent) of the
system, example extensive properties include mass, total volume, total energy, and
• Intensive Property: a property that is independent of the mass of the system.

Example intensive properties include temperature and pressure.
In general, if you cut a system in half and re-measure its properties, intensive properties
remain unchanged, while extensive properties are cut in half. Properties are defined for
systems that are in
• Equilibrium: state in which no spontaneous changes are observed with respect to

time.
We actually never totally achieve equilibrium, we only approximate it. It takes infinite time
to achieve final equilibrium. In this class we will mainly be concerned with two types of
equilibrium:
• Mechanical equilibrium: characterized by equal pressure, and
• Thermal equilibrium: characterized by equal temperature.

A third type of equilibrium is chemical equilibrium, that we will not consider here, and is
characterized by equal chemical potentials.
A difficult conceptual challenge of thermodynamics is to reckon with two systems initially
at their own equilibria, to bring them into contact so that they find a new equilibria. How to
do this without consideration of time can be difficult. Another branch of thermodynamics,
that we will consider only briefly in this course is
• Non-equilibrium thermodynamics: branch of thermodynamics that considers sys-

tems often far from equilibrium and the time-dynamics of their path to equilibrium.
We will go to great effort to construct a thermodynamics that is generally not burdened

with time. Occasionally we will bring time into our problems. Unfortunately, ignoring time
occasionally requires some mental contortions, as seen in the next section.
1.9 Processes and cycles

Often systems undergo a
• Change of State: implies one or more properties of the system has changed.
How these properties would change outside of time is curiously outside the framework of
equilibrium thermodynamics! The best way to think of them is that the changes are slow
relative to the underlying molecular time scales. Fortunately, this will allow us to do a wide
variety of problems of engineering relevance.
We also define a
• Process: a succession of changes of state.
We assume our processes are all sufficiently slow such that each stage of the process is near
,′
equilibrium. Certain common processes are given special names, based on the Greek ι σoς,
isos, meaning “equal”:
• isothermal: constant temperature,
• isobaric: constant pressure, and
• isochoric: constant volume.
An important notion in thermodynamics is that of a
• cycle: series of processes that returns to the original state.
The cycle is a thermodynamic “round trip.”

1.10. FUNDAMENTAL VARIABLES AND UNITS 35
1.10 Fundamental variables and units

We will mainly use the Système International (SI) units in this course. Occasionally, we
will use the English Engineering system of units. As found in US National Institute of
Standards and Technology (NIST) documents, the important fundamental base SI units,
and corresponding English units are
• length:
– meter (m): the length of the path traveled by light in vacuum during a time
interval of 1/299792458 of a second, and
– foot (ft),
• time:
– second: (s) the duration of 9192631770 periods of the radiation corresponding to
the transition between the two hyperfine levels of the ground state of the cesium
133 atom, and
– second: (s) English time units are identical to those of SI,
• mass:
– kilogram (kg): until May 2019 was defined as a mass equal to the mass of the
international prototype of the kilogram (a platinum-iridium bar stored in Paris),
roughly equal to the mass of one liter of water at standard temperature and
pressure; its new definition is such that it is defined by taking the fixed numerical
value of the Planck constant h to be 6.62607015 × 10−34 when expressed in the
unit J s, that is equal to kg m2 s−1 , and
– pound mass: (lbm),
• temperature: an equilibrium property that roughly measures how hot or cold an
object is. Note our senses are poor judges of temperature. Consider snow and air in
thermal equilibrium at 20 ◦ F. Usually, it is possible to keep your bare hands warm for
many hours at 20 ◦ F if you are otherwise dressed warmly. However, if you place your
bare hand in a snow bank you for a few minutes, you have a danger of frostbite. Yet
both are at the same temperature. Why the difference in sense? Our bodies actually
have more sensitivity to heat fluxes instead of temperature; heat leaves our body more
rapidly when in contact with high density objects like snow relative to that of low
density objects like air. More fundamental than common units such as ◦ F are so-called
absolute temperature units:
– Kelvin: (K) until recently, the fraction 1/273.16 of the thermodynamic tem-
perature of the triple point of water, and since May 2019 defined by taking the
fixed numerical value of the Boltzmann constant kB to be 1.380649 × 10−23 when
expressed in the unit J K−1 , that is equal to kg m2 s−2 K−1 , and

– Rankine: (◦ R).
1.11 Zeroth law of thermodynamics

In this class we are taking the axiomatic approach. Recall that an axiom cannot be proven.
It is a statement whose truth can be ascertained only by comparison with experiment.
The axiom can be disproved by a single negative experiment. The so-called zeroth law of
thermodynamics is the axiom that is probably most fundamental. It was formalized after
the so-called first and second laws, and so it is called the zeroth law. Perhaps if a more
fundamental axiom were discovered, it would be called the −1st law of thermodynamics?
• Zeroth law of thermodynamics: When two bodies have equality of temperature
with a third body, then they have equality of temperature.
The origins of the zeroth law are murky. Sommerfeld13 attributes the notion to R. H. Fowler
in a 1931 review of a thermodynamics book. Fowler and Guggenheim explicitly introduce
the term “zeroth law of thermodynamics” later.14 The equivalent statement in mathematical
logic is that if x = y and x = z, then y = z; this is in fact equivalent to the first of Euclid’s
common notions: things that are equal to the same thing are also equal to each other.
Definition of the zeroth law enables the use of a thermometer as a measurement device.
A scale however needs to be defined. The old metric temperature scale, Celsius (◦ C), was
defined so that
• 0 ◦ C is the freezing point of water, and
• 100 ◦ C is the boiling point of water.
These quantities varied with pressure however, so that different values would be obtained on
top of a mountain versus down in the valley, and so this is not a good standard. Until May
2019, the Celsius scale was defined to be nearly the same, but had
• 0.01 ◦ C as the so-called triple point of water, and
• −273.15 ◦ C as absolute zero in K.
The triple point of water is defined at the state where three phases of water (solid, liquid,
and gas) are observed to co-exist. Now it is defined in terms of the Boltzmann constant on
a similar basis as the Kelvin scale, in such as way that the triple point of water remains,
within the limits of experimental error, at 0.01 ◦ C. The transformation between the absolute
Kelvin scale and the Celsius scale is given by
K = ◦ C + 273.15. (1.42)
13
A. Sommerfeld, 1956, Thermodynamics and Statistical Mechanics, Lectures on Theoretical Physics,
Vol. V, Academic Press, New York, p. 1.
14
R. Fowler and E. A. Guggenheim, 1939, Statistical Thermodynamics: A Version of Statistical Mechanics
for Students of Physics and Chemistry, Cambridge University Press, Cambridge, p. 56.

1.12. SECONDARY VARIABLES AND UNITS 37
The English equivalents are degrees Fahrenheit (◦ F) for relative temperature and degrees
Rankine (◦ R) for absolute temperature. The conversions are
T (◦ R) = 1.8T (K), T (◦ F) = 1.8T (◦ C) + 32, T (◦ F) = T (◦ R) − 459.67. (1.43)
1.12 Secondary variables and units

Many units can be derived from the base units. Some important units for thermodynamics
include
P
• Force: This unit is defined from Newton’s second law, m d2 y/dt2 = F.
kg m
– Newton: (N), 1 N = 1 s2
, and
– pound force: (lbf).
Force is straightforward in SI units. It is more confusing in English units, where the

so-called gravitational constant gc is often introduced. In SI units, gc = 1. However in
English units, the law for force is better stated as
1 d2 y X
m = F. (1.44)
gc dt2
Moreover, a gravitational body force is better stated as mg/gc in English units. Now,
1 lbf is induced by a mass of 1 lbm in places where local gravitational acceleration is
g = 32.1740 ft/s2 . Let us consider two important types of problems
– acceleration of a particle in a uniform gravitational field: Here the only force

acting on the particle is the body force, −mg/gc , and Eq. (1.44) reduces to
1 d2 y mg
m 2 = − , (1.45)
gc dt gc
d2 y
= −g, (1.46)
dt2
dy
= −gt + C1 , (1.47)
dt
1
y = − gt2 + C1 t + C2 . (1.48)
2
Note that gc plays no role whatsoever in determining the position of the particle.
– static force balance: Here one wants to determine the force necessary to hold
a particle stationary in a uniform gravitational field. In such a problem the

acceleration is zero, but there are two forces, the gravitational force −mg/gc , and
the counter-balancing force that we will call F . Eq. (1.44) reduces to
1 d2 y mg
m 2 = F− , (1.49)
gc |{z}
dt gc
=0
mg
0 = F− , (1.50)
gc
mg
F = . (1.51)
gc
If we are at a location where g = 32.1740 ft/s2 , we can consider the counter-balancing

force necessary to hold 1 lbm stationary, via Eq. (1.51):

(1 lbm) 32.1740 sft2
F = 1 lbf = . (1.52)
gc
Thus,
lbm ft
gc = 32.1740 . (1.53)
lbf s2
Example 1.3
If the local gravitational acceleration is 32.0 ft/s2 , find the weight W of an object with mass of
m = 1000 lbm.

1 1 ft
W = F = mg = lbm ft
(1000 lbm) 32.0 2 = 994.59 lbf. (1.54)
gc 32.1740 lbf s2
s
Example 1.4
An alternative English unit for mass is the slug with 1 slug = 32.1740 lbm. If local gravitational
acceleration is 32.0 ft/s2 , find the weight W of an object with mass of m = 1 slug.

1 1 32.1740 lbm ft
W =F = mg = lbm ft
(1 slug) 32.0 2 = 32.0 lbf. (1.55)
gc 32.1740 lbf s2
1 slug s

Note, we could also recast gc in terms of slugs as

lbm ft 1 slug slug ft
gc = 32.1740 =1 . (1.56)
lbf s2 32.1740 lbm lbf s2
Thus, we could also say

1 1 ft
W =F = mg = slug ft (1 slug) 32.0 2 = 32.0 lbf. (1.57)
gc 1 lbf s2 s
Because in this unit system, gc has a magnitude of unity, it is analogous to the SI system, and
the slug is more appropriate e lbm as an analog to the kg. Just as we think of N = kg m/s2 ,
we can think of lbf = slug ft/s2 . Mainly so that consistency can be maintained with BS, we will
nearly always use lbm and avoid slugs for mass in the English system.
• Energy: roughly speaking, the ability to do work, found from the product of force
and distance.
– Joule: (J), 1 J = 1 (N m), and
– foot-pound force: (ft lbf).
• Specific Volume: the volume per unit mass, known as v = V /m.
3
– mkg , and
3
ft
– lbm .
• Density: the mass per unit volume, the inverse of specific volume ρ = m/V .

– mkg3 , and

– lbm
ft3
.
Note also that
1 1
v= , ρ= . (1.58)
ρ v
• Pressure: the limA→0 F/A where A is the cross-sectional area and F is the compo-
nent of force acting normal to that area. In thermodynamics, we are almost always
concerned with the absolute pressure as opposed to the gauge pressure. Most common
pressure gauges do not measure the absolute pressure; instead they measure the differ-
ence between the absolute pressure and the atmospheric pressure. The two are related
via the formula
Pgauge = Pabsolute − Patm . (1.59)
We nearly always interpret P as an absolute pressure, so we could also say
Pgauge = P − Patm . (1.60)

– Pascal: (Pa), 1 Pa = 1 N/m2 ; note other common units are 1 bar = 105 Pa,
1 atm = 1.01325 × 105 Pa = 101.325 kPa = 0.101325 MPa, and
– (psia): 1 psia = 1 lbf/in2 . 1 atm = 14.696 psia. The a denotes the “absolute”
pressure as opposed to the “gauge” pressure. The units psig refer to a gauge
pressure.
The SI unit is named after Blaise Pascal, see Fig. 1.20, the French polymath who
Figure 1.20: Blaise Pascal (1623-1662), French scientist and philoso-

pher who considered manometry, among other diverse topics; image from
https://mathshistory.st-andrews.ac.uk/Biographies/Pascal.html.
conducted early scientific experiments with manometers, a common measuring device

for pressure, see Fig. 1.21. There are a variety of styles of manometers. Here, a pipe
mg = Vg = AHg
g H
y
A
fluid at P,
PA
Figure 1.21: Manometer sketch.
containing fluid at pressure P and density ρ has a small tube with cross sectional area

A connecting it to the outside atmosphere at a different pressure Patm . The length H is

easily measured. The gravitational acceleration is g and is in the negative y direction.
Because P > Patm , the manometer fluid is pushed up. However, it finds a mechanical
equilibrium where the weight of the manometer fluid balances the net force induced
by the pressure differential.
The figure includes a cutaway with a free body diagram. The interior fluid exerts a
positive force of P A on the manometer fluid in the cutaway. The atmosphere exerts
another force of Patm A in the negative direction. The third force is the weight of the
fluid: mg. Thus, Newton’s second law tells us
d2 y
m = P A − Patm A − mg. (1.61)
dt2
|{z}
=0
Now, we are concerned with cases that are static, in which case the acceleration
d2 y/dt2 = 0. Thus, we require a force balance, i.e. mechanical equilibrium, that is
achieved when
0 = P A − Patm A − mg, (1.62)

P A = Patm A + mg. (1.63)
Now, mg = ρV g, where V is the volume of the fluid in the cutaway. Obviously from
the geometry, we have V = AH, so mg = ρAHg. Thus,
P A = Patm A + ρAHg, (1.64)

P = Patm + ρgH. (1.65)
Or
∆P = P − Patm = Pgauge = ρgH. (1.66)
Example 1.5
A manometer gives a reading of H = 2 ft in a region where local g = 32.2 ft/s2 . The working fluid
has specific volume v = 0.0164 ft3 /lbm. The atmospheric pressure is Patm = 14.42 lbf/in2 = 14.42 psia.
Find the fluid pressure.
We know that in the SI system.
P = Patm + ρgH. (1.67)
In terms of specific volume, recalling that ρ = 1/v, we have
gH
P = Patm + . (1.68)
v

The challenge here is really the English units. A fair way to approach English units is to replace g by
g/gc in every formula. Thus, in English units, we have
1 g
P = Patm + H. (1.69)
v gc
So our fluid pressure is
! 2
ft
lbf 1 32.2 s2 1 ft lbf
P = 14.42 2 + ft3 lbm ft
(2 ft) = 15.27 = 15.27 psia. (1.70)
in 0.0164 lbm
32.1740 lbf s2
12 in in2

Chapter 2
Properties of a pure substance
Read BS, Chapter 2
2.1 The pure substance

We define a
• Pure substance: a material with homogeneous and invariable composition.
To elaborate,
• Pure substances can have multiple phases: an ice-water mixture is still a pure sub-
stance.
• An air-steam mixture is not a pure substance.
• Air, being composed of a mixture of N2 , O2 , and other gases, is formally not a pure
substance. However, experience shows that we can often treat air as a pure substance
with little error.
2.2 Vapor-liquid-solid phase equilibrium

Often we find that different phases of pure substances can exist in equilibrium with one
another. Let us consider an important gedankenexperiment (Latin-German for “thought
experiment”) in which we boil water. Ordinary water boiling is shown in Fig. 2.1. However,
this ordinary experiment has constraints that are too loose. Most importantly, the mass of
water leaks into the atmosphere; thus, the water vapor and the air become a mixture and
no longer a pure substance.
Let us instead consider a more controlled piston-cylinder arrangement. Inside the cylin-
der, we begin with pure liquid water at T = 20 ◦ C. The piston is free to move in the cylinder,
but it is tightly sealed, so no water can escape. On the other side of the piston is a constant
pressure atmosphere, that we take to be at P = 100 kPa = 0.1 MPa = 105 Pa = 1 bar. We
43
44 CHAPTER 2. PROPERTIES OF A PURE SUBSTANCE
Figure 2.1: Water boiling isobarically in an open environment.
slowly add heat to the cylinder, and observe a variety of interesting phenomena. A sketch
of what we observe is given in Fig. 2.2. We notice the following behavior:
P = 100 kPa P = 100 kPa P = 100 kPa P = 100 kPa P = 100 kPa P = 100 kPa
water vapor
T > 99.62 oC
saturated
vapor,
water vapor
T=99.62 oC
T = 99.62 oC
liquid water saturated liquid,
liquid water T > 20 oC liquid water T = 99.62 oC
T = 20 oC T = 99.62 oC
Q Q Q Q Q
Figure 2.2: Sketch of an experiment in which heat is added isobarically to water in a closed
piston-cylinder arrangement.
• The pressure remains at a constant value of 100 kPa. This is an isobaric process.
• The total volume increases slightly as heat is added to the liquid.
• The temperature of the liquid increases significantly as heat is added to the liquid.
• At a special value of temperature, observed to be T = 99.62 ◦ C, we have all liquid, but

cannot add any more heat and retain all liquid. We will call this state the saturated
liquid state. We call T = 99.62 ◦ C the saturation temperature at P = 100 kPa. As we
continue to add heat,

2.2. VAPOR-LIQUID-SOLID PHASE EQUILIBRIUM 45
– The temperature remains constant (this is isothermal now as well as isobaric).

– The total volume continues to increase.
– We notice two phases present: liquid and vapor, with a distinct phase boundary.
The liquid is dense relative to the vapor. That is ρf > ρg , where f denotes fluid
or liquid and g denotes gas or vapor. Thus, vg > vf .
– As more heat is added, more vapor appears, all while P = 100 kPa and T =
99.62 ◦ C.
– At a certain volume, we have all vapor and no liquid, still at P = 100 kPa,
T = 99.62 ◦ C. We call this state the saturated vapor state.
• As heat is added, we find both the temperature and the volume rise, with the pressure
remaining constant. The water remains in the all vapor state.
We have just boiled water! We sketch this process in the temperature-specific volume plane,
that is, the T − v plane, in Fig. 2.3. Note that the mass m of the water is constant in this
T kP
a
0
10
=
P
saturated two-phase
liquid mixture saturated
vapor
Figure 2.3: Isobar in the T − v plane for our thought experiment in which heat is added
isobarically to water in a piston-cylinder arrangement.
problem, so the extensive V is strictly proportional to specific volume, v = V /m.

We next repeat this experiment at lower pressure (such as might exist on a mountain
top) and at a higher pressure (such as might exist in a valley below sea level). For moderate
pressures, we find qualitatively the exact same type of behavior. The liquid gets hotter,
turns into vapor isothermally, and then the vapor gets hotter as the heat is added. However,
we note the following important facts:
• The saturation temperature (that is the boiling point) increases as pressure increases,
as long as the pressure increase is not too high.

• As pressure increases vf becomes closer to vg .
• Above a critical pressure, P = Pc = 22.089 MPa, there is no phase change ob-

served!1 At the critical pressure, the temperature takes on a critical temperature of
Tc = 374.14 ◦ C. At the critical pressure and temperature, the specific volume takes the
value vf = vg = vc = 0.003155 m3 /kg.
We see how the boiling point changes with pressure by plotting the saturation pressure
as a function of saturation temperature in the T − P plane in Fig. 2.4. This is the so-called
vapor pressure curve. Here, we focus on liquid-vapor mixtures and keep T high enough to
critical
point
rve
compressed
cu
liquid e
ss ur
e
pr
p or
va
superheated
vapor
Figure 2.4: Sketch of saturation pressure versus saturation temperature.
prevent freezing. Note the curve terminates abruptly at the critical point.
We adopt the following nomenclature:
• saturated liquid: the material is at Tsat and is all liquid.
• saturated vapor: the material is at Tsat and is all vapor.
• compressed (subcooled) liquid: the material is liquid with T < Tsat .
• superheated vapor: the material is vapor with T > Tsat .
• two-phase mixture: the material is composed of co-existing liquid and vapor with
both at Tsat .
1
This behavior may have first been carefully documented by T. Andrews, 1869, “The Bakerian lecture:
on the continuity of the gaseous and liquid states of matter,” Philosophical Transactions of the Royal Society
of London, 159: 575-590.

2.2. VAPOR-LIQUID-SOLID PHASE EQUILIBRIUM 47
For two-phase mixtures, we define a new property to characterize the relative concentra-
tions of liquid and vapor. We define the
• quality= x: as the ratio of the mass of the mixture that is vapor (vap) to the total
mixture mass:
mvap
x= . (2.1)
mtotal
We can also take the total mass to be the sum of the liquid and vapor masses:
mtotal = mliq + mvap . (2.2)
So
mvap
x= . (2.3)
mliq + mvap
There are two important limits to remember:
• x = 0: corresponds to mvap = 0. This is the all liquid limit.
• x = 1: corresponds to mvap = mtotal . This is the all gas limit.
We must have
x ∈ [0, 1]. (2.4)
We sketch water’s T − v plane again for a wide variety of isobars in Fig. 2.5. We sketch
T P = 40 MPa
Pc = 22.089 MPa
P = 10 MPa
critical
point
compressed
Tc = 374.14 oC
superheated
vapor
liquid
P = 1 MPa
two-phase
mixture
P = 0.1 MPa
T= 99.62 oC saturated
saturated vapor line
liquid line
vc = 0.003155 m3/kg
v
Figure 2.5: Sketch of T − v plane for water for a variety of isobars.
water’s P − v plane for a wide variety of isotherms in Fig. 2.6. We can perform a similar
thought experiment for ice. We can start with ice at P = 100 kPa and add heat to it. We

critical superheated
point vapor
Pc = 22.089 MPa
Thigh
two-phase o
Tc = 374.14 C
compressed mixture
liquid
saturated saturated
liquid line Tlow
vapor line
vc = 0.003155 m3/kg
v
Figure 2.6: Sketch of P − v plane for water for a variety of isotherms.
observe the ice’s temperature rise until T = Tsat ∼ 0 ◦ C. At that temperature, the ice begins
to melt and the temperature remains constant until all the ice is melted. At this point the
liquid temperature begins to rise. If we continued to add heat, we would boil the water.
We note if we perform this experiment for P < 0.6113 kPa the ice in fact goes directly
to vapor. It is said to have undergone sublimation. There exists a second important point
where ice being heated isobarically can transform into either liquid or gas. This is the so-
called triple point. At the triple point we find the saturation pressure and temperature are
Ptp = 0.6113 kPa and Ttp = 0.01 ◦ C, respectively. It is better described as a triple line,
because in the P − v − T space we will study, it appears as a line with constant P and T , but
variable v. In the P − T projected plane of the P − v − T volume, it projects as a point. We
sketch water’s P − T plane again for a wider range to include the vapor-liquid-solid phase
behavior in Fig. 2.7.
These characteristics apply to all pure substances. For example, nitrogen has a triple
point and a critical point. Table A.2 in BS lists critical constants for many materials. Note
also that phase transitions can occur within solid phases. This involves a re-arrangement of
the crystal structure. This has important implications for material science, but will not be
considered in detail in this course.
2.3 Independent properties

Let us define a
• Simple compressible substance: a material that can be worked upon by pressure

forces.

2.4. THERMAL EQUATIONS OF STATE 49
P (kPa)
supercritical
fluid
critical
ine
22089 point
fusion l
compressed
ine
liquid
l
ion
at
riz
solid po
va
e
0.6113
n lin superheated
tio triple
l i ma point vapor
sub
o
0.01 374.14 T ( C)
Figure 2.7: Sketch of P − T plane for water.
Note we neglect electric, magnetic, and chemical work modes. While this is indeed restrictive,
it will be important for many mechanical engineering applications. The following important
statement can be proved (but will not be so here):
• For a simple compressible substance, two independent intensive thermodynamic prop-

erties define the state of the system.
Consider the implications for
• superheated vapor: If we consider P , T , and v, this states that we must allow one of the
variables to be functions of the other two. We could have P = P (T, v), v = v(T, P ),
or T = T (P, v). All are acceptable.
• two-phase mixture: If we have a two-phase mixture, our experiments show that P and
T are not independent. In this case, we need another property to characterize the
system. That property is the quality, x. So for two-phase mixtures, we might have
v = v(T, x), for example.
2.4 Thermal equations of state

Here, we will describe some of the many different ways to capture the relation between two
independent properties and a third dependent property for a simple compressible substance.
We will focus on a so-called

• Thermal equation of state: an equation that gives the pressure as a function of

two independent state variables. An example is the general form:
P = P (T, v). (2.5)
We will progress from simple thermal equations of state to more complex.
2.4.1 Ideal gas law

For many gases, especially at low density and far from the critical point, it is possible to write
a simple thermal equation of state that accurately describes the relation between pressure,
volume, and temperature. Such equations were developed in the 1600s and early 1800s
based entirely on macroscopic empirical observation. In the late 1800s, statistical mechanics
provided a stronger theoretical foundation for them, but we will not consider that here.
Let us start with the most important equation of state:
• Ideal gas law: This equation, that is a combination of Boyle’s law,2 Charles’ law,3
and Avogadro’s law,4 is most fundamentally stated as
P V = nRT. (2.6)
On the continent, Boyle’s law is sometimes known as Mariotte’s law after Edme Mariotte
(1620-1684), but Boyle published it fourteen years earlier.5 A reproduction of Boyle’s data
is given in Fig. 2.8.6 The data in (V, 1/P ) space is fit well by a straight line with intercept
at the origin; that is 1/P = KV , where K is a constant. Thus, P V = C, where C = 1/K is
a constant.
2
R. Boyle, 1662, New Experiments Physico-Mechanical, Touching the Air: Whereunto is Added A Defence
of the Authors Explication of the Experiments, Against the Obiections of Franciscus Linus and Thomas
Hobbes, H. Hall, Oxford. There exists a 1725 text modified from the 1662 original in The Philosophical
Works of Robert Boyle, Vol. 2, P. Shaw, ed., Innis, Osborn, and Longman, London. Boyle’s law holds that
P V is constant in an isothermal process.
3
attributed to J. A. C. Charles by J. L. Gay-Lussac, 1802, “Recherches sur la dilatation des gaz des
vapeurs,” Annales de Chimie, 43(1): 137-175. Charles’ law holds that V /T is constant for ideal gases
undergoing isobaric processes. Additionally, Guillaume Amontons (1663-1705) performed some of the early
experimentation that led to Charles’ law. John Dalton (1766-1844) is said to have also written on a version
of Charles’ Law in 1801.
4
A. Avogadro, 1811, “Essai d’une manière de déterminer les masses relatives des molécules élémentaires
des corps, et les proportions selon lesquelles elles entrent dans ces combinaisons,” Journal de Physique,
de Chimie et d’Historie Naturelle, 73:58-76. Here, Avogadro hypothesized that “equal volumes of ideal or
perfect gases, at the same temperature and pressure, contain the same number of particles, or molecules.”
5
C. Webster, 1965, “The discovery of Boyle’s law, and the concept of the elasticity of air in the seventeenth
century,” Archive for History of Exact Sciences, 2(6): 441-502. Also described here is how Henry Power
(1623-1668) and Richard Towneley (1629-1707) did important preliminary work that helped Boyle formulate
his law.
6
see J. B. West, 1999, “The original presentation of Boyle’s law, Journal of Applied Physiology, 98(1):
31-39.

a)
b)
0.04
0.03
0.02
0.01
0 10 20 30 40 50
Figure 2.8: a) Boyle’s 1662 data to validate his law (P V is constant for an isothermal
process), b) plot of Boyle’s data: V (the left-most of the columns labelled A) versus reciprocal
of P (reciprocal of column D), demonstrating its near linearity.

Depictions of Boyle, Charles, and Avogadro are given in Fig. 2.9. The ideal gas law was
b) c) a)
Figure 2.9: b) Robert Boyle (1627-1691), Irish scientist who developed an important special
case of the ideal gas law. Image from https://en.wikipedia.org/wiki/Robert Boyle, c)
Jacques Alexandre César Charles (1746-1823), French scientist credited in 1802 by Joseph
Louis Gay-Lussac (1778-1850) for developing an important special case of the ideal gas
law in the 1780s. Image from https://en.wikipedia.org/wiki/Jacques Charles,
a) Lorenzo Romano Amedeo Carlo Avogadro di Quarengna e di Cerroto
(1776-1856), Italian physicist, nobleman, and revolutionary. Image from
https://en.wikipedia.org/wiki/Amedeo Avogadro.
first stated in the form roughly equivalent to Eq. (2.6) by Clapeyron,7 depicted in Fig. 2.10.
It is critical that the temperature here be the absolute temperature. For the original
argument, see Thomson.8 Here, n is the number of moles. Recall there are N = 6.02214179×
1023 molecules in a mole, where N is Avogadro’s number. Also R is the universal gas constant,
whose history is recounted by Jensen.9 From experiment, many performed by Regnault,
depicted in Fig. 2.11, it is determined to be
kJ
R = 8.314472 . (2.7)
kmole K
In this class the over bar notation will denote an intensive property on a per mole basis.
Intensive properties without over bars will be on a per mass basis. Recall the mass-basis
7
É. Clapeyron, 1834, “Mémoire sur la puissance motrice de la chaleur,” Journal de l’École Polytechnique,
14: 153-190.
8
W. Thomson, 1848, “On an absolute thermometric scale founded on Carnot’s theory of the motive power
of heat and calculated from Regnault’s observations, Proceedings of the Cambridge Philosophical Society,
1(5): 66-71.
9
W. B. Jensen, 2003, “The universal gas constant R,” Journal of Chemical Education, 80(7): 731-732.

Figure 2.10: Benoı̂t Paul Émile Clapeyron (1799-1824), French engineer and
physicist who furthered the development of thermodynamics. Image from
https://en.wikipedia.org/wiki/Emile Clapeyron.
Figure 2.11: Henri Victor Regnault (1810-1878), French chemist and physicist who made
careful early measurements of thermodynamic parameters for ideal and non-ideal gases.
https://en.wikipedia.org/wiki/Henri Victor Regnault.
specific volume is v = V /m. Let us define the mole-based specific volume as
V
v= . (2.8)
n
Thus, the ideal gas law can be represented in terms of intensive properties as
V
P = RT, (2.9)
n
|{z}
=v
P v = RT. (2.10)
There are other ways to write the ideal gas law. Recall the molecular mass M is the
mass in g of a mole of substance. More common in engineering, it is the mass in kg of a
kmole of substance. These numbers are the same! From chemistry, for example, we know

the molecular mass of O2 is 32 g/mole = 32 kg/kmole. Symbolically, we can say that

m
M= . (2.11)
n
Now, take the ideal gas law and divide by m:
P V = nRT, (2.12)
V n
P = RT, (2.13)
m
|{z} m
|{z}
=v =1/M
R
Pv = T. (2.14)
M
|{z}
≡R
Now, let us define

R
R≡ . (2.15)
M
Let us check the units:
kJ kmole kJ
[R] = = . (2.16)
kmole K kg kg K
We have actually just lost some universality. Recall R is independent of material. But
because each different gas has a different M, then each gas will have its own R. These values
for various gases are tabulated in Table A.5 of BS.
With this definition, the ideal gas law becomes
P v = RT. (2.17)
This is the form we will use most often in this class. Note the useful fact that
Pv
= R. (2.18)
T
Thus, if an ideal gas undergoes a process going from state 1 to state 2, we can safely say
P1 v1 P2 v2
= . (2.19)
T1 T2
Example 2.1
Find R for air.
We can model air as a mixture of N2 and O2 . Its average molecular mass is known from Table A.5
of BS to be M = 28.97 kg/kmole. So R for air is
kJ
R 8.3145 kmole K kJ
R= = kg
= 0.287 . (2.20)
M 28.97 kmole kg K

Consider some notions from algebra and geometry. The function f (x, y) = 0 describes a
curve in the x − y plane. In special cases, we can solve for y to get the form y = y(x). The
function f (x, y, z) = 0 describes a surface in the x − y − z volume. In special cases, we can
solve for z to get z = z(x, y) to describe the surface in the x − y − z volume.
Example 2.2
Analyze the surface described by f (x, y, z) = z 2 − x2 − y 2 = 0.
Here, we can solve for z exactly to get

p
z = ± x2 + y 2 . (2.21)
This surface is plotted in Fig. 2.12. We can also get three two-dimensional projections of this surface
Figure 2.12: The surface z 2 − x2 − y 2 = 0.
in the x − y plane, the y − z plane, and the x − z plane. Orthographic projections of this surface are
plotted in Fig. 2.13.

y
2
iso-z contours
1
-1
-2
-2 -1 0 1 2
x
z2 z
2
1 1
iso-x
iso-y 0
0 contours
contours
-1 -1
-2 -2
-2 -1 0 1 2 -2 -1 0 1 2
x y
Figure 2.13: Contours of constant x, y and z in orthographic projection planes of the surface
z 2 − x2 − y 2 = 0.

For the x − y plane we consider

p
zo = ± x2 + y 2 . (2.22)
for various values of zo . This yields a family of circles in this plane. For the y − z plane, we consider
p
xo = ± z 2 − y 2 . (2.23)
This gives a family of hyperbolas. For real xo , we require z 2 ≥ y 2 . For the z − x plane, we consider
p
yo = ± z 2 − x2 . (2.24)
This gives a similar family of hyperbolas. For real yo , we require z 2 ≥ x2 .
Similarly, the ideal gas equation P (v, T ) = RT /v describes a surface in the P − v − T

volume. A surface for air is shown in Fig. 2.14. Often, it is easier to understand the behavior
T (K)
200 100
300
400
500
200
P (kPa)
100
0
1
2
3
v (m /kg) 4
3
5
Figure 2.14: Thermodynamic surface for air modeled as an ideal gas.
of the thermodynamic surfaces by projection into various thermodynamic planes and plotting
various iso-contours. Let us do this for an ideal gas.
• isobars:
– Consider curves in the T − v plane on which P is constant. Thus, for the ideal
gas, we consider

P
T = v. (2.25)
R
| {z }
slope
If we insist that P is constant, this gives the equation of an isobar in the T − v

plane. Moreover, for the ideal gas, we see that in the T − v plane isobars are
straight lines with slope P/R. The slope is always positive because P > 0 and
R > 0. So if the pressure is high, the slope is positive and steep. If the pressure
is low, the slope is positive and shallow.

– Consider curves in the P − v plane in which P is constant. Thus, we consider
P = constant, (2.26)
that are straight horizontal lines in the P − v plane.

– Consider curves in the P − T plane in which P is a constant. Thus, we consider
P = constant, (2.27)
that are straight horizontal lines in the P − v plane.
Isobars in various planes are shown in Fig. 2.15.
T Phigh P P
Phigh Phigh
Plow Plow Plow
v v T
Figure 2.15: Isobars for an ideal gas in the T − v, P − v, and P − T planes.
• isotherms
– Consider curves in the T − v plane on which T is constant. Thus, for the ideal
gas, we have
T = constant. (2.28)
These are straight horizontal lines in the T − v plane.
– Consider curves in the P − v plane on which T is a constant. Thus, for the ideal
gas, we have
1
P = (RT ) . (2.29)
v
These are hyperbolas in the P − v plane.
– Consider curves in the P − T plane on which T is a constant. Thus, for the ideal
gas, we have
T = constant. (2.30)
These are straight vertical lines in the P − T plane.
Isotherms in various planes are shown in Fig. 2.16.

T P P Thigh
Tlow
Thigh
Tlow Thigh
Tlow
v v T
Figure 2.16: Isotherms for an ideal gas in the T − v, P − v, and P − T planes.
• isochores
– Consider curves in the T − v plane on which v is constant. Thus, for the ideal
gas, we have
v = constant. (2.31)
These are straight vertical lines in the T − v plane.
– Consider curves in the P − v plane on which v is a constant. Thus, for the ideal
gas, we have
v = constant. (2.32)
These are straight vertical lines in the P − v plane.
– Consider curves in the P − T plane on which v is a constant. Thus, for the ideal
gas, we have
R
P = T. (2.33)
v
| {z }
constant
These are straight lines in the P − T plane with slope R/v. Because R > 0 and
v > 0, the slope is always positive. For large v, the slope is shallow. For small v,
the slope is steep.
Isochores in various planes are shown in Fig. 2.17.
Example 2.3
Given air in a cylinder with stops and a frictionless piston with area A = 0.2 m2 , stop height of
1 m, and total height of 2 m, at initial state P1 = 200 kPa and T1 = 500 ◦ C with cooling, find
• the temperature when the piston reaches the stops, and
• the pressure if the cooling continues to T = 20 ◦ C.

T vlow vlow vlow

vhigh P vhigh P
vhigh
v v T
Figure 2.17: Isochores for an ideal gas in the T − v, P − v, and P − T planes.
1m
air y
1m free body diagram
at the initial state
Figure 2.18: Sketch for example problem of cooling air.
The initial state along with a free body diagram is sketched in Fig. 2.18. We have three distinct
states:
• state 1: initial state
• state 2: piston reaches the stops
• state 3: final state, where T = 20 ◦ C.
At the initial state, the total volume is
V1 = A((1 m) + (1 m)) = (0.2 m2 )(2 m) = 0.4 m3 . (2.34)
We also know that P1 = 200 kPa. For use of the ideal gas law, we must use absolute temperature. So
T1 = 500 + 273.15 = 773.15 K. (2.35)
Now, use the ideal gas law to get v1 :

kJ
RT1 0.287 kg K (773.15 K) m3
v1 = = = 1.10947 . (2.36)
P1 200 kPa kg

Let us check the units:

kJ K kPa kN m m3
kN
= .
kg K kPa m2 kJ kg
As long as we employ kJ for energy and kPa for pressure, we will have few problems with units.
Now, the mass of the air, m, is constant in this problem. Because we have V1 and v1 , we can get
m:
V1 0.4 m3
m= = 3 = 0.360532 kg. (2.37)
v1 1.10947 mkg
We write Newton’s second law for the piston, that we take to have mass mp :
d2 y
mp 2 = P A − Patm A. (2.38)
| {zdt }
∼0
Here we have assumed the piston has not yet reached the stops, that will induce a force as well when
encountered. In a standard assumption in thermodynamics, we neglect the inertia term, mp d2 y/dt2 , as
small. This could be because either the piston mass mp is small or the acceleration d2 y/dt2 is small, or
both could be small. Neglecting inertia, we find the piston to be in a force balance giving Patm A = P A,
giving Patm = P . So our atmosphere must be at Patm = 200 kPa. As the air cools, its temperature
will go down. Because P v = RT , as temperature goes down with constant P , we expect the volume to
decrease. Just when the piston hits the stops, the stops still exert no force on the piston, so
P2 = P1 = 200 kPa. (2.39)
Now
V2 = A(1 m) = (0.2 m2 )(1 m) = 0.2 m3 . (2.40)
So
V2 0.2 m3 m3
v2 = = = 0.554735 . (2.41)
m 0.360532 kg kg
Use the ideal gas law to get T2 :

m3
P2 v2 (200 kPa) 0.554735 kg
T2 = = = 386.575 K. (2.42)
R 0.287 kgkJK
Now, after the piston reaches the stops, the volume is constant. So the process from 2 to 3 is
isochoric, and
V3 = V2 = 0.2 m3 . (2.43)
Thus
m3
v3 = v2 = 0.554735 . (2.44)
kg
So
kJ
RT3 0.287 kg K (20 + 273.15) K)
P3 = = m3
= 151.665 kPa. (2.45)
v3 0.554735 kg
We generate Table 2.1 to summarize the problem. It is usually useful to include sketches of the
process in the various thermodynamic planes. This process is sketched in each of the relevant planes
in Fig. 2.19.

Table 2.1: Numerical values for ideal gas cooling example
variable units state 1 state 2 state 3

T K 773.15 386.575 293.15
P kPa 200 200 151.665
m3
v kg 1.10947 0.554735 0.554735
V m3 0.4 0.2 0.2
T (K) P (kPa) P (kPa)
kg
=
73
m 3/
7
773 - 1 K
5
2 1
0.55
200 - 200 -
2 1
v=
Pa T
1. 6k =
386
15 151.6 - 151.6 -
g
K 3
/k
= 3
m3
2 P T=
293
11
386 -
a
kP
1.
0
=
20
v
293 -
=
3
P
0.555 1.11 0.555 1.11 386 773

v (m3/kg) v (m3/kg) T (K)
Figure 2.19: Sketch of the T − v, P − v, and P − T planes for air-cooling example problem.
2.4.2 Non-ideal thermal equations of state

The ideal gas law is not a good predictor of the P − v − T behavior of gases when
• the gas has high enough density that molecular interaction forces become large and
the molecules occupy a significant portion of the volume; this happens near the vapor
dome typically, or
• the temperature is high enough to induce molecular dissociation (e.g. N2 + N2 ⇌

2N + N2 ).
One alternative is a corrected thermal equation of state.
2.4.2.1 van der Waals

For the van der Waals10 equation of state, that will be studied in more detail in Ch. 9.6, one
has
RT a
P = − 2, (2.46)
v−b v
10
J. D. van der Waals, 1873, Over de Continuiteit van den Gas -en Vloeistoftoestand, Ph.D. Dissertation,
U. Leiden; see also J. D. van der Waals, 1910, Nobel Lecture.

with
27 R2 Tc2 1 RTc
a= , b= . (2.47)
64 Pc 8 Pc
A depiction of van der Waals is given in Fig. 2.20.
Figure 2.20: Johannes Diderik van der Waals (1837-1923), Dutch physicist
and Nobel laureate who developed a corrected state equation. Image from
https://en.wikipedia.org/wiki/Johannes Diderik van der Waals.
2.4.2.2 Redlich-Kwong
For the Redlich-Kwong11 equation of state, one has
RT a
P = − √ , (2.48)
v − b v(v + b) T
with
5/2
R2Tc RTc
a = (0.42748) , b = (0.08664) . (2.49)
Pc Pc
2.4.3 Compressibility factor

In some cases, more detail is needed to capture the behavior of the gas, especially near the
vapor dome. Another commonly used approach to capturing this behavior is to define the
• Compressibility factor: the deviation from ideality of a gas as measured by
Pv
Z= . (2.50)
RT

2
Z = Pv/RT
300 K ideal gas, Z = 1

1 K
200
saturated vapor 130
K
critical
point
liquid
saturated
0
1 10
P (MPa)
Figure 2.21: Sketch of compressibility chart for N2 .
For ideal gases, P v = RT , so Z = 1. Experiments show the behavior of real gases, and this
can be presented in graphical form, as shown for N2 in Fig. 2.21. Note
• for all T , Z → 1 as P → 0. Thus, one has ideal gas behavior at low pressure.
• for T > 300 K, Z ∼ 1 for P < 10 MPa.
• Hold at P = 4 MPa and decrease temperature from 300 K; we see Z decrease below
unity. Now
Pv P P
Z= = , ρ= . (2.51)
RT ρRT ZRT
Because Z < 1, the density ρ is higher than we would find for an ideal gas with Z = 1.
Thus, in this region, there is an attractive force between molecules.
• For P > 30 MPa, we find Z > 1. Thus, a repulsive force exists in this regime. The
forces are complicated.
Note that generalized compressibility charts have been developed for general gases. These
are based on the so-called reduced pressures and temperatures, Pr and Tr , where
P T
Pr = , Tr = . (2.52)
Pc Tc
11
O. Redlich and J. N. S. Kwong, 1949, “On the thermodynamics of solutions. V. an equation of state.
fugacities of gaseous solutions,” Chemical Reviews, 44(1): 233-244.

The reduced pressure and temperature are dimensionless. Values with the c subscript are the
critical constants for the individual gases. Appendix D of BS gives generalized compressibility
charts.
2.4.4 Tabular thermal equations of state

Often equations are too inaccurate for engineering purposes. This is generally because we
may be interested in behavior under a vapor dome. Consider that the surface for steam is
well represented by that shown in Fig. 2.22.
Figure 2.22: P − v − T surface for H2 O, showing solid, liquid, and vapor phases.
In such cases, one should use tables to find a third property, given two independent
properties. We can say that the thermal equation of state is actually embodied in the
tabular data.
We lay down some rules of thumb for this class:
• If steam, use the tables.
• If air or most other gases, use the ideal gas law, but check if the pressure is high or
the properties are near to the vapor dome, in which case use compressibility charts or
non-ideal state equations.
Let us look at how the tables are organized.

2.4.4.1 Saturated liquid-vapor water, temperature tables

For water, the most important table is the saturated steam table. One should go to such
tables first. If the water is a two-phase mixture, tables of this type must be used as the
equation of state. Recall, for two-phase mixtures, pressure and temperature are not inde-
pendent thermodynamic variables. Two properties still determine the state, but quality x is
now important. So for two-phase mixtures we allow
• T = T (v, x),
• P = P (v, x), or
• v = v(T, x),
for example. But P 6= P (T, v) in contrast to ideal gases. or any superheated vapor.
Consider the structure of saturation tables, as shown in Table 2.2, extracted from BS’s
Table B.1.1. Data from the steam tables is sketched in Fig. 2.23. We have the notation:
m3
Specific Volume, kg
Temp. Press. Sat. Liquid Evap. Sat. Vapor

◦
C kPa vf vf g vg
0.01 0.6113 0.001000 206.131 206.132
5 0.8721 0.001000 147.117 147.118
10 1.2276 0.001000 106.376 106.377
15 1.705 0.001001 77.924 77.925
.. .. .. .. ..
. . . . .
35 5.628 0.001006 25.2148 25.2158
40 7.384 0.001008 19.5219 19.5229
.. .. .. .. ..
. . . . .
374.1 22089 0.003155 0 0.00315
Table 2.2: Saturated liquid-vapor water tables, temperature entry, from BS, Table B.1.1.
• f : saturated liquid,
• g: saturated vapor,
• vf : specific volume of saturated liquid, and
• vg : specific volume of saturated vapor.

o
T ( C)
compressed two-phase superheated

liquid mixture vapor
15 oC
f g
vfg = 77.924 m3/kg
vf =0.001000 m3/kg vg =77.925 m3/kg

v (m3/kg)
Figure 2.23: Vapor dome for H2 O with data for vf , vg , and vf g at T = 15◦ C.
Note for liquid-vapor mixtures, this table begins at the triple point temperature 0.01 ◦ C and
ends at the critical temperature 374.1 ◦ C. At P = Pc and T = Tc , we have vf = vg . Note
that
• vf ≃ constant, and
• vg decreases with increasing T .
We define vf g as
vf g ≡ vg − vf . (2.53)
Recall the quality x is

mvap
x= .
mtotal
Consider a mass of fluid m in total volume V . We will take
V = Vliq + Vvap , (2.54)

m = mliq + mvap . (2.55)

Now, use the definition of specific volume and operate on Eq. (2.54) to get
V Vliq Vvap
z}|{ z }| { z }| {
mv = mliq vf + mvap vg , (2.56)
mliq mvap
v = vf + vg , (2.57)
m m
m − mvap mvap
= vf + vg , (2.58)
m m
= (1 − x)vf + xvg , (2.59)
= vf + x (vg − vf ) . (2.60)
| {z }
=vf g
We get the final important results:
v = vf + xvf g , (2.61)
v − vf
x= . (2.62)
vf g
What was presented in this section was a type of mixture theory. This particular theory
describes a so-called Amagat mixture. Amagat mixtures have components that share a
common temperature T and pressure P , but each component has its own volume, a so-called
partial volume. The key equation for Amagat mixtures is Eq. (2.54), V = Vliq + Vvap . Here
the partial volumes are Vliq and Vvap . It describes two-phase gas-liquid mixtures well; it
also most useful for mixtures with two materials with distinct densities and clear boundaries
between materials. In contrast, mixtures of gases are often better described by a so-called
Dalton mixture theory. Components of a Dalton mixture share a common temperature T
and volume V , but each component has its own pressure, a so-called partial pressure. The
key equation for a two-component Dalton mixture is that the total pressure is the sum of
the partial pressures, e.g. P = PA + PB . Such mixtures do not possess distinct boundaries
to distinguish components. Dalton mixture theory is the foundation of theories to describe
gas phase chemistry. In general the theory of mixtures is difficult subject, and there is no
general agreement on what mathematical model best describes general systems. Depictions
of Amagat and Dalton are given in Fig. 2.24.
2.4.4.2 Saturated liquid-vapor water, pressure tables

Sometimes we are given the pressure of the mixture, and a saturation table based on the
pressure is more useful. An example of a portion of such a table is shown in Table 2.3.
Example 2.4
Given a vessel with V = 0.4 m3 filled with m = 2 kg of H2 O at P = 600 kPa, find
• the volume and mass of liquid, and

a) b)
Figure 2.24: a) Émile Hilaire Amagat (1841-1915), French physicist. Image from
https://en.wikipedia.org/wiki/Émile Amagat, b) John Dalton (1766-1844), English
scientist who developed atomic theory and studied color blindness. Image from
https://en.wikipedia.org/wiki/John Dalton.
• the volume and mass of vapor.
The problem is sketched in Fig. 2.25. While the problem statement suggests we have a two-phase
water vapor
P = 600 kPa
m = 2 kg
V = 0.4 m3
liquid water
Figure 2.25: Schematic for liquid-vapor mixture example problem.
mixture, that is not certain until one examines the tables. First, calculate the specific volume of the
water:
V 0.4 m3 m3
v= = = 0.2 . (2.63)
m 2 kg kg
Next go to the saturated water tables with pressure entry to see if the water is a two-phase mixture.

m3
Specific Volume, kg
Press. Temp. Sat. Liquid Evap. Sat. Vapor

kPa ◦
C vf vf g vg
0.6113 0.01 0.001000 206.131 206.132
1.0 6.98 0.001000 129.20702 129.20802
1.5 13.03 0.001001 87.97913 87.98013
2.0 17.50 0.001001 67.00285 67.00385
.. .. .. .. ..
. . . . .
22089 374.1 0.003155 0 0.00315
Table 2.3: Saturated water tables, pressure entry from BS, Table B.1.2.
We find at P = 600 kPa that
m3 m3
vf = 0.001101 , vg = 0.31567 . (2.64)
kg kg
Now, for our mixture, we see that vf < v < vg , so we have a two-phase mixture. Now, apply Eq. (2.62)
to find the quality.
3 3
v − vf v − vf 0.2 m m
kg − 0.001101 kg
x= = = 3 m3
= 0.632291. (2.65)
vf g vg − vf 0.31567 m
kg − 0.001101 kg
Now, from Eq. (2.1), we have x = mvap /mtotal , so
mvap = xmtot = 0.632291(2 kg) = 1.26458 kg. (2.66)
Now, for the liquid mass we have
mliq = mtotal − mvap = (2 kg) − (1.26458 kg) = 0.735419 kg. (2.67)
Most of the mass is vapor, but the fraction that is liquid is large.
Now, let us calculate the volumes.

m3
Vvap = mvap vg = (1.26458 kg) 0.31567 = 0.39919 m3 , (2.68)
kg

m3
Vliq = mliq vf = (0.735419 kg) 0.001101 = 0.000809696 m3 . (2.69)
kg
The volume is nearly entirely vapor.

Temp. v u h s
m3 kJ kJ kJ
◦
C kg kg kg kg K
P = 10 kPa (45.81 ◦ C)
Sat. 14.67355 2437.89 2584.63 8.1501
50 14.86920 2443.87 2592.56 8.1749
100 17.19561 2515.50 2687.46 8.4479
150 19.51251 2587.86 2782.99 8.6881
.. .. .. .. ..
. . . . .
Table 2.4: Superheated water tables, from BS, Table B.1.3.
2.4.4.3 Superheated water tables
The superheat regime is topologically similar to an ideal gas. For a superheated vapor,
the quality x is meaningless, and we can once again allow pressure and temperature to be
independent. Thus, we can have v = v(T, P ). And the tables are in fact structured to give
v(T, P ) most directly. An example of a portion of such a table is shown in Table 2.4. This
portion of the superheated tables focuses on a single isobar, P = 10 kPa. At that pressure,
the saturation temperature is 45.81 ◦ C, indicated in parentheses. As long as T > 45.81 ◦ C,
we can use this table for P = 10 kPa water. And for various values of T > 45.81 ◦ C, we
find other properties, such as specific volume v, and properties we have not yet focused on,
internal energy u, enthalpy h, and entropy s.
2.4.4.4 Compressed liquid water tables
Liquids truly have properties that vary with both T and P . To capture such variation, we
can use compressed liquid tables as an equation of state. An example for water is given
in Table 2.5. If compressed liquid tables do not exist, it is usually safe enough to assume
properties are those for x = 0 saturated liquid at the appropriate temperature.
2.4.4.5 Saturated water, solid-vapor
Other types of saturation can exist. For example, below the triple point temperature, one
can have solid water in equilibrium with water vapor. The process where ice transforms
directly to water vapor is known as sublimation. Saturation tables for ice-vapor equilibrium
exist as well. For example, consider the structure of saturation tables, as shown in Table
2.6, extracted from BS’s Table B.1.5.

Temp. v u h s
m3 kJ kJ kJ
◦
C kg kg kg kg K
P = 500 kPa (151.86 ◦ C)

Sat. 0.001093 639.66 640.21 1.8606
0.01 0.000999 0.01 0.51 0.0000
20 0.001002 83.91 84.41 0.2965
40 0.001008 167.47 167.98 0.5722
.. .. .. .. ..
. . . . .
Table 2.5: Compressed liquid water tables, from BS, Table B.1.4.
m3
Specific Volume, kg
Temp. Press. Sat. Solid Evap. Sat. Vapor

◦
C kPa vi vig vg
0.01 0.6113 0.0010908 206.152 206.153
0 0.6108 0.0010908 206.314 206.315
-2 0.5177 0.0010905 241.662 241.663
-4 0.4376 0.0010901 283.798 283.799
.. .. .. .. ..
. . . . .
Table 2.6: Saturated solid-vapor water tables, temperature entry, from BS, Table B.1.5.

2.4.4.6 Tables for other materials

For many materials similar tables exist, e.g., ammonia, NH3 . Consider the ammonia satura-
tion tables, as shown in Table 2.7, extracted from BS’s Table B.2.1. One also has tables for
m3
Specific Volume, kg

◦
C kPa vf vf g vg
-50 40.9 0.001424 2.62557 2.62700
-45 54.5 0.001437 2.00489 2.00632
-40 71.7 0.001450 1.55111 1.55256
-35 93.2 0.001463 1.21466 1.21613
.. .. .. .. ..
. . . . .
132.3 11333.2 0.004255 0 0.004255
Table 2.7: Saturated liquid-vapor ammonia tables, temperature entry, from BS, Table B.2.1.
superheated ammonia vapor. An example of a portion of such a table is shown in Table 2.8.
Other tables in BS, include those for carbon dioxide, CO2 , a modern refrigerant, R-410a,12
Temp. v u h s
m3 kJ kJ kJ
◦
C kg kg kg kg K
P = 50 kPa (−46.53 ◦ C)
Sat. 2.1752 1269.6 1378.3 6.0839
-30 2.3448 1296.2 1413.4 6.2333
-20 2.4463 1312.3 1434.6 6.3187
-10 2.5471 1328.4 1455.7 6.4006
.. .. .. .. ..
. . . . .
Table 2.8: Superheated ammonia tables, from BS, Table B.2.2.
another common refrigerant, R-134a,13 diatomic nitrogen, N2 , and methane, CH4 .
2.4.4.7 Linear interpolation of tabular data

• Interpolation is often required when exact values are not tabulated.
12
a common cooling fluid invented in 1991, a near-azeotropic mixture of difluoromethane and pentafluo-
roethane.
13
a cooling fluid that became common in the 1990s, 1,1,1,2-tetrafluoroethane.

• In this course we will primarily use linear interpolations.
• Use extrapolations only if there is no other choice.
• Occasionally double interpolations will be necessary.
2.4.4.7.1 Single interpolation The most common interpolation is the single interpo-
lation of variables. We give an example here.
Example 2.5
Given water at T = 36.7 ◦ C, with v = 10 m3 /kg, find the pressure and the quality if a two-phase
mixture.
A wise first step is to go to the saturated tables. We check Table B.1.1 from BS and find there are
no values at T = 36.7 ◦ C. So we must create our own personal steam table values at this temperature,
just to determine if where we are on the thermodynamic surface. We list the important part of the
saturated water liquid-vapor tables in Table 2.9.
m3
Specific Volume, kg

◦C kPa vf vf g vg
.. .. .. .. ..
. . . . .
35 5.628 0.001006 25.2148 25.2158
36.7 ? ? ? ?
40 7.384 0.001008 19.5219 19.5229
.. .. .. .. ..
. . . . .
Table 2.9: Relevant portion of saturated liquid-vapor water tables, temperature entry, from
BS, Table B.1.1.
We seek to get appropriate values for P , vf , vf g , and vg at T = 36.7 ◦ C. Let us find P first. The
essence of linear interpolation is to fit known data to a straight line, then use the formula of that line to
predict intermediate values of variables of interest. We know values of P at T = 35 ◦ C and T = 40 ◦ C.
In fact we have two points: (T1 , P1 ) = (35 ◦ C, 5.628 kPa), and (T2 , P2 ) = (40 ◦ C, 7.384 kPa). This lets
us fit a line using the familiar point-slope formula:

P2 − P1
P − P1 = (T − T1 ). (2.70)
T2 − T1
| {z }
slope
We could have used the other point. Note when T = T1 , that P = P1 . Also, when T = T2 , P = P2 .

Substituting numbers, we get

(7.384 kPa) − (5.628 kPa)
P = (T − (35 ◦ C)) + (5.628 kPa), (2.71)
(40 ◦ C) − (35 ◦ C)

kPa
= 0.3512 ◦ (T − (35 ◦ C)) + (5.628 kPa), (2.72)
C

kPa
= 0.3512 ◦ ((36.7 ◦ C) − (35 ◦ C)) + (5.628 kPa), (2.73)
C
= 6.225 kPa. (2.74)
The interpolation is sketched in Fig. 2.26.
P (kPa)
7.384 2
6.225
5.628 1
35 36.7 40 T (oC)
Figure 2.26: Sketch of linear interpolation to find P when T = 36.7 ◦ C, v = 10 m3 /kg for
water.
Now, we need to interpolate for vf and vg as well. Let us apply the same technique. For vf , we
have  
3 3

m3
0.001008 mkg − 0.001006 mkg
vf − 0.001006 =  (T − (35 ◦ C)). (2.75)
kg (40 ◦ C) − (35 ◦ C)
When T = 36.7 ◦ C, we get

m3
vf = 0.00100668 . (2.76)
kg
For vg , we get
 3

m3

m3
19.5229 mkg − 25.2158 kg
vg − 25.2158 =  (T − (35 ◦ C)). (2.77)
kg (40 ◦ C) − (35 ◦ C)
When T = 36.7 ◦ C, we get

m3
vg = 23.2802 . (2.78)
kg

Knowing vf and vg , we do not need to interpolate for vf g . We can simply use the definition:

m3 m3 m3
vf g = vg − vf = 23.2802 − 0.00100668 = 23.2792 . (2.79)
kg kg kg
Now, v = 10 m3 /kg. Because at T = 36.7◦ C, we have vf < v < vg , we have a two-phase mixture.
Let us get the quality. From Eq. (2.62), we have

m3 m3
v − vf 10 kg − 0.00100668 kg
x= = m 3 = 0.429525. (2.80)
vf g 23.2792 kg
Thus
mvap
x= = 0.429525. (2.81)
mtot
2.4.4.7.2 Double interpolation Sometimes, we need to do two linear interpolations.

We call this double interpolation. Say we are given superheated water with vo and To and
we are asked to find Po . But neither vo nor To are listed in the tables. Then we need to do
a multi-step procedure.
• Go to the tables and for the given To and vo , estimate approximately the value of Po
by visual examination.
• For a nearby value of P = P1 , get a linear interpolation of the form T = T (v, P1). Use
this to get T1 = T (vo , P1 ).
• For a different nearby value of P = P2 , get another linear interpolation of the form
T = T (v, P2). Use this to get T2 = T (vo , P2 ). We now have two points (T1 , P1 ) and
(T2 , P2 ), both valid at v = vo .
• Use the two points (T1 , P1 ), (T2 , P2 ) to develop a third interpolation P = P (T, vo ).
Estimate Po by Po = P (To , vo ).
Example 2.6
Consider m = 1 kg of H2 O initially at T1 = 110 ◦ C, x1 = 0.9. The H2 O is heated until T2 = 200 ◦ C.
As sketched in Fig. 2.27, the H2 O is confined in a piston-cylinder arrangement, where the piston
is constrained by a linear spring with dP/dv = 40 kPa/m3 /kg. At the initial state, the spring is
unstretched. Find the final pressure.
While this problem seems straightforward, there are many challenges. Let us first consider what we
know about the initial state. Because we have a numerical value for x1 , we know state 1 is a two-phase
mixture. From the tables, we find that
m3 m3 m3
P1 = 143.3 kPa, vf 1 = 0.001052 , vg1 = 1.2101 , vf g1 = 1.20909 . (2.82)
kg kg kg

water
y
free body diagram
Figure 2.27: Sketch of a piston-cylinder-spring arrangement.
We can then calculate v1 for the mixture:

m3 m3 m3
v1 = vf 1 + x1 vf g1 = 0.001052 + (0.9) 1.20909 = 1.08923 . (2.83)
kg kg kg
We now know everything we need about state 1.

At state 2, we only know one intensive thermodynamic property, the temperature, T2 = 200 ◦ C.
• To get a second, and thus define the final state, we will need to bring in information about the process.
Now, we will need to consider a force balance on the piston. Newton’s second law for the piston says
d2 y X
mpiston = Fy . (2.84)
dt2
From our free body diagram, we note three forces:
• force due to the interior pressure from the water,
• force due to the exterior pressure from the atmosphere,
• force due to the linear spring, that we call Fs = k(y − y1 ) where k is the spring constant, y is the
position of the piston, and y1 is the initial position of the piston. Note that Fs = 0 when y = y1 .
We write this as
d2 y
mpiston = P A − Patm A − k(y − y1 ). (2.85)
dt2
Now, in classical thermodynamics, we make the assumption that the inertia of the piston is so small
that we can neglect its effect. We are really requiring that a force balance exist for all time. Thus,
even though the piston will move, and perhaps accelerate, its acceleration will be so small that it can
be neglected relative to the forces in play. We thus take
d2 y
mpiston ≪ |Fs |, |Patm A|, |P A|. (2.86)
dt2
With this assumption, we have
0 ≃ P A − Patm A − k(y − y1 ). (2.87)

Solve for P , the water pressure, to get

k
P = Patm + (y − y1 ). (2.88)
A
Now, V = Ay and V1 = Ay1 , so we can say
k
P = Patm + (V − V1 ). (2.89)
A2
Let us use the fact that V = mv and V1 = mv1 to rewrite as
km
P = Patm + (v − v1 ). (2.90)
A2
This equation is highlighted because it provides an algebraic relationship between two intensive thermo-
dynamic properties, P and v, and such a tactic will be useful for many future problems. Using numbers
from our problem, with dP/dv = km/A2 , we can say
!
kPa m3
P = (143.3 kPa) + 40 m3 v − 1.08923 , (2.91)
kg
kg
!
kPa
P = (99.7308 kPa) + 40 m3 v. (2.92)
kg
| {z }
linear spring rule
Now, at state 1, we have V = V1 and P = P1 = 143.3 kPa, so we must have Patm = 143.3 kPa for this
problem.
Let us now consider the possibilities for state 2. We are constrained to be on the line in P − v space
given by our force balance, Eq. (2.92). We are also constrained to be on the T = 200 ◦ C isotherm,
that is also a curve in P − v space. So let us consider the P − v plane, as sketched in Fig. 2.28. The
P P
km/A2 ~ large
km/A2 ~ small
2 2
T2 =200 oC T2 =200 oC
1 T1 =110 oC 1 T1 =110 oC
v v
Figure 2.28: Sketch of two configurations of the P − v plane for a piston-cylinder-linear

spring problem for water.
isotherms for T1 = 110 ◦ C and T2 = 200 ◦ C are set in both parts of Fig. 2.28. Because both T1 and
T2 are well below Tc , both isotherms pierce the vapor dome. Our final state has a line in P − v space
from the force balance intersecting the state 2 isotherm. There are two distinct possibilities for the
final state:

• for a stiff spring, i.e. large km/A2 , our line will intersect the isotherm within the vapor dome, or
• for a loose spring, i.e. small km/A2 , our line will intersect the isotherm in the superheated vapor
region.
Let us consider the first possibility: state 2 is under the vapor dome. If that is the case, then the
tables tell us that P2 = 1553.8 kPa. At this pressure, Eq. (2.92) gives us v = (1553.8 − 99.7308)/40 =
36.3517 m3 /kg. However, at this pressure vg = 0.12736 m3 /kg. Because we just found v > vg , our
assumption that the final state was under the dome must be incorrect!
Therefore, let us go to the more difficult case posed by the second possibility: state 2 is a superheated
vapor. In general the intersection of the straight line with the isotherm is difficult. We can use
linearization to assist us. Let us choose a small region of the tables, and locally fit the 200 ◦ C isotherm
to a straight line. This will give us a second independent equation in P − v space. We will then solve
two equations in two unknowns for the final state.
Our initial pressure, P1 = 143.3 kPa lies between 100 kPa and 200 kPa. We have values from the
superheat tables at these pressures for v at 200 ◦ C. So, we approximate the isotherm by the line

(200 kPa) − (100 kPa) m3
P − (100 kPa) = v − 2.17226 . (2.93)
1.08034 m
3
− 2.17226 m3 kg
kg kg
!
kPa
P = (298.939 kPa) − 91.5818 m3 v. (2.94)
kg
| {z }
linear approximation of isotherm from tabular thermal EOS
We simultaneously solve the two linear equations, Eqs. (2.92, 2.94), and get the unique solution
m3
P2 = 160.289 kPa, v2 = 1.151395 . (2.95)
kg
Because we found 100 kPa < P2 = 160.289 kPa < 200 kPa, we made a good assumption on the final
pressure, and our interpolation values from the tables are acceptable. Lastly, we sketch the process in
the T − v and P − T planes in Fig. 2.29.

T P
Pa
k
9
28
60.
1
= 2 liquid
P2
solid 2
Pa
k
3.3 1
14
=
P1 vapor
1
v T
Figure 2.29: Sketch of the T − v and P − T planes for piston-cylinder-linear spring problem
for water.

Chapter 3
The first law of thermodynamics
Read BS, Chapter 3
Recall in the example problem of Chapter 1, p. 24, we saw that mechanical energy was
conserved for a mass falling under the sole influence of a gravitational force. We took the
mechanical energy to be the sum of the kinetic and potential energy of the system. But
when we included a drag force, we found that mechanical energy was no longer conserved,
but in fact dissipated with time. There was considerable discussion in the 17th and 18th
centuries that pitted advocates of a so-called vis viva (“force of life,” a type of kinetic energy)
against those who argued for the primacy of momentum conservation. Leibniz led the vis viva
camp, and Newton and Descartes led the momentum camp. Both ultimately are equivalent
formulations when analyzed carefully.
In this chapter, we will expand our notion of energy and in so doing recover a new
conservation principle. This new principle, not known to Newton, is the first law of thermo-
dynamics. It takes many equivalent forms, and relies at a minimum on the introduction of
a new type of energy, thermal energy, that is necessary to conserve the total energy.
Thermal energy is actually a macro-scale representation of micro-scale mechanical energy.
Recall that at the micro-scale, molecules are in random motion. This random motion has
kinetic energy associated with it. But we cannot hope to keep track of it all for each
individual particle. So we surrender knowledge of the micro-scale motions, and allow the
temperature to be a measure of the average micro-scale kinetic energy. We can also take the
historical approach and develop the principle of energy conservation without further appeal
to micro-scale arguments. Let us begin that approach here.
3.1 Mathematical preliminaries: exact differentials

We first review some notions from calculus of many variables. Recall in thermodynamics,
we are often concerned with functions of two independent variables, e.g. P = P (v, T ), as is
found in an equation of state. Here, let us consider w = w(x, y) for a general analysis. Here
x, y, and w are simply general mathematical variables with no thermodynamic meaning.
81
82 CHAPTER 3. THE FIRST LAW OF THERMODYNAMICS
Later, we shall reserve the variable w for thermodynamic work.
3.1.1 Partial derivatives

Recall if w = w(x, y), then the partial derivative of w can be taken if one of the variables is
held constant.
Example 3.1 p
If w = x2 + y 2 , find the partial of w with respect to x and then with respect to y.
First let us get the derivative with respect to x. We take

∂w x
= p . (3.1)
∂x y x2 + y 2
Next for the derivative with respect to y, we have
∂w y
= p . (3.2)
∂y x x + y2
2
3.1.2 Total derivative

If w = w(x, y), for every x and y, we have a w. Here, we can think of w as a scalar field.
The field here is the (x, y) plane, and w varies over this field. If we change x only, we expect
a change in w. If we change y only, we expect a change in w. If we change both x and y, we
also expect a change in w. This ability of w to change in response to changes in both x and
y is reflected in the total differential:
∂w ∂w
dw = dx + dy. (3.3)
∂x y ∂y x
On a particular path C in the x − y plane along which we know y = y(x), we also have the
total derivative
dw ∂w ∂w dy
= + . (3.4)
dx ∂x y ∂y x dx
Now, we can integrate dw along a variety of paths C in the x − y plane. Two paths from
(x1 , y1 ) to (x2 , y2 ) are shown in Fig. 3.1. At the respective points, the function takes on the
values of w1 and w2 . Integrating Eq. (3.3), we get
Z 2 Z !
∂w ∂w
dw = dx + dy . (3.5)
1 C ∂x y ∂y x

3.1. MATHEMATICAL PRELIMINARIES: EXACT DIFFERENTIALS 83
path B
path A
Figure 3.1: Sketch of two paths from (x1 , y1 ) to (x2 , y2 ).
Now, because w = w(x, y), it will not matter which path we choose. The integral is said to
be path-independent.
Conversely, if we were given
dw = M(x, y) dx + N(x, y) dy, (3.6)
the associated integrals are path-independent iff w(x, y) can be found by solving.
∂w ∂w
M= , N= . (3.7)
∂x y ∂y x
One easy way to check this is to form the following two partial derivatives of Eqs. (3.7):
∂M ∂2w ∂N ∂2w
= , = . (3.8)
∂y x ∂y∂x ∂x y ∂x∂y
Now, if w(x, y) and all its partial derivatives are continuous and differentiable, it is easy
to prove the order of differentiation does not matter: ∂ 2 w/∂x∂y = ∂ 2 w/∂y∂x. Thus, if
w = w(x, y), we must insist that
∂M ∂N
= . (3.9)
∂y x ∂x y
We define the following:
• exact differential: a differential that yields a path-independent integral.
Example 3.2
If dw = x dx + y dy, is dw exact?

Here,
∂w
M= = x. (3.10)
∂x y
Thus
1 2
w= x + f (y). (3.11)
2
Thus
∂w df
= = y. (3.12)
∂y x dy
Thus
1 2
f (y) = y + C, (3.13)
2
and
1 2
w(x, y) = (x + y 2 ) + C. (3.14)
2
Examine a difference in w:
1 2
w2 − w1 = (x − x21 + y22 − y12 ). (3.15)
2 2
Note the constant C drops out and that the difference in w only depends on the end points and not
the path between points 1 and 2. Here, dw is exact. Note also that
∂M ∂N
= 0, = 0. (3.16)
∂y x ∂x y
Thus, our condition for exactness, Eq. (3.9), is satisfied.
Example 3.3
If dw = y dx − x dy, is dw exact?
Here,
∂w
M= = y. (3.17)
∂x y
Thus
w = yx + f (y). (3.18)
So
∂w df
=x+ = −x. (3.19)
∂y x dy
Thus
df
= −2x. (3.20)
dy
But functions of y cannot be functions of x. Therefore, we have generated nonsense. We cannot find
w(x, y)! Thus, w is not a state variable, and the differential of w is inexact. We give a new notation
for such differentials: δw, and we say more accurately for this inexact differential
δw = y dx − x dy. (3.21)
If we choose a path, we can still find differences in w. Let us examine the integral of w along two paths,
illustrated in Fig. 3.2.

path B, C2 (x,y)=(1,1)
(x,y)=(0,1)
2
path B, C1
h
at
p
(x,y)=(0,0)
1 x
Figure 3.2: Sketch of two paths in the x − y plane.
• Path A: Integrate from (x, y) = (0, 0) to (x, y) = (1, 1) along the path x = y and find w2 − w1 .
On path A, x = y and dx = dy. So eliminate y to get
δw = x dx − x dx = 0. (3.22)
Thus, δw = 0, and w = C after integrations. Thus w2 = w1 = C, and w2 − w1 = 0.

• Path B: Integrate from (x, y) = (0, 0) to (x, y) = (1, 1) along the path given first by the y-axis from 0
to 1, then by the line y = 1 from x = 0 to x = 1 and find w2 − w1 .
We have Z Z Z
2
δw = y dx − x dy + y dx − x dy. (3.23)
1 C1 C2
On C1 , we have x = 0, and dx = 0. On C2 , we have y = 1 and dy = 0. So we get

Z 2 Z Z Z 1
δw = y(0) − (0) dy + (1) dx − x(0) = dx. (3.24)
1 C1 C2 0
Thus
w2 − w1 = 1. (3.25)
We chose a different path and found a different difference in w. Note also that
∂M ∂N
= 1, = −1; (3.26)
∂y x ∂x y
thus, our condition for exactness, Eq. (3.9), is not satisfied.

As an aside, we note that the notions of exact differentials and path-independence are
sometimes presented in other terms, especially in mathematics texts. We review some of
these alternates here. We begin by restricting attention to a function of one variable, w(x),
and take that
dw
= w ′ (x), (3.27)
dx
where w ′ (x) is simply a known function of x. We can thus say
dw = w ′(x) dx. (3.28)
We can integrate both sides to get
Z 2 Z 2
dw = w ′ (x) dx. (3.29)
1 1
Now at point 1, we have x = x1 and w = w1 . Similarly, at point 2, we have x = x2 and

w = w2 . So
Z x2
dw
w2 − w1 = dx, (3.30)
x1 dx
Z x2
dw
w(x2 ) − w(x1 ) = dx. (3.31)
x1 dx
Equation (3.31 is the so-called fundamental theorem of calculus.

For functions of two independent variables, w(x, y), such as is common in thermody-
namics, we can perform an equivalent analysis. Let us begin with Eq. (3.3), that is the
two-dimensional analog of Eq. (3.28):
∂w ∂w
dw = dx + dy. (3.32)
∂x y ∂y x
Recalling that the superscript (·)T denotes the transpose, take now the two-dimensional
column vector dx and the two-dimensional row vector (∇w)T as

dx T ∂w ∂w
dx = , (∇w) = ∂x y ∂y , (3.33)
dy x
so that we can rewrite Eq. (3.32) in vector form as

∂w ∂w dx
dw = ∂x y ∂y = (∇w)T · dx. (3.34)
x dy
Let us now integrate both sides on a contour C, that is a specified path in the x − y plane:
Z Z
dw = (∇w)T · dx, (3.35)
C Z C
w(x2 , y2 ) − w(x1 , y1 ) = (∇w)T · dx. (3.36)

C

The end points of C are (x1 , y1)T and (x2 , y2)T . Here Eq. (3.36) is the two-dimensional
equivalent of Eq. (3.31) and is sometimes called the fundamental theorem for line integrals.
Because we have required w = w(x, y), the integral on any C is path-independent..
If the curve C is a closed contour, then its beginning point is its end point. Thus
(x1 ,Hy1 )T = (x2 , y2)T , and so w(x1 , y1) = w(x2 , y2). We denote closed contour integrals
by C . In this case Eq. (3.36) reduces to
I
0= (∇w)T · dx. (3.37)
C
We may confirm this by recalling Stokes’ theorem:

I ZZ
T
f · dx = (∇ × f)T · n dA. (3.38)
C A
Here f is any vector field. For a vector field, each component of the vector may vary over
the field. For example, for a two-dimensional vector field, f(x), we can expect to have

fx x
f= , x= , (3.39)
fy y
and then
fx = fx (x, y), fy = fy (x, y). (3.40)
We also may recall from calculus that iff f has the special property that it is the gradient of
a scalar field, f = ∇w, then the vector field f is curl-free; that is
∇ × f = ∇ × ∇w = 0. (3.41)
Thus we have for this special case of Stokes’ theorem that

I ZZ
T
(∇w) · dx = 0T · n dA = 0, (3.42)
C S
thus confirming Eq. (3.37).
Example 3.4
If w = w(x, y), verify ∇w is curl-free.
We have
∂w !
∂x y
∇w = ∂w . (3.43)
∂y x

And we find
i j k
∂ ∂
0 ∂2w ∂2w
∇ × ∇w = ∂x ∂y = − k = 0. (3.44)
∂w ∂w ∂x∂y ∂y∂x
∂x y ∂y x 0
This holds for all w(x, y) that are continuous and differentiable because under these conditions the
operators ∂/∂x and ∂/∂y commute. For example, if w = x2 + xy 3 , we find
∂w ∂w
= 2x + y 3 , = 3xy 2 . (3.45)
∂x y ∂y x
We then see that

∂2w ∂2w
= 3y 2 , = 3y 2 . (3.46)
∂y∂x ∂x∂y
The two mixed partials are the same, despite the fact they were taken in different order.
Example 3.5
Consider the two-dimensional scalar field w(x) defined by
w(x) = w(x, y) = x2 + y 2 . (3.47)
A contour plot of curves of constant w is given in Fig. 3.3. We see the magnitude w is zero at the
4
3
1 2
y 0
-1
-2
-2 -1 0 1 2
x
Figure 3.3: Scalar field for w(x, y) = x2 + y 2 .

origin. We also see that the gradient of w is

∂ 2 ∂ 2 2x
∇w = (x + y 2 )i + (x + y 2 )j = 2xi + 2yj = . (3.48)
∂x ∂y 2y
The gradient vector points away from the origin and is in the direction of most rapid increase of w.
The gradient vector is normal to the contours of w.

Example 3.6
Consider the two-dimensional vector field f (x) defined by
fx (x, y) = x, fy (x, y) = y. (3.49)
A plot of the vector field is given in Fig. 3.4. We see the magnitude f is zero at the origin. We also
y 0
-1
-2
-2 -1 0 1 2
x
Figure 3.4: Vector field for fx = x, fy = y.
see that
∂fx ∂fy
∇·f = + = 1 + 1 = 2. (3.50)
∂x ∂y
And we see that
i j k
∂ ∂
∇×f = ∂x ∂y 0 = 0. (3.51)
x y 0
The vector field has a non-zero divergence, but is curl-free.
Example 3.7
Consider the two-dimensional vector field f (x) defined by
fx (x, y) = −x + y, fy (x, y) = x + y. (3.52)

y 0
-1
-2
-2 -1 0 1 2
x
Figure 3.5: Vector field for fx = −x + y, fy = x + y.
A plot of the vector field is given in Fig. 3.5. We see the magnitude f is zero at the origin. We also
see that
∂fx ∂fy
∇·f = + = −1 + 1 = 0. (3.53)
∂x ∂y
And we see that
i j k
∂ ∂
∇×f = ∂x ∂y 0 = 0. (3.54)
−x + y x+y 0
The vector field is both divergence-free and curl-free.
As an aside that will not be deeply explored further, for functions of three independent
variables, w(x, y, z), such as might occur in thermodynamics with additional work modes
like electric work, we can perform an equivalent analysis. Let us begin with
∂w ∂w ∂w
dw = dx + dy + dz, (3.55)
∂x y,z ∂y x,z ∂z x,y
that is the three-dimensional analog of Eq. (3.3): Take now the three-dimensional column
vector dx and the three-dimensional row vector (∇w)T as
 
dx
dx = dy  , (∇w)T = ∂w ∂x y,z
∂w
∂y
∂w
∂z x,y , (3.56)
x,z
dz

so that we can rewrite Eq. (3.55) in vector form as

 
dx
dw = ∂w ∂x y,z
∂w
∂y
∂w
∂z x,y
dy  = (∇w)T · dx. (3.57)
x,z
dz
Let us now integrate both sides on a contour C, that is a specified path in the x − y − z
volume:
Z Z
dw = (∇w)T · dx, (3.58)
C C
Z
w(x2 , y2 , z2 ) − w(x1 , y1 , z1 ) = (∇w)T · dx. (3.59)
C
The end points of C are (x1 , y1 , z1 )T and (x2 , y2 , z2 )T . Because we have required w =
w(x, y, z), the integral on any C is path-independent. If the curve C is a closed con-
tour, then its beginning point is its end point. Thus (x1 , y1 , z1 )T = (x2 , y2 , z2 )T , and so
w(x1 , y1 , z1 ) = w(x2 , y2, z2 ). In this case Eq. (3.59) reduces to
I
0 = (∇w)T · dx. (3.60)
C
We now apply these ideas to thermodynamics.
Example 3.8
Examine the ideal gas law P v = RT in the context of exact differentials.
We have P = RT /v; thus, in terms of differentials, we can say
∂P ∂P
dP = dv + dT, (3.61)
∂v T ∂T v
RT R
= − 2 dv + dT. (3.62)
v v
Now RT /v 2 > 0 and R/v > 0. Thus
• For an isothermal process, dv > 0 induces dP < 0; volume increase corresponds to pressure decrease.
• For an isochoric process, dT > 0 induces dP > 0; temperature increase corresponds to pressure
increase.
Let us verify the condition for an exact differential holds. We should have

∂ RT ∂ R
− 2 = , (3.63)
∂T v ∂v v
R R
− 2 = − 2. (3.64)
v v
The differential is exact, as it must be because we actually began with a path-independent state function.

Changes in v can induce changes in P . Changes in T can induce changes in P . And changes in
both v and T can induce changes in P . Let us see if there is a special path in T − v space that does
not induce changes in P , so dP = 0. This would be an isobaric path. On an isobaric path for an ideal
gas we have
RT R
dP = − 2
dv + dT, (3.65)
|{z} v v
=0
RT R dT
0 = − 2
+ , (3.66)
v v dv
T dT
= − + , (3.67)
v dv
dT T
= , (3.68)
dv v
dT dv
= , (3.69)
T v
T v
ln = ln , (3.70)
To vo
T v
= . (3.71)
To vo
Isobars, curves with dP = 0, are straight lines in the T − v plane for an ideal gas.
In closing, we note that an equivalent language is often used in the dynamics and robotics
literature. Differentials that are exact are associated with so-called holonomic systems.
And those that are inexact and cannot be rendered exact by use of an integrating factor
are associated with non-holonomic systems. So the state of a holonomic system is path-
independent, while that of a non-holonomic system is path-dependent.
3.2 Work
3.2.1 Definitions
From Newtonian mechanics, we know going from state 1 to state 2, that the work 1W2 is
done by a force moving through a distance. The word “work” was first used in this sense by
the French mechanician Gaspard-Gustave Coriolis, depicted in Fig. 3.6. Work is defined as
Z 2
1W2 = F · dx. (3.72)
1
R2
Formally, we should probably use the transpose notation and say 1W2 = 1 FT ·dx. However,
from here on, we adopt the common convention that the transpose is understood to be part
of the vector dot product. In differential form, we have
δW = F · dx. (3.73)

3.2. WORK 93
Figure 3.6: Gaspard-Gustave Coriolis (1792-1843), French physicist who used to

word “work” to characterize a force acting through a distance. Image from
https://en.wikipedia.org/wiki/Gaspard-Gustave Coriolis.
In one-dimensional systems, we have

Z 2
1W2 = F dx, (3.74)
1
δW = F dx. (3.75)
Note that we have anticipated that the work differential is inexact. This is an important
point, as work integrals will be path-dependent, and work will not be a state variable for a
system. Here, F is a three-dimensional force vector, x is a three-dimensional distance vector,
and · is the dot product operator. Recall that the dot product of two vectors yields a scalar.
The terms F and x are scalar equivalents valid for one-dimensional systems. The units of
force are N, those of distance are m, so the units of work are N m, that have been defined
as Joules (J).
Work is done by a system if the sole effect on the surroundings (i.e. everything external
to the system) could be the raising of a weight. We take the following sign convention:
• + work done by the system,
• − work done on the system.
This sign convention is not universal. Many physicists use precisely the opposite convention.
Probably the reason for this convention is that thermodynamics is a science that was invented
by engineers in the nineteenth century. And those engineers wanted to produce work from
steam engines. Systems doing work were viewed favorably and endowed with a positive sign.
We associate energy with the ability to do work. We define
• Power: the time rate of doing work = δW/dt.
• Specific work: the work per unit mass 1w2 = 1W2 /m. Because work is path-
dependent, the intensive quantity 1w2 is not a thermodynamic state variable.

3.2.2 Work for a simple compressible substance

Consider the scenario sketched in Fig. 3.7. In state 1, we have a material at pressure P
A A
P+dP x+dx
x
P
state 1 state 2
Figure 3.7: Sketch of a piston-cylinder arrangement as work is done as the material expands
when weights are removed.
confined in a cylinder of cross-sectional area A. The height of the piston in the cylinder is
x. The pressure force of the material on the piston is just balanced by weights on top of the
piston.
Now, remove one of the weights. We notice a motion of the piston to a new height x+ dx.
We let a long time elapse so the system comes to rest at its new equilibrium. We notice
there is a new pressure in the chamber, P + dP sufficient to balance the new weight force.
Obviously work was done as a force acted through a distance. Let us calculate how much
work was done. The differential work is given from Eq. (3.75) as
δW = F dx. (3.76)
Now, F varies during the process. At state 1, we have F = P A. At state 2, we have

F = (P + dP )A. Let us approximate F by its average value:
1 A
F ∼ (P A + (P + dP )A) = P A + dP. (3.77)
2 2
So
A A
δW = P A + dP dx = P A dx + dP dx . (3.78)
2 2 | {z }
∼0
Let us only retain terms that are differential and neglect the square of differential terms, so
δW = P A dx. (3.79)

3.2. WORK 95
Now, because A dx = dV , the differential volume, we get the important formula:
δW = P dV. (3.80)
We can integrate this and get the equally important

Z 2
1W2 = P dV. (3.81)
1
Note we employ the unusual notation 1W2 to emphasize that the work R 2 depends on the path
from state 1 to state 2. We are tempted to write the incorrect form 1 δW = W2 − W1 , but
this would imply the work is a state function, which it is not, as shown directly.
Example 3.9
Show that work is path-dependent.
We have δW = P dV. This notation presumes path-dependence. Let us presume however path-
independence and replace δW by dW . In terms of intensive variables, assuming path-independence, we
would have
dw = P dv. (3.82)
If w were a path-independent property, we could have w = w(P, v). That would admit the exact
∂w ∂w
dw = dv + dP. (3.83)
∂v P ∂P v
| {z } | {z }
=P =0
Our physics of dw = P dv + 0 dP tells us by comparison that we would need
∂w ∂w
= P, and = 0. (3.84)
∂v P ∂P v
Integrating the first gives

w = P v + f (P ). (3.85)
Differentiating with respect to P gives
∂w df (P )
=v+ = 0. (3.86)
∂P v dP
Thus
df (P )
= −v. (3.87)
dP
Functions of P cannot
R2 be functions of v if P and v are independent. Therefore, dw is not exact,
w 6= w(P, v), and 1 P dv is path-dependent.

P P
1 1
pa
th
A
2 2
path B
areaA
areaB
V V
Figure 3.8: P − V diagrams for work for two different processes connecting the same states.
R2
We can also see the path-dependence of 1W2 by realizing that 1W2 = 1 P dV represents
the area under a curve in a P − V diagram. Consider two paths, A and B from the same
points 1 to 2 as depicted in the P − V space of Fig. 3.8. The area under the curve defined
by Path A is clearly different from that under the curve defined by Path B. Clearly, the
work 1W2 depends on the path selected, and not simply the end points. Obviously then, to
calculate the work, we will need full information on P (V ) for the process under consideration.
Many processes in thermodynamics are well modeled as a
• Polytropic process: a process that is described well by an equation of the form
P V n = constant = C.
Here, n is known as the polytropic exponent.
Example 3.10
Find the work for a gas undergoing a polytropic process with n 6= 1.
A polytropic process has

C
P (V ) = . (3.88)
Vn
So the work is
Z 2 Z 2
C dV
1W2 = dV = C , (3.89)
1 Vn 1 Vn
C 2
= V 1−n 1 , (3.90)
1−n
C
= V21−n − V11−n . (3.91)
1−n
Now, C = P1 V1n = P2 V2n , so
P2 V2 − P1 V1
1W2 = . (3.92)
1−n
Note this formula is singular if n = 1.

3.2. WORK 97
Now, if n = 1, we have P V = C, that corresponds to an isothermal process if the material

is also an ideal gas. Note that non-ideal gases can also have P V = C; they just are not
isothermal. We need to be able to analyze polytropic processes with n = 1.
Example 3.11
Find the work for a gas undergoing a polytropic process with n = 1.
For this process, we have

C
P (V ) = . (3.93)
V
Therefore the work is Z 2
dV V2
1W2 = C = C ln . (3.94)
1 V V1
Because P1 V1 = C, we can say
V2
1W2 = P1 V1 ln . (3.95)
V1
Example 3.12
Find the work for an isobaric process.
An isobaric process is a polytropic process with n = 0. Thus

P (V ) = C = P1 . (3.96)
We also have P2 = P1 . The work is
Z 2 Z 2
1W2 = P1 dV = P1 dV = P1 (V2 − V1 ). (3.97)
1 1
Example 3.13
Find the work for an isochoric process.
An isochoric process has dV = 0. Thus

Z 2
1W2 = P |{z}
dV = 0. (3.98)
1
=0
There is no work for an isochoric process. This also corresponds to a polytropic process with n → ∞.

A family of paths in the P − V plane for a set of polytropic processes of varying n is

shown in Fig. 3.9.
1
n=0 (isobaric)
n<1 (polytropic)
n=1 (isothermal for ideal gases)
(isochoric) n>1 (polytropic)
Figure 3.9: P − V diagram for various polytropic processes.
Example 3.14
An ideal gas undergoes a two-step process. Beginning at state 1, it is isothermally compressed to
state 2. Then it is isobarically compressed to state 3. Find the work.
The process is sketched in Fig. 3.10. We have
3 2
Figure 3.10: Sketch of a two-step, isothermal-isobaric, compression of an ideal gas.

3.2. WORK 99
1W3 = 1W2 + 2W3 , (3.99)

Z 2 Z 3
= P dV + P dV, (3.100)
1 2
Z 2 Z 3
dV
= mRT1 + P2 dV, (3.101)
1 V 2
V2
= mRT1 ln + P2 (V3 − V2 ), (3.102)
V1
V2
= P1 V1 ln + P2 (V3 − V2 ). (3.103)
V1
Note that 1W3 < 0, because V2 < V1 and V3 < V2 . So work is done on the system in compressing it.
Note also that even though we do not know T , we can solve the problem. This is because work only
requires information about the P -V state of the system and the nature of the process.
Example 3.15
Consider the piston-cylinder arrangement sketched in Fig. 3.11. Here, m = 2 kg of water is initially
water
water m = 2 kg
m = 2 kg o
T3 = 200 C
V1 = 0.02 m3
o
T1 = 50 C
Q Q
Figure 3.11: Sketch of a piston-cylinder arrangement for water heating example.
at V1 = 0.02 m3 , T1 = 50 ◦ C. When P = 100 kPa, the piston leaves the stops. The water is heated
from its initial state to a final state of 200 ◦ C. Find diagrams for the process in the P − T , T − v, and
P − v planes and the work done by the water.
At state 1, we have
V1 0.02 m3 m3
v1 = = = 0.01 . (3.104)
m 2 kg kg
Go to the saturated water tables, temperature entry. At T1 = 50 ◦ C, we find vf = 0.001012 m3 /kg
and vg = 12.0318 m3 /kg. This gives vf g = 12.0308 m3 /kg. Because vf < v1 < vg , we have a two-phase
mixture at the initial state. This fixes P1 = 12.350 kPa.

Now, we have

m3 m3
v1 − vf 0.01 kg − 0.001012 kg
x1 = = m3
= 0.000747083. (3.105)
vf g 12.0308 kg
Next, heat at constant volume until

• the piston leaves the stops at P = 100 kPa, or
• the fluid becomes saturated.
We search the saturated water tables and examine the state at P = 100 kPa. We see at that
pressure that vf = 0.001043 m3 /kg and vg = 1.69400 m3 /kg. So when P reaches 100 kPa, we still have
vf < v < vg , so the water is still a two-phase mixture.
We define then state 2 as the state where P2 = 100 kPa with v2 = v1 = 0.01 m3 /kg. From here on
the process is isobaric. It is useful at this stage to consider a sketch of the processes given in Fig. 3.12.
Now, at P2 = 100 kPa, we find that T2 = 99.62 ◦ C. And we have v2 = v1 = 0.01 m3 /kg. Using vf and
T P P
3
2 3
2 3 2
1
1 1
v v T
Figure 3.12: Sketch of T − v, P − v, and P − T diagrams.
vg at P = 100 kPa, we find

m3 m3
v2 − vf 0.01 kg − 0.001043 kg
x2 = = m3
= 0.00529073. (3.106)
vf g 1.69296 kg
Now, we heat isobarically until T3 = 200 ◦ C, with P3 = P2 = 100 kPa. This gives us two properties,
so we know the state. The superheat tables give us v3 = 2.17226 m3 /kg. Now, the final volume is

m3
V3 = mv3 = (2 kg) 2.17226 = 4.34452 m3 . (3.107)
kg
Let us get the work.

1W3 = 1W2 +2W3 . (3.108)
|{z}
=0
But 1W2 = 0 because this is an isochoric process. So

Z 3 Z 3
1W3 = 2W3 = P dV = P2 dV = P2 (V3 − V2 ) = P2 m(v3 − v2 ). (3.109)
2 2

3.2. WORK 101
Table 3.1: Numerical values for water heating example
variable units state 1 state 2 state 3

P kPa 12.350 100 100
m3
v kg 0.01 0.01 2.17226
T ◦C 50 99.62 200
x - 0.000747083 0.00529073 -
V m3 0.02 0.02 4.34452
Substituting numbers, we find

m3 m3
1W3 = (100 kPa)(2 kg) 2.17226 − 0.01 = 432.452 kJ. (3.110)
kg kg
A summary is given in Table 3.1.
Example 3.16
Measured P − V data for an internal combustion engine is obtained. Estimate the work.
The data is given in Table 3.2. Here, we have N = 6 points. The best way to address this problem
Table 3.2: Values for P and V in an expansion process.
P (bar) V (cm3 )
20.0 454
16.1 540
12.2 668
9.9 780
6.0 1175
3.1 1980
is via numerical integration. We could use a variety of methods like Simpson’s rule. Let use a simpler
and robust method based on the areas of an assembly of trapezoids. We will estimate
Z 2 N
X −1
1W2 = P dV ≃ Piave ∆Vi . (3.111)
1 i=1
Here, we are taking the area of trapezoid i as the product of the base,
∆Vi = Vi+1 − Vi , i = 1, . . . , N − 1, (3.112)

and the average pressure of trapezoid i,
Pi + Pi+1
Piave = , i = 1, . . . , N − 1. (3.113)
2
If ∆Vi were uniform, one could show this is equivalent to the well-known trapezoidal method. We can
summarize the calculations in Table 3.3. We see
Table 3.3: Tabular calculation of work.

i P V Piave ∆Vi Piave ∆Vi
bar cm3 bar cm3 bar cm3
1 20.0 454 18.05 86 1552.3
2 16.1 540 14.15 128 1811.20
3 12.2 668 11.05 112 1237.6
4 9.9 780 7.95 395 3140.25
5 6.0 1175 4.55 805 3662.75
6 3.1 1980 - - -
11404.1
Z 2 N
X −1
1W2 = P dV ≃ Piave ∆Vi = 11404.1 bar cm3 . (3.114)
1 i=1
Let us convert to kJ:

100 kPa m3 kJ
11404.1 bar cm3 = 1.14 kJ. (3.115)
bar (100 cm)3 kPa m3
A sketch of the process is given in Fig. 3.13.
Example 3.17
We are given air in the spring-restrained piston-cylinder arrangement of Fig. 3.14 with P1 = 100 kPa,
V1 = 0.002 m3 , x1 = 0 m, no force on the piston at state 1, Patm = 100 kPa, and A = 0.018 m2 . The
air expands until V2 = 0.003 m3 . We know the spring is linear with Fspring = kx with k = 16.2 kN/m.
Find the final pressure of the air and the work done by the air on the piston.
First note here that x is distance and not quality! The free body diagram is sketched in Fig. 3.14.
For the piston to be in mechanical equilibrium, we require
PA = Patm A + kx, (3.116)

k
P = Patm + x. (3.117)
A

3.2. WORK 103
P (bar)
25
20
15
10
0 500 1000 1500 2000

V (cm3)
Figure 3.13: Sketch of P − V diagram.

x
air
k
P, V
Figure 3.14: Sketch of piston-spring problem.
This gives us P (x). So at the initial state, where x1 = 0, we have P1 = Patm = 100 kPa. We also need
V (x):
V (x) = V1 + Ax. (3.118)
Let us eliminate x. From Eq. (3.118), we get
V − V1
x= . (3.119)
A
Substitute Eq. (3.119) into Eq. (3.117) to get

k V − V1 k
P = Patm + = Patm + 2 (V − V1 ). (3.120)
A A A
Note, when V = V1 , we find P = Patm .
Now, to get the work, we take
Z 2 Z V2
k
1W2 = P dV = Patm + 2 (V − V1 ) dV. (3.121)
1 V1 A
| {z }
=P

We integrate this to find

V2
k 2
1W2 = Patm V + (V − V1 ) , (3.122)
2A2 V1
 
k 
= Patm (V2 − V1 ) + (V2 − V1 )2 − (V1 − V1 )2  , (3.123)
2A2 | {z }
=0
2
1 V2 − V1
= Patm (V2 − V1 ) + k (3.124)
| {z } 2 A
work done on atmosphere | {z }
work done on spring
The P − V diagram is sketched in Fig. 3.15. Let us calculate the numerical values.
P
2
k/A2
1 1
P1=Patm
V1 V2 V
Figure 3.15: Sketch of P − V diagram in piston-linear spring problem.

kN 1
P2 = (100 kPa) + 16.2 2 2
(0.003 m3 ) − (0.002 m3 ) = 150 kPa. (3.125)
m (0.018 m )
2
1 kN (0.003 m3 ) − (0.002 m3 )
1W2 = (100 kPa)((0.003 m3 ) − (0.002 m3 )) + 16.2 , (3.126)
| {z } 2 m 0.018 m2
0.1 kJ | {z }
0.025 kJ
= 0.125 kJ. (3.127)
Example 3.18
A gas is compressed from state (P1 , V1 ) to (P2 , V2 ) via two different paths, A and B:

3.2. WORK 105
• Path A: a polytropic process in which P V n = C.

• Path B: an isobaric compression to (P1 , V2 ) followed by an isochoric compression to (P2 , V2 ).
Determine the work along both paths.
First consider Path A. On Path A, we have
P V n = C = P1 V1n = P2 V2n . (3.128)
So
1
P =C . (3.129)
Vn
Now, the work is
Z 2 Z V2 V2
dV CV 1−n V21−n − V11−n
1W2 = P dV = C = =C . (3.130)
1 V1 Vn 1−n V1 1−n
Now, using C = P1 V1n = P2 V2n , we get
P2 V2 − P1 V1
1W2 = . (3.131)
1−n
Now, for Path B, we denote the intermediate state as ∗ and quickly calculate
Z 2 Z ∗ Z 2
W
1 2 = P dV = P dV + P dV. (3.132)
1 1 ∗
Now, the ∗ state has V∗ = V2 , so

Z 2 Z V2 Z V2
1W2 = P dV = P dV + P dV . (3.133)
1 V1 V2
| {z }
=0
Now, because the first part of the process is isobaric with P = P1 , the first integral is easy. And the
second integral is zero, because the integral has no width. So we get
Z V2
W
1 2 = P 1 dV = P1 (V2 − V1 ) = P1 V2 − P1 V1 . (3.134)
V1
Note the work for the different paths is different!

P2 V2 − P1 V1
6= P1 V2 − P1 V1 . (3.135)
1−n
The two different paths are sketched in the P − V diagrams of Fig. 3.16. In both processes, A and B,
the work is negative. The gas is worked upon; it is thus doing negative work.
Example 3.19
A spherical balloon contains air at P1 = 150 kPa and is placed in a vacuum. It has an initial
diameter of D1 = 0.3 m. The balloon is heated until its diameter is D2 = 0.4 m. It is known that the
pressure in the balloon is proportional to its diameter. Calculate the work of expansion.

P P
2 2
P2 P2
P1 1 P1 1
*
V2 V1 V V2 V1 V
Figure 3.16: Sketch of P − V diagrams for compression on two different paths.

P
vacuum
2
P2
P1 + heat P2
P1 1
state 1
state 2 V
V1 V2
Figure 3.17: Diagrams for balloon heating problem
The process is diagrammed in Fig. 3.17. We are told the pressure is proportional to the diameter.
For a sphere we have
3
4 D π 3
V = π = D . (3.136)
3 2 6
So 1/3
6V
D= . (3.137)
π
Because the volume is proportional to the cube of the diameter, the diameter is proportional to the
cube root of the volume. So the pressure is proportional to the cube root of the volume:
1/3
6V
P = kD = k . (3.138)
π
Here, we have defined the proportionality constant as k. The free body diagram is unusual and is
sketched in Fig. 3.18. Here, we give a sketch only for a sector of the balloon. The interior air pressure
exerts a net upwards force on the balloon surface. The lateral pressure forces cancel each other. The
net upwards force on this sector is balanced by a net downwards force exerted by the surface tension
force of the balloon material.

3.2. WORK 107
atmospheric pressure
surface air pressure net surface net air

surface pressure net surface
tension tension tension tension
force force
force
Figure 3.18: Free body diagram for a sector of the spherical balloon surface.
Let us calculate the work:

Z 2
1W2 = P dV, (3.139)
1
1/3 Z V2
6
= k V 1/3 dV, (3.140)
π V1
1/3 h iV2
6 3
= k V 4/3 , (3.141)
π 4 V1
1/3
6 3 4/3 4/3
= k V2 − V1 , (3.142)
π 4
1/3 4/3 !
6 3 4/3 V2
= k V1 −1 . (3.143)
π 4 V1
Now, we know state 1, so this lets us determine k:

−1/3
6V1
k = P1 (3.144)
π
Thus, the work is
−1/3 1/3 4/3 !
6V1 6 3 4/3 V2
1W2 = P1 V1 −1 , (3.145)
π π 4 V1
| {z
}
k
4/3 !
3 V2
= P1 V1 −1 , (3.146)
4 V1
3 ! 3 4/3 !
4 D1 3 D2
= P1 π −1 , (3.147)
3 2 4 D13
| {z }
V1
3 4 !
D1 D2
= P1 π −1 , (3.148)
2 D1
3 4 !
0.3 m 0.4 m
= (150 kPa)π −1 , (3.149)
2 0.3 m
= 3.43612 kJ. (3.150)

3.2.3 Other forms of work

We note that there are other forces besides pressure forces, and those forces can also do
work. Consider
• a stretching wire stretched by tension force T through length change dL. The differ-
ential work is
δW = −T dL. (3.151)
• a surface with surface tension S. The differential work is
δW = −S dA. (3.152)
• a system with electric work where E is the electric field strength, q is the particle
charge, and x is the distance:
δW = −qE dx. (3.153)
In total, for materials that are more than simple compressible substances, we have
δW = P dV − T dL − S dA − qE dx − . . . (3.154)
It can be shown that the more work modes we include, the more independent thermodynamic
variables are necessary to specify the state of the system. R2
Lastly we note that a gas expanding into a vacuum has 1W2 6= 1 P dV because it is
inherently a non-equilibrium process.
3.3 Heat
Let us make the following definition:
• Heat: a form of energy transferred across the boundary of a system at a given tem-
perature to another system (or the surroundings) at a different temperature by virtue
of the temperature difference between the two.
We adopt the notion that bodies do not contain heat, but that heat only has relevance as a
type of energy that crosses system boundaries. Note that work is in a similar class; it is not
contained within a system, but can be identified when it crosses system boundaries. We will
make a distinction between heat and work energy transfers.
We also note that when two bodies are at the same temperature, there can be no heat
transferred between the two bodies. The subject of heat transfer considers the details of the
heat transfer process. There are three fundamental classes of heat transfer:

3.3. HEAT 109
• heat diffusion, also called conduction. Physically this is due to local effects. Bacon fried
in a skillet on a stove via conduction effects is a culinary example. This is characterized
by Fourier’s law1    ∂T 
qx ∂x
q = −k∇T = qy  = −k  ∂T ∂y
, (3.155)
qz ∂T
∂z
where q is the heat flux vector with units J/s/m2 = W/m2 , k is the thermal conductivity
with units J/s/m/K = W/m/K, and ∇T is the vector representing the spatial gradient
of temperature. This version of Fourier’s law is valid for a material that is isotropic;
one for which there is no preferential direction for heat transfer. A more sophisticated
version is required for layered materials like lasagna.
While an argument exists for labeling this vector q̇ because it is a rate, we adhere to
the common notation of q for this rate. Recall that ∇T is a vector pointing in the
direction in which T rises most steeply with respect to changes in position. Because
of the minus sign, we see that the thermal energy flows in the direction of maximum
temperature decrease. This law was developed by Joseph Fourier, who built an elegant
and correct theory of a special case of non-equilibrium thermodynamics before the laws
of equilibrium thermodynamics were formulated, let alone fully understood. Fourier is
depicted in Fig. 3.19.
Figure 3.19: Jean Baptiste Joseph Fourier (1768-1830), French physicist and math-
ematician who developed a correct theory of heat conduction. Image from
https://mathshistory.st-andrews.ac.uk/Biographies/Fourier.html.
In one dimension and also neglecting any time-dependency, we get

dT
q = −k . (3.156)
dx
1
J. B. J. Fourier, 1822, Théorie Analytique de la Chaleur, Chez Firmin Didot, Paris.
1878 English translation.

If we multiply by the local cross-sectional area, we find Q̇ = qA, and

dT Thot − Tcold
Q̇ = −kA ∼ −kA . (3.157)
dx L
Here, Q̇ has units J/s or W (Watts).
• convection. This is actually a version of conduction, albeit enhanced by fluid flow.
Bacon baked in a pan in an oven via convection effects is a culinary example. For some
systems, convective effects are well modeled by Newton’s law of cooling2,3 :
q = h(Thot − Tcold ), (3.158)
Q̇ = qA = hA(Thot − Tcold ). (3.159)
Here, h is often taken to be a constant with units W/m2 /K. More sophisticated
approximations for h can be obtained when the motion of the surrounding fluid is
characterized by a detailed model of fluid physics.
• thermal radiation. Physically this is due to remote effects. The earth is heated by the
sun via radiation effects, not conductive energy diffusion. Bacon cooked on a platter
in a microwave oven via radiation effects is a culinary example. For some systems, the
radiative heat transfer rate is well modeled by
4 4
q = σ(Thot − Tcold ), (3.160)
4 4
Q̇ = qA = σA(Thot − Tcold ). (3.161)
Here, σ is the Stefan-Boltzmann constant, σ = 5.67 × 10−8 W/m2 /K4 .
We adopt the traditional engineering sign convention for heat:
• + heat enters the system,
• - heat leaves the system.
The sign convention again is motivated by nineteenth century steam engines. Heat had to
be added into a system to get work out of the system. Because engineers were and are
concerned with this problem, this convention is taken.
We define a special kind of process in which Q̇ = 0 as
• Adiabatic: a type of process for which there is no heat transfer.
,
The word “adiabatic” was first used by Rankine.4 It is from the Greek α διάβατ oς: not
,
to be passed through; in detail, α (not) + διά (through) + βατ óς (passable). An image of
Rankine’s text containing the first use of the word is shown in Fig. 3.20.
2
Anonymous, 1701, “Scala graduum caloris,” Philosophical Transactions, 270: 824-829; often attributed
to I. Newton.
3
J. A. Ruffner, 1963, “Reinterpretation of the genius of Newton’s ‘law of cooling,’ ” Archive for History
of Exact Sciences, 2(2):138-152.
4
W. J. M. Rankine, 1859, A Manual of the Steam Engine and Other Prime Movers, Griffin, London,
p. 302.

3.3. HEAT 111
Figure 3.20: Image of the first modern use of the word “adiabatic” from Rankine’s 1859 text.
As is work, heat transfer is a path-dependent function characterized by inexact differen-

tials. We take
Z 2
1 Q2 = δQ, (3.162)
1
δQ
Q̇ = , (3.163)
dt
1Q2
1q2 = . (3.164)
m
Here, 1q2 is the specific thermal energy transfer. It has units J/kg. Note 1q2 6= q, where q is
the heat flux with units W/m2 . In this thermodynamics course, we will mainly be concerned
R2
with 1q2 . In a heat transfer course, q is more important. Now, 1W2 = 1 P dV . We will see
R2
in future chapters that there is an equivalent for heat in that 1 Q2 = 1 T dS, where S is the
entropy, to be defined later in Ch. 6.
We finish with some notes of comparison:
• 1Q2 and 1W2 as well as 1q2 and 1w2 are affiliated with transient phenomena; both cross
boundaries when the system changes state.
• 1Q2 and 1W2 as well as 1q2 and 1w2 only exist at system boundaries.
• 1Q2 and 1W2 as well as 1q2 and 1w2 are both path-dependent, have inexact differentials,
and are not properties of the system.
In later chapters, we will occasionally dispense with the subscripts and simply use Q, W , q,
and w to denote the heat transfer and the work. Again Q and W will have units of J; q and
w will have units of J/kg. They will still be path-dependent.

3.4 Representations of the first law

There are a variety of ways to represent the first law of thermodynamics, also known as
the principle of conservation of energy. Some of them are not obvious, but have withstood
the scrutiny of detailed experiment. Perhaps the simplest, but also the most obtuse, is the
following.
3.4.1 Cycle
• First law of thermodynamics: During any cycle, the cyclic integral of heat added
to a system is proportional to the cyclic integral of work done by the system.
H
If we denote a cyclic integral, also known as a closed contour integral, by , the mathematical
representation of this law is
I I
J δQ = δW, (δQ in cal, δW in J). (3.165)
Here J is a proportionality constant, sometimes known as the mechanical equivalent of heat.

Now, during the development of thermodynamics, δQ was measured in calories (cal), where
1 cal represented the energy necessary to raise the temperature of 1 g of water 1 ◦ C, and
δW was measured in J that represented the work done in moving a 1 kg mass against a
gravitational field of strength g = 1 m/s2 through a distance of 1 m.
In the now-discredited caloric theory, heat was thought to be a fluid and not explicitly
related to work. This theory began to lose credibility with the experiments conducted in
Bavaria by the colorful American scientist Benjamin Thompson5 (Count Rumford). By doing
work in boring a cannon immersed in water and boiling the water, Thompson was able to
demonstrate that the work of boring was converted into heat. Thompson’s image is shown
in Fig. 3.21a. Thompson’s etching of the cannon used in his experiment is reproduced in
Fig. 3.21b.
In the 1840s considerable effort was made to relate mechanical and thermal energy and
thus measure J. There is some controversy over who first quantified this value. By many
accounts Julius Robert von Mayer achieved the first success in 1842,6 though his exposi-
tion often lacked the mathematical and experimental support that many scientists demand.
Mayer is pictured in Fig. 3.22. Contemporaneously, and with more publicity, Joule spent
considerable effort in carefully measuring J.7,8 He estimated J = 4.41 J/cal, that has since
5
B. Thompson (Count Rumford), 1798, “An inquiry concerning the source of the heat that is excited by
friction,” Philosophical Transactions of the Royal Society of London, 88: 80-102.
6
J. R. Mayer, 1842, “Bemerkungen über die Kräfte der unbelebten Natur,” Annalen der Chemie und
Pharmacie, 42: 233-240.
7
J. P. Joule, 1845, “On the existence of an equivalent relation between heat and the ordinary forms of
mechanical power,” The London, Edinburgh and Dublin Philosophical Magazine and Journal of Science,
Ser. 3, 27: 205-207.
8
J. P. Joule, 1850, “On the mechanical equivalent of heat,” Philosophical Transactions of the Royal So-
ciety of London, 140: 61-82.

3.4. REPRESENTATIONS OF THE FIRST LAW 113
a) b)
Figure 3.21: a) Benjamin Thompson (Count Rumford) (1753-1814) in a 1783 portrait by

Thomas Gainsborough. American scientist whose cannon-boring experiments discredited
the caloric theory. Image from https://en.wikipedia.org/wiki/Benjamin Thompson, b)
Image of cannon from B. Thompson (Count Rumford), 1798, “An inquiry concerning the
source of the heat which is excited by friction,” Philosophical Transactions of the Royal
Society of London, 88: 80-102.
been corrected to
J
J = 4.184 . (3.166)
cal
Small deviations in the value of J exist in the literature today. The value used here is based on
the so-called thermochemical calorie. We give a portrait of Joule in Fig. 3.23a. A nineteenth
century etching of Joule’s device is given in Fig. 3.23b. A modern full-scale replica of Joule’s
apparatus designed and constructed by Mr. Leon Hluchota and Prof. Patrick F. Dunn, based
upon Joule’s original experimental display in the Science Museum, London, and formerly in
use in undergraduate laboratories at the University of Notre Dame, is shown in Fig. 3.23c.

Figure 3.22: Julius Robert von Mayer (1814-1878). German physician and physi-
cist who in 1842 said “Energy can be neither created nor destroyed.” Image from
https://en.wikipedia.org/wiki/Julius Robert von Mayer.

a) b) c)
Figure 3.23: a) James Prescott Joule (1818-1889). English experimen-

talist who demonstrated the mechanical equivalent of heat. Image from
https://en.wikipedia.org/wiki/James Prescott Joule, b) Sketch of Joule’s origi-
nal apparatus, from Harper’s New Monthly Magazine, No. 231, August 1869, c) Operational
full-scale replication of Joule’s apparatus at the University of Notre Dame.
While Joule performed the key experiments, the critical acceptance of the first law is
attributed by many to the work of Hermann von Helmholtz,9 pictured in Fig. 3.24. However,
Figure 3.24: Hermann Ludwig Ferdinand von Helmholtz (1821-1894). German physi-
cian and physicist who impacted nearly all of nineteenth century mechanics. Image from
https://mathshistory.st-andrews.ac.uk/Biographies/Helmholtz.html.
Truesdell notes that in this work Helmholtz restricts his conservation principle to kinetic and
potential energies.10 The classical theoretical framework for the first law and more was firmly
solidified by Rudolf Clausius.11 Clausius is depicted in Fig. 3.25.
9
H. Helmholtz, 1847, Über die Erhaltung der Kraft, Reimer, Berlin.
10
C. Truesdell, 1980, The Tragicomical History of Thermodynamics 1822-1854, Springer, New York, p. 161.
11
R. Clausius, 1850, “Ueber die bewegende Kraft der Wärme und die Gesetze, welche sich daraus für die
Wärmelehre selbst ableiten lassen,” Annalen der Physik und Chemie 79: 368-397.

Figure 3.25: Rudolf Julius Emmanuel Clausius (1822-1888). German theoreti-

cian who systematized classical thermodynamics into a science. Image from
https://mathshistory.st-andrews.ac.uk/Biographies/Clausius.html.
Now, in this class, we will not bother much with the mechanical equivalent of heat, and
simply insist that δQ be measured in units of work. When δQ has units of J, then J = 1,
and we recover from Eq. (3.165) our preferred form of the first law:
I I
δQ = δW, (δQ in J, δW in J). (3.167)
3.4.2 Process
We arrive at an alternate representation of the first law by the following analysis. Consider
the sketch of Fig. 3.26. Now, consider two cycles, each passing through points 1 and 2, albeit
2
A
1 C
Figure 3.26: Sketch of P − V diagram for various combinations of processes forming cyclic
integrals.

via different paths:
• Cycle I: 1 to 2 on Path A followed by 2 to 1 on Path B,
• Cycle II: 1 to 2 on Path A followed by 2 to 1 on Path C.
The only difference

H between
H Cycles I and II is they take different return paths. Now, write
the first law δQ = δW for Cycle I:
Z 2 Z 1 Z 2 Z 1
δQA + δQB = δWA + δWB , Cycle I. (3.168)
1 2 1 2
For Cycle II, we have similarly

Z 2 Z 1 Z 2 Z 1
δQA + δQC = δWA + δWC , Cycle II. (3.169)
1 2 1 2
Now, subtract Eq. (3.169) from Eq. (3.168) to get

Z 1 Z 1 Z 1 Z 1
δQB − δQC = δWB − δWC . (3.170)
2 2 2 2
Rearrange Eq. (3.170) to get

Z 1 Z 1
(δQ − δW )B = (δQ − δW )C . (3.171)
2 2
Now, B and C are arbitrary paths; Eq. (3.171) asserts that the integral of δQ − δW from 2
to 1 is path-independent. This is in spite of the fact that both δW and, as we will see later,
δQ are path-dependent quantities. Therefore, we can deduce that this defines
• energy: a thermodynamic property that is a theoretical construct suggested by the

first law of thermodynamics as something to account for the difference between heat
transfer and work in any process between the same start and end states.
Energy is a new extensive property of the system denoted by E. While we like to think we
have intuition for what constitutes energy, it really is an elusive quantity. Viewed at another
way, the Newtonian mechanical energy is easily visualized in terms of kinetic and potential
energy, but it is not always conserved! Our new energy includes thermal energy, that we
think we can easily feel, so we still have a good intuition for it. So we have generalized
energy so that it is always conserved, at the expense of losing the ability to easily visualize
it.
Recall that properties depend only on the state and not the path taken to arrive at the
state. Let us then take the following definition for the differential of E:
dE = δQ − δW. (3.172)

If we integrate from 1 to 2, we get

Z 2 Z 2 Z 2
dE = δQ − δW, (3.173)
1 1 1
yielding
E2 − E1 = 1Q2 − 1W2 . (3.174)
Equation (3.174) is the alternate representation of the
• First law of thermodynamics: For a system undergoing a process, the change in
energy is equal to the heat added to the system minus the work done by the system.
As mentioned at the beginning of this section, the first law of thermodynamics is often
referred to as conservation of energy. Literally, conservation implies that there is no change.
In order for energy not to change, and thus be formally conserved, we would have to insist
that either 1Q2 = 1W2 , or more restrictively, that the process be adiabatic and isochoric so
that 1Q2 = 1W2 = 0. In either case E2 = E1 . Thus, the first law, E2 − E1 = 1Q2 − 1W2 , is
actually a statement of evolution of energy in response to heat and work. However, this is
unusual usage, and most people use the first law interchangeably with energy conservation.
Now, we consider E to represent the total energy of the system. It has units of J. It
includes energy that is
• potential,
• kinetic,
• thermal,
• chemical,
• electric,
• magnetic,
• etc.
We will find it useful to lump all of the types of energy that are not potential or kinetic into
a single term U, that we call
• internal energy: that portion of total energy E that is not kinetic or potential
energy. It includes thermal, chemical, electric, magnetic, and other forms of energy.
We take U to have units of J. We call the kinetic energy KE and the potential energy P E.
So we take
E
|{z} = U
|{z} + KE
|{z} + PE
|{z} . (3.175)
total energy internal energy kinetic energy potential energy
In this course we shall mainly be concerned with changes of U that are associated with
changes of the thermal energy of the system. A useful way to think of thermal energy is

• Thermal energy: kinetic energy associated with random motions of molecules at

the microscale.
We can only observe this microscale kinetic energy with great difficulty. We usually have no
hope of having any detailed knowledge of it, and so only consider it in the average. In fact,
the temperature is a measure of the average microscale kinetic energy. We distinguish the
thermal energy from KE, that we take to exist at the observable macroscale.
Each form of energy is an extensive property of the system. Taking differentials of
Eq. (3.175), we get
dE = dU + d(KE) + d(P E). (3.176)
So the first law, dE = δQ − δW , can be written as
dU + d(KE) + d(P E) = δQ − δW. (3.177)
In the next two examples, let us consider two special cases of Eq. (3.177), that are familiar
from Newtonian mechanics.
Example 3.20
Consider a system of mass m for which dU = 0, d(P E) = 0, and δQ = 0. Physically, this
might correspond to a system with constant internal energy, held at constant height, and with no heat
exchanges with its surroundings. For this system, the first law gives a balance between changes in
kinetic energy and work. And let us assume that the system is being accelerated by a horizontal force
F as it acts through a distance in the x direction. Assume there is no friction force.
We give a sketch in Fig. 3.27. The first law tells us
F m x
Figure 3.27: Sketch of a system being accelerated by a force.
dU +d(KE) + d(P E) =
|{z} δQ −δW, (3.178)
| {z } |{z}
=0 =0 =0
d(KE) = −δW. (3.179)
Newtonian mechanics tells us
δW = −F dx. (3.180)
The sign is negative because the system is being worked upon. So we get
d(KE) = − (−F dx) = F dx. (3.181)
| {z }
δW
Newton’s second axiom tells us F = m(d2 x/dt2 ). And because dx/dt = v, we can say F = m(dv/dt),
so
dv
d(KE) = m dx. (3.182)
dt

Note we use v for velocity and distinguish that from v, the specific volume.
Let us divide both sides by dt to get
d(KE) dv dx
= m , (3.183)
dt dt |{z}
dt
=v
dv
= m v, (3.184)
dt
dv
= mv , (3.185)
dt
d v2
= m , (3.186)
dt 2
2
d v
= m , (3.187)
dt 2
1 2 1 2
KE2 − KE1 = mv − mv . (3.188)
2 2 2 1
If we take v1 = 0 and define the KE1 as zero at this point, and take 2 to just be a general state, we
get the commonly used
1
KE = mv2 . (3.189)
2
Note that
d(KE) = mv dv. (3.190)
Example 3.21
Consider a system of mass m for which dU = 0, d(KE) = 0, and δQ = 0. Physically, this might
correspond to a system with constant internal energy, held at constant upward velocity, with no heat
exchanges with its surroundings. For this system, the first law gives a balance between changes in
potential energy and work. And let us assume that the system is being raised at constant velocity v
through distance z against a gravitational force, where the gravitational acceleration g is constant.
We give a sketch in Fig. 3.28. Newton’s second axiom says that m(d2 z/dt2 ) = F − mg. Now, we are
v = constant m z
mg
g
Figure 3.28: Sketch of a system being raised at constant velocity v in a gravitational potential
field, where g is constant.

told the velocity is constant, so the acceleration here is zero, so Newton’s second axiom gives a force
balance, 0 = F − mg, thus
F = mg. (3.191)
Now, the first law tells us, when dU = 0, d(KE) = 0, and δQ = 0, that
dU + d(KE) +d(P E) = δQ −δW, (3.192)
| {z } |{z}
=0 =0
d(P E) = −δW. (3.193)
Now
δW = −F dz. (3.194)
We have a negative sign because the system is having work done on it by F . Thus
d(P E) = − (−F dz) = F dz. (3.195)
| {z }
δW
Because F = mg, we get

d(P E) = mg dz. (3.196)
Integrating, we find
P E2 − P E1 = mg(z2 − z1 ). (3.197)
If we take P E1 to be zero at z1 = 0, and consider a general z, we get the commonly used
P E = mgz. (3.198)
Note that
d(P E) = mg dz. (3.199)
Now, because dE = dU + d(KE) + d(P E) from Eq. (3.175), we can substitute from
Eqs. (3.190, 3.199) to get
dE = dU + mv dv + mg dz. (3.200)
Integrate Eq. (3.200) from state 1 to state 2 to get
1
E2 − E1 = U2 − U1 + m(v22 − v12 ) + mg(z2 − z1 ). (3.201)
2
Now, substitute Eq. (3.201) into Eq. (3.174) to obtain
1
U2 − U1 + m(v22 − v12 ) + mg(z2 − z1 ) = 1Q2 − 1W2 . (3.202)
| 2 {z }
=E2 −E1
Now, if m is a constant, as it will be for a system, we can divide both sides by m to get
1 2
u2 − u1 + v2 − v12 + g(z2 − z1 ) = 1q2 − 1w2 . (3.203)
2
Here, we have employed the intensive variables

• u ≡ U/m, the internal energy per unit mass, also known as the specific internal energy.
It has units kJ/kg. It is an intensive thermodynamic property.
• 1q2 ≡ 1Q2 /m, the heat transfer per unit mass. It has units kJ/kg and is not a thermo-
dynamic property.
• 1w2 ≡ 1W2 /m, the work per unit mass. It has units kJ/kg and is not a thermodynamic
property.
3.5 Specific internal energy for general materials

Just as P , v, and T are thermodynamic properties, so is u. In fact, it can be considered to
be one of the necessary two properties necessary to define a third. So, for example, if we are
given P and u, we could find v = v(P, u) or T = T (P, u).
More importantly, let us consider the most general form for u; a form where u is a function
of at most two independent thermodynamic variables, say T and v,
u = u(T, v). (3.204)
For materials such as water, u(T, v) is tabulated. Note that the tables must presume a
reference value for energy so as to give it an absolute nature. However, as long as we confine
our thermodynamics to a single substance, differences in energy will be the only quantities
that have relevance in determining physical quantities of interest. That is to say, the reference
state will not be important for single material problems. This is not true for multiple material
problems such as when chemical reactions are present.
Similar to vf and vg , the tables have
• uf : the specific internal energy of a saturated liquid, and
• ug : the specific internal energy of a saturated vapor.
Also similar to v, we have
uf g = ug − uf . (3.205)
We also get a similar analysis for quality x as for volume. For a two-phase mixture, the total
energy of the mixture is the sum of the energies of the components:
U = Uliq + Uvap , (3.206)
mu = mliq uf + mvap ug , (3.207)
mliq mvap
u = uf + ug , (3.208)
m m
m − mvap mvap
= uf + ug , (3.209)
m m
= (1 − x)uf + xug , (3.210)
= uf + x(ug − uf ), (3.211)
= uf + xuf g . (3.212)

3.6. SPECIFIC ENTHALPY FOR GENERAL MATERIALS 123
We can solve for x by inverting Eq. (3.212) to get

u − uf
x= . (3.213)
uf g
Let us consider the heat transfer for an isochoric

R 2 process in which we also have ∆KE =
∆P E = 0. Because the process is isochoric, 1W2 = 1 P dV = 0. So the first law, Eq. (3.202),
reduces to
U2 − U1 = 1Q2 , (3.214)
1Q2 = U2 − U1 = ∆U. (3.215)
The change in U gives the heat transfer for isochoric processes.
3.6 Specific enthalpy for general materials

Let us define a new thermodynamic property, enthalpy, in terms of known thermodynamic
properties. The extensive total enthalpy H and intensive specific enthalpy h are defined as
H ≡ U + P V, (3.216)
H U V
h = = +P . (3.217)
m m m
Thus,
h = u + P v. (3.218)
Its most important feature is its utility in control volume analysis that will be fully discussed
in Ch. 4.1.2.2.4; its underlying mathematical rationale will be given in Ch. 9.3.
The first written use of the word “enthalpy” is given by Porter,12 who notes the term was
introduced by the Dutch physicist and Nobel laureate Heike Kamerlingh Onnes (1853-1926).
,
The word is from the Greek ǫνθάλπǫιν, meaning “to warm in.” We give an image of Porter’s
citation of Onnes’ usage in Fig. 3.29.
Eq. (3.218) is valid for general materials. It will be seen to be useful for many problems,
though in principle, we could get by with u alone just as well. Now, because u, P , and v are
thermodynamic properties, so then is h:
h = h(T, P ). (3.219)
Sometimes tables give h and we need to find u; thus,
u = h − P v. (3.220)
12
A. W. Porter, 1922, “The generation and utilisation of cold,” Transactions of the Faraday Society, 18:
139-143.

Figure 3.29: Image of the first known printed use of the word “enthalpy” from Porter, 1922.
Similar to u, we can easily show

h − hf
h = hf + xhf g , x= . (3.221)
hf g
The enthalpy is especially valuable for analyzing isobaric processes. Consider a special
isobaric process in which P1 = P2 = P , ∆KE = ∆P E = 0. Then the first law, Eq. (3.202),
reduces to
U2 − U1 = 1Q2 − 1W2 . (3.222)
R2
Because 1W2 = 1 P dV = P (V2 − V1 ) for the isobaric process, the first law reduces to
U2 − U1 = 1Q2 − P (V2 − V1 ), (3.223)
= 1Q2 − P2 V2 + P1 V1 , (3.224)
1Q2 = (U2 + P2 V2 ) − (U1 + P1 V1 ), (3.225)
| {z } | {z }
=H2 =H1
= H2 − H1 = ∆H. (3.226)
The change in H gives the heat transfer for isobaric processes.
3.7 Specific heat capacity

Sometimes when heat is added to a body, the temperature does not change, but the phase
does, as in a vaporization or melting process. Such heat is called latent heat. Usually
however, when heat is added to a body, we observe a temperature rise. We call this type
of heat sensible heat, because we can sense it. The same amount of heat will raise the
temperature of some materials more than others. When the sensible heat required to raise a
unit mass by one degree is determined, it is called the specific heat. These notions were first
systematized in the 1700s by the Scottish physicist Joseph Black,13 pictured in Fig. 3.30.
13
J. Black, 1803, Lectures on the Elements of Chemistry delivered in the University of Edinburgh by the
Late Joseph Black, M.D., John Robinson, University of Edinburgh.

3.7. SPECIFIC HEAT CAPACITY 125
Figure 3.30: Joseph Black (1728-1799). Scottish physicist and chemist who pioneered
calorimetry, discovered CO2 , and financially supported James Watt’s research on steam
power. Image from https://en.wikipedia.org/wiki/Joseph Black.
We loosely define the
• Specific heat capacity: the amount of heat needed to raise the temperature of a
unit mass of material by one degree.
The word is a relic of the discredited caloric theory of heat in which heat was thought to be
a fluid that could somehow fill its container. We often simply call it the “specific heat.” We
give the specific heat the symbol c. It has units kJ/kg/K. Its loose mathematical definition
is
1 δQ
c≃ . (3.227)
m dT
We also define the extensive heat capacity as C = δQ/dT , where C has units kJ/K. We will
not use C explicitly from here on.
It turns out that because δQ is path-dependent, so is c. So let us specify two common
paths:
• specific heat at constant volume: cv . We determine this on a path that is

isochoric. On such a path, the first law holds that δQ = dU, because δW = 0. So we
take
1 ∂U
cv = . (3.228)
m ∂T v
Because u = U/m, and we take m to be constant, we get
∂u
cv = . (3.229)
∂T v

Now, for general materials u = u(T, v), so we see that

cv = cv (T, v), (3.230)
that is to say, cv (T, v) is itself a thermodynamic property for general materials. It can
vary with two independent variables. We shall see later for some materials it varies
only with T , and for other materials, it is actually a constant.
• specific heat at constant pressure: cP . We determine this on a path that is
isobaric. On such a path, the first law holds that δQ = dH. So we take
1 ∂H
cP = . (3.231)
m ∂T P
Because h = H/m, and we take m to be constant, we get
∂h
cP = . (3.232)
∂T P
Now, for general materials h = h(T, P ), so we see that

cP = cP (T, P ), (3.233)
that is to say, cP (T, P ) is itself a thermodynamic property for general materials. It can
vary with two independent variables. We shall see later for some materials it varies
only with T , and for other materials, it is actually a constant.
We also note that
• The work required for the isobaric expansion results in δQ|P > δQ|v for a unit change
in temperature. From this, it is easy to conclude that
cP > cv . (3.234)
• specific heat for incompressible materials: c. Most liquids and solids under
moderate to low pressure conditions (P < 1 GPa) are well modeled as incompressible.
Thus, in any heating process, there will be little if any associated work of compression.
For such a material, there is no need to distinguish cv and cP , so we simply use c for
the specific heat. We thus take
du
c(T ) = . (3.235)
dT
Often, especially if the temperature changes are small, we can ignore the temperature
variation of c for incompressible materials and simply take
du
c= . (3.236)
dT
More rigorous mathematical discussion of specific heat capacity will be given in Ch. 9.4.

3.8. CALORIC EQUATIONS OF STATE 127
3.8 Caloric equations of state

Recall that thermal equations of state are given by P = P (T, v). We also have equations of
state for the energy. We call such a relation a
• Caloric equation of state: an equation that gives the energy as a function of two
independent state variables. An example is the general form:
u = u(T, v). (3.237)
In Ch. 9.4, we shall see there are restrictions on the form u(T, v) can take. In a complicated
fashion, the form of the caloric equation of state for u(T, v) has some dependency on the
thermal state equation P = P (T, v).
One of the more confusing notions to beginning students of thermodynamics is which
forms of energy and specific heat are appropriate for which materials. Here, we discuss them
in more detail, moving from simple to complex.
3.8.1 Ideal gases

For ideal gases, we have P v = RT . Ideal gases can be either calorically perfect or calorically
imperfect. For all ideal gases, be they calorically perfect or calorically imperfect, it will be
proved in Ch. 9.4 that the caloric equation of state takes on a simpler form:
u = u(T ). (3.238)
Now, we can specify h for an ideal gas. From Eq. (3.218), h = u + P v, and the ideal gas
law, P v = RT , we get
h = u(T ) + RT. (3.239)
Thus, the enthalpy of an ideal gas is a function of T only:
h = h(T ). (3.240)
Now, for the specific heats of an ideal gas, Eq. (3.229) gives
∂u d
cv (T, v) = = (u(T )) = cv (T ). (3.241)
∂T v dT
Separating variables in Eq. (3.241), we can also say for an ideal gas
du = cv (T ) dT. (3.242)
For cP , Eq. (3.232) gives
∂h d
cP (T, P ) = = (h(T )) = cP (T ). (3.243)
∂T P dT

Separating variables in Eq. (3.243), we get then
dh = cP (T ) dT. (3.244)
Now, we can differentiate Eq. (3.239) to get
dh = du + R dT. (3.245)
Now, substitute Eqs. (3.242,3.244) into Eq. (3.245) to get
cP (T ) dT = cv (T ) dT + R dT, (3.246)
cP (T ) = cv (T ) + R. (3.247)
cP (T ) − cv (T ) = R. (3.248)
This is sometimes known as Mayer’s relation. Last, let us define the ratio of specific heats,
k, as
cP
k= . (3.249)
cv
For general materials k = k(T, v). For an ideal gas, we have
cv (T ) + R R
k= =1+ . (3.250)
cv (T ) cv (T )
So k = k(T ) for an ideal gas. We will see that k(T ) is often nearly constant. Because R > 0
and cv (T ) > 0, we must have k > 1 for an ideal gas. This is also a consequence of Eq. (3.234).
Later in Ch. 9.4, it will be rigorously shown for general materials.
3.8.1.1 Calorically perfect

A calorically perfect ideal gas (CPIG) has constant specific heat. Examples of CPIGs include
noble and monatomic gases (e.g. He, Ne, Ar, O, H, N) over a wide range of temperatures
and pressures, and more complex molecules (e.g. O2 , N2 , CO2 , CH4 ) over narrower bands of
temperatures and pressures.
For the CPIG, cv is a constant, so
∂u
= cv . (3.251)
∂T v
But for the ideal gas, u = u(T ), so the partial derivatives become total derivatives and
du
= cv . (3.252)
dT
Integrating, we get the simple caloric equation of state:
u(T ) = uo + cv (T − To ), valid for CPIG. (3.253)

Note that
u = uo + cv T − cv To , (3.254)
u
| +{zP v} = uo + cv T − cv To + |{z}
Pv , (3.255)
=h =RT
h = uo + cv T − cv To + RT, (3.256)
= uo + cv T − cv To + RT + RTo − RTo , (3.257)
= uo + RTo + (cv + R) T − (cv To + RTo ), (3.258)
| {z } | {z } | {z }
=ho =cP =cP To
= ho + cP T − cP To . (3.259)
So
h(T ) = ho + cP (T − To ), valid for CPIG. (3.260)
For a CPIG,
cP cP
k= = = constant. (3.261)
cv cv
Example 3.22
Given Ar at P1 = 140 kPa, T1 = 10 ◦ C, V1 = 200 ℓ that undergoes a polytropic compression to
P2 = 700 kPa, T2 = 180 ◦ C, find 1Q2 .
To calculate 1Q2 , we will need to invoke the first law: U2 − U1 = 1Q2 − 1W2 . Now, for the CPIG, we
will be able to calculate ∆U from knowledge of ∆T , and we can easily compute 1W2 from its definition
R2
as 1 P dV . This will let us calculate the heat transfer.
Ar is a noble gas. Because it is noble, it is well modeled as a CPIG over a wide range of tempera-
tures., as long as it is far from the vapor dome. From Table A.5 in BS, we find the data
kg kJ kJ
M = 39.948 , R = 0.2081 , cv = 0.312 . (3.262)
kmole kg K kg K
Note that
cv 0.312 kgkJK 3
= = 1.49928 ∼ . (3.263)
R 0.2081 kgkJK 2
This result will be seen to be valid for all monatomic gases, such as He, Ne, H, O, etc.
Also note that

kg kJ kJ
M R = 39.948 0.2081 = 8.31318 = R. (3.264)
kmole kg K kmole K
Now, we need to have absolute temperatures. Adding 273.15 to both temperatures, we get
T1 = 283.15 K, T2 = 453.15 K. (3.265)
Also
0.001 m3
V1 = (200 ℓ) = 0.2 m3 . (3.266)
ℓ

Let us compute the mass m. From one incarnation of the ideal gas law, we have
P1 V1 (140 kPa)(0.2 m3 )
m= = = 0.475192 kg. (3.267)
RT1 0.2081 kgkJK (283.15 K)
Now, at state 2, we have

P2 V2 = mRT2 , (3.268)

kJ
mRT2 (0.475192 kg) 0.2081 kg K (453.15 K)
V2 = = = 0.0640155 m3 . (3.269)
P2 700 kPa
Now, for the polytropic process, we have
P1 V1n = P2 V2n , (3.270)
n
P1 V2
= , (3.271)
P2 V1
n
P1 V2
ln = ln , (3.272)
P2 V1

V2
= n ln , (3.273)
V1

P1
ln P 2
n = , (3.274)
V2
ln V1
140 kPa

ln 700 kPa
= 0.0640155 m3
= 1.41279. (3.275)
ln 0.2 m3
We have calculated the work for a polytropic process before. Repeating, we find
Z 2 Z 2 1−n V2
dV V P1 V1n 1−n
1W2 = P dV = P1 V1n = P 1 V1
n
= (V − V11−n ). (3.276)
1 1 V
n 1 − n V1 1−n 2
But because P1 V1n = P2 V2n , the work reduces to simply
P2 V2 − P1 V1 (700 kPa)(0.0640155 m3 ) − (140 kPa)(0.2 m3 )
1W2 = = = −40.7251 kJ. (3.277)
1−n 1 − 1.41279
The work is negative, so the gas was worked upon in compression. Now, the first law tells us
U2 − U1 = 1Q2 − 1W2 , (3.278)
1Q2 = U2 − U1 + 1W2 , (3.279)
= mcv (T2 − T1 ) + 1W2 , (3.280)

kJ
= (0.475192 kg) 0.312 ((453.15 K) − (283.15 K)) + (−40.7251 kJ) , (3.281)
kg K
= −15.5209 kJ. (3.282)
The heat transfer is negative, so heat was lost from the system, even though the temperature went up.
The reason the temperature went up is that the internal energy was raised more by work than it lost
by heat transfer. A plot of this process is given in Fig. 3.31.

P (kPa)
1000
800
2
600
400
200 T2 = 453.15 K
1 T1 = 283.15 K
0
0.10 0.20 0.30
v (m3/kg)
Figure 3.31: Plot of polytropic compression of Ar.
3.8.1.2 Calorically imperfect

For calorically imperfect ideal gases (CIIG), e.g. O2 at moderate to high temperatures
(300 K < T < 6000 K):
• u = u(T ),
• cv = cv (T ),
• h = h(T ),
• cP = cP (T ).
For such temperatures, our assumption of constant cv is not as valid. But for ideal gases,
we can still take cv = cv (T ), so
du
= cv (T ). (3.283)
dT
We can integrate via separation of variables to get
du = cv (T ) dT, (3.284)
Z 2 Z 2
du = cv (T ) dT, (3.285)
1 1
Z 2
u2 − u1 = cv (T ) dT. (3.286)
1
We can interpret the difference in u as the area under the curve in a plot of cv (T ) versus T
as plotted in Fig. 3.32. More generally, we could say

T1 T2 T
Figure 3.32: Relation between u2 − u1 and area under the curve in a plot of cv (T ) for a
calorically imperfect ideal gas.
Z T
u(T ) = uo + cv (T̂ ) dT̂ , valid for all ideal gases. (3.287)
To
Here, T̂ is a dummy variable of integration. Similarly, we could show

Z T2
h2 − h1 = cP (T ) dT, (3.288)
T1
and more generally,

Z T
h(T ) = ho + cP (T̂ ) dT̂ , valid for all ideal gases. (3.289)
To
Now, cv , cP and R all have units of kJ/kg/K. Let us consider the ratio
cv
cv cv M M cv
= = R
= . (3.290)
R RM M R
The ratio is now in terms of molar specific properties with cv and R having units of
kJ/kmole/K. Note that R is the universal gas constant. A plot of cv /R versus T for a
variety of simple molecules is given in Fig. 3.33. We note some remarkable facts:
• For monatomic gases, such as Ar, O, and H, cv /R = 3/2 for a wide variety of temper-
atures.
• For diatomic gases, such as O2 and H2 for T < 600 K, cv /R ∼ 5/2, and for T > 600 K,
cv /R → 7/2.
• For larger molecules such as CO2 or H2 O, cv /R is larger still.

6 CO2
H2O
4 O2
7/2
H2
5/2
2 O
3/2
Ar, H
0
0 2000 4000 6000 T (K)
Figure 3.33: cv /R as a function of T for several molecules.
What we are seeing actually reflects some fundamental physics. We first note that sta-
tistical thermodynamics proves
• Temperature is a measure of the average translational kinetic energy of a set of molecules.
Now, we consider some features of Fig. 3.33.
• Monatomic molecules, such as Ar, O or H have three fundamental modes of kinetic

energy: translation in the x, y, and z directions. Each mode contributes 1/2 to cv /R,
that sums to 3/2.
• For diatomic molecules, we summarize the behavior in the sketch given in Fig. 3.34.
– At low temperatures, diatomic molecules, such as H2 or O2 , act like monatomic

molecules.
– At low temperatures, diatomic molecules begin to rotate, and the rotational en-
ergy becomes an important component. In fact when energy is added to diatomic
molecules, some is partitioned to translation and some is partitioned to rota-
tion. There are two non-trivial axes of rotation, each adding 1/2 to cv /R, giving
cv /R ∼ 5/2.
– At higher temperatures, diatomic molecules begin to vibrate as well, and this
energy becomes an important component. There are two vibrational modes, one

7/2
5/2
3/2 model diatomic

ideal gas
1 3 600 2000 T (K)

k ~1.4 variable k
trans trans+rot trans+rot+vib
Figure 3.34: Sketch of cv /R as a function of T for a model diatomic gas. Sketch adapted
from J. D. Anderson, 1989, Hypersonic and High Temperature Gas Dynamics, McGraw-Hill,
New York, p. 442. Note that a real gas would liquefy in the low temperature region of the
plot! So this model is really for a non-existent gas that has no liquid-inducing intermolecular
forces.
for kinetic energy and one for potential energy. Each adds another 1/2 to cv /R,
giving cv /R ∼ 7/2 at high temperature.
– At higher temperatures still, the diatomic molecules begin to dissociate, e.g. O2 +
O2 → 2O + O2 .
– At even higher temperatures, its electrons are stripped, and it becomes an ionized
plasma. This is important in engineering applications ranging from welding to
atmospheric re-entry vehicles.
• For triatomic molecules such as H2 O or CO2 , there are more modes of motion that can
absorb energy, so the specific heat is higher still.
The dissociation and ionization behavior of the components of air is clearly displayed in
Fig. 3.35. Here we see at low temperatures, T < 1000 K, diatomic N2 and O2 are dominant
in air. Both of these major components begin to dissociate at higher temperatures. For
T > 6000 K, we no longer find diatomic N2 and O2 , but instead find their monatomic
components N and O. At higher temperatures still, the molecule loses electrons, and positive
ions remain.
Feynman14 summarizes the argument that this preference for one type of energy over
14
R. P. Feynman, R. B. Leighton, and M. Sands, 1963, The Feynman Lectures on Physics, Volume 1,
Addison-Wesley, Reading, Massachusetts, pp. 40-7–40-10.

Figure 3.35: Equilibrium composition of air at low density and various temperatures. Figure
from W. E. Moeckel and K. C. Weston, 1958, Composition and thermodynamic properties
of air in chemical equilibrium, NACA Technical Note 4265.
another (translation, rotational, vibrational) depending on temperature is surprising to those

not versed in quantum mechanics and violates standard assumptions of classical statistical
mechanics. In fact, he notes that Maxwell had a hint of the problem as early as 1859, and
stated this concern more directly in 1869. Maxwell summarized those concerns in an 1875
lecture, transcribed in Nature.15 Feynman argues that the reason for the energy partition
observed in diatomic gases is a “failure of classical physics” and instead is a pure effect
of quantum mechanics; that is to say k = cP (T )/cv (T ) = k(T ) is a non-classical result!
Employment of the theories of quantum and statistical mechanics allows an accounting for
the observation that there is a preference of molecules to exist in lower energy states, and at
those states, the discrete quantization is important. High energy vibrational states are less
likely than translational states at low temperature. At higher temperature, there is a higher
probability that all states are populated, and one recovers results consistent with classical
physics.
Let us also recall that cP (T ) − cv (T ) = R; thus, cP (T ) − cv (T ) = R. Let us summarize
15
J. C. Maxwell, 1875, “On the dynamical evidence of the molecular constitution of bodies,” Nature,
11(280): 374-377.

• for monatomic gases,

3
cv = R, (3.291)
2
5
cP = cv + R = R, (3.292)
2
5
cP R 5
= k = 23 = = 1.6667. (3.293)
cv 2
R 3
• for diatomic gases at moderate temperature, 50 K < T < 600 K,

5
cv = R, (3.294)
2
7
cP = cv + R = R, (3.295)
2
7
cP R 7
= k = 25 = = 1.4. (3.296)
cv 2
R 5
To summarize, usually the most problematic case is whether or not specific heats vary
with temperature in ideal gases. For low temperatures, the specific heat is well modeled as a
constant; here the internal energy change is strictly proportional to the temperature change.
For moderate to high temperatures, a temperature-variation of the specific heat is observed.
Changes in internal energy are no longer strictly proportional to changes in temperature.
The behavior is analogous to solid mechanics. At low strain ǫ, stress σ is proportional to
strain, and the constant of proportionality is the modulus of elasticity E. For high strains,
the linearity is lost; we could say the elastic modulus becomes a function of strain. We give
a sketch in Fig. 3.36 of the comparison to solid mechanics
u
calorically
imperfect
E region higher
lower
plastic
region
elastic calorically
region perfect
E region
1 1
Figure 3.36: Sketch of comparison of stress-strain behavior in solids with ideal gas internal
energy-temperature behavior.
There are four main ways to calculate changes in enthalpy for ideal gases:

• assumption of constant cP evaluated at 298 K,
• assumption of constant cP evaluated at an intermediate temperature,

R2
• using a known analytic form of cP (T ) in the direct integration of 1 cP (T ) dT , or
• estimation using the ideal gas tables.
Example 3.23
Calculate the heat transferred per unit mass to N2 in an isobaric process that starts at T1 = 300 K
and finishes at T2 = 1000 K. Use the four different means of calculating enthalpy changes to estimate
the heat transfer.
The first law on a per mass basis gives
u2 − u1 = 1q2 − 1w2 . (3.297)

R2
For an isobaric process, 1w2 = 1 P dv = P (v2 − v1 ) = P2 v2 − P1 v1 . So
Z 2
1q2 = u2 − u1 + P2 v2 − P1 v1 = h2 − h1 = ∆h = cP (T ) dT. (3.298)
1
• Constant cP at 298 K. From Table A.5 in BS, we find cP = 1.042 kJ/kg/K. Thus, we estimate

kJ kJ
∆h = h2 − h1 = cP (T2 − T1 ) = 1.042 ((1000 K) − (300 K)) = 729.4 . (3.299)
kg K kg
• Constant cP at the average temperature. The average temperature is

T1 + T2 300 K + 1000 K
Tave = = = 650 K. (3.300)
2 2
Table A.6 in BS has polynomial curve fits for cP . For N2 , we find from Table A.6 that
T
cP = C0 + C1 θ + C2 θ2 + C3 θ3 , θ≡ . (3.301)
1000 K
The numbers Co , . . . , C3 actually have units and are designed to yield a value for cP in kJ/kg/K. We
get
2
kJ kJ T kJ T
cP = 1.11 − 0.48 + 0.96
kg K kg K 1000 K kg K 1000 K
3
kJ T
− 0.42 , (3.302)
kg K 1000 K
2
kJ kJ 650 K kJ 650 K
= 1.11 − 0.48 + 0.96
kg K kg K 1000 K kg K 1000 K
3
kJ 650 K
− 0.42 , (3.303)
kg K 1000 K
kJ
= 1.08826 . (3.304)
kg K

So the improved prediction of the enthalpy change is

kJ kJ
∆h = cP (T2 − T1 ) = 1.08826 ((1000 K) − (300 K)) = 761.78 . (3.305)
kg K kg
• Direct integration. We know

Z T2
∆h = h2 − h1 = cP (T ) dT. (3.306)
T1
Let us change variables from T to θ. We have θ = T /(1000 K), so dθ = dT /(1000 K), and dT =
(1000 K) dθ. Thus,
Z θ2

∆h = C0 + C1 θ + C2 θ2 + C3 θ3 ((1000 K) dθ), (3.307)
θ1
θ2
θ2 θ3 θ4
= (1000 K) C0 θ + C1 + C2 + C3 , (3.308)
2 3 4 θ1
2
kJ kJ θ
= (1000 K) 1.11 θ − 0.48
kg K kg K 2
3 4 1000 K
kJ θ kJ θ 1000 K
+ 0.96 − 0.42 . (3.309)
kg K 3 kg K 4 300 K
1000 K
The final value is

kJ
∆h = 765.81 . (3.310)
kg
• Use of ideal gas tables. Lastly, we can use the ideal gas tables. For us, Table A.8 of BS is best. We
find h(300 K) = 311.67 kJ/kg and h(1000 K) = 1075.91 kJ/kg. So

kJ kJ kJ
∆h = h2 − h1 = 1075.91 − 311.67 = 764.24 . (3.311)
kg kg kg
Example 3.24
A calorically imperfect ideal gas is known to have
cP (T ) = cP o + aT. (3.312)
A sample of this gas begins at P1 , T1 . It is heated isobarically to T2 , and expanded isochorically to T3 .

Find the change in internal energy of the gas, u3 − u1 .
Because internal energy changes are path-independent, we need not worry about any details of
the process, including its path. That is because we know the end states. So we can use a version of
Eq. (3.286) to get
Z T3
u3 − u1 = cv (T ) dT. (3.313)
T1

We are given cP , but need cv . We can use Mayer’s relation, Eq. (3.248) to get
cv (T ) = cP (T ) − R = (cP o − R) + aT. (3.314)
So
Z T3
u3 − u1 = ((cP o − R) + aT ) dT. (3.315)
T1 | {z }
cv (T )
a
u3 − u1 = (cP o − R)(T3 − T1 ) + (T32 − T12 ). (3.316)
2
Had we been asked to find the heat transfer, 1Q3 , we would have had to calculate the work 1W3 , and
the details of the path would have been required. Because internal energy is a state function, only the
end states are required to evaluate the change in internal energy.
3.8.2 Liquids and solids

Most liquids and solids for moderate pressures or below, P < 1 GPa, are well modeled as
incompressible. For such materials, we have caloric equations of state and specific heats of
the form
• u = u(T ),
• c = c(T ).
For such a material Z T
u(T ) = uo + c(T̂ ) dT̂ , (3.317)
To
and
du
= c(T ). (3.318)
dT
Often, we can take a calorically perfect model in which c loses its temperature variation, and
get the commonly used equations
u(T ) = uo + c(T − To ), (3.319)

du
= c. (3.320)
dT
Example 3.25
A mass of cast iron of m = 1 kg is heated from T1 = 300 K to T2 = 400 K. Determine the thermal
energy required.

The first law is

U2 − U1 = 1Q2 − 1W2 . (3.321)
Incompressible solids have effectively zero work, so
U2 − U1 = 1Q2 . (3.322)
On a per mass basis, we have

1Q2 = m(u2 − u1 ). (3.323)
Now, invoke the caloric equation of state for the solid, u2 − u1 = c(T2 − T1 ), to get
1Q2 = mc(T2 − T1 ). (3.324)
For cast iron, Table A.3 of BS tells us c = 0.42 kJ/kg/K, so

kJ
Q
1 2 = (1 kg) 0.42 ((400 K) − (300 K)) = 42 kJ. (3.325)
kg K
Example 3.26
A tall thermally insulated shaft contains 100 kg of water at its base at 20 ◦ C. At the top of the
shaft, at z = 100 m, is a 5 kg mass of aluminum, also at 20 ◦ C. The mass is dropped to the base, comes
to rest, and comes to equilibrium with the water. Local g = 9.81 m/s2 . Find the final temperature. A
sketch of this process is given in Fig. 3.37.
z = 100 m
g = 9.81 m/s2
z=0m
Figure 3.37: Sketch of potential to thermal energy conversion problem.

We can consider the masses of the water and the aluminum to constitute our combined system.
The combined system is isolated; that is there is no heat or work crossing the system boundary. We
cannot neglect mechanical energy in this problem. So the first law is
E2 − E1 = 1Q2 − 1W2 , (3.326)

| {z }
=0
E2 − E1 = 0, (3.327)
   
U2 + KE2 +P E2  − U1 + KE1 +P E1  = 0. (3.328)
| {z } | {z }
=0 =0
Now, at the beginning and end of the process neither water nor aluminum is in motion; hence, KE1 =
KE2 = 0. So
(U2 − U1 ) + (P E2 −P E1 ) = 0. (3.329)
|{z}
=0
Now, at state 2, neither mass retains any potential energy, so
U2 − U1 = P E1 . (3.330)
Now, U represents an extensive energy of the combined system. And we have two components. The
total is the sum of the parts, that leads to
mwater cwater (T2 − T1 ) + mAl cAl (T2 − T1 ) = mAl gH. (3.331)
Solve for T2 and get

mAl gH
T2 = T1 + . (3.332)
mwater cwater + mAl cAl
Take data for c from Tables A.3,4 of BS and get

◦
kJ
(5 kg) 9.81 sm2 (100 m) 1000 J
T2 = (20 C) + = 20.0116 ◦ C. (3.333)
kJ kJ
(100 kg) 4.18 kg K + (5 kg) 0.90 kg K
Note because we are dealing with temperature differences, K plays the same role as ◦ C, and there is no
need to transform.
3.8.3 General materials

Examples of general materials include water and van der Waals gases. For such materials
we have caloric equations of state and specific heats
• u = u(T, v),
• cv = cv (T, v),
• h = h(T, P ),

u (kJ/kg)
v (m3/kg)
v (m3/kg)
/k
T (K) T (K)
Figure 3.38: Plot of u = u(T, v), cv = cv (T, v) using a Redlich-Kwong caloric state equation
model for superheated water.
• cP = cP (T, P ).
A plot of an estimate for u(T, v) and cv (T, v) for superheated water is given in Fig. 3.38.
Here, we selected a Redlich-Kwong model obeying

kJ kJ
u(T, v) = 2936.35 + 1.41 (T − (673.15 K))
kg kg K

kJ K1/2 v
56186.9 kg ln 3
v+0.00117008 mkg
+ √ , (3.334)
T
kJ v
28093.4 kg K1/2 ln 3
kJ v+0.00117008 m
kg
cv (T, v) = 1.41 + . (3.335)
kg K T 3/2
This particular model ignores some of the potential temperature variation of cv , but is useful
for illustration purposes. It gives results not unlike those in portions of the steam tables. As
an aside, the Redlich-Kwong thermal equation of state, see p. 63, for superheated water is

0.461504 kgkJK T kJ K1/2 m3
43.8287 kg kg
P = − √ . (3.336)
m3 m3
v − 0.00117008 kg v v + 0.00117008 kg T
Even more generally, u and h are tabulated for materials such as water.
Example 3.27
Consider water in a fixed total volume of V = 5 m3 . Initially there is Vf = 0.05 m3 of liquid water
and Vg = 4.95 m3 of water vapor. The initial pressure is P1 = 100 kPa. Heat is added isochorically
until the saturated vapor state is reached. Determine 1Q2 .

A sketch of this process is given in Fig. 3.39. The first law tells us
state 1 state 2
P1 = 100 kPa
water vapor water vapor
V = 4.95 m3 V = 5 m3
x2 = 1
liquid water
V = 0.05 m3
Figure 3.39: Sketch of an isochoric water heating problem.
U2 − U1 = 1Q2 − 1W2 . (3.337)

|{z}
=0
Because the process is isochoric, the work is zero. So
1Q2 = U2 − U1 , (3.338)
= m(u2 − u1 ). (3.339)
So, we need to find u1 and u2 from caloric equations of state to solve this problem. Here, however, our
equations reside in tabular form. Let us see what we know. At state 1, we have P1 = 100 kPa. We
need another property to fix state 1. We can get the quality, x1 , by the following analysis. First, we
know that state 1 is a two-phase mixture. The saturated water tables tell us
m3 m3
T1 = 99.62 ◦ C, vf = 0.001043 , vg = 1.69400 . (3.340)
kg kg
So we get at state 1
Vf 0.05 m3 Vg 4.95 m3
mf = = 3 = 47.9386 kg, mg = = 3 = 2.92208 kg. (3.341)
vf 0.001043 m
kg
vg 1.694 mkg
So the total mass is
m = mf + mg = (47.9386 kg) + (2.92208 kg) = 50.8607 kg. (3.342)
So the quality at state 1 is

mg 2.92208 kg
x1 = = = 0.0574526. (3.343)
m 50.8607 kg
So we get
V 5 m3 m3
v1 = = = 0.0983077 . (3.344)
m 50.8607 kg kg

Now, for the internal energy at state 1, we can say

u1 = uf + x1 uf g . (3.345)
We find uf and ug from the saturated water tables, Table B.1.2 of BS, giving us

kJ kJ kJ
u1 = 417.33 + (0.0574526) 2088.72 = 537.332 . (3.346)
kg kg kg
At state 2, we know that v2 = v1 = 0.0983077 m3 /kg. We also know that x2 = 1, because we were
told in the problem statement the final state was saturated vapor. We interpolate Table B.1.2 of BS
between P = 2000 kPa and P = 2250 kPa to find
 

kJ
2601.98 kJ
kg − 2600.26 kg
kJ
u2 = 2600.26 +  
kg 0.08875 m
3
− 0.09963 m3
kg kg

m3 m3
× 0.0983077 − 0.09963 , (3.347)
kg kg
kJ
= 2600.47 . (3.348)
kg
So the heat transfer is

kJ kJ
1Q2 = m(u2 − u1 ) = (50.8607 kg) 2600.47 − 537.332 = 104993 kJ. (3.349)
kg kg
One could also interpolate for the final pressure and temperature and find them to be P2 = 2030.38 kPa,
and T2 = 213.153 ◦ C. Sketches of this process are given in Fig. 3.40.
T P P
2
2 2
1 1
v v T
Figure 3.40: Sketches of the process of isochoric water heating problem in various planes.
3.9 Time-dependency
We venture gently away from classical thermodynamics into non-equilibrium thermodynam-
ics. Let us admit time t into the differential form of the first law by scaling Eq. (3.177) by
dt:
dU d d δQ δW
+ (KE) + (P E) = − . (3.350)
dt dt dt dt dt

3.9. TIME-DEPENDENCY 145
Because total energy E = U + KE + P E, we could also say

dE δQ δW
= − . (3.351)
dt dt dt
Often, we will use the Newtonian “dot” notation for time derivatives. A common form is
dE
= Q̇ − Ẇ . (3.352)
dt
In this course, we will often neglect changes in KE and P E, in which case the time-dependent
first law reduces to
dU
= Q̇ − Ẇ . (3.353)
dt
Many times in this course, we will treat Q̇ and Ẇ as constants. Problems become more
interesting when they are variable. Such problems are also important. Now, Newton did
consider some thermal problems. In fact Newton’s law of cooling (see Eq. (3.159)) tells us
that Q̇ is proportional to the surface area of a body and the temperature difference between
the body and its environment:
Q̇ = −hA(T − T∞ ). (3.354)
Note that heat flows into a body when it has a temperature less than its surroundings,
T < T∞ . Here, we have the
• Heat transfer coefficient, h, with units W/m2 /K.
Note h 6= h. Enthalpy is a different physical quantity with different units. While Eq. (3.354)
is given the elevated name of “law,” one must realize that it is by no means a law of the same
status as the first law of thermodynamics. Rather, it is actually only a useful but fallible
approximation; often h is not a constant but rather a complicated function of the local
material’s state and geometrical configuration. Its nuances are the subject of the discipline
of convective heat transfer.
Example 3.28
Use Newton’s law of cooling along with the first law of thermodynamics and a caloric state equation
to estimate the time necessary to bake a potato. See Fig. 3.41 showing a potato in a modern convection
oven.
Let us describe the potato as a sphere with surface area A, volume V , temperature T , density ρ,
initial temperature To , immersed in an oven at constant T∞ . The first law tells us that
dU
= Q̇ − Ẇ . (3.355)
dt
We can take Ẇ = 0 because of incompressibility, so
dU
= Q̇. (3.356)
dt

Figure 3.41: Unbaked potato in a convection oven, t = 0 min.
Now, the caloric state equation is U = Uo + mc(T − To ). Take c constant, so dU/dt = mc(dT /dt); thus,
the first law becomes
dT
mc = Q̇. (3.357)
dt
Now, invoke Newton’s law of cooling to get
dT
mc = −hA(T − T∞ ). (3.358)
dt
Now, m = ρV , so
dT
ρV c = −hA(T − T∞ ). (3.359)
dt
So
dT hA
=− (T − T∞ ). (3.360)
dt ρcV
Now, this is a first order ordinary differential equation for T (t). We can separate variables to get
dT hA
= − dt, (3.361)
T − T∞ ρcV
Z Z
dT hA
= − dt, (3.362)
T − T∞ ρcV
hA
ln(T − T∞ ) = − t + C, (3.363)
ρcV

′ hA
T − T∞ = C exp − t . (3.364)
ρcV
Here, C ′ = exp(C) and is also a constant. Now, when t = 0, we have T = To , so

′ hA
To − T∞ = C exp − 0 , (3.365)
ρcV
| {z }
=1
= C′. (3.366)

Thus
hA
T (t) = T∞ + (To − T∞ ) exp − t . (3.367)
ρcV
Notice that T (0) = To and T (∞) = T∞ . And notice that when the argument of the exponential is
−1, giving exp(−1) = 0.357, we get a good estimate of the time it takes to get most of the way to
equilibrium. We define this as the time constant, τ of the system. Here, we can get τ via
hA
− τ = −1, (3.368)
ρcV
ρcV
τ = . (3.369)
hA
So we get fast cooking (small τ ) if
• ρ is small (a light potato),
• c is small,
• h is large (the heat transfer rate is fast), or
• A/V is large (the surface to volume ratio is large).
For our potato, let us model it as a sphere of liquid water with R = 0.05 m, ρ = 997 kg/m3 , c =
4.18 kJ/kg/K. Let us take h = 0.012 kW/m2 /K. Let us take To = 20 ◦ C, T∞ = 200 ◦ C. For a sphere,
we have the surface to volume ratio of
A 4πR2 3
= 3
= . (3.370)
V (4/3)πR R
With this the time constant τ is

ρcR
τ= . (3.371)
3h
Small radius leads to fast cooking. Our temperature varies with time as

3h
T (t) = T∞ + (To − T∞ ) exp − t , (3.372)
ρcR
 
kW
3 0.012 2
m K
= (200 ◦ C) + ((20 ◦ C) − (200 ◦ C)) exp − t , (3.373)
kg kJ
997 m 3 4.18 kg K (0.05 m)

−t
= (200 ◦ C) − (180 ◦ C) exp . (3.374)
5788.14 s
The time constant τ = 5788.14 s = 96.469 min. When T = 100 ◦ C, the potato is probably cooked
enough. This occurs at
min
t = (3402.19 s) = 56.7032 min. (3.375)
60 s
If we leave the potato in the oven too long, it will get too hot, and all its water will boil away. Note that
our simple analysis does not account for latent heat and mass loss of vaporization. The temperature
history is plotted in Fig. 3.42.
This example used what is known as the lumped capacitance method for analysis of T (t). In actuality,
T is a function of space and time: T = T (x, y, z, t). It will be shown in later courses that the lumped
capacitance method is valid when the so-called Biot number, Bi, is much less than unity. Biot himself
was an important physicist who played a small role in thermal sciences. He is depicted in Fig. 3.43.
The Biot number for our potato problem is defined as

o
T ( C)
200
150
100
cooking time
50 t = 3402.19 s = 56.7 min
0 10000 20000 30000 t (s)
Figure 3.42: Temperature versus time for the potato in the oven.
Figure 3.43: Jean-Baptiste Biot (1774-1862). French physicist, astronomer, and mathemati-
cian. Image from https://en.wikipedia.org/wiki/Jean Baptiste Biot.
hR
Bi ≡ , (3.376)
k
where k is the thermal conductivity. For the lumped capacitance to be valid, we need
hR
Bi = ≪ 1, thus k ≫ hR. (3.377)
k
For liquid water, k ∼ 0.006 kW/m/K, so for this problem

0.012 mkW2 K (0.05 m)
Bi = = 0.1. (3.378)
0.006 mkWK
This is a small enough Biot number that our lumped capacitance method is acceptable. Physically, if
the thermal conductivity is high relative to the product of h and R, thermal energy diffuses rapidly in
the solid, and internal temperature gradients are small.
The predictions of this theory were tested in a domestic laboratory. The specimen of Fig. 3.41 was
subjected to the culinary regimen suggested by the analysis: 57 min at 200 ◦ C in a convection oven.

a) b)
Figure 3.44: a) Baked potato following an oven-based heat transfer process, t = 57 min;
b) the same potato subjected to a routine post-heating, pre-human-ingestion mashing and
seasoning procedure.
The results can be seen in Fig. 3.44. Qualitative testing by a human subject following treatment with
NaCl, ground black pepper, and bovine butter revealed the potato induced a high degree of gastronomic
satisfaction.
Let us consider an excursion into thermochemistry in which exothermic reaction plays a

role. This will provide a mathematical model that can be shown to model to growth and
decay of temperature. Such models are analogs to models of growth and decay of disease
such as might be present in a pandemic.
Consider an exothermic chemical reaction in which generic species A reacts irreversibly
to form generic species B, converting chemical to thermal energy in so doing.
A → B. (3.379)
The species A and B are taken to have identical specific heats and identical molecular
masses. We characterize the degree of reaction by the so-called reaction progress variable, λ.
We restrict λ such that
λ ∈ [0, 1], (3.380)
with λ = 0, corresponding to all of A being present, and λ = 1 corresponding to all of B

being present. The reaction dynamics can be described by the so-called law of mass action
with a temperature-dependent Arrhenius kinetic rate, so that

dλ −E
= a(1 − λ) exp , λ(0) = 0. (3.381)
dt RT
Here a > 0 is a reaction constant with units s−1 , and E > 0 is the activation energy, with
units kJ/kg. Note

• When T is small, dλ/dt is small.
• When the reaction is complete, λ = 1, we get dλ/dt = 0.
Now the caloric equation of state for such a mixture must be adjusted to account for exother-
mic heat release. It can be shown that a rational model for this, in the limit where A and
B have the same specific heats c, is
u(T, λ) = c(T − To ) − λq . (3.382)

| {z } | {z } |{z}
internal energy thermal energy chemical energy
We have partitioned the internal energy into thermal energy and chemical energy. Here
q > 0 is the exothermic chemical energy release per unit mass with units kJ/kg. If the
system were adiabatic with no work, u must be constant, u = uo , and for such a system, we
would get
uo = c(T − To ) − λq. (3.383)
Because we have at the initial state that T = To and λ = 0, we also have uo = 0, thus giving
λq
T (λ) = To + . (3.384)
c
So at the initial state, before reaction commences, we would have
T (λ = 0) = To . (3.385)
At the complete reaction state, we would have
q
T (λ = 1) = To + . (3.386)
c
Because q > 0 and c > 0, the temperature at complete reaction increases due to the exother-
mic nature of the reaction.
Now let us consider the first law for an incompressible system with reaction and convective
heat loss to the surroundings. One might consider this a “burning potato” problem, relative
to our previous example of a “baking potato.” Start with our previous expression of the first

law, Eq. (3.353) and analyze using our new caloric state equation that accounts for reaction:
dU
= Q̇ − |{z}
Ẇ , (3.387)
dt
0
d
(mu) = Q̇, (3.388)
 dt 
d 
m (c(T − To ) − λq) = −hA(T − T∞ ), (3.389)
dt | {z } | {z }
u Q̇
d q hA
T − To − λ = − (T − T∞ ), (3.390)
dt c mc
dT q dλ hA
− = − (T − T∞ ), (3.391)
dt c dt mc
dT hA q dλ
= − (T − T∞ ) + . (3.392)
dt mc c dt
Now use Eq. (3.381) to eliminate dλ/dt to get

dT hA q −E
=− (T − T∞ ) + a(1 − λ) exp . (3.393)
dt |mc {z } c RT
| {z }
effect from heat transfer increase from exothermic reaction
As before, we take our mass to be a sphere with radius R and density ρ, giving
hA 3h
= . (3.394)
mc ρcR
Now both T and λ vary with time t, and the evolution equations are both coupled and
non-linear. The mathematical problem is

dλ −E
= a(1 − λ) exp , λ(0) = 0, (3.395)
dt RT

dT 3h q −E
= − (T − T∞ ) + a(1 − λ) exp , T (0) = To . (3.396)
dt ρcR c RT
There is no exact solution to this coupled system of non-linear ordinary differential equations.
In the non-reactive limit, when a = 0, we recover our baked potato problem, and there is
an exact solution. Let us consider an example identical to our baked potato problem, but
allow for the potato to react exothermically. The parameters we use for reaction are not
tightly linked to any experiment. While they will induce the correct qualitative behavior,
the predictions will not match quantitative experimental results.
Example 3.29
Consider a “burning potato” identical to the “baking potato” of the previous example, but with
combustion thermo-kinetic parameters E = 1400 kJ/kg, R = 0.287 kJ/kg/K, q = 2000 kJ/kg. Explore

results for three rates, fast reaction with a = 10 s−1 , moderate reaction with a = 1 s−1 , and slow
reaction with a = 0.1 s−1 . Explore analogs with “social distancing” in epidemiology.
Let us first examine the highest possible temperature our system could reach. Let us say it got
to T∞ by convective heat transfer, then burned suddenly so that the temperature rose by q/c. So the
maximum possible temperature is
q
Tmax = T∞ + , (3.397)
c
kJ
2000 kg
= (200 ◦ C) + , (3.398)
4.18 kgkJK
= 678.469 ◦ C. (3.399)
We convert all temperatures to an absolute scale in K, as that is what is required by the Arrhenius
models of chemical kinetics. Our system is thus
 
kJ
dλ −1400 kg
= a(1 − λ) exp  , λ(0) = 0, (3.400)
dt kJ
0.287 kg K T

dT 3 0.012 mkW2 K
= − (T − 473 K),
dt kg
997 m 3 4.18 kgkJK (0.05 m)
 
2000 kJ
kg −1400 kJkg
+ a(1 − λ) exp   , T (0) = 293 K. (3.401)
4.18 kgkJK 0.287 kgkJK T
These two coupled non-linear ordinary differential equations are solved numerically for both values of
kinetic rate constant a. Any standard numerical method can handle this, such as the forward Euler
method. Here, the routine NDSOLVE was used within the Mathematica software tool.
We first consider fast reaction with a = 10 s−1 . Plots of T and λ are given in Fig. 3.45. For
T (oC)
700
1.0
Tmax
600
0.8
500 burning potato
400 0.6
300
0.4
200
baking potato 0.2
100
0 10 000 20 000 30 000 10 000 20 000 30 000

t (s) t (s)
Figure 3.45: Temperature and reaction progress variation with time for a potato model in
the presence of heat transfer and reaction with fast reaction kinetics, a = 10 s−1 .
comparison, the temperature profile of the “baking potato” with q = 0 is given as well. We note the
following

• The reaction initially is slow, and λ ∼ 0.

• At a time near t = 4000 s, the reaction accelerates, and λ begins a rapid transition to λ ∼ 1.
• At early time, the temperature of the burning potato is nearly the same as the baking potato. That
is because heat transfer dominates reaction at early time.
• When the reaction accelerates, the temperature of the burning potato rises sharply relative to the
baking potato.
• Convective heat transfer prevents the maximum temperature of Tmax = 678.469 ◦ C from being real-
ized.
• After the reaction has went to near completion, there is no more chemical energy release, and the
burnt potato comes to thermal equilibrium with its surroundings, similar to the baked potato.
We next consider moderate reaction with a = 1 s−1 . Plots of T and λ are given in Fig. 3.46. For
T (oC)
700
1.0
Tmax
600
0.8
500
burning potato
400 0.6
300
0.4
200
baking potato 0.2
100
0 10 000 20 000 30 000 10 000 20 000 30 000

t (s) t (s)
the presence of heat transfer and reaction with moderate reaction kinetics, a = 1 s−1 .
this moderate rate, the temperature profile of the “baking potato” with q = 0 is given as well. We note
the following, relative to the fast reaction,
• The reaction remains initially slow, and λ ∼ 0.
• At a later time, near t = 15000 s, the reaction accelerates, and λ begins a less rapid transition to
λ ∼ 1.
• At early time, the temperature of the burning potato is nearly the same as the baking potato. That
is because heat transfer dominates reaction at early time.
• When the reaction accelerates, the temperature of the burning potato rises, but not as sharply relative
to the baking potato.
• Relative to the fast reaction, the peak temperature is lower, and the time at which the peak is reached
is delayed.
• After the reaction has went to near completion, there is no more chemical energy release, and the
burnt potato comes to thermal equilibrium with its surroundings, similar to the baked potato.
We last consider slow reaction with a = 0.1 s−1 . Plots of T and λ are given in Fig. 3.47. We note
the following, relative to the two faster reactions,

T (oC)
700 1.0
600 Tmax
0.8
500
400 0.6
300 0.4
baking and burning potato
200
0.2
100
0 200 000 400 000 600 000 800 000 1 × 106 200 000 400 000 600 000 800 000 1 × 106
t (s) t (s)
the presence of heat transfer and reaction with slow reaction kinetics, a = 0.1 s−1 .
• The reaction is much slower, reaching completion near t = 106 s.

• The burning and baking potatoes have nearly identical temperature profiles. This is because energy
released by slow reaction is soon convected away by the faster convection process.
• There is a low peak temperature, relative to the two faster reactions.
While the analogy is not perfect, there is some relevance to epidemiology. One might want to think
of T as a measure of the severity of a pandemic, and a as a so-called “social distancing” parameter.
Large a corresponds to a tightly packed social network, where rapid transmission can occur. Small a
corresponds to a network with more “social distance’’ between entities. It is possible via reduction of
a to effectively mitigate entirely the peak temperature.

3.10. FINAL COMMENTS ON CONSERVATION 155
3.10 Final comments on conservation

We note the first law is often re-stated as energy is conserved. Let us reconcile this with our
mathematical statement, dE/dt = Q̇ − Ẇ . This equation tells us that the total energy E can
change in response to heat and work crossing the system boundary. But conservation implies
that a quantity does not change. We can recover the proper notion of conservation if we
speak of an isolated system, that we recall is one that is not influenced by its surroundings.
So for an isolated system there can, by definition be no work or heat transfer, so
dE
= 0, isolated system. (3.402)
dt
E = constant, isolated system. (3.403)
Even more fundamentally, we can say, by its definition, that the mass m of a system is
constant; thus, the mass of an isolated system must also be constant.
This theoretical formulation is often successful in describing nature for a wide variety
of scenarios. However, it does not always succeed. It has been observed to fail for systems
that move at a velocity close to the speed of light. For such systems, not only do energy
and mass conservation principles fail, so do Newton’s laws of mechanics. To realign theory
with observation, it was necessary for Einstein, depicted in Fig. 3.48, to re-formulate a new,
Figure 3.48: Albert Einstein (1879-1955), German theoretical physicist who de-
veloped theories that explained data better than those of Newton. Image from
https://mathshistory.st-andrews.ac.uk/Biographies/Einstein.html.
modified set of axioms. In a certain sense the new axioms are simple. For example one
can replace Newton’s second law with the seemingly simpler dv/dt = 0, where v is a new

velocity vector; however, the coordinate axes associated with this vector are complicated.
Another consequence of Einstein’s reformulation was the remarkable results of mass-energy
equivalence via the famous relation
E = mc2 , (3.404)
where c is the speed of light in a vacuum. Another way of viewing Einstein’s contributions
is via a new conservation property: the mass-energy of an isolated system is constant. It is
the conservation of mass-energy that is the key ingredient in both nuclear weapon systems
as well as nuclear power generation.

Chapter 4
First law analysis for a control volume
Read BS, Chapter 4

Problems in previous chapters have focused on systems. These systems always were com-
posed of the same matter. However, for a wide variety of engineering devices, for example
• flow in pipes,
• jet engines,
• heat exchangers,
• gas turbines,
• pumps,
• furnaces, or
• air conditioners,
a constant flow of new fluid continuously enters and exits the device. In fact, once the fluid
has left the device, we often are not concerned with that fluid, as far as the performance
of the device is concerned. Of course, we might care about the pollution emitted by the
device and the long term fate of expelled particles. Pollution dispersion, in contrast to
pollution-creation, is more a problem of fluid mechanics than thermodynamics.
Analysis of control volumes is slightly more complicated than for systems, and the equa-
tions we will ultimately use are slightly more complex. Unfortunately, the underlying math-
ematics and physics that lead to the development of our simplified control volume equations
are highly challenging! Worse still, most beginning thermodynamics texts do not expose the
student to all of the many nuances required for the simplification.
In the previous chapter, we considered system of fixed mass, so the axiom of mass con-
servation was trivially satisfied. And we developed the energy conservation axiom for fixed
mass systems. In contrast, the focus of this chapter is to formulate mass and energy conser-
vation axioms for a control volume configuration. A sketch of a generic apparatus for control
volume analysis is given in Fig. 4.1.
157
158 CHAPTER 4. FIRST LAW ANALYSIS FOR A CONTROL VOLUME
Figure 4.1: Sketch of a generic configuration for control volume analysis.
4.1 Detailed derivations of control volume equations

This section will give a summary of the necessary mathematical operations necessary to cast
the conservation of mass and energy axioms in a traditional control volume formulation.
The analysis presented has been amalgamated from a variety of sources. Most directly, it is
a specialization of analysis given by Powers.1,2 Basic mathematical foundations are covered
well by Kaplan.3 A detailed and readable description, that has a stronger emphasis on fluid
mechanics, is given in the undergraduate text of Whitaker.4 A rigorous treatment of the de-
velopment of all equations presented here is included in the graduate text of Aris.5 Popular
aerospace and mechanical engineering undergraduate fluids texts have closely related expo-
sitions.6,7 However, despite their detail, these texts have some minor flaws! The treatment
given by BS is not as detailed. This section will use a notation generally consistent with BS
and show in detail how to arrive at its results.
4.1.1 Relevant mathematics

We will use several theorems that are developed in vector calculus. Here, we give short
motivations and presentations. The reader should consult a standard mathematics text for
1
J. M. Powers, 2024, Mechanics of Fluids, Cambridge, Cambridge.
2
J. M. Powers, 2024, Lecture Notes on Intermediate Fluid Mechanics, University of Notre Dame,
https://www3.nd.edu/∼powers/ame.60635/notes.pdf.
3
W. Kaplan, 2003, Advanced Calculus, Fifth Edition, Addison-Wesley, New York.
4
S. Whitaker, 1992, Introduction to Fluid Mechanics, Krieger, Malabar, Florida.
5
R. Aris, 1962, Vectors, Tensors, and the Basic Equations of Fluid Mechanics, Dover, New York.
6
J. D. Anderson, 1990, Modern Compressible Flow with Historical Perspective, Second Edition, McGraw-
Hill, New York, Ch. 2.4.
7
F. M. White, 1979, Fluid Mechanics, McGraw-Hill, New York, Ch. 3.4.

4.1. DETAILED DERIVATIONS OF CONTROL VOLUME EQUATIONS 159
detailed derivations.
4.1.1.1 Fundamental theorem of calculus

Restating here in new variables what was earlier expressed in Eq. (3.31), the fundamental
theorem of calculus is as follows
Z x=b Z x=b
dφ
ψ(x) dx = dx = φ(b) − φ(a). (4.1)
x=a x=a dx
It effectively says that to find the integral of a function ψ(x), that is the area under the
curve, it suffices to find a function φ, whose derivative is ψ, i.e. dφ/dx = ψ(x), evaluate φ
at each endpoint, and take the difference to find the area under the curve.
4.1.1.2 Divergence theorem

The divergence theorem, often known as Gauss’s8 theorem, is the analog of the fundamental
theorem of calculus extended to volume integrals. Gauss is depicted in Fig. 4.2. While it is
Figure 4.2: Johann Carl Friedrich Gauss (1777-1855), German mathematician; image from
https://mathshistory.st-andrews.ac.uk/Biographies/Gauss.html.
often attributed to Gauss, who reported it in 1813, it is said that it was first discovered by
Joseph Louis Lagrange in 1762.9
Let us define the following quantities:
• t → time,
• x → spatial coordinates,
• Va (t) → arbitrary moving and deforming volume,

8
Carl Friedrich Gauss, 1777-1855, Brunswick-born German mathematician, considered the founder of
modern mathematics. Worked in astronomy, physics, crystallography, optics, bio-statistics, and mechanics.
Studied and taught at Göttingen.
9
https://en.wikipedia.org/wiki/Divergence theorem.

• Aa (t) → bounding surface of the arbitrary moving volume,

• n → outer unit normal to moving surface, and
• φ(x, t) → arbitrary vector function of x and t.
The divergence theorem is as follows:
Z Z
∇ · φ dV = φ · n dA. (4.2)
Va (t) Aa (t)
The surface integral is analogous to evaluating the function at the end points in the funda-
mental theorem of calculus.
If φ(x, t) has the form φ(x, t) = cφ(x, t), where c is a non-zero constant vector and φ is
a scalar function, then the divergence theorem, Eq. (4.2), reduces to
Z Z
∇ · (cφ) dV = (cφ) · n dA, (4.3)
Va (t) Aa (t)
Z ! Z
| {z· c} +c · ∇φ
φ∇ dV = φ (c · n) dA, (4.4)
Va (t) =0 Aa (t)
Z Z
c· ∇φ dV = c · φn dA, (4.5)
Va (t) Aa (t)
Z Z
c· ∇φ dV − φn dA = 0. (4.6)
Va (t) Aa (t)
| {z }
=0
Now, because c is non-zero, the term in parentheses must be zero. Thus,

Z Z
∇φ dV = φn dA. (4.7)
Va (t) Aa (t)
If we take φ to be the scalar of unity (whose gradient must be zero), the divergence theorem
reduces to
Z Z
∇(1) dV = (1)n dA, (4.8)
Va (t) Aa (t)
Z
0 = (1)n dA, (4.9)
Aa (t)
Z
n dA = 0. (4.10)
Aa (t)
That is, the unit normal to the surface, integrated over the surface, cancels to zero when the
entire surface is included.
We will use the divergence theorem, Eq. (4.2), extensively. It allows us to convert some-
times difficult volume integrals into more easily interpreted surface integrals. It is often
useful to use this theorem as a means of toggling back and forth from one form to another.

4.1.1.3 Leibniz’s rule

Leibniz’s10 rule relates time derivatives of integral quantities to a form that distinguishes
changes that are happening within the boundaries to changes due to fluxes through bound-
aries. Leibniz is depicted in Fig. 4.3.
Figure 4.3: Gottfried Wilhelm von Leibniz (1646-1716), German mathe-

matician, philosopher, and polymath who co-invented calculus; image from
https://mathshistory.st-andrews.ac.uk/Biographies/Leibniz.html.
Let us consider the scenario sketched in Fig. 4.4. Say we have some value of interest, Φ,
that results from an integration of a kernel function φ over Va (t), for instance
Z
Φ= φ dV. (4.11)
Va (t)
We are often interested in the time derivative of Φ, the calculation of which is complicated
by the fact that the limits of integration are time-dependent. From the definition of the
derivative, we find that
Z R R
dΦ d V (t+∆t)
φ(t + ∆t) dV − Va (t)
φ(t) dV
= φ dV = lim a . (4.12)
dt dt Va (t) ∆t→0 ∆t
Now, we have
Va (t + ∆t) = Va (t) + VII (∆t) − VI (∆t) . (4.13)
| {z } | {z } | {z } | {z }
new old captured abandoned
Here, VII (∆t) is the amount of new volume captured in time increment ∆t, and VI (∆t) is
the amount of volume abandoned in time increment ∆t. So we can split the first integral in
10
Gottfried Wilhelm von Leibniz, 1646-1716, Leipzig-born German philosopher and mathematician. In-
vented calculus independent of Newton and employed a superior notation to that of Newton.

AII
AI
wII
dAI
wI VII
VI
nI nII
I
dAII
II
Figure 4.4: Sketch of the motion of an arbitrary volume Va (t). The boundaries of Va (t) move
with velocity w. The outer normal to Va (t) is Aa (t). Here, we focus on just two regions: I,
where the volume is abandoning material, and II, where the volume is capturing material.
the last term of Eq. (4.12) into

Z Z Z Z
φ(t + ∆t) dV = φ(t + ∆t) dV + φ(t + ∆t) dV − φ(t + ∆t) dV ,
Va (t+∆t) Va (t) VII (∆t) VI (∆t)
| {z } | {z } | {z } | {z }
new intrinsically evolved captured abandoned
(4.14)
that gives us then
Z
d
φ dV =
dt Va (t)
R R R R
V (t)
φ(t + ∆t) dV + VII (∆t)
φ(t + ∆t) dV − VI (∆t)
φ(t + ∆t) dV − Va (t)
φ(t) dV
lim a .
∆t→0 ∆t
(4.15)
Rearranging Eq. (4.15) by combining terms with common limits of integration, we get
Z R
d V (t)
(φ(t + ∆t) − φ(t)) dV
φ dV = lim a
dt Va (t) ∆t→0 ∆t
R R
V (∆t)
φ(t + ∆t) dV − VI (∆t)
φ(t + ∆t) dV
+ lim II . (4.16)
∆t→0 ∆t
Let us now further define
• w → the velocity vector of points on the moving surface Va (t). Here we use the bold
notation to indicate w is a velocity vector, in contrast to the scalar w we would use
for velocity magnitude.

Now, the volume swept up by the moving volume in a given time increment ∆t is
dVII = w · n} ∆t dAII = wII ∆t dAII ,
| {z (4.17)
| {z }
positive distance
and the volume abandoned is

dVI = |w{z· n} ∆t dAI = − wI ∆t dAI . (4.18)
| {z }
negative distance
Substituting into our definition of the derivative, Eq. (4.16), we get

Z Z
d (φ(t + ∆t) − φ(t))
φ dV = lim dV
dt Va (t) ∆t→0 V (t)
a
∆t
R R
AII (∆t)
φ(t + ∆t)wII ∆t dAII + AI (∆t) φ(t + ∆t)wI ∆t dAI
+ lim . (4.19)
∆t→0 ∆t
Now, we note that
• We can use the definition of the partial derivative to simplify the first term on the right
side of Eq. (4.19),
• The time increment ∆t cancels in the surface integrals of Eq. (4.19), and
• Aa (t) = AI + AII ,
so that
Z Z Z
d ∂φ
φ dV = dV + φw · n dA . (4.20)
dt Va (t) Va (t) ∂t Aa (t)
| {z } | {z } | {z }
total time rate of change intrinsic change within volume net capture by moving volume
This is the three-dimensional scalar version of Leibniz’s rule. Say we have the special case
in which φ = 1; then Leibniz’s rule, Eq. (4.20), reduces to
Z Z Z
d ∂
dV = (1) dV + (1)w · n dA, (4.21)
dt Va (t) ∂t{z }
Va (t) | Aa (t)
=0
Z
d
Va (t) = w · n dA. (4.22)
dt Aa (t)
This simply says the total volume of the region, that we call Va (t), changes in response to
net motion of the bounding surface.
Example 4.1
Consider a sphere whose initial radius is r(0) = ro and whose outer boundary is moving outwards,
normal to the surface, with a uniform and constant velocity normal to the surface of magnitude w.
Find V (t).

w
n w
w
w w
Figure 4.5: Sketch of a uniformly expanding sphere.
The scenario is depicted in Fig. 4.5. Here the volume is no longer arbitrary, as we have specified
it, so it is V (t). And the bounding surface is A(t). The velocity of the outer boundary is uniform and
aligned with the outer surface normal n,
w = wn. (4.23)
Thus,
w · n = w |n{z
· n} = +w. (4.24)
=1
Thus Eq. (4.22) reduces to

Z
dV
= w dA, (4.25)
dt A(t)
Z
= w dA, (4.26)
A(t)
= wA(t). (4.27)
For the sphere, we have V = 4πr3 /3 and A = 4πr2 , so we get

d 4 3
πr = w4πr2 , (4.28)
dt 3
1 d 3
r = wr2 , (4.29)
3 dt
dr
r2 = wr2 , (4.30)
dt
dr
= w. (4.31)
dt
This is not a surprise and likely could be inferred by simple intuition. So integrating and applying the
initial condition r(0) = ro , we get
r(t) = wt + ro . (4.32)

Thus,
4 3
V (t) = π (wt + ro ) . (4.33)
3
Leibniz’s rule, Eq. (4.20), reduces to a more familiar result in the one-dimensional limit.
We can then say
Z x=b(t) Z x=b(t)
d ∂φ db da
φ(x, t) dx = dx + φ(b(t), t) − φ(a(t), t). (4.34)
dt x=a(t) x=a(t) ∂t dt dt
As in the fundamental theorem of calculus, Eq. (4.1), for the one-dimensional case, we do
not have to evaluate a surface integral; instead, we simply must consider the function at
its endpoints. Here, db/dt and da/dt are the velocities of the bounding surface and are
equivalent to w. The terms φ(b(t), t) and φ(a(t), t) are equivalent to evaluating φ on Aa (t).
We can also apply the divergence theorem, Eq. (4.2), to Leibniz’s rule, Eq. (4.20), to
convert the surface integral into a volume integral to get
Z Z Z
d ∂φ
φ dV = dV + ∇ · (φw) dV. (4.35)
dt Va (t) Va (t) ∂t Va (t)
Combining the two volume integrals, we get

Z Z
d ∂φ
φ dV = + ∇ · (φw) dV. (4.36)
dt Va (t) Va (t) ∂t
4.1.1.4 General transport theorem

Let B be an arbitrary extensive thermodynamic property, and β be the corresponding in-
tensive thermodynamic property so that
dB = β dm. (4.37)
The product of a differential amount of mass dm with the intensive property β gives a
differential amount of the extensive property. Because
dm = ρ dV, (4.38)
where ρ is the mass density and dV is a differential amount of volume, we have
dB = βρ dV. (4.39)

If we take the arbitrary φ = ρβ, Leibniz’s rule, Eq. (4.20), becomes our general transport
theorem: Z Z Z
d ∂
ρβ dV = (ρβ) dV + ρβ (w · n) dA. (4.40)
dt Va (t) Va (t) ∂t Aa (t)
Applying the divergence theorem, Eq. (4.2), to the general transport theorem, Eq. (4.40),
we find the alternate form
Z Z
d ∂
ρβ dV = (ρβ) + ∇ · (ρβw) dV. (4.41)
dt Va (t) Va (t) ∂t
4.1.1.5 Reynolds transport theorem

Osborne Reynolds11 made many pioneering contributions to fluid mechanics. He is depicted
in Fig. 4.6. Among other things, he wrote a treatise on the development of conservation
Figure 4.6: Osborne Reynolds (1842-1912), Anglo-Irish engineer; image from

https://mathshistory.st-andrews.ac.uk/Biographies/Reynolds.html.
principles in a general sense.12 From this work, and after employing more modern notation,
we arrive at what is now known as the Reynolds transport theorem if we force the arbitrary
velocity of the moving volume to take on the velocity of a fluid particle, i.e. take
w = v. (4.42)
Here v is the velocity vector of the fluid particle, in contrast to the scalar velocity magnitude
v. In this case, our arbitrary volume is no longer arbitrary. Instead, it always contains the
same fluid particles. We call this volume a
11
Osborne Reynolds, 1842-1912, Belfast-born British engineer and physicist, educated in mathematics at
Cambridge, first professor of engineering at Owens College, Manchester, did fundamental experimental work
in fluid mechanics and heat transfer.
12
O. Reynolds, 1903, Papers on Mechanical and Physical Subjects, Volume III, The Sub-Mechanics of the
Universe, Cambridge, Cambridge.

• material volume, Vm (t): a volume that always contains the same fluid particles.
Corresponding to this we also have a
• material surface, Am (t): a closed surface that always encloses the same fluid par-
ticles.
The proper way to generalize laws of nature that were developed for point masses is to
consider collections of fixed point masses, that will always reside within a material volume.
That said, it is simple to specialize the general transport theorem to obtain the Reynolds
transport theorem. Here, we give two versions, the first using volume and surface integrals,
and the second using volume integrals only. In this special case, Eqs. (4.40) and (4.41)
become, respectively,
Z Z Z
d ∂
ρβ dV = (ρβ) dV + ρβ (v · n) dA, (4.43)
dt Vm (t) Vm (t) ∂t Am (t)
Z Z
d ∂
ρβ dV = (ρβ) + ∇ · (ρβv) dV. (4.44)
dt Vm (t) Vm (t) ∂t
The implications of Eq. (4.43) are summarized in the words of Reynolds in Fig. 4.6.
Figure 4.7: Image from Reynolds’ 1903 study, p. 9, giving his key general axiom.
4.1.1.6 Fixed (control) volumes

If we take our arbitrary volume to be fixed in space, it is most often known as a
• control volume: a fixed volume in space.
For control volumes

w = 0. (4.45)

Thus, the arbitrary volume loses its time-dependency, so that
Va (t) = V, Aa (t) = A, (4.46)
and the general transport theorem, Eq. (4.40), reduces to

Z Z
d ∂
ρβ dV = (ρβ) dV. (4.47)
dt V V ∂t
4.1.2 Conservation axioms

A fundamental goal of mechanics is to convert the verbal notions that embody the basic
axioms into usable mathematical expressions. First, we must list those axioms. The axioms
themselves are simply principles that have been observed to have wide validity as long as
length scales are sufficiently large to contain many molecules. Many of these axioms can be
applied to molecules as well. The axioms cannot be proven. They are simply statements
that have been useful in describing the universe.
A summary of the axioms in words is as follows
• Mass conservation principle: The time rate of change of mass of a material region is
zero.
• Linear momenta principle: The time rate of change of the linear momenta of a material
region is equal to the sum of forces acting on the region. This is Euler’s generalization
of Newton’s second law of motion.
• Angular momenta principle: The time rate of change of the angular momenta of a
material region is equal to the sum of the torques acting on the region. This was first
formulated by Euler.
• Energy conservation principle: The time rate of change of energy within a material
region is equal to the rate that energy is received by heat and work interactions. This
is the first law of thermodynamics.
• Entropy inequality: The time rate of change of entropy within a material region is
greater than or equal to the ratio of the rate of heat transferred to the region and the
absolute temperature of the region. This is the second law of thermodynamics.
Here, we shall systematically convert two of these axioms, the mass conservation principle
and the energy conservation principle, into mathematical form.
4.1.2.1 Mass
Mass is an extensive property for which we have
B = m, β = 1. (4.48)

dV
dA
w=v
Figure 4.8: Sketch of a finite material region Vm (t), infinitesimal mass element ρ dV , and
infinitesimal surface element dA with unit normal n, and general velocity w equal to fluid
velocity v.
The mass conservation axiom is simple to state mathematically. It is
d
m = 0. (4.49)
dt
A relevant material volume is sketched in Fig. 4.8. We can define the mass enclosed within
a material volume based upon the local value of density:
Z
m= ρ dV. (4.50)
Vm (t)
So the mass conservation axiom is

Z
d
ρ dV = 0. (4.51)
dt Vm (t)
d
R R ∂
R
Invoke the Reynolds transport theorem, Eq. (4.43), dt Vm (t)
[ ] dV = Vm (t) ∂t
[ ] dV + Am (t)
v·
n[ ] dA, to get
Z Z Z
d ∂ρ
ρ dV = dV + ρv · n dA = 0. (4.52)
|{z}
| {z } physics
mathematics
The first equality of Eq. (4.52) is simply a mathematical statement involving definitions;
forcing either of the terms to Requal zero is a statement
R of physics. Now, we invoke the
divergence theorem, Eq. (4.2) V (t) ∇ · [ ] dV = A(t) n · [ ] dA, to convert a surface integral

to a volume integral to get the mass conservation axiom to read as

Z Z
∂ρ
dV + ∇ · (ρv) dV = 0, (4.53)
Vm (t) ∂t Vm (t)
Z
∂ρ
+ ∇ · (ρv) dV = 0. (4.54)
Vm (t) ∂t
| {z }
=0
Now, in an important step, we realize that the only way for this integral, that has arbitrary
limits of integration, to always be zero, is for the integrand itself to always be zero. Hence,
we have
∂ρ
+ ∇ · (ρv) = 0. (4.55)
∂t
This is the important differential form of the mass conservation principle. We can expand
this via the product rule to also say
∂ρ
+ v · ∇ρ +ρ∇ · v = 0. (4.56)
|∂t {z }
=dρ/dt
It can be shown that if we travel at the same velocity of a fluid particle, that any changes
with time are properly described by the so-called material derivative:
d ∂
≡ + v · ∇. (4.57)
dt ∂t
So the mass conservation axiom, Eq. (4.56), can be recast as
dρ
+ ρ∇ · v = 0, (4.58)
dt
so long as d/dt is interpreted as the derivative following a particle: the material derivative.
Rearranging, we see
dρ
= −ρ∇ · v, (4.59)
dt
telling us that a fluid particle’s density changes in response to the divergence of its velocity
vector field. Moreover, a fluid particle that is incompressible, forcing dρ/dt = 0, must be
associated with a divergence-free velocity field, ∇ · v = 0.
Example 4.2
For a fluid in the velocity vector field
   x
vx vo L
v =  0 =  0 , (4.60)
0 0

find the time rate of change of its density as well as its acceleration.
For this one-dimensional flow that only varies in the x direction, we have
∂vx ∂vy ∂vz ∂ x vo

∇·v = + + = vo +0+0= . (4.61)
∂x ∂y ∂z ∂x L L
So we have from mass conservation that

dρ
= −ρ∇ · v. (4.62)
dt
This gives
dρ ρvo
=− . (4.63)
dt L
For ρ, vo , L > 0, this flow has density decreasing with time of a fluid particle. That is because the
boundaries of the material region are being stretched by the velocity field.
The acceleration of the fluid particle is
dv ∂v
= + v · ∇v. (4.64)
dt ∂t
Because we have strictly one-dimensional motion, this reduces to
dvx ∂vx ∂vx

= + vx , (4.65)
dt ∂t ∂x

x 1
= 0 + vo vo . (4.66)
L L
This gives
dvx v2 x
= o2 . (4.67)
dt L
The fluid particle is accelerating to the right. Even though the velocity field does not change with time,
the position and velocity of an individual fluid particle does change as it travels. The configuration is
sketched in Fig. 4.9.
Example 4.3
Show by considering ρ = ρ(t, x, y, z) that the material derivative dρ/dt = ∂ρ/∂t + v · ∇ρ.
If ρ = ρ(t, x, y, z), then by the definition of a total differential of a function of many variables, see
Eq. (3.3),
∂ρ ∂ρ ∂ρ ∂ρ
dρ = dt + dx + dy + dz. (4.68)
∂t ∂x ∂y ∂z

x
L
Figure 4.9: Sketch of a fluid element in a velocity field with non-zero divergence. The fluid
is stationary at x = 0 and has velocity vo when x = L. This stretches the material volume,
causing the density of the material volume to drop. And the fluid particle accelerates as
its x-coordinate increases. Also sketched is the velocity as a function of x; the slope of this
curve is the velocity gradient, which here is constant.
Let us divide by a time differential dt to get then

dρ ∂ρ ∂ρ dx ∂ρ dy ∂ρ dz
= + + + . (4.69)
dt ∂t ∂x dt ∂y dt ∂z dt
Now with
dx dy dz
= vx , = vy , = vz , (4.70)
dt dt dt
we get after small rearrangement
dρ ∂ρ ∂ρ ∂ρ ∂ρ
= + vx + vy + vz , (4.71)
dt ∂t ∂x ∂y ∂z
 
∂ρ
∂ρ  ∂x
∂ρ 
= + vx vy vz  ∂y , (4.72)
∂t | {z } ∂ρ
v ∂z
| {z }
∇ρ
∂ρ
= + v · ∇ρ. (4.73)
∂t
It is a material derivative because we are insisting that dx/dt = vx , dy/dt = vy , dz/dt = vz , the material
particle velocity.
Example 4.4
Prove for a general intensive thermodynamic property β that
Z Z
d dβ
ρβ dV = ρ dV. (4.74)
dt Vm (t) Vm (t) dt

We have from the Reynolds transport theorem, Eq. (4.43), that

Z Z Z
d ∂
ρβ dV = (ρβ) dV + ρβ (v · n) dA, (4.75)
Z
∂
= (ρβ) + ∇ · (ρβv · n) dV, (4.76)
Vm (t) ∂t
    
Z
  ∂ρ   ∂β 
=   ∂t + ∇ · (ρv · n) + ρ  ∂t + v · ∇β  dV,
β     (4.77)
Vm (t) | {z } | {z }
=0 =dβ/dt
Z
dβ
= = ρ dV. (4.78)
Vm (t) dt
We can get a useful control volume formulation by integrating the mass conservation
principle, Eq. (4.55), over a fixed volume V :
Z Z
∂ρ
+ ∇ · (ρv) dV = 0 dV . (4.79)
V ∂t V
| {z }
=0
Now,
Rb the integral of 0 over a fixed domain must be zero. This is equivalent to saying
a
0 dx = 0, where the area under the curve of 0 has to be zero. So we have
Z
∂ρ
+ ∇ · (ρv) dV = 0. (4.80)
V ∂t
Next apply the divergence theorem, Eq. (4.2), to Eq. (4.80) to get
Z Z
∂ρ
dV + ρv · n dA = 0. (4.81)
V ∂t A
Applying now the result from Eq. (4.47) to Eq. (4.81), we see for the fixed volume that
Z Z
d
ρ dV + ρv · n dA = 0. (4.82)
dt V A
We note now that at an inlet that v points in an opposite direction to n, so we have
v · n < 0, at inlets. (4.83)
At exits,v and n point in the same direction so that
v · n > 0, at exits. (4.84)

If now, we take the simplifying assumption that ρ and v have no spatial variation across
inlets and exits, we get for a control volume with one inlet and one exit that
Z
d
ρ dV + ρe |ve |Ae − ρi |vi |Ai = 0. (4.85)
dt V
Here, the subscript i denotes inlet, and the subscript e denotes exit. Rearranging Eq. (4.85),
we find Z
d
ρ dV = ρi |vi |Ai − ρe |ve |Ae . (4.86)
dt V
We now define the mass in the control volume mcv as
Z
mcv = ρ dV. (4.87)
V
Here, Eq. (4.87) is equivalent to the equation on p. 144 of BS. Thus, we have
dmcv
= ρi |vi |Ai − ρe |ve |Ae . (4.88)
| dt
{z } | {z }
mass rate in
| {z }
mass rate out
rate of change of mass
Here, mcv is the mass enclosed in the control volume. If there is no net rate of change of
mass the control volume is in steady state, and we can say that the mass flow in must equal
the mass flow out:
ρi |vi |Ai = ρe |ve |Ae . (4.89)
We define the mass flow rate ṁ as
ṁ = ρ|v|A. (4.90)
For steady flows with a single entrance and exit, we have
ṁ = constant. (4.91)
For unsteady flows with a single entrance and exit, we can rewrite Eq. (4.88) as
dmcv
= ṁi − ṁe . (4.92)
dt
For unsteady flow with many entrances and exits, we can generalize Eq. (4.88) as
dmcv X X
= ρi |vi |Ai − ρe |ve |Ae , (4.93)
dt }
| {z | {z } | {z }
rate of change of mass mass rate in mass rate out
dmcv X X
= ṁi − ṁe . (4.94)
dt }
| {z | {z } | {z }

Note that Eq. (4.94) is fully equivalent to BS’s Eq. (4.1) (p. 143), but that it actually takes
a good deal of effort to get to this point with rigor! For steady state conditions with many
entrances and exits we can say
X X
ρi |vi |Ai = ρe |ve |Ae . (4.95)
Thus
X X
ṁi = ṁe . (4.96)
Here, Eq. (4.96) is the same as BS’s Eq. (4.9), p. 149.
Example 4.5
Consider a control volume with three openings. A sketch is given in Fig. 4.10. The system is in a
water
Figure 4.10: Sketch for a steady state control volume mass conservation example.
steady state, and water is the working fluid. Measurements at the openings shows conditions at each
to be
1. inlet with T1 = 200 ◦ C, P1 = 7 bar, ṁ1 = 40 kg/s
2. inlet with A2 = 25 cm2 , T2 = 40 ◦ C, P2 = 7 bar.
3. exit with x3 = 0, P3 = 7 bar, (Av)3 = 0.06 m3 /s.

Find ṁ3 , ṁ2 and v2 .
Our mass conservation equation is

dmcv X X
= ṁi − ṁe , (4.97)
dt
= ṁ1 + ṁ2 − ṁ3 . (4.98)
Now, because the problem is steady, d/dt = 0. Thus, there is no net accumulation of mass in the
control volume, and we must have
0 = ṁ1 + ṁ2 − ṁ3 , (4.99)
ṁ2 = ṁ3 − ṁ1 . (4.100)
Now, ṁ = ρvA = Av/v. So
(Av)3
ṁ3 = . (4.101)
v3
Now, we have two properties at state 3, so we know the state. The water tables thus give us
m3
v3 = vf = 0.001108 . (4.102)
kg
so
m3
0.06
s kg
ṁ3 = 3 = 54.1516 . (4.103)
0.001108 mkg
s
So
kg kg kg
ṁ2 = ṁ3 − ṁ1 = 54.1516 − 40 = 14.1516 . (4.104)
s s s
Now,
A2 v2
ṁ2 = , (4.105)
v2
ṁ2 v2
v2 = . (4.106)
A2
Now, state 2 is a compressed liquid. Let us estimate v2 by vf at T2 = 40 ◦ C. We find this is
v2 = 0.001008 m3 /kg. So

m3
14.1516 kg
s 0.001008 kg
v2 = 2 , (4.107)
2 1 m
(25 cm ) 100 cm
m
= 5.70594 . (4.108)
s
Example 4.6
A tank is initially empty. A liquid with ρ = 62.4 lbm/ft3 is poured into the tank at a constant mass
flow rate of ṁi = 7 lbm/s. The tank has cross-sectional area A = 0.2 ft2 , and the fluid in the tank has

Figure 4.11: Sketch of a tank filling problem.
a variable height H(t). There is a hole at the bottom of the tank. The fluid flows out of the tank at a
rate proportional to the fluid height: ṁe = kH, where k = 1.4 lbm/ft/s. Find H(t). A sketch is given
in Fig. 4.11.
Our mass conservation law for the control volume says
dmcv
= ṁi − ṁe . (4.109)
dt
Now, mcv = ρV = ρAH(t). Substituting, we get
d
(ρAH) = ṁi − kH, (4.110)
dt
dH
ρA = ṁi − kH, (4.111)
dt
dH ṁi k
= − H, (4.112)
dt ρA ρA

dH ṁi k
= 1− H . (4.113)
dt ρA ṁi
The only variable here is H(t). Every other parameter is a known constant. We can separate variables
to get
dH ṁi
= dt, (4.114)
1 − ṁki H ρA
Z Z
dH ṁi
= dt, (4.115)
1 − ṁki H ρA

ṁi k ṁi
− ln 1 − H = t + C. (4.116)
k ṁi ρA

H (ft)
5
t (s)
10 20 30 40 50
Figure 4.12: Plot of H(t) for a transient tank filling problem.
Now, when t = 0, we have H = 0, so C = 0 and

ṁi k ṁi
− ln 1 − H = t, (4.117)
k ṁi ρA

k k
ln 1 − H = − t, (4.118)
ṁi ρA

k k
1− H = exp − t , (4.119)
ṁi ρA

ṁi k
H = 1 − exp − t . (4.120)
k ρA
Note
ṁi
lim H(t) = . (4.121)
t→∞ k
In the long time limit, a steady state height is reached where the flow out balances the flow in. If we
increase ṁi , the steady state height increases proportionally.
We also see by inspection that the time constant is
ρA
τ= . (4.122)
k
This gives an estimate of how long it takes to reach the steady state height.
! !!
7 lbm
s 1.4 lbm
ft s
H(t) = 1 − exp − t , (4.123)
1.4 lbm
ft s 62.4 lbm
ft3 (0.2 ft2 )

t
= (5 ft)(1 − exp − . (4.124)
8.91429 s
The time constant is τ = 8.91429 s; H(∞) is 5 ft. A plot of pH(t) is shown in Fig. 4.12. Note that the
physics of this problem would be better modeled by ṁe = k H(t), as we shall see later in Ch. 7.2.3.
However, near the equilibrium state, our linear analysis can be shown to be appropriate. A better
capture of the initial transients would require the indicated modifications, and would necessitate a
numerical solution, rather than the closed form analytic solution given here.

4.1.2.2 Energy
For energy, we must consider the total energy that includes internal, kinetic, and potential.
Our extensive property B is thus
1
B = E = U + mv · v + mgz. (4.125)
2
Here, we have assumed the fluid resides in a gravitational potential field in which the gravi-
tational potential energy varies linearly with height z. The corresponding intensive property
β is
1
β = e = u + v · v + gz. (4.126)
2
We recall the first law of thermodynamics, that states the change of a material volume’s
total energy is equal to the heat transferred to the material volume less the work done by
the material volume. Mathematically, this is stated as Eq. (3.172):
dE = δQ − δW. (4.127)
We recall the total derivative is used for dE, because energy is a property and has an exact
differential, while both heat transfer and work are not properties and do not have exact
differentials. It is more convenient to express the first law as a rate equation, that we get by
dividing Eq. (4.127) by dt to get
dE δQ δW
= − . (4.128)
dt dt dt
Recall that the upper case letters denote extensive thermodynamic properties. For example,
E is total energy, inclusive of internal and kinetic and potential13 , with SI units of J. Let us
consider each term in the first law of thermodynamics in detail and then write the equation
in final form.
4.1.2.2.1 Total energy term For a fluid particle, the differential amount of total energy
is

1
dE = ρβ dV = ρ u + v · v + gz dV, (4.129)
2
| {z }
β

1
= ρ dV u + v · v + gz . (4.130)
| {z } 2
mass | {z }
internal +kinetic +potential
13
Strictly speaking our derivation will only be valid for potentials that are time-independent. This is the
case for ordinary gravitational potentials. The modifications for time-dependent potentials are straightfor-
ward, but require a more nuanced interpretation than space permits here.

4.1.2.2.2 Work term Let us partition the work into work WP done by a pressure force
FP and work done by other sources, that we shall call Wmv , where the subscript “mv”
indicates “material volume.”
W = WP + Wmv . (4.131)
Taking a time derivative, we get
δW δWP
= + Ẇmv . (4.132)
dt dt
The work done by other sources is often called shaft work and represents inputs of such
devices as compressors, pumps, and turbines. Its modeling is often not rigorous.
Recall that work is done when a force acts through a distance, and a work rate arises when
a force acts through a distance at a particular rate in time (hence, a velocity is involved).
Recall also that it is the dot product of the force vector with the position or velocity that
gives the true work or work rate. In shorthand, we can say that the differential work done
by the pressure force FP is
δWP = FP · dx, (4.133)

δWP dx
= FP · = FP · v. (4.134)
dt dt
Here, W has the SI units of J, and FP has the SI units of N. Now, let us consider the work
done by the pressure force. In a piston-cylinder arrangement in which a fluid exists with
pressure P within the cylinder and the piston is rising with velocity v, the work rate done
by the fluid is positive. We can think of the local stress vector in the fluid as pointing in
the same direction as the fluid is moving at the piston surface, so that the dot product is
positive. Now, we can express the pressure force in terms of the pressure by
FP = P An. (4.135)
Substituting Eq. (4.135) into Eq. (4.134), we get
δWP
= P An · v. (4.136)
dt
It is noted that we have been a little loose distinguishing local areas from global areas. Better
stated, we should say for a material volume that
Z
δWP
= P n · v dA. (4.137)
dt Am (t)
This form allows for P and v to vary with location. This is summarized in the sketch of
Fig. 4.13.

v
v n v
Figure 4.13: Sketch of a fluid doing work.
4.1.2.2.3 Heat transfer term If we were considering temperature fields with spatial
dependency, we would define a heat flux vector. This approach is absolutely necessary to
describe many real-world devices, and is the focus of a standard undergraduate course in
heat transfer. Here, we will take a simplified assumption that the only heat fluxes are easily
specified and are all absorbed into a lumped scalar term we will call Q̇mv . This term has
units of J/s = W in SI. So we have then
δQ
= Q̇mv . (4.138)
dt
4.1.2.2.4 The first law of thermodynamics Putting the words of the first law into
equation form, we get
Z
d 1 δQ δW
ρ u + v · v + gz dV = − . (4.139)
dt Vm (t) 2 dt dt
| {z }
Emv
We next introduce our simplification of heat transfer, Eq. (4.138), and partition of work,
Eq. (4.132), along with Eq. (4.137) into Eq. (4.139) to get
Z Z
d 1
ρ u + v · v + gz dV = Q̇mv − Ẇmv + P n · v dA . (4.140)
dt Vm (t) 2 Am (t)
| {z }
=δW/dt
Now, we bring the pressure work integral to the right side of Eq. (4.140) to get
Z Z
d 1
ρ u + v · v + gz dV + P n · v dA = Q̇mv − Ẇmv . (4.141)
dt Vm (t) 2 Am (t)

We next invoke the Reynolds transport theorem, Eq. (4.43), into Eq. (4.141) to expand the
derivative of the first integral so as to obtain
Z Z
∂ 1 1
ρ u + v · v + gz dV + ρ u + v · v + gz v · n dA
Vm (t) ∂t 2 Am (t) 2
| {z }
dEmv /dt
Z
+ P n · v dA = Q̇mv − Ẇmv . (4.142)
Am (t)
We next note that the two surface integrals have the same limits and can be combined to
form
  
Z Z
∂ 1 ρ u + P + 1 v · v + gz  v · n dA
  
ρ u + v · v + gz dV +
Vm (t) ∂t 2 ρ 2
  
Am (t)
| {z }
h
= Q̇mv − Ẇmv .
(4.143)
We recall now the definition of enthalpy h, Eq. (3.218),

P
h=u+ = u + P v. (4.144)
ρ
Invoking Eq. (4.144) into Eq. (4.143), we get

Z Z
∂ 1 1
ρ u + v · v + gz dV + ρ h + v · v + gz v · n dA
Vm (t) ∂t 2 Am (t) 2
= Q̇mv − Ẇmv . (4.145)
Next use the divergence theorem, Eq. (4.2), to rewrite Eq. (4.145) as
Z Z
∂ 1 1
ρ u + v · v + gz dV + ∇ · ρv h + v · v + gz dV
Vm (t) ∂t 2 Vm (t) 2
= Q̇mv − Ẇmv . (4.146)
Now, for convenience, let us define the specific heat transfer and work, qmv and wmv , each
with SI units J/kg such that
Z
∂
Q̇mv = (ρqmv ) dV, (4.147)
Vm (t) ∂t
Z
∂
Ẇmv = (ρwmv ) dV, (4.148)
Vm (t) ∂t

so that by substituting Eq. (4.147) and Eq. (4.148) into Eq. (4.146), we get
Z Z
∂ 1 1
ρ u + v · v + gz dV + ∇ · ρv h + v · v + gz dV
Vm (t) ∂t 2 Vm (t) 2
Z Z
∂ ∂
= (ρqmv ) dV − (ρwmv ) dV. (4.149)
Vm (t) ∂t Vm (t) ∂t
Now, all terms in Eq. (4.149) have the same limits of integration, so they can be grouped to
form
Z
∂ 1 1
ρ u + v · v + gz + ∇ · ρv h + v · v + gz
Vm (t) ∂t 2 2

∂ ∂
− (ρqmv ) + (ρwmv ) dV = 0. (4.150)
∂t ∂t
As with the mass equation, because the integral is zero, in general we must expect the
integrand to be zero, giving us

∂ 1 1 ∂ ∂
ρ u + v · v + gz + ∇ · ρv h + v · v + gz − (ρqmv ) + (ρwmv ) = 0.
∂t 2 2 ∂t ∂t
(4.151)
To get the standard control volume form of the equation, we then integrate Eq. (4.151)
over a fixed control volume V to get
Z
∂ 1 1
ρ u + v · v + gz + ∇ · ρv h + v · v + gz
V ∂t 2 2

∂ ∂
− (ρqmv ) + (ρwmv ) dV = 0. (4.152)
∂t ∂t
Now, defining the heat transfer and work rates, Q̇cv and Ẇcv ,
Z
∂
Q̇cv = (ρqmv ) dV, (4.153)
∂t
ZV
∂
Ẇcv = (ρwmv ) dV, (4.154)
V ∂t
we employ Eq. (4.153) and Eq. (4.154) in Eq. (4.152) to get

Z
∂ 1 1
ρ u + v · v + gz + ∇ · ρv h + v · v + gz dV = Q̇cv − Ẇcv .
V ∂t 2 2
(4.155)

Applying the divergence theorem, Eq. (4.2), to Eq. (4.155) to convert a portion of the volume
integral into a surface integral, and Eq. (4.47) to bring the time derivative outside the integral
for the fixed volume, we get
Z Z
d 1 1
ρ u + v · v + gz dV + ρv · n h + v · v + gz dA = Q̇cv − Ẇcv . (4.156)
dt V 2 A 2
| {z }
Ecv
We now define the total energy in the control volume as

Z
1
Ecv = ρ u + v · v + gz dV. (4.157)
V 2
Next assume that all properties across entrances and exits are uniform so that the surface
integral in Eq. (4.155) reduces to
Z
1
ρv · n h + v · v + gz dA =
A 2
X
X 1 1
ṁe he + ve · ve + gze − ṁi hi + vi · vi + gzi . (4.158)
2 2
Substituting Eq. (4.157) and Eq. (4.158) into Eq. (4.156), we get
X
dEcv X 1 1
+ ṁe he + ve · ve + gze − ṁi hi + vi · vi + gzi = Q̇cv − Ẇcv . (4.159)
dt 2 2
Rearranging Eq. (4.159), we get
dEcv
= Q̇cv − Ẇcv
dt }
| {z |{z} |{z}
CV heat transfer rate CV shaft work rate
rate of CV energy change
X 1
X
1

+ ṁi hi + vi · vi + gzi − ṁe he + ve · ve + gze .
2 2
| {z } | {z }
total enthalpy rate in total enthalpy rate out
(4.160)
Here, Eq. (4.160) is equivalent to BS’s Eq. (4.7), p. 148. Note that the so-called total enthalpy
is often defined as
1
htot = h + v · v + gz. (4.161)
2
Employing Eq. (4.161) in Eq. (4.160), we find
dEcv X X
= Q̇cv − Ẇcv + ṁi htot,i − ṁe htot,e . (4.162)
dt

4.2. MASS CONSERVATION IN BRIEF 185
Here, Eq. (4.162) is equivalent to BS’s (4.8), p. 148.

If there is a single entrance and exit, we lose the summation, so that Eq. (4.160) becomes

dEcv 1 1
= Q̇cv − Ẇcv + ṁi hi + vi · vi + gzi − ṁe he + ve · ve + gze . (4.163)
dt 2 2
If the flow is steady, we have dEcv /dt = 0 and ṁi = ṁe = ṁ, so the first law with a single
entrance and exit becomes

1
0 = Q̇cv − Ẇcv + ṁ hi − he + (vi · vi − ve · ve ) + g(zi − ze ) . (4.164)
2
Defining the specific heat transfer and work as
Q̇cv Ẇcv
q= , w= , (4.165)
ṁ ṁ
and substituting Eq. (4.165) into Eq. (4.164), we get
1
0 = q − w + hi − he + (vi · vi − ve · ve ) + g(zi − ze ). (4.166)
2
Now, Eq. (4.166) can be rearranged to form BS’s Eq. (4.13), p. 149:
1 1
q + hi + vi · vi + gzi = w + he + ve · ve + gze . (4.167)
2 2
This looks more like the first law when we rearrange as

1 1
he + ve · ve + gze − hi + vi · vi + gzi = q − w. (4.168)
2 2
If the flow is adiabatic, steady, has one entrance and one exit, and there is no shaft work,
we find that the total enthalpy must remain constant:
1 1
hi + vi · vi + gzi = he + ve · ve + gze . (4.169)
2 2
4.2 Mass conservation in brief

Here, we summarize the key equations for mass conservation derived in the previous section.
We consider the mass enclosed in a fixed control volume V , Eq. (4.87):
Z
mcv = ρ dV. (4.170)
V
The density ρ can vary throughout V . In this class, we will nearly always take it to be
constant throughout the volume. If ρ is constant throughout V , then it can be brought

R
outside the integral operator, yielding mcv = ρ V dV = ρV. Our control volume will have a
finite number of openings where fluid can enter and exit.
We state mass conservation for a control volume as
dmcv X X
= ρi |vi |Ai − ρe |ve |Ae , (4.171)
dt }
| {z | {z } | {z }
dmcv X X
= ṁi − ṁe . (4.172)
dt }
| {z | {z } | {z }
Equations (4.171, 4.172) were fully developed in the previous section where they appeared
as Eqs. (4.93, 4.94). Here, the fluid at an inlet i has density ρi , velocity vector vi and flows
through cross-sectional area Ai . An analogous set of variables exists at each exit e. Let us
look at the units of the important quantity ρ|v|A:
kg m m2 kg
ρ|v|A → 3
= . (4.173)
m s 1 s
Obviously, it is a rate of mass flow; consequently, we define the mass flow rate as
ṁ ≡ ρ|v|A. (4.174)
Often we will neglect the vector notation and take |v| = v. Equation (4.172) expresses
mathematically the notion of mass conservation for the control volume:
• The time rate of accumulation of mass within the control volume is equal to the net
rate of mass flow into the control volume.
In short
accumulation = in - out.
In the important case in which there is no net accumulation rate, the so-called steady state
limit, we get
dmcv X X
= ṁi − ṁe . (4.175)
| dt
{z }
=0
X X
ṁi = ṁe . (4.176)
Equation (4.176) is the same as Eq. (4.96). If there is a single entrance and exit, then we
simply get
ṁi = ṁe = ṁ = constant. (4.177)

4.3. ENERGY CONSERVATION IN BRIEF 187
4.3 Energy conservation in brief

We can state the first law of thermodynamics for a control volume as
dEcv
= Q̇cv − Ẇcv
dt }
| {z |{z} |{z}
CV heat transfer rate CV shaft work rate
rate of CV energy change
X
X 1 2 1 2
+ ṁi hi + vi + gzi − ṁe he + ve + gze . (4.178)
2 2
| {z } | {z }
total enthalpy rate in total enthalpy rate out
Here, Eq. (4.178) is equivalent Eq. (4.160) and to BS’s Eq. (4.7), p. 148. Note that the
so-called total enthalpy is often defined as
1 1
htot = h + v2 + gz = u + v2 + gz + P v. (4.179)
2 2
Note that in this context, total enthalpy is on a per mass basis. The “total” comes from
summing internal, kinetic, potential, and P v terms. Employing Eq. (4.179) in Eq. (4.178),
we find
dEcv X X
= Q̇cv − Ẇcv + ṁi htot,i − ṁe htot,e . (4.180)
dt
Here, Eq. (4.180) is equivalent to Eq. (4.162) and to BS’s Eq. (4.8), p. 148. Equation (4.180)
expresses mathematically the notion of energy conservation for the control volume:
• The time rate of accumulation of total energy within the control volume is equal the
rate of heat transfer into the control volume minus the rate of work done leaving the
control volume plus the net rate of total enthalpy entering the control volume.
The new terms here are attributable to total enthalpy entering and exiting the control
volume.
Again, the total enthalpy is the sum of the specific internal energy, the specific kinetic
energy, the specific potential energy and the term P v. It is easy to imagine that Ecv , that
itself is composed of u, KE, and P E, is affected by the flow of u, KE, and P E into and out
of the control volume. However the term P v is unusual. It is multiplied by ṁ. Let us check
the units: 3
kg kN m kJ
ṁP v → 2
= = kW. (4.181)
s m kg s
It has the units of power. As shown in detail in the previous section,
• The term ṁP v embedded within the control volume energy equation within htot accounts
for the net work rate done by the fluid as it enters and exits the control surface bounding
the control volume.

• The term Ẇcv represents so-called shaft work and does not include work associated with
the expansion of the working fluid.
If there is a single entrance and exit, we lose the summation, so that Eq. (4.178) becomes

dEcv 1 2 1 2
= Q̇cv − Ẇcv + ṁi hi + vi + gzi − ṁe he + ve + gze . (4.182)
dt 2 2
If the flow is steady, we have dEcv /dt = 0 and ṁi = ṁe = ṁ, so the first law with a single
entrance and exit becomes

1 2 2
0 = Q̇cv − Ẇcv + ṁ hi − he + (vi − ve ) + g(zi − ze ) . (4.183)
2
Defining the specific heat transfer and work as
Q̇cv Ẇcv
q= , w= , (4.184)
ṁ ṁ
and substituting Eq. (4.184) into Eq. (4.183), we get
1
0 = q − w + hi − he + (vi2 − ve2 ) + g(zi − ze ). (4.185)
2
Now, Eq. (4.185) can be rearranged to form BS’s Eq. (4.13), p. 149:
1 1
q + hi + vi2 + gzi = w + he + ve2 + gze . (4.186)
2 2
This looks more like the first law when we rearrange as

1 2 1 2
he + ve + gze = hi + vi + gzi + q−w . (4.187)
2 2 | {z }
| {z } | {z } CV heat and work
outlet inlet
If the flow is adiabatic, steady, has one entrance and one exit, and there is no shaft work,
we find that the total enthalpy must remain constant:
1 1
hi + vi2 + gzi = he + ve2 + gze . (4.188)
2 2
Example 4.7
A well insulated chamber with V = 1 ft3 initially contains air at P = 14.7 lbf/in2 , T = 100 ◦ F. See
Fig. 4.14. Intake and exhaust valves are opened, and air enters and exits at 1 lbm/min through each
valve. The entering air is at P = 30 lbf/in2 , T = 200 ◦ F. Assume the air is well mixed so that P and
T are uniform throughout the chamber. Find P (t), T (t).

4.3. ENERGY CONSERVATION IN BRIEF 189
air initially at
1 P = 14.7 psia 2
T = 100 oF
P = 30 psia
T. = 200 oF
m = 1 lbm/min
air
Figure 4.14: Sketch of a transient thermal mixing problem.
The temperatures are moderate, so it is not a problem to model air as a CPIG. For a CPIG, we
have
V
P = RT, (4.189)
m
dh = cP dT, (4.190)
du = cv dT. (4.191)
The mass and energy evolution equations for the control volume are, respectively,
dmcv
= ṁ1 − ṁ2 , (4.192)
dt
dEcv v2 v2
= Q̇cv − Ẇcv + ṁ1 h1 + 1 + gz1 − ṁ2 h2 + 2 + gz2 . (4.193)
dt 2 2
Let us find the initial mass in the control volume. At t = 0 s, we have

lbf
in2

PV PV 14.7 in 2 144 ft 2 1 ft3
m= = = lbf ft = 0.0709 lbm. (4.194)
RT R
T 1545 lbmole ◦R
(100 + 460 ◦ R)
M 28.97 lbm
lbmole
Now, the mass balance gives

dmcv
= ṁ1 − ṁ2 , (4.195)
dt
with ṁ1 = ṁ2 = (1 lbm/min)(min/60 s) = (1/60) lbm/s.

dmcv 1 lbm 1 lbm
= − = 0, (4.196)
dt 60 s 60 s
mcv = constant = 0.0709 lbm. (4.197)
Now, for the energy equation, we can neglect changes in kinetic and potential energies. The volume is
insulated, so Q̇cv = 0. And there is no shaft work, so Ẇcv = 0. So the energy balance reduces to
dEcv
= ṁ1 h1 − ṁ2 h2 . (4.198)
dt
Now, because ṁ1 = ṁ2 = ṁ, and because Ecv = Ucv = mcv ucv when KE and P E are neglected, the
energy equation reduces to
d
(mcv ucv ) = ṁ(h1 − h2 ). (4.199)
dt

Because mcv is constant, and because du = cv dT , where cv is constant for a CPIG, we get
dTcv
mcv cv = ṁ(h1 − h2 ). (4.200)
dt
Now, h1 is known from the inlet conditions. After mixing, the exit enthalpy takes on the value of the
enthalpy in the chamber. So we have from the caloric equation of state for h that
h1 − h2 = cP (T1 − Tcv ). (4.201)
Thus, the energy equation becomes
dTcv
mcv cv = ṁcP (T1 − Tcv ). (4.202)
dt
We also have the initial condition T (0) = To . We rewrite the energy equation, Eq. (4.202), as
dTcv ṁ cP
= (T1 − Tcv ), (4.203)
dt mcv cv
|{z}
=k
ṁ
= k(T1 − Tcv ). (4.204)
mcv
Note the temperature is equilibrated when Tcv = T1 . We expect after a long time that the entire system
acquires the character of the inlet after all the old material is flushed from the control volume. The
question now is how long is the flushing time? To answer this, we must solve a differential equation.
Let us separate variables to get
dTcv ṁ
= k dt, (4.205)
T1 − Tcv mcv
ṁ
− ln(T1 − Tcv ) = kt + C, (4.206)
mcv

ṁ
T1 − Tcv = exp − kt − C , (4.207)
mcv

ṁ
= C ′ exp − kt . (4.208)
mcv
At the initial state, we get
T1 − To = C ′. (4.209)
Thus
ṁ
Tcv = T1 − (T1 − To ) exp − kt . (4.210)
mcv
Note that as t → ∞, Tcv → T1 . Also note by inspection, the time constant of relaxation is
mcv 1
τ= . (4.211)
ṁ k
So the time to equilibrium is short if
• the input forcing ṁ is large, or
• the mass in the control volume, mcv , is small.

4.4. SOME DEVICES 191
For us, the time constant is

0.0709 lbm
τ= 7
1 lbm = 3.04 s. (4.212)
5 60 s
So the temperature variation is

◦ ◦ ◦ t
Tcv (t) = (660 R) − ((660 R) − (560 R)) exp − . (4.213)
3.04 s

t
Tcv (t) = (660 ◦ R) − (100 ◦ R) exp − . (4.214)
3.04 s
For the pressure, we have P V = mRT , so

mRTcv (t)
Pcv (t) = , (4.215)
V
ft lbf
1545 lbmole t
(0.0709 lbm) lbm
28.97 lbmole
◦R
(660 ◦ R) − (100 ◦ R) exp − 3.04 s ft2
= . (4.216)
1 ft3 144 in2
So the pressure is

t
Pcv (t) = (17.33 psia) − (2.6258 psia) exp − . (4.217)
3.04 s
A plot of T (t) and P (t) is given in Fig. 4.15. Note, the final pressure is not the inlet pressure.
P (psia)
T (oR) 18
640
16
600
14
0 4 8 12 0 4 8 12
t (s) t (s)
Figure 4.15: Plots of T (t) and P (t) for a transient thermal mixing problem.
4.4 Some devices

Here, we will consider rudiments of control volume analysis for some common engineering
devices.

4.4.1 Throttling device

A flow is throttled when, for example, it flows through a partially open valve. When it does
so, we notice that there can be a significant pressure loss from one side of the partially open
valve to the other. A sketch of a throttling device is given in Fig. 4.16.
1 2
Figure 4.16: Sketch of a throttling device.
We model a throttling device as steady with one entrance and exit, with no control
volume work or heat transfer. We neglect changes in area as well as potential energy. Mass
conservation tells us
dmcv
= ṁ1 − ṁ2 , (4.218)
dt }
| {z
=0
0 = ṁ1 − ṁ2 , (4.219)
ṁ1 = ṁ2 = ṁ. (4.220)
Energy conservation tells us that

dEcv v12 v22
= Q̇cv − Ẇcv +ṁ1 h1 + + gz1 − ṁ2 h2 + + gz2 , (4.221)
dt }
| {z |{z} |{z} 2 2
=0 =0
=0
 
v2 v2
0 = ṁ h1 − h2 + 1 − 2 + gz1 − gz2  .
 
(4.222)
|2 {z 2} | {z
=0
}
∼0
Now, in throttling devices there may be a change in velocity due to compressibility effects,
but it is observed to be small when the flow velocity is much less than the speed of sound.
We shall assume here the velocity is small relative to the speed of sound so as to recover
v1 ∼ v2 and thus
h1 = h2 . (4.223)
So, we can say that such a throttling device is one in which pressure drops and enthalpy
remains constant.

Example 4.8
Let us throttle steam from P1 = 3 bar, T1 = 200 ◦ C to P2 = 1 bar. Find T2 .
We know two properties at state 1; therefore, its state is determined. At state 2, we know the
pressure and something about the path of the process that brought us to state 2. The process was
iso-enthalpic, so h2 = h1 . From the tables, we find
kJ
h1 = 2865.5 = h2 . (4.224)
kg
So at state 2, we have P2 = 1 bar, h2 = 2865.5 kJ/kg. We interpolate the steam tables to find
T2 = 195.04 ◦ C. (4.225)
Note for steam, the enthalpy remained constant, but the temperature dropped. If the material had
been an ideal gas, the temperature drop would have been zero, because for ideal gases, the enthalpy is
related only to temperature.
Example 4.9
Consider calorically perfect ideal air flowing in a duct at P1 = 100000 Pa, T1 = 300 K, v1 = 10 m/s.
Take cP = 1000 J/kg/K. The air pressure is throttled down by a valve to P2 = 90000 Pa. Consider
fully the effects of compressibility and kinetic energy, find the downstream state after throttling, and
show the assumption that h2 ∼ h1 (giving for the CPIG T2 ∼ T1 ) is a good approximation for this flow.
From the ideal gas law we get
P1 100000 Pa kg
ρ1 = = = 1.16144 3 . (4.226)
RT1 J
287 kg K (300 K) m
We write the mass, energy, thermal and caloric state equations as
ρ2 v2 = ρ1 v1 , (4.227)
v2 v2
h2 + 2 = h1 + 1 , (4.228)
2 2
P2 = ρ2 RT2 , (4.229)
h2 − h1 = cP (T2 − T1 ). (4.230)
This simplifies to
ρ2 v2 = ρ1 v1 , (4.231)
v2 v2
cP T 2 + 2 = cP T 1 + 1 , (4.232)
2 2
P2 = ρ2 RT2 . (4.233)

Substituting numbers, we get the system of non-linear algebraic equations

kg m kg
ρ2 v2 = 1.16144 3 10 = 11.6144 2 , (4.234)
m s m s
2
J v22 J 10 ms J
1000 T2 + = 1000 (300 K) + = 300050 , (4.235)
kg K 2 kg K 2 kg

J
90000 Pa = ρ2 287 T2 . (4.236)
kg K
This forms three equations in the three unknowns ρ2 , T2 , v2 . Detailed manipulation can reduce this to
a quadratic equation, with two roots. The first root yields
kg
ρ2 = 1.04533 , (4.237)
m3
T2 = 299.988 K, (4.238)
m
v2 = 11.1107 . (4.239)
s
This is the physical root. We see that
T2 ∼ T1 . (4.240)
Thus, from the ideal gas law
ρ2 T 2 ρ1 T 1
= , (4.241)
P2 P1
T1 P2
ρ2 = ρ1 , (4.242)
T2 P1
P2
∼ ρ1 , (4.243)
P1
90000 Pa
∼ ρ1 , (4.244)
100000 Pa
∼ 0.9ρ1 , (4.245)

kg
∼ 0.9 1.16144 3 , (4.246)
m
kg
∼ 1.0453 3 . (4.247)
m
Then from mass conservation, we get
ρ2 v2 = ρ1 v1 , (4.248)
ρ1
v2 = v1 , (4.249)
ρ2
1
∼ v1 , (4.250)
0.9
m 1
∼ 10 , (4.251)
s 0.9
m
∼ 11.11 . (4.252)
s
There is also an obviously non-physical root that is ρ2 = −0.000215037 kg/m3 , T2 = −1458300 K,
v2 = −54011.1 m/s. Negative density and absolute temperature are physically unacceptable.

As an aside, detailed calculation reveals the exact solution for v2 is

r !
P12 v12 (2cP T1 +v12 ) v 
cP P2 T1 c2P P22 T12
+1−1
u
u 2 v2 1 + v12
P2 cP T 1  t P1 1 2cP T1 
v2 = = v1 2  1+4 − 1 (.4.253)
P1 v1 P1 v1 P2 2cP T1
This is a complicated expression that is difficult to interpret. But Taylor series expansion for v12 /(cP T1 ) ≪
1 yields
2 ! !
P1 P1 v12
v2 = v1 1+ 1− + ... . (4.254)
P2 P2 2cP T1
This is an excellent approximation that yields results fully equivalent to what was obtained earlier.
Similarly, Taylor series of the solution for T2 and ρ2 in the same limit yields the approximations
2 ! !
P1 v12
T2 = T1 1 + 1− + ... , (4.255)
P2 2cP T1
2 ! !
P2 P1 v12
ρ2 = ρ1 1− 1− + ... . (4.256)
P1 P2 2cP T1
Direct substitution of the parameter values reveals that the approximations, relative to the exact
solution, agree up to the fifth decimal place for v2 , T2 , and ρ2 . And at leading order, we get the simple
approximations of the response to the pressure difference of a throttling device as

P1 P2
v2 ∼ v1 , T2 ∼ T1 , ρ2 ∼ ρ1 . (4.257)
P2 P1
Physically, we can say that in this limit, the pressure difference induces negligible change in thermal
energy and finite change in mechanical energy.
Example 4.10
A spacecraft returning to earth is traveling at v = 10000 m/s. It is in a region of the atmosphere
where T = 100 K, cP = 1000 J/kg/K. Neglecting any effects of shock waves, provide a simple estimate
of the temperature at the nose of the vehicle.
In the laboratory frame, the atmosphere is at rest and the spacecraft is moving. Transform to the
reference frame in which the spacecraft is at rest and is encountering air moving at 10000 m/s. As the
air approaches the spacecraft, it decelerates to zero velocity at the nose. Take the 1 state to be the
moving atmosphere and the 2 state to be the nose of the spacecraft. The first law of thermodynamics

yields
v12 v22
h1 + = h2 + , (4.258)
2 2
|{z}
0
v12
h2 − h1 = , (4.259)
2
v12
cP (T2 − T1 ) = , (4.260)
2
v12
T2 = T1 + , (4.261)
2cP

m 2
104 s
= (100 K) + , (4.262)
J
2 103 kg K
= 50100 K. (4.263)
This is a high temperature. It is not a bad first estimate; however the actual temperature would be
different because of neglected physics such as a) shock waves, b) temperature-dependent specific heat,
c) dissociation and ionization of molecules.
4.4.2 Nozzles and diffusers

Similar to a throttling device, we model nozzles and diffusers as steady with one entrance
and exit, with no control volume work. We may or may not neglect heat transfer. We neglect
potential energy changes but take kinetic energy changes into account.
A nozzle is a device that induces a velocity increase; a diffuser is a device that induces
a velocity decrease. For flows with subsonic velocities, nozzles have area decrease in the
flow direction, while diffusers have area increases with the flow direction. We sketch these
common configurations in Fig. 4.17. If one systematically applied the conservation of mass,
momentum, and energy principles, after detailed analysis, one finds the converse state of
affairs for supersonic flow conditions. Supersonic nozzles have increasing area; supersonic
diffusers have decreasing area. This is why in the design of rocket nozzles, the cross-sectional
area broadens at the base. The broadening area induces a higher velocity, and induces a
higher thrust for a supersonic rocket engine. A brief analysis of this will be given in Ch. 9.9.3.
We analyze nozzles and diffusers as follows. Mass conservation tells us
dmcv
= ṁ1 − ṁ2 , (4.264)
dt }
| {z
=0
ṁ1 = ṁ2 = ṁ. (4.265)

1
2
2 1
subsonic subsonic
nozzle diffuser
Figure 4.17: Sketch of a subsonic nozzle and diffuser.
Energy conservation tells us

dEcv v12 v22
= Q̇cv − Ẇcv +ṁ1 h1 + + gz1 − ṁ2 h2 + + gz2 , (4.266)
dt }
| {z
|{z} 2 2
=0
=0
 
v12 v22
0 = Q̇cv + ṁ h1 − h2 + − + gz1 − gz2  , (4.267)
2 2 | {z }
=0
Q̇cv v12 v22

= + h1 − h2 + − . (4.268)
ṁ 2 2
If the nozzle or diffuser is also adiabatic, we get
v12 v2
h1 + = h2 + 2 . (4.269)
2 2
4.4.3 Turbine
A turbine is a device in which work is generated through expansion of a fluid as it passes
through a fan-like device. The fluid interacts with the blades and turns the fan. Ultimately
thermal and mechanical energy is transferred from the fluid into the rotational kinetic energy
of the fan blades. A sketch of a turbine is given in Fig. 4.18. For a turbine, we typically
neglect kinetic and potential energy changes of the fluid. We may or may not neglect heat
transfer. We also neglect any unsteady effects. Mass conservation tells us
dmcv
= ṁ1 − ṁ2 , (4.270)
dt }
| {z
=0
ṁ1 = ṁ2 = ṁ. (4.271)

Figure 4.18: Highly simplified sketch of a turbine.

dEcv v12 v22
= Q̇cv − Ẇcv + ṁ1 h1 + + gz1 − ṁ2 h2 + + gz2 , (4.272)
dt }
| {z 2 2
=0
 
v2 v2
0 = Q̇cv − Ẇcv + ṁ h1 − h2 + 1 − 2 + gz1 − gz2  ,
 
(4.273)
|2 {z 2} | {z
=0
}
=0
= Q̇cv − Ẇcv + ṁ(h1 − h2 ), (4.274)

Ẇcv = Q̇cv + ṁ(h1 − h2 ). (4.275)
We often neglect Q̇cv to get
Ẇcv = ṁ(h1 − h2 ). (4.276)
On a per mass basis, we can scale by ṁ to get
w = h1 − h2 . (4.277)
For turbines, h1 > h2 , so we get w > 0. The device is doing work.
4.4.4 Pumps and compressors

The analysis for a pump or compressor is effectively identical to that for a turbine. However
the device operates in an opposite sense. Mechanical energy from either rotating (like a

compressor in a jet engine) or reciprocating machinery (like a piston-cylinder arrangement)

is transferred to the working fluid, raising its energetic state. We typically neglect changes
in kinetic and potential energy of the fluid and consider the device to be in a steady state.
We sometimes neglect heat transfer to the device.
The analysis is as follows. Mass conservation tells us
dmcv
= ṁ1 − ṁ2 , (4.278)
dt }
| {z
=0
ṁ1 = ṁ2 = ṁ. (4.279)

dEcv v12 v22
= Q̇cv − Ẇcv + ṁ1 h1 + + gz1 − ṁ2 h2 + + gz2 , (4.280)
dt }
| {z 2 2
=0
 
v2 v2
0 = Q̇cv − Ẇcv + ṁ h1 − h2 + 1 − 2 + gz1 − gz2  ,
 
(4.281)
|2 {z 2} | {z
=0
}
=0
= Q̇cv − Ẇcv + ṁ(h1 − h2 ), (4.282)

Ẇcv = Q̇cv + ṁ(h1 − h2 ). (4.283)
We often neglect Q̇cv to get
Ẇcv = ṁ(h1 − h2 ). (4.284)
On a per mass basis, we can scale by ṁ to get
w = h1 − h2 . (4.285)
For pumps and compressors, h1 < h2 , so we get w < 0. The device requires an input of
work.
4.4.5 Heat exchanger

A heat exchanger is a device in which a working fluid trades its thermal energy with another
working fluid. A sketch of a heat exchanger is given in Fig. 4.19. For heat exchangers, we
typically neglect all work, as well as changes in kinetic and potential energy. Also
• there will be exchange of thermal energy between individual flow streams, but
• globally for the entire device, there will be no heat transfer with the environment.

Figure 4.19: Sketch of a counterflow heat exchanger.
Let us consider a counterflow heat exchanger. The mass balance for steady flow is trivial.
The energy balance, neglecting changes in KE and P E states
dEcv X X
= Q̇cv − Ẇcv + ṁi hi − ṁe he . (4.286)
dt } |{z} |{z}
| {z i e
=0 =0
=0
Applying this to the counterflow heat exchanger gives
ṁ1 h1,hot + ṁ2 h2,cold = ṁ1 h1,cold + ṁ2 h2,hot , (4.287)

ṁ1 (h1,hot − h1,cold ) = ṁ2 (h2,hot − h2,cold ) . (4.288)
Example 4.11
Given an adiabatic air turbine with ṁ = 1.5 kg/s with the following inlet and exit conditions
• Pi = 1000 kPa,
• Ti = 1200 K,
• Pe = 100 kPa,
• Te = 700 K,
calculate the work output of the turbine assuming i) CPIG, ii) CIIG, and neglecting changes in kinetic
and potential energy. We have a simple sketch in Fig. 4.20.
i e
turbine
Figure 4.20: Sketch of a turbine problem.

The first law for both CPIG and CIIG is the same. It is, after neglecting changes of kinetic and
potential energy,
dEcv
= Q̇cv −Ẇcv + ṁhi − ṁhe . (4.289)
| dt
{z } |{z} =0
=0
Because this problem is adiabatic and steady, we get
Ẇcv = ṁ(hi − he ). (4.290)
Our estimate of the work output will depend on which caloric equation of state we choose.
• CPIG. For a CPIG, we have
hi − he = cP (Ti − Te ). (4.291)
For ideal gases we have, cP −cv = R and k = cP /cv . The second of these gives cv = cP /k. Substituting
into the first gives
cP
cP − = R, (4.292)
k
1
cP 1 − = R, (4.293)
k

k−1
cP = R, (4.294)
k
k
cP = R. (4.295)
k−1
For diatomic gases at moderate temperatures, we have k = 7/5; thus,
7
k 5 7 1 1 5
cP = R= 7 R= R, cv = R= 7 R = R. (4.296)
k−1 5 −1 2 k−1 5 −1 2
So
7
5 kJ kJ
cP = 7 0.287 = 1.0045 , (4.297)
5 −1
kg K kg K

1 kJ kJ
cv = 7 0.287 = 0.7175 . (4.298)
5 − 1 kg K kg K
Note these are the same values listed in Table A.5 of BS. So for the turbine work output, we get

kg kJ
Ẇcv = ṁcP (Ti − Te ) = 1.5 1.0045 ((1200 K) − (700 K)) = 753.375 kW. (4.299)
s kg K
• CIIG. For the CIIG, we have a few choices. We could use Table A.6 of BS to estimate R e cP at an
intermediate temperature, and then treat it as a constant. We could form the integral i cP (T ) dT .
Or we could use the ideal gas tables, Table A.7.1 of BS. Let us do the third method. At Ti = 1200 K,
we find
kJ
hi = 1277.81 . (4.300)
kg
At Te = 700 K, we find
kJ
he = 713.56 . (4.301)
kg
So

kg kJ kJ
Ẇcv = ṁ(hi − he ) = 1.5 1277.81 − 713.56 = 846.375 kW. (4.302)
s kg kg

The two methods yield similar results. Use of the more accurate CIIG reveals there is more useful work
that can be expected when we take actual material behavior into account. This is because the high
temperature gas has some extra energy stored in its vibrational modes that was unaccounted for by
the CPIG model. Recall our CPIG model did account for rotational modes in taking cv = 5/2R. But
as can be seen from examining Fig. 3.33, for diatomic molecules in air, such as O2 , cv /R > 5/2 for
our temperatures of T ∈ [700 K, 1200 K]. Presumably, molecular vibration is relevant at these elevated
temperatures; in particular, the enhanced work predicted by the CIIG model is a manifestation of
accounting for the additional energy available that was initially bound into vibrational modes at high
temperature. The CIIG model properly accounts for this.
Example 4.12
Given a steam turbine with ṁ = 1.5 kg/s, Q̇cv = −8.5 kW with the following inlet and exit
conditions
• Pi = 2 MPa,
• Ti = 350 ◦ C,
m
• vi = 50 s ,
• zi = 6 m,
• Pe = 0.1 MPa,
• xe = 1,
m
• ve = 200 s ,
• ze = 3 m,
find the power output. The same simple sketch of Fig. 4.20 applies.
The first law states

dEcv vi2 ve2
= Q̇cv − Ẇcv + ṁ hi + + gzi − ṁ he + + gze , (4.303)
dt }
| {z 2 2
=0

1 2 2
Ẇcv = Q̇cv + ṁ hi − he + (vi − ve ) + g(zi − ze ) . (4.304)
2
From the steam tables, we learn
kJ kJ
hi = 3137 , he = 2675.5 . (4.305)
kg kg

4.5. INTRODUCTION TO THE RANKINE CYCLE 203
So

kg kJ kJ
Ẇcv = (−8.5 kW) + 1.5 3137 − 2675.5
s kg kg
!
1 kJ m 2 m 2 9.81 sm2
+ 50 − 200 + J
((6 m) − (3 m)) , (4.306)
2 1000 J s s 1000 kJ
 

kg 
 kJ kJ kJ 

= (−8.5 kW) + 1.5  461.6 − 18.75 + 0.0294 , (4.307)
s  kg kg kg 
| {z } | {z } | {z }
=∆h =∆KE =∆P E
= 655.7 kW. (4.308)
Note
• The dominant term is the ∆h term. Kinetic and potential energy changes, as well as heat transfer
effects are small in comparison. This is typical for turbines.
• The factor of 1000 is necessary to give the units of specific kinetic and potential energy changes kJ/kg.
Let us quickly check the units for kinetic energy, here without the factor of 1000:
J Nm kg m2 m2
= = 2 = 2. (4.309)
kg kg s kg s
4.5 Introduction to the Rankine cycle

Consider the
• Rankine cycle: a thermodynamic cycle that forms the foundation for most steam
power plants.
The cycle was studied analytically by Rankine,14 depicted in Fig. 4.21. We will consider
additional nuances of the Rankine cycle in Ch. 8.1. The key features of the Rankine cycle
are
• 1 → 2: compression of a liquid by a pump,
• 2 → 3: boiling of the liquid to form a vapor,
• 3 → 4: expansion of the vapor through a turbine, and
• 4 → 1: condensation of the vapor to liquid in a condenser.

Figure 4.21: William John Macquorn Rankine (1820-1872), Scottish engineer who sys-
tematically studied and published discussion of steam power cycles. Image from
https://en.wikipedia.org/wiki/William John Macquorn Rankine.
A sketch of the Rankine cycle in the P − v plane is given in Fig. 4.22. The Rankine
cycle forms the cornerstone of a wide variety of power generating devices in the world today.
Whether the heat source comes from burning coal, natural gas, fuel oil, garbage, nuclear
fission, solar energy, or some other source, it can always be used to boil water, that is
the key feature of the Rankine cycle. Most modern power plants are considerably more
complicated than the simple outline given here. Some are equipped to use a variety of
fuels. Often coal burning components are used continuously for so-called base loads and
are supplemented during peak consumption hours by natural gas. Some modern plants use
natural gas for base loads. Some cycles are equipped for district heating and cooling, some
for electric power generation, some for nautical propulsion.
Example 4.13
Consider the Rankine cycle sketched in Fig. 4.23 for generation of power. The conditions of the
steam are indicated in Fig. 4.23.
The cycle includes some of the effect of losses; thus, it is not an ideal Rankine cycle. But it is close.
Find, on a per unit mass basis,
• the heat transfer in the line between the boiler and the turbine,
• the turbine work,
• the heat transfer in the condenser, and
14
W. J. M. Rankine, 1859, A Manual of the Steam Engine and Other Prime Movers, Griffin, London.

2 boiler 3
pump turbine
condenser
1 4
Figure 4.22: Sketch of a Rankine cycle in the P − v plane.
• the heat transfer in the boiler.
For illustration, Fig. 4.24 shows photographs of some elements of the University of Notre Dame
Power Plant configured in a way that reflects the rudimentary Rankine cycle. The actual plant is more
complex than indicated in Fig. 4.24. It contains additional elements for cooling, heating with gas and
oil, and pollution removal. A full analysis is beyond the scope of these notes.
We use the steam tables for our equation of state. We can model this as a steady state system
composed of devices with one inlet and one exit. We can neglect changes in KE and P E. For any such
device, the mass flow rate ṁ is constant, and the energy balance gives
dEcv
= Q̇cv − Ẇcv + ṁ(hi − he ), (4.310)
| dt
{z }
=0
0 = q − w + hi − he . (4.311)
• Work in pump. Let us assume an ideal pump, that is adiabatic, 1q2 = 0. Work is done by the pump
on the fluid in compressing it. A pump restores the high pressure to the fluid to compensate for
frictional losses.
1w2 = hi − he = h1 − h2 . (4.312)
We go to the tables and find h1 = 188.45 kJ/kg. Now, we know 1w2 = −4 kJ/kg, so this lets us
calculate h2 via

kJ kJ kJ
h2 = h1 − 1w2 = 188.45 − −4 = 192.45 . (4.313)
kg kg kg
Note:
– The pump work is relatively small, as will be seen by comparison to later calculations of work
and heat transfers for other devices.
• Heat transfer in boiler. A boiler is an active device in which thermal energy of perhaps combustion
or nuclear reaction is used to boil water. The tables tell us that h3 = 3023.5 kJ/kg. There is no work

3 3a
turbine
4
boiler
1
condenser
pump
Figure 4.23: Sketch of a steam power cycle, known as the Rankine cycle.
in a boiler, so 2w3 = 0, and the first law becomes
2q3 = he − hi = h3 − h2 , (4.314)

kJ kJ kJ
= 3023.5 − 192.45 = 2831.1 . (4.315)
kg kg kg
Note
– The heat transfer to the boiler, 2q3 > 0, so the thermal energy is entering the device.
• Heat transfer in line between boiler and turbine. For the flow in the pipe, there is no external shaft
work, so 3w3a = 0, and thus
3q3a = he − hi = h3a − h3 . (4.316)
From double interpolation of the steam tables, h3a = 3002.5 kJ/kg, so

kJ kJ kJ
3q3a = 3002.5 − 3023.5 = −21 . (4.317)
kg kg kg
– Because the thermal energy leaves the pipe, 3q3a < 0. In general we will find that q will be
path-dependent. For this type of analysis, it reduces to differences in the end states.
– Note that P drops because of friction in the pipes.
– Note that T drops because of heat transfer to the surroundings.
– Most importantly, note that we easily determined a quantity of global importance with data
that was easily measured at the inlet and exit of the pipe. We did not need to consider the
detailed fluid and thermal fields within the pipe.
• Turbine work. We assume an ideal turbine, which is adiabatic, 3aq4 = 0, so
3aw4 = hi − he = h3a − h4 . (4.318)

boiler turbine
pump condenser
Figure 4.24: Elements of the University of Notre Dame Power Plant, configured in an over-
simplified manner, 14 June 2010. At this time, the energy source was coal combustion, that
is no longer the case.
Now, we know h3a already. For h4 , we have
h4 = hf + x4 hf g , at P4 = 15 kPa, (4.319)

kJ kJ
= 225.94 + (0.9) 2373.1 , (4.320)
kg kg
kJ
= 2361.7 . (4.321)
kg
Thus,

kJ kJ kJ
3aw4 = 3002.5 − 2361.7 = 640.8 . (4.322)
kg kg kg
Note
– the pressure P drops as the steam expands through the turbine,

– the work 3aw4 > 0, that indicates the turbine is doing work, and
– the function of a turbine is precisely that: to do work.

• Heat transfer in condenser. A condenser is just a place for steam to convert to liquid water. It is a
passive device. So 4w1 = 0. The first law reduces to
4q1 = he − hi = h1 − h4 . (4.323)
We know h4 . And h1 is a compressed liquid state. Let us approximate h1 using the saturated liquid
value at T = 45 ◦ C; thus, h1 = 188.45 kJ/kg. So

kJ kJ kJ
4q1 = 188.45 − 2361.7 = −2173.25 . (4.324)
kg kg kg
Note:
– the heat transfer 4q1 < 0 because thermal energy leaves the condenser to the surroundings. This
is why cooling lakes near power plants are warmer than they otherwise would be.
This problem is an example of a thermodynamic cycle. This particular cycle is important because it is
the foundation of most electric power generation in the world, as well as many other applications. We
can introduce the concept of a cycle thermal efficiency, η as the ratio of what we want to what we pay
for. Here, we want net work, that is the difference of the turbine work and the pump work, and we pay
for the heat transfer into the boiler. Mathematically, we have
what we want
η = , (4.325)
what we pay for
wcycle
= , (4.326)
qin
wturbine + wpump
= , (4.327)
qboiler
3aw4 + 1w2
= , (4.328)
q
2 3
640.8 kJ kg + −4
kJ
kg
= kJ
, (4.329)
2831.1 kg
= 0.225. (4.330)
This is a low efficiency by modern standards. Later we shall use the second law of thermodynamics to
calculate the peak efficiency of a steam power plant. That analysis will reveal that the peak efficiency
for this plant is η = 0.445.
In a spirit similar to that of double-entry accounting, we summarize the key energy balances in
Table 4.1. Note there is some small round-off error due mainly due to significant digits and interpolation
that has been ignored in Table 4.1. Each row must maintain the control volume balance ∆h = q − w.
Each column must add to form its total. Because h is a state property, the net ∆h for the cycle must
be zero. And to satisfy the first law for the cycle, the total q must equal the total w.
There is another useful way to formulate the thermal efficiency. We can begin with Eq. (4.328) and
say
3aw4 + 1w2
η= . (4.331)
q
2 3

4.6. PREVIEW: EQUATIONS OF CONTINUUM MECHANICS 209

kJ kJ kJ
Process ∆h kg q kg w kg
1→2 4 0 -4
2→3 2831.1 2831.1 0
3 → 3a -21 -21 0
3a → 4 -640.8 0 640.8
4→1 -2173.25 -2173.25 0
Total 0 636.8 636.8
Table 4.1: First law analysis summary for our non-ideal Rankine cycle.
Using Eqs. (4.312,4.314), we can reform Eq. (4.331) as

(h3a − h4 ) + (h1 − h2 )
η = , (4.332)
h3 − h2
h3a − h3 + h3 − h4 + h1 − h2
= , (4.333)
h3 − h2
(h3a − h3 ) + (h3 − h2 ) + (h1 − h4 )
= , (4.334)
h3 − h2
(h3 − h3a ) + (h4 − h1 )
= 1− . (4.335)
h3 − h2
And with the net heat entering as qin = h3 − h2 and the net heat loss as qout = (h3 − h3a ) + (h4 − h1 ),
we can express Eq. (4.335) as
qout qin − qout wcycle
η =1− = = . (4.336)
qin qin qin
The local steam power plant at the University of Notre Dame runs on a variant of the Rankine
cycle. This particular power plant relies on a local source, St. Joseph’s Lake, for its cooling water. On
a cold winter day, the effect of the power plant on St. Joseph’s Lake, relative to the nearby St. Mary’s
Lake, that has no connection to the power plant, is dramatic. Fig. 4.25 shows the two lakes on the
same day. The power plant also generates steam for heating of the buildings on campus that requires
many underground steam lines. The effect of heat loss is the lines is obvious on many winter days on
the Notre Dame campus as is evident from Fig. 4.26.
4.6 Preview: equations of continuum mechanics

This course focuses on mass and energy conservation coupled with equations of state for
systems that are well modeled as quasi-equilibrium processes. We can do many important
problems with these tools. However, there are many problems that we cannot do with these
tools, e.g. problems with coupled time- and space-dependency, or problems with detailed
material motion.
Let us, as a preview for future courses, write various related sets of partial differential
equations that can couple equilibrium thermodynamics with mechanics.

Figure 4.25: St. Joseph’s Lake (left) and St. Mary’s Lake (right) on the University of Notre
Dame campus, 14 February 2010, demonstrating the effect of exchanging heat with the
condenser of the Notre Dame power plant near St. Joseph’s Lake.
Figure 4.26: Zones of melted snow in regions near underground steam lines on the University
of Notre Dame campus, 14 February 2010.

4.6.1 Full set

We first give a summary of a reasonably complete and general set of equations for a continuum
material. One way to write these equations is as follows:
∂ρ
+ ∇ · (ρv) = 0, mass, (4.337)
∂t
∂
(ρv) + ∇ · (ρvv) = ρg − ∇P + ∇ · τ ,
∂t
linear momenta, (4.338)
T
τ = τ, angular momenta, (4.339)

∂ 1 1
ρ u+ v·v + ∇ · ρv u + v · v = −∇ · q − ∇ · (P v) + ∇ · (τ · v)
∂t 2 2
+ρv · g, energy. (4.340)
Equations (4.337-4.340) are the axioms of mass conservation, linear momenta conservation,
angular momenta conservation, and energy conservation, respectively. They are valid for
any pure material, be it solid, liquid, or gas, as long as we are at velocities small relative to
the velocity of light. New variables here include the deviatoric stress tensor τ , and the heat
flux vector q. The vector g is the constant gravitational acceleration. It is easily modified
for variable gravitational acceleration.
The conservation axioms are necessary but insufficient to determine all flow variables.
They must be supplemented by constitutive relations. Constitutive relations specify the
actual material. A general set is given here.
P = P (ρ, T ), thermal EOS, (4.341)

u = u(ρ, T ), caloric EOS, (4.342)
q = q(T, ∇T, . . . ), heat flux, (4.343)
τ = τ (T, ∇v, ∇x, . . . ), stress. (4.344)
Equation (4.341) is a thermal equation of state. An example is the ideal gas law P = ρRT .
Equation (4.342) is a caloric equation of state. An example is a calorically perfect ideal
gas, u = cv (T − To ) + uo . Equation (4.343) is a relation between the heat flux vector and
other state variables. An example is Fourier’s law, q = −k∇T. Equation (4.344) is a relation
between the deviatoric stress and a variety of variables. For example, a Newtonian fluid
obeying Stokes’ assumption has τ = µ((∇v + (∇v)T ) − (1/3)(∇ · v)I). This relates stress to
strain rate. On the other hand, the stress in a solid would be related to the strain, instead
of the strain rate.
4.6.2 Static solids equations

For a static solid, we take v = 0 and the density constant. The mass equation becomes
irrelevant, as does the angular momenta equation. The linear momenta equation reduces to

a force balance, because inertia is zero. We take the total stress tensor σ = τ − P I, where
I is the identity matrix.
∇ · σ = −ρg, linear momenta, (4.345)

E ν
σ = e+ tr(e)I , stress-strain relation. (4.346)
1+ν 1 − 2ν
Here, E is the modulus of elasticity, ν is Poisson’s ratio (−1 ≤ ν ≤ 1/2), and e is the strain
tensor.
4.6.3 Incompressible fluid mechanics equations

In the discipline of incompressible fluid mechanics, we typically take ρ to be a constant along
a given fluid particle pathline, consider fluid properties such as viscosity and conductivity to
be constant, and reduce our equations to the following set for an ideal gas
∇ · v = 0, mass, (4.347)

∂v
ρ + v · ∇v = ρg − ∇P + µ∇2 v, linear momenta, (4.348)
∂t
τT = τ, angular momenta, (4.349)

∂T ∂P
ρcP + v · ∇T = + v · ∇P + k∇2 T + 2µD : D, energy, (4.350)
∂t ∂t
1
D = ∇v + (∇v)T , strain rate, (4.351)
2
τ = 2µD, stress-strain rate, (4.352)
q = −k∇T, Fourier’s law. (4.353)
The mass and linear momenta equations form an independent set. The energy equation is
coupled to to mass and momenta equations because of the velocity vector.
4.6.4 Compressible fluid mechanics equations

In compressible fluid mechanics, we account for density changes and thermodynamics, but
usually neglect gravity, viscosity and heat conduction. Our equations reduce to, for a CPIG,

∂ρ
+ v · ∇ρ + ρ∇ · v = 0, mass, (4.354)
∂t

∂v
ρ + v · ∇v = −∇P, linear momenta, (4.355)
∂t

∂u ∂v
+ v · ∇u = −P + v · ∇v , energy, (4.356)
∂t ∂t
P = ρRT, thermal state, (4.357)
u = cv T + u o , caloric state. (4.358)

Figure 4.27: Sketch of one-dimensional energy diffusion.
Notice the energy equation is simply the familiar du/dt = −P dv/dt, when d/dt is interpreted
as the material derivative ∂/∂t + v · ∇. This implies that du = −P dv and that the flow
is adiabatic; that is, reversible work is the only available mechanism to change the internal
energy. Also note that the mass equation is dρ/dt = −ρ∇ · v. Thus, the incompressibility
condition ∇ · v = 0 implies dρ/dt = 0.
4.6.5 Heat transfer in incompressible static solids

For heat transfer in static solids, we require v = 0. Moreover, there is no work. We take
a calorically perfect solid with constant thermal conductivity k that obeys Fourier’s law
q = −k∇T and get the first law of thermodynamics to reduce to the simple scalar equation
known to Fourier in the early nineteenth century:
∂T
ρc = k∇2 T, energy. (4.359)
∂t
This equation is known as the heat equation. Let us see how it arises for a stationary solid
whose properties only vary with scalar distance x and time t.
Let us perform a physical derivation of the heat equation, a partial differential equation
that is a manifestation of the first law of thermodynamics in combined with an experimentally
known relationship for the heat flux. The equation will describe the process of energy
diffusion. As depicted in Fig. 4.27 consider a volume V of dimension A by ∆x. We describe
the heat flux in the x direction as qx . At the left boundary x1 , we have diffusive heat flux
that we notate as qx |x1 . At the right boundary at x1 + ∆x, we have diffusive heat flux out
of qx |x1 +∆x . Recall the units of heat flux are J/m2 /s. We assume the walls are thermally
insulated; thus, there is no heat flux through the walls. The total energy within the volume
is U with units of J. We also have the specific energy u = U/m with units of J/kg, where m
is the mass enclosed within V . We suppress any addition of mass into V , so that m can be
considered a constant.

We allow for the specific energy u to vary with space and time: u = u(x, t). Certainly it
is possible to define an average specific energy within the volume, denoted with an over-bar:
Z x1 +∆x
1
u(t) = u(x, t) dx. (4.360)
∆x x1
Our physical principle is the first law of thermodynamics, that being
| {z out} .
change in total energy = heat in − work (4.361)
=0
There is no work for our system. But there is heat flux through the system boundaries. In
a combination of symbols and words, we can say
total energy @ t + ∆t − total energy @ t = energy flux in − energy flux out . (4.362)
| {z } | {z }
unsteady | {z } and diffusion
advection
=0
Advection is associated with velocity, which is zero, so there is no advection. Mathematically,

we can say
U|t+∆t − U|t = − (Uf lux out − Uf lux in ) , (4.363)

ρA∆x u|t+∆t − u|t = − qx |x1 +∆x − qx |x1 |{z} A∆t , (4.364)
| {z } | {z } | {z } 2
kg (m s)
J/kg J/m2 /s

u|t+∆t − u|t qx |x1 +∆x − qx |x1
ρ = − . (4.365)
∆t ∆x
Now, let ∆x → 0 and ∆t → 0 so as to induce mean values to be local values, and finite
differences to be derivatives, yielding a differential representation of the first law of thermo-
dynamics of
∂u ∂qx
ρ =− . (4.366)
∂t ∂x
Now, let us invoke some relationships known from experiment. First, the specific internal
energy of many solids is well modeled by a calorically perfect state equation:
u = cT + u0 . (4.367)
Here c is the constant specific heat with units J/kg/K, T is the temperature with units of
K, and u0 is a constant with units of J/kg whose value is unimportant, as for nonreactive
materials, it is only energy differences that have physical importance. The caloric state equa-
tion simply states the specific internal energy of a material varies linearly with temperature.

Experiment reveals that Fourier’s law is a good model for the heat flux in many materials.
In one dimension, this is
∂T
qx = −k . (4.368)
∂x
Here k is the so-called thermal conductivity of a material. It has units J/s/m/K. It is
sometimes dependent on T , but we will take it as a constant here. For agreement with
experiment, we must have k ≥ 0. The equation reflects the fact that heat flow in the positive
x direction is often detected to be proportional to a field in which temperature is decreasing
as x increases. In short, thermal energy flows from regions of high temperature to low
temperature.
Equation (4.368) along with the caloric state equation, Eq. (4.367) when substituted into
Eq. (4.366) yields

∂ ∂ ∂T
ρ (cT + u0 ) = − −k , (4.369)
∂t ∂x ∂x
∂T ∂2T
ρc = k 2, (4.370)
∂t ∂x
∂T k ∂2T
= . (4.371)
∂t ρc ∂x2
|{z}
α
Here we have defined the thermal diffusivity α = k/ρ/c. Thermal diffusivity has units of
m2 /s. In final form we have
∂T ∂2T
=α 2. (4.372)
∂t ∂x
Equation (4.372) is known as the heat equation. It really is a manifestation of the first law
of thermodynamics.


Chapter 5
The second law of thermodynamics
Read BS, Chapter 5
Conservation of mass and energy are fine concepts that allow us to quantify and predict well
many phenomena that are observed in nature. And if a phenomenon can be repeated, it
becomes subject to prediction, and can be thought of as a science.
However, conservation of mass and energy, by themselves, admit as possibilities phenom-
ena that are not observed in nature! For instance consider an isolated system composed of
two equal masses of liquid water. See Fig. 5.1. The first is at TA = 310 K, the second is at
A B
T = 320 K T = 280 K
A B
not observed in nature
T = 310 K T = 290 K
A B
T = 300 K T = 300 K
observed in nature
Figure 5.1: Sketch of two scenarios, both of which satisfy mass and energy conservation.
TB = 290 K. A long time elapses. Because the combined system is isolated, there are no
217
218 CHAPTER 5. THE SECOND LAW OF THERMODYNAMICS
external heat or work exchanges with the environment. But we will allow heat exchanges
between mass A and mass B. Consider two possibilities, both admitted by mass and energy
conservation, as t → ∞:
• TA → 320 K, TB → 280 K. The thermal energy that is gained by A is lost by B, such

that the net energy is conserved and the first law is satisfied. This is never observed
in nature.
• TA → 300 K, TB → 300 K. The thermal energy that is lost by A is gained by B, so

once again the first law is satisfied. This is always observed in nature.
So mass conservation and the first law of thermodynamics, both of which speak to this
gedankenexperiment, are insufficient to guarantee that we will predict what is observed in
nature. We need another axiom!
In a similar way, there are a variety of phenomena that may satisfy mass and energy
conservation, but are not observed in nature. Some include
• water running uphill without an external assist,
• CO2 and H2 O reacting spontaneously to form CH4 and O2 , and
• air separating into its constituents spontaneously.
5.1 Statements of the second law

The second law of thermodynamics is an attempt to provide a single all-encompassing state-
ment that expands our thermodynamic theory so as to predict the just-described behavior.
Though our statement of the second law will be simple enough, it will be obtuse and some-
times difficult to reconcile with nature. It is also a profound concept that has wide ranging
ramifications. Its origins are firmly rooted in the engineering sciences, as it was motivated
by optimization of steam engines. However, it has found applications in many realms of
physics, chemistry, ecology, economics, computer science, and other fields.
In his influential essay contrasting scientific and humanistic cultures, C. P. Snow1 sees
understanding of the second law as an indicator of scientific literacy and goes on to chide
his humanist colleagues:
A good many times I have been present at gatherings of people who, by the stan-
dards of the traditional culture, are thought highly educated and who have with
considerable gusto been expressing their incredulity at the illiteracy of scientists.
Once or twice I have been provoked and have asked the company how many of
them could describe the Second Law of Thermodynamics, the law of entropy. The
response was cold: it was also negative. Yet I was asking something which is
1
C. P. Snow, 1959, The Two Cultures, Cambridge U. Press, Cambridge (reprinted 1998).

5.1. STATEMENTS OF THE SECOND LAW 219
about the scientific equivalent of: ‘Have you read a work of Shakespeare’s?’ I
now believe that if I had asked an even simpler question —such as, What do you
mean by mass, or acceleration, which is the scientific equivalent of saying, ‘Can
you read?’ —not more than one in ten of the highly educated would have felt that
I was speaking the same language. So the great edifice of modern physics goes up,
and the majority of the cleverest people in the western world have about as much
insight into it as their Neolithic ancestors would have had.
Let us summarize some more reasons for studying the second law:
• It predicts the direction in time of processes.
• It aids in determining equilibrium conditions.
• It allows one to determine peak performance of practical devices.
• It enables one to frame analysis of the factors that inhibit the realization of peak
performance.
• It allows a rational definition of the absolute temperature scale.
• It has implications beyond engineering in physics, philosophy, economics, computer

science, etc.
5.1.1 Entropy-based statement

There are many ways to state the second law of thermodynamics. One statement is as
follows:
• Second law of thermodynamics: The entropy of an isolated system can never

decrease with time.
This definition begs the question, what is entropy? A formal definition will be deferred to
the next chapter. Let us loosely define it here as a measure of the so-called randomness (or
disorder) of a system, with high randomness corresponding to high entropy. Low randomness
or low disorder often corresponds to low entropy.
Interpreted in another way, structure or order requires energy input to be realized, while
over time, without continued maintenance, structure and order decay. The formulation of
the second law we adopt will be robust enough to prevent us from predicting water to run
uphill, methane to spontaneously form from carbon dioxide and water, or air to separate into
its constituents. It will also be seen to be an important principle for predicting the optimal
behavior of a wide variety of engineering devices.
All that said, it should be noted that the equivalence of entropy with disorder, while
useful and common, is likely not universal. Certainly Wright2 characterizes it as “a highly
2
P. G. Wright, 1970, “Entropy and disorder,” Contemporary Physics, 11(6): 581-588.

contentious opinion” and discusses counter-examples, especially as related to molecular level

phenomena. Wright’s arguments are reinforced by Styer,3 who concludes that the notion of
“entropy as disorder” be used only in conjunction with the notion of “entropy as freedom.”
Freedom here is to be interpreted as the ability to acquire a variety of states: if only one state
is available, freedom is severely restricted, and entropy is low; if many states are available,
freedom is widespread, and entropy is high. Both terms, “disorder” and “freedom” are shown
by Styer to have alternate interpretations that render both imperfect metaphors for entropy.
5.1.2 Clausius statement

Clausius, the German mathematical physicist who probably did the most to cast thermody-
namics on a scientific basis, gives a more precise statement of the second law:
• Second law of thermodynamics: “Heat cannot, of itself, pass from a colder to a

hotter body.”4
The original German version appeared earlier5 and is reproduced in Fig. 5.2.
Figure 5.2: Image of the original 1854 appearance of the Clausius form of the second law.
The Clausius formulation of the second law is easy to understand in engineering terms
and is illustrated schematically in Fig. 5.3. Note that air conditioners move heat from cold
regions to hot regions, but that work input is required.
5.1.3 Kelvin-Planck statement

Another statement of the second law is inspired by statements of the nineteenth century
scientists, Kelvin and Planck, depicted in Fig. 5.4. The so-called Kelvin-Planck statement,
a modern rendition of earlier statements, is often given as
• Second law of thermodynamics: It is impossible for any system to operate in

a thermodynamic cycle and deliver a net amount of work to its surroundings while
receiving an energy transfer by heat from a single thermal reservoir.
3
D. F. Styer, 2000, “Insight into entropy,” American Journal of Physics, 68(12): 1090-1096.
4
R. Clausius, 1879, The Mechanical Theory of Heat, Macmillan, London, p. 78.
5
R. Clausius, 1854, “Ueber eine veränderte Form des zweiten Hauptsatzes der mechanischen Wärmethe-
orie,” Annalen der Physik und Chemie, 169(12): 481-506.

Q Q
cold hot cold hot
is impossible without W is possible without W
Figure 5.3: Schematic of the Clausius statement of the second law of thermodynamics.
Thomson’s original 1851 statement6 is somewhat different. Planck’s 1897 statement7 is

closer. Both are reproduced in Fig. 5.5.
The Kelvin-Planck formulation of the second law is easy to understand in engineering
terms and is illustrated schematically in Fig. 5.6. In this figure and chapter, we often
simplify the notation for heat transfer and work. It is understood here that Q and W are
path-dependent. For the schematic of Fig. 5.6, the first law, neglecting changes in kinetic
and potential energy, states that
U2 − U1 = Q − W. (5.1)
But we have specified that the process is a cycle, so U1 = U2 , and thus the first law holds
Q = W. (5.2)
Now, the second law, for this scenario, holds that positive Q cannot be delivered, that gives,
for an engine in contact with a single thermal reservoir,
Q ≤ 0, W ≤ 0. (5.3)
In informal language, the Kelvin-Planck statement says
• you can turn all the work into heat, but
• you cannot turn all the heat into work.
5.1.4 Carathéodory statement

Another form of the second law was given by the mathematician and advocate for the ax-
iomatic approach to thermodynamics, Carathéodory, depicted in Fig. 5.7. The Carathéodory
6
W. Thomson (later Lord Kelvin), 1851, “On the dynamical theory of heat, with numerical results de-
duced from Mr. Joule’s equivalent of a thermal unit, and M. Regnault’s observations on steam,” Transactions
of the Royal Society of Edinburgh, 20: 261-268; 289-298.
7
M. Planck, 1897, Vorlesungen über Thermodynamik, Walter de Gruyter, Berlin; reprinted in
English translation as Treatise on Thermodynamics, Dover, New York, p. 89.

a) b)
Figure 5.4: a) William Thomson (Lord Kelvin) (1824-1907), Ulster-born Scottish sci-
entist who had profound impact on nineteenth century science including thermodynam-
ics. Image from https://mathshistory.st-andrews.ac.uk/Biographies/Thomson.html,
b) Max Karl Ernst Ludwig Planck (1858-1947), German physicist. Image from
https://commons.wikimedia.org/wiki/File:Max Planck (1858-1947).jpg.
a) Kelvin b) Planck
Figure 5.5: Images of a) Thomson’s 1851 and b) Planck’s 1897 statements of the second law
of thermodynamics.
statement8 is
• Second law of thermodynamics: In the neighborhood of any equilibrium state

of a thermodynamic system, there are equilibrium states that are adiabatically inac-
cessible.
Demonstration of its equivalence to other statements is not straightforward, and we shall

not consider it in any further detail in these notes; nevertheless, the notions embodied in
this statement as well as the school of thought in thermodynamics that has grown around
the work of its author has benefits associated with casting thermodynamics in the context of
more general mathematical ideas, thus making results from modern mathematics more easily
applicable to thermodynamics. The work of Carathéodory is often dismissed as somehow too
mathematical. Indeed, his approach to thermodynamics in general requires a mathematical
8
C. Carathéodory, 1909, “Untersuchungen über die Grundlagen der Thermodynamik,” Mathematische
Annalen, 67: 355-386.

thermal reservoir thermal reservoir
Q Q
cyclic W cyclic W
engine engine
first law: Q=W, because ΔU=0 for cycle first law: Q=W, because ΔU=0 for cycle
second law: This scenario cannot be. second law: This scenario can be.
Figure 5.6: Schematic of the Kelvin-Planck statement of the second law of thermodynamics.
Figure 5.7: Constantin Carathéodory (1873-1950), Greek mathematician. Image from

https://en.wikipedia.org/wiki/Constantin Caratheodory.
sophistication beyond that needed to understand the more common Clausius or Kelvin-
Planck formulations. The interested reader can consult a relevant discussion.9
5.1.5 Judeo-Christian statement

One finds in Genesis 3:19 the admonition given by the Catholic church in its Ash Wednesday
services,
• Modulated second law of thermodynamics: Remember man that thou art

dust, and unto dust thou shalt return.
9
M. W. Zemansky, 1966, “Kelvin and Caratheodory–a reconciliation,” American Journal of Physics,
34(10): 914-920.

5.1.6 Historian-inspired statement

The great American historian and generalist, Henry Brooks Adams (1838-1918) grandson
and great-grandson of American presidents, wrote a detailed and entertaining essay on the
second law. Among his many comments is included10
• Adams’ Second Law of Thermodynamics: “...but to the vulgar and ignorant

historian it meant only that the ash-heap was constantly increasing in size.”
5.1.7 Literature-inspired statement

Though he probably did not intend it for engineering, the Nobel literature laureate Chinua
Achebe’s most famous novel has a title that also serves as a rough-and-ready statement of
the second law:
• Informal second law of thermodynamics: Things fall apart.
The title of the novel is drawn from a line in Nobel literature laureate William Butler Yeats’
apocalyptic 1921 poem, The Second Coming.
5.1.8 Food service-inspired statement

• Second law of thermodynamics, special case: Untended food rots.
5.2 Reversible and irreversible processes

We shall find it useful to have in hand definitions for so-called reversible and irreversible
processes. Let us take
• Reversible process: A process in which it is possible to return both the system

and surroundings to their original states.
• Irreversible process: A process in which it is impossible to return both the system

and surroundings to their original states.
Now, it may be possible to restore the system to its original state but not the surroundings
(or the surroundings to its original state but not the system). Such a process is irreversible.
We shall often study reversible processes as they represent an ideal of the most we can
ever hope to achieve. Some common engineering idealizations of reversible processes include
• frictionless motion,
• ideal inviscid flow of a fluid over an airfoil.

10
H. B. Adams, 1910, A Letter to American Teachers of History, J. H. Furst, Washington.

5.3. ANALYSIS OF CARNOT HEAT ENGINES 225
Now, everything in the real world deviates from the ideal. In flow over a wing, friction in
the form of viscosity causes local irreversible heating of the air near the wing and the wing
itself. Often in the real world these irreversibilities are confined to small regions and often
do not largely affect the motion of the body.
If the world in which we live were reversible, we would realize some benefits, but ulti-
mately life would be impossible. In a reversible world
• Pendulum clocks would never need to be rewound.
• Electricity would flow without generating heat; thus, computer fans would be unnec-
essary, among other things, but
• We would not be able to walk!
5.3 Analysis of Carnot heat engines

Here, we will present what amounts to a version of the discussion of the 1850s inspired by
the original work of Carnot11 depicted along with his writings in Fig. 5.8, for heat engines.
Carnot’s 1824 work was done before any formal notions of the first and second laws had been
systematized.
The informal statement of the Kelvin-Planck version of the second law is that you cannot
turn all the heat into work. Now, an engineer often wants to harvest as much as possible
of the thermal energy of combustion and convert it into mechanical energy. Kelvin-Planck
simply says we cannot have it all. But it lets us have some! In fact if we only harvest a
portion of the thermal energy for work and reject the rest in the form of thermal energy, we
can satisfy the second law. We show this schematically in Fig. 5.9. The first law for this
system is
U2 − U1 = QH − QL − W. (5.4)
Note that here, we are thinking of W , QH , and QL as all positive. If we were rigorous with
our sign convention, we would have reversed the arrow on QL because our sign convention
always has positive work entering the system. However,
• following a common practice, the rigorous sign convention is traditionally abandoned

for analysis of heat engines!
Now, we are requiring a cyclic process, so U1 = U2 ; thus, Eq. (5.4) reduces to
W = QH − QL . (5.5)
11
S. Carnot, 1824, Réflexions sur la Puissance Motrice du Feu et sur les Machines propres à Développer
cette Puissance, Bachelier, Paris. (English translation, 2005, Reflections on the Motive Power of Fire, Dover,
Mineola, New York). 1897 English translation.

Figure 5.8: Sadi Nicolas Léonard Carnot (1796-1832), French engineer

whose analysis formed the basis for modern thermodynamics; image from
https://mathshistory.st-andrews.ac.uk/Biographies/Carnot Sadi.html and the
title page from his magnum opus.
Now, recall we previewed the idea of thermal efficiency, η, in Eq. (4.330):
what you want

η= . (5.6)
what you pay for
We recast it for the scenario of Fig. 5.9, where we want W and we pay for QH :
W
η= . (5.7)
QH
Now, we use Eq. (5.5) to eliminate W in Eq. (5.7) and get
QH − QL QL
η= =1− . (5.8)
QH QH
Equation (5.8) is analogous to the earlier Eq. (4.336). Now, if QL = 0, we get η = 1, and
our engine does a perfect job in converting all the heat into work. But if QL = 0, we violate
the Kelvin-Planck version of the second law! So we must reject some heat. Another version

5.3. ANALYSIS OF CARNOT HEAT ENGINES 227
hot thermal reservoir
QH
cyclic W
engine
QL
cold thermal reservoir
first law: W = QH - QL, because ΔU = 0 for cycle
second law: This scenario can be.
Figure 5.9: Schematic of a realizable heat engine.
of the Kelvin-Planck statement then is

QL
η =1− < 1. (5.9)
QH
It is possible to prove the following corollary to the Kelvin-Planck statement:
ηirreversible ≤ ηreversible , (5.10)
for cycles operating between the same thermal reservoirs.

And there is a second corollary, applicable for two different cycles, both reversible, and both
operating between the same thermal reservoirs:
ηreversible,1 = ηreversible,2. (5.11)
Example 5.1
Prove the first corollary, Eq. (5.10).
We show this schematically in Fig. 5.10. In Fig. 5.10a, we have two cyclic engines operating
between the same thermal reservoirs. One is reversible, and extracts reversible work, WR from an input
of QH . The other engine is irreversible and extracts irreversible work WI from the same heat input
QH . Because WI 6= WR , the two engines reject a different amount of heat to satisfy the first law. The
reversible engine rejects QL , and the irreversible engine rejects Q′L .

hot thermal reservoir hot thermal reservoir
cyclic cyclic cyclic cyclic

engine engine engine engine
cold thermal reservoir cold thermal reservoir
a) original configuration b) reversed configuration
Figure 5.10: Schematic of a heat engine used in proof of the first corollary of the Kelvin-
Planck statement of the second law of thermodynamics.
In Fig. 5.10b, we take formal advantage of the reversibility of one of the engines to reverse all
processes, as shown by the different nature of all the arrows. In fact we have created a refrigerator, a
device in which a work input is used to move thermal energy from a cold region to a hot region. Now,
in this configuration, an identical QH is added and removed from the hot thermal reservoir. So the net
effect into the system defined by the combined two cyclic engines is that there is no net heat transfer to
the combined system. We show this schematically in Fig. 5.11. Now, the combined effect of Fig. 5.11 is
cyclic cyclic
engine engine
Figure 5.11: Net effect of the configuration from Fig. 5.10b.
in a form suitable for application of the Kelvin-Planck form of the second law. We have the net work
as
W = WI − WR . (5.12)
And the Kelvin-Planck statement, from Eq. (5.3) holds that
W ≤ 0. (5.13)

5.4. THE ABSOLUTE TEMPERATURE SCALE 229
Imposing Eq. (5.12) into Eq. (5.3), we get
WI − WR ≤ 0, (5.14)
WI ≤ WR . (5.15)
Now, the efficiency of the reversible engine from Fig. 5.10a is

WR
ηreversible = . (5.16)
QH
The efficiency of the irreversible engine from Fig. 5.10a is
WI
ηirreversible = . (5.17)
QH
And because QH is the same in both, and WI ≤ WR , we have proved the first corollary:
ηirreversible ≤ ηreversible . (5.18)
The second corollary, Eq. (5.11), can be proved in a similar manner.
5.4 The absolute temperature scale

The second corollary to the Kelvin-Planck statement holds that all reversible engines op-
erating between the same thermal reservoirs have the same η. This is independent of any
details of the cycle or the materials involved. This implies that
• The thermal efficiency, η, should depend only on the character of the reservoirs in-
volved.
Specifically, we will define η in terms of what we will call the temperature of the reservoir.
This is the classical macroscopic interpretation of temperature. Later statistical theories give
it the additional interpretation as a measure of the average translational kinetic energy of
molecules of the system. But that is not our approach here! Now, we might suppose that
this new thermodynamic property, temperature, should somehow be a measure of how much
heat is transferred from one reservoir to another. Moreover, each reservoir will have its
own temperature. The hot reservoir will have temperature TH ; the cold reservoir will have
temperature TL . So we are then saying that
η = η(TH , TL ). (5.19)
As of yet, this functional form is unspecified. Substituting this form into our earlier Eq. (5.8),
we get
QL
η(TH , TL ) = 1 − . (5.20)
QH

This can only be true if QL and QH have some relation to TL and TH . So let us propose a
useful definition. We insist that our temperatures take the form of that for a Carnot cycle
TL QL
= . (5.21)
TH QH
This is just a definition that cannot be argued. Its utility will be seen as its justification,
but nothing more. Eq. (5.21) is valid only in the context of a Carnot cycle, and not for other
cycles.
Our logic train is that we observe heat engines, such as steam engines seen by Carnot
in the early 1800s, doing work as a result of heat transfers. That effect, work, must have a
cause. And we are going to assert that the cause is affiliated with a temperature difference.
So far our temperature has been defined only in terms of a ratio. Let us make an arbitrary
choice to avoid ratios. We take, for convenience, the temperature of the triple point of water
to be 273.15 K. Thus for any system, the local T is

Q
T = (273.15 K) . (5.22)
Qtriple point reversible cycle
This implies we can connect our heat engine to a reservoir maintained at the triple point
temperature of water, and measure the associated Qs for the heat engine. With our definition,
Eq. (5.21), our thermal efficiency, Eq. (5.8), becomes
TL
η =1− . (5.23)
TH
This famous formula is the thermal efficiency for an idealized heat engine; it is not valid for
other heat engines. This formula is not found in Carnot’s original work; nor is it straightfor-
wardly presented in the later works of the 1850s. Clausius puts most of the pieces in place,12
so much so that Müller and Müller13 attribute the formula to him; but it is not directly seen
in his 1854 study; see Truesdell.14
Note
lim η = 1, (5.24)
TL →0
lim η = 1. (5.25)
TH →∞
These two statements have practical importance. While we would like to drive our efficiency
to be as close to unity as possible, nature limits us. Generally, we have little to no control
12
R. Clausius, 1854, “Ueber eine veränderte Form des aweiten Hauptsatzes der mechanischen Wärmethe-
orie,” Annalen der Physik und Chemie, 169(12): 481-506.
13
I. Müller and W. H. Müller, 2009, Fundamentals of Thermodynamics and Applications with Historical
Annotations and many Citations from Avogadro to Zermelo, Springer, Berlin, p. 131.
14
C. Truesdell, 1980, The Tragicomical History of Thermodynamics, 1822-1854, Springer, New York,
p. 330.

5.5. ANALYSIS OF CARNOT REFRIGERATORS AND HEAT PUMPS 231
over the environmental temperature TL , so it is a lower bound, usually around TL ∼ 300 K.

And material properties for engines limit TH . For many metals, TH ∼ 1500 K is approaching
values where material strength is lost. So a practical upper bound based on these numbers
tells us η ∼ 1 − (300 K)/(1500 K) = 0.8 is maybe the most we can expect. We plot η as a
function of TH for fixed TL = 300 K in Fig. 5.12. For real systems, with irreversible features,
maximum efficiency = 1
1.0
0.8
0.6
0.4
0.2
0.0
0 2000 4000 6000 8000 10000
Figure 5.12: Plot of idealized thermal efficiency as a function of hot thermal reservoir tem-
perature with TL = 300 K.
the values are much worse.
5.5 Analysis of Carnot refrigerators and heat pumps

A refrigerator or heat pump is a device that, with work input, moves thermal energy from
cold regions to hot regions. Without the work input, this could not be achieved, as it
would violate the Clausius statement of the second law of thermodynamics. We show this
schematically in Fig. 5.13. For a refrigerator, we define a coefficient of performance β,
sometimes called COP , as
what you want QL
β= = , Carnot refrigerator. (5.26)
what you pay for W
Now, the first law for this cycle gives W = QH − QL , so
QL 1 1
β= = QH
= TH
, Carnot refrigerator. (5.27)
QH − QL QL
−1 TL
−1
Note that
β ≥ 0, (5.28)
for TH /TL ≥ 1. In addition, it is possible to have β > 1 if TH /TL < 2. Because we reserve
efficiencies to have 0 ≤ η ≤ 1, the COP is not really an efficiency. But it is a useful measure
that is used as an industry standard for refrigerators.

cyclic W
engine
Figure 5.13: Schematic of a realizable refrigerator or heat pump.
For heat pumps, we want to bring QH into a warm room from a cold outdoors to make
the warm room warmer. So it has a related COP , that we define as β ′ :
QH
β′ = , Carnot heat pump. (5.29)
W
Again the first law gives W = QH − QL , so
QH 1 1
β′ = = QL
= , Carnot heat pump. (5.30)
QH − QL 1 − QH 1 − TTHL
For TL /TH ≤ 1, we have β ′ ≥ 1.

Note for both refrigerators and heat pumps, as W → 0, both β → ∞ and β ′ → ∞, and
we transfer thermal energy from a cold reservoir to a hot one, in violation of the Clausius
statement of the second law.
Example 5.2
The inside of a refrigerator is held at TL = 3 ◦ C. The surroundings are at TH = 27 ◦ C. We must
remove 1.25 kW of thermal energy from the inside of the refrigerator to balance the thermal energy
entering from the environment. Find the best possible coefficient of performance β and the minimum
power requirement to operate the refrigerator.
First, we need to convert to an absolute scale. So TL = 3 + 273.15 = 276.15 K, and TH = 300.15 K.

The problem is sketched in Fig. 5.14. We have
1 1
β= TH
= 300.15 K
= 11.5063. (5.31)
TL −1 276.15 K −1

5.6. REJECTED THERMAL ENERGY ON A NATIONAL SCALE 233
cyclic
engine
Figure 5.14: Schematic of a Carnot refrigerator diagram for our example.
Now,
what we want Q̇L
β= = . (5.32)
what we pay for Ẇ
Thus
Q̇L 1.25 kW
Ẇ = = = 0.108367 kW. (5.33)
β 11.5063
Note the rules of heat transfer determine the power load needed to keep the temperature constant. We
have not considered those here. This is the smallest possible engine that would be needed. Inefficiencies
would cause the actual size needed to be higher.
5.6 Rejected thermal energy on a national scale

We see that the second law characterizes the necessary rejection of heat in processes involving
heat and work. Let us briefly examine this on a national scale. Figure 5.15 reports US energy
usage in 2022 from a wide variety of sources directed to a wide variety of applications. The
basic unit of energy here is the quad where 1 quad = 1015 Btu = 1.055 × 1018 J = 1.055 EJ,
where EJ is an “exajoule.” Much can be gleaned from this chart. Overall US energy use
is estimated at 100.3 quad for the year indicated. As far as the second law is concerned,
electricity generation rejects 24.2 quad waste heat per annum and transportation rejects
21.7 quad waste heat per annum. In total, 67.3 quad is rejected, and 33.0 quad is directed
towards a useful intended purpose. Thus, the thermal efficiency of the US in 2022 was
33.0 quad
ηU S = = 0.329. (5.34)
(33.0 quad) + (67.3 quad)

Figure 5.15: Chart of distribution of energy sources and usage in the US in 2022. Data from
Lawrence Livermore National Laboratory, https://flowcharts.llnl.gov/.
Figure 5.16 shows a comparable plot for 2011. One can see evolution of the energy
budget.
Example 5.3
If all the waste heat in the US in 2022 were directed into Lake Michigan, find its temperature rise.
In more convenient units, the waste heat for a given year is

1.055 × 1018 J
Q = (67.3 quad) = 7.10 × 1019 J. (5.35)
quad
Now, Lake Michigan has a volume of 4900 km3 . Therefore the mass of water in Lake Michigan is
roughly
103 m 3

kg
m = ρV = 997 3 4900 km3 = 4.88 × 1015 kg. (5.36)
m km
If all the waste energy were dumped into Lake Michigan, we could expect from a first law analysis to
find a temperature rise of
Q 7.10 × 1019 J
∆T = = = 3.5 K. (5.37)
mcP (4.88 × 1015 kg) 4180 J
kg K

5.6. REJECTED THERMAL ENERGY ON A NATIONAL SCALE 235
Figure 5.16: Chart of distribution of energy sources and usage in the US in 2011. Data from
Lawrence Livermore National Laboratory, https://flowcharts.llnl.gov/.
Locally on the University of Notre Dame campus, both St. Mary’s and St. Joseph’s Lakes would be
vaporized many times over.


Chapter 6
Entropy
Read BS, Chapter 6
Much as the new property energy arose via consideration of the first law of thermodynamics,
we will find it useful to introduce
• Entropy: a thermodynamic property that provides a quantitative measure of the

disorder of a given thermodynamic state,
from consideration of the second law of thermodynamics. The word itself was coined by
Clausius,1 who based it on the combination of ǫν- (en-) “to put into,” and τ ρoπή (tropē),
“turn” or “conversion.” The Greek here is a modification of the original Greek of Clausius,
who employed ὴ τ ρoπὴ. An image of the first use of the word is given in Fig. 6.1.
In some ways entropy is simply a mathematical convenience and a theoretical construct.
However, its resilience is due to the fact that it is useful for engineers to summarize important
design concepts for thermal systems such as steam power plants, automobile engines, jet
engines, refrigerators, heat pumps, and furnaces.
6.1 Theoretical development

Let us motivate the property of entropy by considering Fig. 6.2. Here, we perform our
analysis on a differential basis. We have a thermal reservoir at Tres that delivers a small
amount of heat δQ′ to a reversible cyclic engine, labeled “1.” This engine delivers a small
amount of work δW ′ and rejects a small amount of heat δQ to another reservoir at variable
T , labeled “2.” Let us examine the implications of our temperature definition and the second
law of thermodynamics on this scenario.
1
R. Clausius, 1865, “Ueber verschiedene für die Anwendung bequeme Formen der Hauptgleichungen der
mechanischen Wärmetheorie,” Annalen der Physik und Chemie, 125(7): 353-390.
237
238 CHAPTER 6. ENTROPY
Figure 6.1: Image capturing the first use of the word entropy, from R. Clausius, 1865.
reversible
cyclic
engine
1
T
2
combined system boundary
Figure 6.2: Sketch of heat engine configuration to motivate the development of entropy.

6.1. THEORETICAL DEVELOPMENT 239
We first apply Eq. (5.21) to the configuration of Fig. 6.2:
δQ′ Tres
= . (6.1)
δQ T
Thus
δQ′ δQ
= . (6.2)
Tres T
Now, let us take the combined system, enclosed within the dotted box, to be composed
of 1 and 2. The first law in differential form for the combined system is
dE = δQ′ − δW ′ . (6.3)
Note that we have not yet required the process be cyclic. Also note that δQ is internal and
so does not cross the boundary of the combined system and is not present in our first law
formulation. Rearrange Eq. (6.3) to get
δW ′ = δQ′ − dE. (6.4)
Now, use Eq. (6.2) to eliminate δQ′ in Eq. (6.4):

δQ
δW ′ = Tres − dE. (6.5)
T
Now,
H let us let this configuration undergo a thermodynamic cycle, indicated by the operation
applied to Eq. (6.5): I I I
′ δQ
δW = Tres − dE . (6.6)
T
| {z }
=0
Because E is a thermodynamic property, its cyclic integral is zero. But Q and W are not
properties, so they have non-zero values when integrated through a cycle. Performing the
integration of Eq. (6.6) and realizing that, by definition, Tres is a constant, we get
I
′ δQ
W = Tres . (6.7)
T
Now, we can apply the Kelvin-Planck form of the second law of thermodynamics, Eq. (5.3),
to the configuration of Fig. 6.2; thus, we require simply that
W ′ ≤ 0. (6.8)
That is, we cannot convert all the heat to work, but we can convert all the work to heat.
Because Kelvin-Planck tells us W ′ ≤ 0, Eq. (6.7) tells us
I
δQ
Tres ≤ 0. (6.9)
T

And because Tres > 0, we can divide Eq. (6.9) by it without changing the sense of the
inequality to get a mathematical representation of the second law of thermodynamics:
I
δQ
≤ 0, (6.10)
T
second law of thermodynamics.
If all processes are reversible, we lose the inequality, and get simply
I
δQ
= 0, (6.11)
T
all processes reversible.
Now, let us reconsider Fig. 3.26, recast here as Fig. 6.3, that was used for development
of the path-independent thermodynamic property E. Here, we will use similar concepts to
develop the thermodynamic property of entropy. Let us restrict our discussion to reversible
P
2
A
1 C
Figure 6.3: Sketch of a P − V diagram for various combinations of processes forming cyclic
integrals.
processes, that are the best we could hope for in an ideal world. So we demand that Eq. (6.11)
holds.
Now, from Fig. 6.3, consider starting
H from 1, proceeding on path A to 2, and returning
to 1 via path B. The cyclic integral δQ/T = 0 decomposes to
Z 2 Z 1
δQ δQ
+ = 0. (6.12)
1 T A 2 T B
Now, perform the same exercise going from 1 to 2 on path A and returning on path C,
yielding
Z 2 Z 1
δQ δQ
+ = 0. (6.13)
1 T A 2 T C

6.1. THEORETICAL DEVELOPMENT 241
Now, subtract Eq. (6.13) from Eq. (6.12) to get

Z 1 Z 1
δQ δQ
− = 0. (6.14)
2 T B 2 T C
Thus
Z 1 Z 1
δQ δQ
= . (6.15)
2 T B 2 T C
R1 R2
We can reverse direction and recover the same result, because 2 = − 1 :
Z 2 Z 2
δQ δQ
= . (6.16)
1 T B 1 T C
R2
Because paths B and C are different and arbitrary, but 1 δQ/T is the same on either path,
the integral must be path-independent. It therefore defines a thermodynamic property of the
system. We define that property as entropy, S, an extensive thermodynamic property:
Z 2
δQ
S2 − S1 = . (6.17)
1 T
Note the units of S must be kJ/K in the SI system. We also can scale by the constant mass
m to get the corresponding intensive property s = S/m:
Z 2
δq
s2 − s1 = . (6.18)
1 T
The units for s are kJ/kg/K; note they are the same as cP , cv , and R. In differential form,
we can say
δq
ds = . (6.19)
T
This leads us to
δq = T ds. (6.20)
Integrating Eq. (6.20), we get
Z 2 Z 2
δq = T ds. (6.21)
1 1
Thus, we get
Z 2
1q2 = T ds. (6.22)
1
R2
This is the heat transfer equivalent to 1w2 = 1 P dv. So we see the heat transfer per unit
mass for a process from 1 to 2 is given by the area under the curve in the T − s plane; see
Fig. 6.4. Note if our process lies on a so-called

T T
1 1
pa t
pa
hA
th
B
2 2
areaA areaB
s s
Figure 6.4: Sketch of processes in the T − s plane, with the associated heat transfer.
• Isentrope: a curve on which entropy s is constant,

then by Eq. (6.22), 1q2 = 0; thus, the process is adiabatic. Now, Eq. (6.22) only applies for a
reversible process. Combining these notions, we summarize with the important equivalence:
isentropic = adiabatic + reversible.

For problems in which no chemical reactions are present, we will find ourselves interested
only in entropy differences. For problems with chemical reactions, the absolute values of
entropy will be important. Such values can be obtained by consideration of the
• Third law of thermodynamics: “every substance has a finite positive entropy,
but at the absolute zero of temperature the entropy may become zero, and does so
become in the case of perfect crystalline substances,”
quoted here from Lewis and Randall.2 The law, another axiom of thermodynamics, was
developed over several years by Nernst,3 depicted in Fig. 6.5. It will not be considered
further here.
Because entropy is a thermodynamic property, it can be used to help determine the state.
That is we can say any of the following:
s = s(T, v), s = s(T, P ), s = s(v, x), P = P (T, s), v = v(P, s), ... (6.23)
For two-phase mixtures, we have, as expected,

s − sf
s = sf + xsf g , s = sf + x(sg − sf ), s = (1 − x)sf + xsg , x= . (6.24)
sf g
2
G. N. Lewis and M. Randall, 1923, Thermodynamics and the Free Energy of Chemical Substances,
McGraw-Hill, New York, p. 448.
3
e.g. W. H. Nernst, 1906, Ueber die Berechnung chemischer Gleichgewichte aus thermischen Messungen,
Nachrichten von der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-physikalische
Klasse, Weidmannsche Buchhandlung, Berlin.

6.2. SECOND LAW IN TERMS OF ENTROPY 243
Figure 6.5: Walther Hermann Nernst (1864-1941), German physical

chemist who developed the third law of thermodynamics; image from
https://en.wikipedia.org/wiki/Walther Nernst.
6.2 Second law in terms of entropy

H
We now have a statement of the second law, δQ/T ≤ 0, valid for reversible or irreversible
R2
heat transfer, and a definition of entropy S2 − S1 = 1 δQ/T , provided the heat transfer is
reversible. The two seem similar. Let us combine them to cast the second law in terms of
entropy. Consider the cycle in the T − S diagram of Fig. 6.6. We start at 1, and proceed to
2 along path I, that represents an irreversible process. We return from 2 to 1 along path R,
that represents a reversible process. The second law, Eq. (6.10), holds
I
δQ
≤ 0, (6.25)
T
I
δQ
0 ≥ . (6.26)
T
The equality implies all processes are reversible; the inequality implies some portion of the
process is irreversible. Now, for a reversible process we also have
Z 2
δQ
S2 − S1 = . (6.27)
1 T
Because the process is reversible, we can reverse to get

Z 1
δQ
S1 − S2 = . (6.28)
2 T

I 2
R
1
Figure 6.6: Sketch of a cycle in the T − S plane composed of irreversible process I from 1
to 2, followed by reversible process R from 2 back to 1.
Now, apply the second law, Eq. (6.10), to the scenario of Fig. 6.6:
Z 2 Z 1
δQ δQ
0≥ + . (6.29)
1 T I 2 T R
Now, substitute Eq. (6.28) into Eq. (6.29) to eliminate the integral along R to get
Z 2
δQ
0 ≥ + S1 − S2 , (6.30)
1 T I
Z 2
δQ
S2 − S1 ≥ . (6.31)
1 T I
More generally, we can write the second law of thermodynamics as
Z 2
δQ
S2 − S1 ≥ . (6.32)
1 T
If 1 → 2 is reversible, the equality holds; if 1 → 2 is irreversible, the inequality holds.
Now, if the system is isolated, there can be no heat transfer interactions and δQ = 0. So
S2 − S1 ≥ 0. (6.33)
isolated system
This implies 2 occurs later in time than 1. Thus, for isolated systems, the entropy increases
as time moves forward.
Example 6.1
Two large thermal reservoirs, one at TA and the other at TB , exchange a finite amount of heat Q
with no accompanying work exchange. The reservoirs are otherwise isolated and thus form their own

universe, when considered as a combined system. Consider the implications for entropy and the second
law.
The scenario is sketched in Fig. 6.7. Assume for now that positive Q leaves A and enters B. Both A
Q
A B
Figure 6.7: Sketch of heat transfer from A to B.
and B are so massive that the respective loss and gain of thermal energy does not alter their respective
temperatures. Consider the entropy changes for each system:
Z 2 Z 2
δQ 1 Q
SA2 − SA1 = = δQ = − , (6.34)
1 T TA 1 TA
Z 2 Z 2
δQ 1 Q
SB2 − SB1 = = δQ = . (6.35)
1 T TB 1 TB
The entropy change for A is negative because Q was negative for A; the entropy change for B is positive
because Q for B was positive. Now, our universe is the combination of A and B, so the entropy change
of the universe is found by adding the entropy changes of the components of the universe:
Q Q
(SA2 + SB2 ) − (SA1 + SB1 ) = − . (6.36)
| {z } | {z } TB TA
=SU 2 =SU 1
With the universe entropy SU as SU = SA + SB , we get

1 1
SU2 − SU1 = Q − . (6.37)
TB TA
The universe is isolated, so the second law holds that SU2 − SU1 ≥ 0; thus,

1 1
Q − ≥ 0. (6.38)
TB TA
Now, we have assumed Q > 0; therefore, we can divide by Q without changing the sense of the
inequality:
1 1
− ≥ 0, (6.39)
TB TA
TA − TB
≥ 0. (6.40)
TA TB

Because TA > 0 and TB > 0, we can multiply both sides by TA TB without changing the sense of the
inequality to get
TA − TB ≥ 0, (6.41)
TA ≥ TB . (6.42)
We have thus confirmed that our mathematical formulation of the second law in terms of entropy yields
a result consistent with the Clausius statement of the second law. We must have TA ≥ TB in order to
transfer positive heat Q from A to B.
Example 6.2
We have m = 10 kg of liquid water at T1 = 0 ◦ C. The water freezes, and its final state is solid
water at T2 = 0 ◦ C. Give a second law analysis.
Let us assume reversible processes. For the water, we have

Z 2
δq
S2 − S1 = m . (6.43)
1 T
While we can use the steam tables to find the heats of vaporization and sublimation for water, there
are no values listed for the liquid-solid transition process. We consult other sources, and quickly find
the heat of fusion for water estimated as ∆hf usion = 333.55 kJ/kg. We will ignore work because the
density change is not great (though it is not zero). Now, the water will have 1q2 = −∆hf usion as it
solidifies, because it is losing energy. The temperature is constant, so we get

kJ
m 2
Z
m1q2 (10 kg) −333.55 kg kJ
S2 − S1 = δq = = = −12.2112 . (6.44)
T 1 T (0 + 273.15 K) K
Note that ice has more structure, i.e. less randomness, than liquid water. This is reflected in the drop
of entropy.
In order for this process to have taken place, the surroundings must have received a transfer of
(10 kg)(333.55 kJ/kg) = 3335.5 kJ of energy. For this to have occurred, we would need Tsurr ≤ 0 ◦ C.
Say the surroundings were at −10 ◦ C. Then their entropy would have increased via
3335.5 kJ kJ
∆Ssurr = = 12.6753 . (6.45)
(−10 + 273.15) K K
Thus, the entropy of the universe would have went up by

kJ kJ kJ
∆Suniv = −12.2112 + 12.6753 = 0.4641 . (6.46)
K K K

Figure 6.8: Schematic for adiabatic reversible (isentropic) compression of ammonia problem.
Example 6.3
Given saturated ammonia vapor at P1 = 200 kPa compressed by a piston to P2 = 1.6 MPa in a
reversible adiabatic process, find the work done per unit mass.
Consider the diagram of in Fig. 6.8. The fact that the compression is reversible and adiabatic
implies that it is an isentropic process. If this were a CPIG, we could use algebraic equations, to be
developed, to calculate state 2. But these equations are not valid for materials such as ammonia (NH3 )
near the vapor dome. So, we consult the tables.
Interpolation of Table B.2.1 of BS tells us that
kJ kJ m3
s1 = 5.6034 , u1 = 1301.3 , T1 = −18 ◦ C, v1 = 0.598902 . (6.47)
kg K kg kg
Because the process is isentropic,
kJ
s2 = s1 = 5.6034 . (6.48)
kg K
The saturation tables, BS’s B.2.1, tell us that near P2 = 1600 kPa, that sg ∼ 4.8 kJ/kg/K. Because
s2 > sg , the ammonia is superheated at this state. Knowing P2 and s2 , we turn to BS’s Table B.2.2 to
find by linear interpolation that
kJ m3
T2 = 134.9 ◦ C, u2 = 1548.4 , v2 = 0.118337 . (6.49)
kg kg
Note that T2 is just above the critical temperature of ammonia, that is Tc = 132.3 ◦ C. But P2 is well
below the critical pressure, that is Pc = 11333.2 kPa. Note also that v2 < v1 .
Now, the first law of thermodynamics tells us that
u2 − u1 = 1q2 −1w2 . (6.50)

|{z}
=0

Because the process is isentropic, 1q2 = 0, and

kJ kJ kJ
w
1 2 = u 1 − u 2 = 1301.3 − 1548.4 = −247 . (6.51)
kg kg kg
With considerable effort we could have also constructed 1w2 by integration of the P − v diagram along
an isentrope.
The process is sketched in Fig. 6.9.
T P=1600 kPa P
2
T=
134.9 o
C
P=200 kPa
1 1
T=
-18.8 o
C
s v
Figure 6.9: T − s and P − v diagrams for adiabatic reversible (isentropic) compression of

ammonia problem.
Example 6.4
Liquid water at P1 = 100 kPa, x1 = 0 is heated isobarically until T2 = 200 ◦ C. Find the heat
added.
The tables tell us that at state 1, we have T1 = 99.62 ◦ C, v1 = 0.001043 m3 /kg, h1 = 417.44 kJ/kg,
and s1 = 1.3025 kJ/kg/K. We note an intermediate stage i when xg = 1 that Tg = T1 = 99.62 ◦ C,
hg = 2675.46 kJ/kg, and sg = 7.3593 kJ/kg/K. At the final state, the superheat tables give us
v2 = 2.17226 m3 /kg, h2 = 2875.27 kJ/kg, and s2 = 7.8342 kJ/kg/K.
For the isobaric process, we have the heat transfer given by ∆h. For the part of the process under
the dome, we have

kJ kJ kJ
1qg = hg − h1 = 2675.46 − 417.44 = 2258.02 . (6.52)
kg kg kg
This quantity is the latent heat. It represents heat added without
Rg increase in temperature. It is easily
calculated by an independent method. We also have 1qg = 1 T ds. Because an isobaric process under
the vapor dome is also isothermal, we can say 1qg = T (sg − s1 ), so

kJ kJ kJ
1qg = (99.62 + 273.15) (K) 7.2593 − 1.3025 = 2257.79 . (6.53)
kg K kg K kg

6.3. THE GIBBS EQUATION 249
The difference in the two predictions can be attributed to rounding error.

For the part of the process outside of the dome, we have

kJ kJ kJ
gq2 = h2 − hg = 2875.27 − 2675.46 = 199.81 . (6.54)
kg kg kg
This quantity is the sensible heat. It represents heat added with an increase in temperature. We can
R2
sense it. It could also be calculated via numerical integration of gq2 = g T ds for this non-isothermal
process; we omit this calculation.
The total heat added, work, and total entropy change are

kJ kJ kJ
1q2 = h2 − h1 = 2875.27 − 417.44 = 2457.83 , (6.55)
kg kg kg

m3 m3 kJ
1w2 = P (v2 − v1 ) = (100 kPa) 2.17226 − 0.001043 = 217.12 , (6.56)
kg kg kg

kJ kJ kJ
s2 − s1 = 7.8342 − 1.3025 = 6.5317 . (6.57)
kg K kg K kg K
Note that 1w2 6= 1q2 because there is a change in internal energy for this process. The process is
sketched in Fig. 6.10.
T P
2
1 P=100 kPa g 2
P=100 kPa g
1 sensible
heat
latent work
heat
s v
Figure 6.10: T − s and P − v diagrams for isobaric heating of water problem.
6.3 The Gibbs equation

We are now in a position to obtain one of the most important relations in thermodynamics,
the

• Gibbs equation: a foundational thermodynamic relation between state variables

consistent with the first law of thermodynamics,
named for the great nineteenth century American engineer, physicist, chemist, and mathe-
matician, depicted in Fig. 6.11. The Gibbs equation is a re-capitulation of the first law of
Figure 6.11: Josiah Willard Gibbs (1839-1903), American mechanical engineer who
revolutionized the science of classical and statistical thermodynamics; images from
https://mathshistory.st-andrews.ac.uk/Biographies/Gibbs.html.
thermodynamics. It is taken in the limit that

• all processes are assumed to be reversible.
And we shall only consider the Gibbs equation for simple compressible substances, neglecting
changes in kinetic and potential energy, though other forms are possible.
We first recall the differential form of the first law, Eq. (3.172), dE = δQ − δW , neglect4
changes in KE and P E, so that dE = dU, and then consider this on a per mass basis,
obtaining
du = δq − δw. (6.58)
Now, for a simple compressible substance undergoing pressure-volume work, we have Eq. (3.80)
cast on a per mass basis, δw = P dv. For the same substance undergoing simultaneous re-
versible heat transfer, we have from Eq. (6.20), δq = T ds. So we can recast Eq. (6.58) as
4
The justification of neglecting changes in KE and P E is rarely stated. If we retained the total en-
ergy, we would be led ultimately to ds = de/T + (P/T ) dv. Moreover, we would conclude that changes
in KE or P E could lead to a change in entropy. However, we choose to specifically retain an account-
ing for mechanical energy via Newton’s second law. Detailed analysis of Newton’s second law would show
that work done by certain classes of forces, e.g. gravity forces and forces due to pressure differences, was
reversible, and does not dissipate mechanical energy. That class of change of KE and P E should not
be thought of as entropy-generating, because of its reversibility. However, work done by other types of
forces, e.g. viscous shear forces, does dissipate mechanical energy into thermal energy. Such a conver-
sion is irreversible, and should contribute to an entropy change. These notions are best understood in
the context of the full mass, momenta, and energy equations for a continuum. Details may be found in
https://www3.nd.edu/∼powers/ame.60635/notes.pdf.

6.3. THE GIBBS EQUATION 251
the Gibbs equation5 :
du = T
| {zds} − P
| {zdv} . (6.59)
δq δw
Gibbs presented this now famous equation in an obscure journal, that was the only journal
in which he ever published. A reproduction from that journal of the page where his equation
first appeared6 is given in Fig. 6.12.
Figure 6.12: An image of the 1873 first appearance of the Gibbs equation in print.
Note that the Gibbs equation is generally valid for all materials. We have said nothing
about either a thermal or caloric equation of state. It is thus valid for ideal gases, non-ideal
gases, liquids, solids, or two-phase mixtures. And though we considered reversible processes
in its derivation, once we have it, we note that only properties are defined, and so we do not
have to restrict ourselves to reversible processes. It is commonly rearranged in a way that
allows the entropy to be determined:
T ds = du + P dv. (6.60)
5
In this usage, similar to that given by BS, we are not alone, but may be in the minority. Some texts
call Eq. (6.59) the “first Gibbs equation.” Perhaps a more common name for a variant of Eq. (6.59) is
the “Fundamental Thermodynamic Relation,” that is commonly described for the extensive analog, dU =
T dS − P dV .
6
J. W. Gibbs, 1873, “Graphical methods in the thermodynamics of fluids,” Transactions of the Connecti-
cut Academy of Arts and Sciences, 2: 309-342.

The Gibbs equation can also be written in terms of enthalpy. Recall Eq. (3.218), h = u+P v,
also valid for general materials. Elementary calculus then tells us
h = u + P v, (6.61)
dh = du + P dv + v dP, (6.62)
du = dh − P dv − v dP. (6.63)
Substitute Eq. (6.63) into the Gibbs relation Eq. (6.60) to yield
T ds = |dh − P dv
{z − v dP} +P dv, (6.64)
du
= dh − v dP. (6.65)
So we can say7
dh = T ds + v dP. (6.66)
Example 6.5
Given the enthalpy change, calculate the entropy change for water going from saturated liquid to
saturated vapor along a T = 100 ◦ C isotherm.
Under the vapor dome, an isotherm is an isobar, so dP = 0. So Eq. (6.66) reduces to

dP} = T ds.
dh = T ds + v| {z (6.67)
=0
Because T is constant, the integration is easy,

Z g Z g
dh = T ds, (6.68)
f f
Z g
= T ds, (6.69)
f
hg − hf = T (sg − sf ), (6.70)
hf g = T sf g , (6.71)
hf g
sf g = . (6.72)
T
From Table B.1.1 of BS, we find at T = 100 ◦ C, hf g = 2257.03 kJ/kg. Thus, we must have
kJ
2257.03 kg kJ
sf g = = 6.04859 . (6.73)
(100 + 273.15) K kg K
We compare this to the value listed in BS’s Table B.1.1 of sf g = 6.0480 kJ/kg/K. The difference is
likely due to round-off errors. Note that the entropy of the vapor is greater than that of the liquid,
sf g = sg − sf > 0. We easily envision the vapor as having less structure or less order than the liquid.
7
Equation (6.66) is called in some texts the “second Gibbs equation.” BS considers it to be a “Gibbs
equation.”

6.4. ENTROPY FOR IDEAL GASES 253
6.4 Entropy for ideal gases

Let us find how to write the entropy for an ideal gas. We first use the Gibbs equation,
Eq. (6.60), to solve for ds to get
du P
ds = + dv. (6.74)
T T
|{z}
=R/v
Now, ideal gases have P v = RT , so P/T = R/v, so we have

du R
ds = + dv. (6.75)
T v
Now, ideal gases also have from Eq. (3.242), du = cv (T ) dT , so
cv (T ) dT dv
ds = +R . (6.76)
T v
We will first consider a CPIG, and then a CIIG.
6.4.1 Calorically perfect

For a CPIG, cv is a constant, and Eq. (6.76) reduces to
cv dT dv
ds = +R . (6.77)
T v
This is easily integrated as follows
Z 2 Z 2 Z 2
cv dT dv
ds = + R , (6.78)
1 1 T 1 v
Z 2 Z 2
dT dv
= cv +R . (6.79)
1 T 1 v
So we get
T2 v2
s2 − s1 = cv ln + R ln . (6.80)
T1 v1
In general for a CPIG we can say
T v
s(T, v) = so + cv ln + R ln , (6.81)
To vo
where o denotes a reference state. We can invert to solve for T /To as follows

T (s, v) vo R/cv s − so
= exp . (6.82)
To v cv

Because R/cv = (cP − cv )/cv = k − 1, we also write

T (s, v) vo k−1 s − so
= exp . (6.83)
To v cv
Now, we can also rearrange Eq. (6.66) to get

dh v
ds = − dP. (6.84)
T T
Now, the ideal gas gives us v/T = R/P , and the calorically perfect assumption gives us
dh = cP dT , with cP a constant. Thus, Eq. (6.84) reduces to
dT dP
ds = cP −R . (6.85)
T P
We integrate Eq. (6.85) to get
Z 2 Z 2 Z 2
dT dP
ds = cP −R . (6.86)
1 1 T 1 P
Thus,
T2 P2
s2 − s1 = cP ln − R ln . (6.87)
T1 P1
In general, for a CPIG we can say
T P
s(T, P ) = so + cP ln − R ln , (6.88)
To Po
where o denotes a reference state. We can invert Eq. (6.88) to form

R/cP
T (s, P ) P s − so
= exp . (6.89)
To Po cP
Because R/cP = (cP − cv )/cP = 1 − 1/k = (k − 1)/k, we have
(k−1)/k
T (s, P ) P s − so
= exp . (6.90)
To Po cP
Straightforward algebra also shows that
v P
s(P, v) = so + cP ln + cv ln , (6.91)
vo Po

k s − so
Pv = Po vok exp . (6.92)
cv

6.4.2 Calorically imperfect

For the CIIG, we integrate Eq. (6.76) to get
Z 2 Z 2 Z 2
cv (T ) dT dv
ds = +R . (6.93)
1 1 T 1 v
This gives
Z 2
cv (T ) dT v2
s2 − s1 = + R ln . (6.94)
1 T v1
For the CIIG, we have Eq. (3.244), dh = cP (T ) dT , along with the ideal gas result
v/T = R/P ; thus, Eq. (6.84) reduces to
cP (T ) dT dP
ds = −R . (6.95)
T P
Integrating, we get
Z 2
cP (T ) dT P2
s2 − s1 = − R ln . (6.96)
1 T P1
In general we can say

Z T
cP (T̂ ) dT̂ P
s = so + −R ln . (6.97)
To T̂ Po
| {z }
=soT
Here, the “hat” notation indicates a dummy variable of integration. Here, soT is a function
of temperature and represents the entropy when the pressure is evaluated at its reference
value of P = Po . In BS, Table A.8 gives values of soT . Note that
o
• in this course a superscript denotes evaluation at a reference pressure. Typically
Po = 100 kPa.
So for the CIIG, we have

Z T
cP (T̂ ) dT̂ P P
s(T, P ) = so + −R ln = soT − R ln . (6.98)
To T̂ Po Po
| {z }
=soT
We could also say

P2
s2 − s1 = soT2 − soT1 − R ln . (6.99)
P1

Example 6.6
Consider a mass of air, m = 10 kg cooled isochorically from P1 = 1 MPa, T1 = 500 K to the
temperature of the surroundings, T2 = 300 K. Find the entropy change of the air, the surroundings,
and the universe. Assume a CPIG model and then compare to results for a CIIG model.
The scenario is sketched in Fig. 6.13. For air, we take R = 0.287 kJ/kg/K, cv = 0.7175 kJ/kg/K.
state 1 state 2
Figure 6.13: Sketch of air cooling in a large surroundings.

Thus, cP = R + cv = 1.0045 kJ/kg/K.
CPIG model
Let us consider the CPIG model. Now, v1 = v2 for the isochoric problem. The ideal gas law gives
P2 v2 P1 v1
= , (6.100)
T2 T1
P2 v1 P1 v1
= , (6.101)
T2 T1
T2
P2 = P1 , (6.102)
T1

300 K
= (1000 kPa) , (6.103)
500 K
= 600 kPa. (6.104)
The first law for the isochoric process tells us
U2 − U1 = 1Q2 − 1W2 , (6.105)
|{z}
=0
mcv (T2 − T1 ) = 1Q2 , (6.106)

kJ
1Q2 = (10 kg) 0.7175 ((300 K) − (500 K)), (6.107)
kg K
= −1435 kJ. (6.108)
The negative sign indicates the energy left the system. This is the energy that entered the surroundings.
We get the entropy change of the air via the non-specific version of Eq. (6.80):

T2 v2
S2 − S1 = m cv ln + R ln , (6.109)
T1 v1
 

kJ 300 K kJ
= (10 kg)  0.7175 ln + 0.287 ln 1  , (6.110)
kg K 500 K kg K |{z}
=0

kJ
S2 − S1 = −3.66517 . (6.111)
K
The entropy went down for the system. Note this is consistent with ds = δq/T . When δq < 0, the
entropy drops. The system becomes more ordered as it cools.
Now, the surroundings gained thermal energy, Q = 1435 kJ. However, they are so massive that
the surroundings’ temperature remained constant. So the entropy change of the surroundings is found
from
Z 2
δQ
S2,surr − S1,surr = , (6.112)
1 T surr
Z 2
1
= δQ, (6.113)
Tsurr 1
Qsurr
= , (6.114)
Tsurr
1435 kJ
= , (6.115)
300 K
kJ
S2,surr − S1,surr = 4.78333 . (6.116)
K
So the energy addition to the surroundings raised its entropy.

The entropy change of the universe is that of the combination of the surroundings’ change and the
system’s change:
∆Suniv = (S2,surr − S1,surr ) + (S2 − S1 ), (6.117)

kJ kJ
= 4.78333 + −3.66517 , (6.118)
K K
kJ
∆Suniv = 1.11816 . (6.119)
K
The universe is isolated, and its entropy went up in this process, consistent with the second law of
thermodynamics.
CIIG model
Let us repeat the analysis with a CIIG model. First, P2 = 600 kPa, unchanged from the CPIG
model. The first law for the isochoric, work-free process gives
1Q2 = m(u2 − u1 ). (6.120)
But here, we use Table A.7.1 from BS to obtain the internal energies. Doing so, we get

kJ kJ
1Q2 = (10 kg) 214.36 − 359.84 , (6.121)
kg kg
= −1454.8 kJ. (6.122)
Note this value is close to the CPIG prediction of −1435 kJ.

We can adapt Eq. (6.99) to calculate the entropy change of the system:


P2
S2 − S1 = m soT2 − soT1 − R ln . (6.123)
P1
Table A.7.1 from BS gives us soT2 = 6.86926 kJ/kg/K and soT1 = 7.38692 kJ/kg/K. Thus

kJ kJ kJ 600 kPa
S2 − S1 = (10 kg) 6.86926 − 7.38692 − 0.287 ln , (6.124)
kg K kg K kg K 1000 kPa
kJ
= −3.71053 . (6.125)
K
This is close to the CPIG prediction of −3.66517 kJ/K.
For the entropy change of the surroundings, we get
Qsurr
S2,surr − S1,surr = , (6.126)
Tsurr
1454.8 kJ
= , (6.127)
300 K
kJ
= 4.84933 . (6.128)
K
Combining the entropy changes of the surroundings and system, we get that for the universe,

kJ kJ
∆Suniv = 4.84933 + −3.71053 , (6.129)
K K
kJ
= 1.1388 . (6.130)
K
This is close to the CPIG estimate of 1.11816 kJ/K.
6.5 Entropy for an incompressible solid or liquid

For an incompressible solid or liquid we have dv = 0, so the Gibbs equation, Eq. (6.59),
reduces to
T ds = du. (6.131)
If we take du = c dT , we get
T ds = c dT, (6.132)
c dT
ds = , (6.133)
T
Z T
c(T̂ ) dT̂
s − so = . (6.134)
To T̂
And if the solid or liquid is calorically perfect with c a true constant, we get
T
s − so = c ln . (6.135)
To

6.6. ISO- CURVES 259
6.6 Iso- curves

Let us use the Gibbs equation in its various forms to identify a few important curves.
6.6.1 Isochores
To identify isochores, let us consider Eq. (6.60),
T ds = du + P |{z}
dv , (6.136)
=0
T ds = du, (6.137)
∂u
T = . (6.138)
∂s v
This is valid for a general material. Iff we have an ideal gas, then du = cv (T ) dT , and on an
isochore, Eq. (6.60) becomes
T ds = du
|{z} +P |{z}
dv , (6.139)
=cv (T ) dT =0
T ds = cv (T ) dT, (6.140)
T ∂T
= . (6.141)
cv (T ) ∂s v
Thus, the slope of an isochore in the T − s plane for an ideal gas is T /cv .
6.6.2 Isobars
To identify isobars, let us consider Eq. (6.65),
T ds = dh − v |{z}
dP , (6.142)
=0
T ds = dh, (6.143)
∂h
T = . (6.144)
∂s P
This is valid for a general material. Iff we have an ideal gas, then dh = cP (T ) dT , and
on an isobar, Eq. (6.65) becomes
T ds = dh
|{z} −v |{z}
dP , (6.145)
=cP (T ) dT =0
T ds = cP (T ) dT, (6.146)
T ∂T
= . (6.147)
cP (T ) ∂s P

Thus, the slope of an isobar in the T −s plane for an ideal gas is T /cP . Because cP (T ) > cv (T ),
the slope of the isochore is greater than the slope of an isobar at a given point.
For air as a CPIG with k = 7/5, R = 0.287 kJ/kg/K, the scenario is sketched in Fig. 6.14.
For materials such as water, the behavior is similar. The slope of the isochore is greater at
T (K)
1000
kPa
/kg
Pa
/kg
kPa
g
00
0k
m /k
10 m 3
=1
1 m3
100 3
=1
=1
800 P
v=
v=
P
v=
600
400
200
slope isochore > slope isobar
1 2 s (kJ/kg/K)
Figure 6.14: Sketch of isochores and isobars in the T − s plane for CPIG air, k = 7/5,
R = 0.287 kJ/kg/K, so = 0 kJ/kg/K.
a given point than that of an isobar.
6.6.3 Isentropes
We introduce an
• Isentrope: a curve on which entropy is constant.
For general materials, we identify isentropes by considering Eq. (6.60) with ds = 0:
T |{z}
ds = du + P dv, (6.148)
=0
0 = du + P dv, (6.149)
du = −P dv. (6.150)
Because there is no heat transfer on an isentrope, for such a process, all of the P dv work
goes into changing the internal energy of the system. We could also say
∂u
= −P. (6.151)
∂v s

6.7. ISENTROPIC RELATIONS FOR AN IDEAL GAS 261
Similarly,
T |{z}
ds = dh − v dP, (6.152)
=0
0 = dh − v dP, (6.153)
dh = v dP. (6.154)
We could also say

∂h
= v. (6.155)
∂P s
6.7 Isentropic relations for an ideal gas

Here, we will consider algebraic relations for ideal gases undergoing isentropic processes. The
results are simple for CPIGs and a little more complicated for CIIGs.
6.7.1 Calorically perfect

Let us consider the important case of a CPIG undergoing an isentropic process. Start with
the Gibbs equation, Eq. (6.60), T ds = du + P dv. Now, for an isentropic CPIG, we have
ds = 0, P = RT /v, and du = cv dT , so we get
RT
0 = cv dT + dv, (6.156)
v
dT dv
= cv +R , (6.157)
T v
dT dv
−cv = R , (6.158)
T v
Z 2 Z 2
dT dv
−cv = R , (6.159)
1 T 1 v
T2 v2
−cv ln = R ln , (6.160)
T1 v1
T2 R v1
ln = ln , (6.161)
T1 cv v2
R/cv
v1
= ln , (6.162)
v2
k−1
v1
= ln , (6.163)
v2
k−1
T2 v1
= . (6.164)
T1 v2

Now, because T = P v/R, we can also say

P2 v2 k−1
R v1
P1 v1
= , (6.165)
R
v2
k−1
P2 v2 v1
= , (6.166)
P1 v1 v2
k
P2 v1
= , (6.167)
P1 v2
k−1 k−1
P2 k
v1
= . (6.168)
P1 v2
We can summarize by combining Eqs. (6.164, 6.168) to get an important result, that we
emphasize is valid for isentropic calorically perfect ideal gases only:
k−1 k−1
T2 P2 k
v1
= = . (6.169)
T1 P1 v2
Another useful form is given by rearranging Eq. (6.168) to get the result, again valid for
isentropic calorically perfect ideal gases only:
P1 v1k = P2 v2k . (6.170)
We see that the isentropic relation between P and v is that for a polytropic process, see
p. 96, with the polytropic exponent n = k. Recall for an ideal gas undergoing an isothermal
process, we have P1 v1 = P2 v2 , that is polytropic with n = 1.
Example 6.7
Compare the slope of an isentrope and isotherm through the same point in the P − v plane for a
CPIG.
For an isotherm in the P − v plane for a CPIG, we can say
Pv = Po vo , (6.171)
1
P = Po vo , (6.172)
v
∂P 1
= −Po vo , (6.173)
∂v T v2
∂P Po
=− . (6.174)
∂v T v=vo vo

For an isentrope in the P − v plane for a CPIG, we can say
P vk = Po vok , (6.175)
1
P = Po vok k , (6.176)
v
∂P 1
= −kPo vok , (6.177)
∂v s v 1+k
∂P Po
= −k . (6.178)
∂v s v=vo vo
Both slopes are negative, because Po > 0 and vo > 0. Because k > 1, the magnitude of the isentrope’s
slope is steeper than that of the isotherm. We give a sketch of the behavior of an isentrope and isotherm
for a CPIG, each passing through the point (Po , vo ) in the (P, v) plane, in Fig. 6.15.
n=1, isotherm
n=k, isentrope
Figure 6.15: Sketch of an isentrope and an isotherm for a CPIG in the P − v plane.
Example 6.8
Show that dT /dz is constant in an isentropic atmosphere, and give an estimate for it.
We must consider the forces acting on a differential slice of the atmosphere. Consider the free body
diagram in Fig. 6.16. Newton’s second law for the fluid gives rise to
d2 z
ρA dz 2 = P A − (P + dP )A − ρgA dz. (6.179)
| {z } dt
dm
| {z }
=0

dz
PA
Figure 6.16: Free body diagram for a differential slice of the atmosphere.
For an atmosphere in mechanical equilibrium, we neglect fluid acceleration and arrive at a force balance
of surface forces against the weight of the fluid:
P A = (P + dP )A + ρgA dz, (6.180)

= P A + A dP + ρgA dz, (6.181)
0 = A dP + ρgA dz, (6.182)
= dP + ρg dz, (6.183)
dP
= −ρg. (6.184)
dz
As ρ > 0 and g > 0, we see that the pressure gradient dP/dz < 0. That is, the pressure goes down
as we go up in height. If the density were constant, as it would be for liquid water, we could easily
integrate this to get a linear relationship between P and z. In fact we did assume constant density in
an earlier analysis, that led to the simple Eq. (1.66), ∆P = ρgH. However, for air, ρ is not constant;
it too varies with z. So a more complicated analysis is required.
Let us eliminate both P and ρ from Eq. (6.184) in favor of T because we are seeking to show dT /dz
is a constant. Let us first impose the CPIG assumption. Thus
P = ρRT. (6.185)
Now, let us invoke an isentropic relation for a CPIG. From Eq. (6.169), we can easily deduce that
k
k−1
T
P = Po . (6.186)
To
Now, take d/dz of Eq. (6.186), using the chain rule, to get
k
k−1 −1
dP Po k T dT
= . (6.187)
dz To k − 1 To dz
Now, use Eq. (6.185) to eliminate ρ in Eq. (6.184) and use Eq. (6.187) to eliminate dP/dz in Eq. (6.184),
yielding
k−1 1
Po k T dT P
=− g. (6.188)
To k − 1 To dz RT
|{z}
| {z }
ρ
dP/dz

Now, use Eq. (6.186) to eliminate P to get

k
k−1
1 T
Po k

T
k−1
dT Po To
= − g, (6.189)
To k − 1 To dz RT
k
k−1
1
k−1 T
k T dT To
= − g, (6.190)
k−1 To dz R TTo
k−11
T
To
= − g, (6.191)
R
dT k−1 g
= − , (6.192)
dz k R
cP
−1 g
= − cv cP , (6.193)
cv cP − cv
g
= − . (6.194)
cP
So dT /dz is indeed a constant. Let us use numerical values to estimate the gradient:
m
dT 9.81 s2 K K
=− J
= −0.009766 = −9.766 . (6.195)
dz 1004.5 kg K
m km
Commercial airplanes fly at altitude near 10 km. So we might expect temperatures at this altitude to
be about 100 K less than on the ground with this simple estimate. Mountain climbers of Denali have
to rise 5.5 km from its base to its peak; thus, they might expect to experience a 53 K temperature drop
during the ascent, that is a drop of roughly 95 ◦ F. Climbers of Pike’s Peak have only a 1.6 km rise,
and so can expect at 15 K (27 ◦ F) drop. There are other mitigating effects that can modulate these
temperature changes that we have not included.
The predictions of Eq. (6.195) are compared to data for a standard atmosphere from Kuethe and
Chow8 in Fig. 6.17. Obviously, the theory and data have some mismatch. For z < 10000 m, the slopes
T (K)
300 troposphere stratosphere
200
100 data
theory: adiabatic dry air
0 4000 8000 12000 z (m)
Figure 6.17: Atmospheric temperature as a function of altitude: predictions of Eq. (6.195)

compared with data from Kuethe and Chow.
8
A. M. Kuethe and C.-Y. Chow, 1998, Foundations of Aerodynamics, Fifth Edition, John Wiley, New
York, p. 533.

of the curves are different, with our theory predicting a drop of roughly 10 K/km. The observed rate
of drop in this region is linear but at a rate of 6.5 K/km. Kundu and Cohen9 attribute the discrepancy
to the neglect of the influence of moisture in the air in the atmosphere. For z > 10000 m, the difference
becomes more pronounced. At this stage, we have reached the end of the troposphere and entered the
stratosphere, where different physics are at work. The predictions of an adiabatic theory for pressure
variation with altitude compare some what better to the data as shown in Fig. 6.18.
P (kPa)
100
data
80 theory: adiabatic dry air
60
40
20
0 4000 8000 12000 z (m)
Figure 6.18: Atmospheric pressure as a function of altitude: predictions of an adiabatic

theory compared with data from Kuethe and Chow.
Example 6.9
We are given a cylinder divided by a frictionless, mobile, adiabatic piston, with air on one side
and water on the other side as shown in Fig. 6.19. The air is thermally insulated from the water
and the surroundings. The water is in thermal contact with its surroundings. Initially Va1 = 0.1 m3 ,
Vw1 = 0.1 m3 , Ta1 = 40 ◦ C, Tw1 = 90 ◦ C, xw1 = 0.1, where a stands for air and w for water. Heat is
added until xw2 = 1. Find the final pressure and the heat added.
Because the piston is frictionless and mobile, we can say at all times that the air and water are in
mechanical equilibrium:
Pa = Pw . (6.196)
The pressure will vary throughout the process, but air and water will always have the same pressure.
We also can see that the total volume will be constant:
Va + Vw = 0.2 m3 . (6.197)
Now, the air undergoes a reversible adiabatic process, i.e. isentropic. The water undergoes a non-
adiabatic process; hence, it is not isentropic.
9
P. K. Kundu and I. M. Cohen, 2008, Fluid Mechanics, Fourth Edition, Academic Press, Amsterdam,
p. 605.

adiabatic wall
conducting wall
adiabatic, mobile piston
air water
Figure 6.19: Schematic of a piston-cylinder arrangement with air on one side and water on
the other.
At the initial state the water is saturated at 90 ◦ C, x = 0.1. We learn from the tables that
Pw1 = 70.14 kPa, (6.198)

vw1 = vf + x1 vf g , (6.199)

m3 m3
= 0.001036 + (0.1) 2.35953 , (6.200)
kg kg
m3
= 0.2370 , (6.201)
kg
uw1 = uf + x1 uf g , (6.202)

kJ kJ
= 376.85 + (0.1) 2117.7 , (6.203)
kg kg
kJ
= 588.62 , (6.204)
kg
sw1 = sf + x1 sf g , (6.205)

kJ kJ
= 1.1925 + (0.1) 6.2866 , (6.206)
kg K kg K
kJ
= 1.8212 . (6.207)
kg K
Now, we can find the mass of the water via
Vw1 0.1 m3
mw = = 3 = 0.4219 kg. (6.208)
vw1 0.2370 m
kg
Now, because of pressure equilibrium,
Pa1 = Pw1 = 70.14 kPa. (6.209)
From the ideal gas law,

kJ
RTa1 0.287 kg K (40 + 273) K m3
va1 = = = 1.281 . (6.210)
Pa1 70.14 kPa kg

3 3
Pw (kPa) vg m
kg RHS m kg
100 1.6940 0.5923
300 0.6058 0.0910
400 0.4625 0.0240
450 0.4140 0.0011
453 0.4133 0.0008
500 0.3749 -0.0172
Table 6.1: Iteration results for water-air problem.
So the mass of the air is

VA1 0.1 m3
ma = = 3 = 0.07808 kg. (6.211)
va1 1.281 m
kg
Now, for the isentropic process for the air, we have

k k
Pa1 Va1 = Pa2 Va2 , (6.212)
1/k
Pa1
Va2 = Va1 . (6.213)
Pa2
Now, at state 2, Eq. (6.197) gives Va2 + Vw2 = 0.2 m3 . Substitute Eq. (6.213) into this and eliminate
Vw2 in favor of mass and specific volume to get
1/k
Pa1
Va1 + mw vw2 = 0.2 m3 . (6.214)
Pa2 | {z }
| {z } =Vw2
=Va2
Next use Pa = Pw to get

1/k
Pw1
Va1 + mw vw2 = 0.2 m3 . (6.215)
Pw2
Substitute numbers we know to get

1/1.4
3 70.14 kPa
(0.1 m ) + (0.4219 kg)vw2 − 0.2 m3 = 0. (6.216)
Pw2
Now, this is a relation between Pw2 and vw2 . It is not sufficient to fix the state. However, we know
x2 = 1. This gives us another curve in the P − v space. Let us guess a value of Pw2 , read the value of
vg from the tables, substitute both into Eq. (6.216), see how close the right hand side (RHS) comes to
zero, and iterate until we have an acceptable estimate. We generate Table 6.1. After the iteration, we
see that
m3
Pw2 = Pa2 = 453 kPa, vw2 = 0.4133 . (6.217)
kg
The tables then give
kJ kJ
Tw2 = 148 ◦ C, sw2 = 6.856 , uw2 = 2557.7 . (6.218)
kg K kg

Now
m3
Vw2 = mw vw2 = (0.4219 kg) 0.4133 = 0.1744 m3 . (6.219)
kg
Thus,
Va2 = (0.2 m3 ) − Vw2 = (0.2 m3 ) − (0.1744 m3 ) = 0.0256 m3 . (6.220)
Thus,
Va2 0.0256 m3 m3
va2 = = = 0.3282 . (6.221)
ma 0.07808 kg kg
Now, we know two properties of the air, so we can get

m3
Pa2 va2 (453 kPa) 0.3282 kg
Ta2 = = = 518.1 K. (6.222)
R 0.287 kgkJK
Now, consider the first law of thermodynamics for the entire system. The work is zero because the
entire system is isochoric. So the first law gives
U2 − U1 = 1Q2 − 1W2 , (6.223)

|{z}
=0
1Q2 = ma cv (Ta2 − Ta1 ) + mw (uw2 − uw1 ), (6.224)

kJ
= (0.07808 kg) 0.7175 ((518.1 K) − (313 K))
kg K

kJ kJ
+(0.4219 kg) 2557.7 − 588.62 , (6.225)
kg kg
1Q2 = 842.2 kJ. (6.226)
Let us keep on analyzing! Consider the air alone, that is adiabatic. The first law for the air alone
says
Ua2 − Ua1 = a1Qa2 −a1Wa2 , (6.227)

| {z }
=0
a1Wa2 = −(Ua2 − Ua1 ), (6.228)
= ma cv (Ta1 − Ta2 ), (6.229)

kJ
= (0.07808 kg) 0.7175 ((313 K) − (518.1 K)), (6.230)
kg K
= −11.49 kJ. (6.231)
The work is negative as the air was compressed.

Let us calculate the entropy changes. For the combined system, we have
∆Ssystem = ∆Sa +∆Sw , (6.232)

|{z}
=0
= mw (sw2 − sw1 ), (6.233)

kJ kJ
= (0.4219 kg) 6.856 − 1.8212 , (6.234)
kg K kg K
kJ
= 2.124 . (6.235)
K

Now, in order for the process to occur, the surroundings must at least have been Tsurr = 148 ◦ C =
421 K. The surroundings must have had Q = −842.2 kJ, so their entropy change was
Q −842.2 kJ kJ
∆Ssurr = = = −2.00 . (6.236)
Tsurr 421 K K
So the entropy change of the universe is

kJ kJ kJ
∆Suniv = ∆Ssys + ∆Ssurr = 2.124 + −2.00 = 0.124 . (6.237)
K K K
Example 6.10
Consider the cannon shown in Fig. 6.20. At t = 0 s, the bore of the cannon between the projectile
m A
Figure 6.20: Sketch of a cannon.
and its base is filled with a CPIG at Po . The projectile is located at x = xo . The projectile fits snugly in
the cannon, has mass m, and cross-sectional area A. The ambient pressure is P∞ . Find the equations of
motion for the projectile. Assume the gas expands isentropically, and the local pressure is P . Assume
a drag force, proportional to the cube of projectile velocity, retards its motion.
This is a problem that is not in mechanical equilibrium. Newton’s law of motion for the projectile
states
3
d2 x dx
m 2 = (P − P∞ )A − C . (6.238)
dt dt
Here, the product of mass and acceleration is balanced by the net pressure force and the drag force.
Let us define the velocity as
dx
v= . (6.239)
dt
So our equation of motion becomes
dv
m = (P − P∞ )A − Cv3 . (6.240)
dt

Now, for an isentropic expansion of a CPIG we have
Po Vok = P V k, (6.241)
Po (Axo )k = P (Ax)k , (6.242)
x k
o
P = Po . (6.243)
x
Thus, we can say

dv xo k
m = Po − P∞ A − Cv3 . (6.244)
dt x
We can thus form a system of two non-linear ordinary differential equations:

dv P∞ A Po xo k C
= − 1 − v3 , v(0) = 0, (6.245)
dt m P∞ x m
dx
= v, x(0) = xo . (6.246)
dt
For general parameters, an analytic solution is unlikely. However, we can form a numerical solution.
Our equations are of the form
dv
= f (v, x), v(0) = 0, (6.247)
dt
dx
= g(v, x), x(0) = xo . (6.248)
dt
We can use a finite difference method, here the first order explicit Euler method, to discretize these
equations as
vn+1 − vn
= f (vn , xn ), (6.249)
∆t
xn+1 − xn
= g(vn , xn ). (6.250)
∆t
Here, n represents an old value and n + 1 represents a new value. We solve for the new to obtain
vn+1 = vn + ∆tf (vn , xn ), (6.251)

xn+1 = xn + ∆tg(vn , xn ). (6.252)
We start with known values of vn and xn . We use the above formulæ to get new values, then we repeat
the process as long as necessary until we have a full solution. When we choose ∆t sufficiently small,
we should get a stable solution that is a good approximation of the actual solution. Linearization near
the equilibrium point gives an estimate for ∆t:
s
1+k
mxo Po k
∆t < . (6.253)
kPo A P∞
This estimate is only valid in the limit as t → ∞, and so may not work for early time, where ∆t may
need to be much smaller. For early time, we can linearize near the initial point and find that we need
r
mxo
∆t < , (6.254)
kPo A
to capture the early time dynamics.

x (m) v (m/s)
2
300
200
1
100
0.002 0.004 0.006 0.008 0.010 0.002 0.004 0.006 0.008 0.010
t (s) t (s)
Figure 6.21: Position and velocity of projectile in a cannon at early time.
For m = 1 kg, A = 0.01 m2 , C = 0.01 N/(m/s)3 , P∞ = 105 Pa, Po = 108 Pa, xo = 0.1 m, k = 7/5,
we have obtained a numerical solution. Note that at such a high pressure, not unlike those in real
cannons, the ideal gas law is likely not an accurate assumption.
A plot at early time for x(t) and v(t) is given in Fig. 6.21. This shows a window in time that is
consistent with a realistic barrel length. For these values, our estimate, Eq. (6.254), for the time step
to capture the early time dynamics gives ∆t < 2.6 × 10−4 s. Note this is consistent with the rise time
results shown in Fig. 6.21 that show the peak velocity of near 300 m/s reached in under 0.001 s. These
muzzle velocities are a little low for actual cannons, that also may have higher initial pressures. Around
t = 0.01 s, we have x ∼ 2 m, that is not unlike a cannon barrel length.
Let us now assume an unrealistically long barrel length and calculate for a long time. A plot at
late time for x(t) and v(t) is given in Fig. 6.22. At late time we see the projectile is approaching an
x (m)
v (m/s)
15 300
10 200
5 100
1 2 3 1 2 3
t (s) t (s)
Figure 6.22: Position and velocity of projectile in a cannon at late time.
equilibrium where the velocity goes to zero due to drag forces and the position oscillates about an
equilibrium where there is a pressure balance. As t → ∞, our estimate, Eq. (6.253) for ∆t is applicable.
For these values, our estimate for the time step near equilibrium, Eq. (6.253), gives ∆t < 0.10 s. This
time scale actually captures the period of oscillation effectively. It is surprising that the system came
to equilibrium at such a short distance x = 15 m, and it is hard to imagine an actual artillery shell
behaving in this way. It is likely that the drag model is highly inaccurate.
A clearer picture of the early, intermediate, and late time dynamics can be seen in the log-log scale
plot of x(t) given in Fig. 6.23. For early time t < 10−4 s, the piston is effectively stationary. We call
this a time of inertial confinement. For t > 10−4 s, we see that x grows exponentially as time grows,
reflected in the positive sloped line on the log-log plot of Fig. 6.23. During this phase of the expansion,

x (m)
100 oscillatory
relaxation to
equilibrium
10
exponential
growth
1
inertial
confinement
0.1
t (s)
Figure 6.23: Position of projectile in a cannon versus time, plotted on a log-log scale to
reveal the regimes where different physics are dominant.
the pressure force of the compressed air dominates the resistive force of the atmospheric pressure and
the drag force. As x rises, and as dx/dt rises, the resistance of the atmosphere and the drag force
both become more prominent, until around t ∼ 1 s, when the forces begin to balance one another. At
this point, there is overcompensation due to the inertia of the piston, and the atmospheric pressure
dominates the charge pressure. This sets up an oscillation, that ultimately decays due to the drag force.
Example 6.11
For the previous example problem, show details of the linearized analysis in the late time limit so
as to derive Eqs. (6.253).
At equilibrium, we must have dv/dt = 0 and dx/dt = 0. So, as with all equilibrium states, we solve
the corresponding algebra problem deduced from Eqs. (6.245), 6.246):
k !
P∞ A Po xo C 3
0 = −1 − v , (6.255)
m P∞ xeq m eq
0 = veq . (6.256)
Equations (6.255, 6.256) form two algebraic equations for the two unknowns veq and xeq at the equi-
librium state. One root is obvious, and the other is easily found with simple algebra. The equilibrium
state is thus seen to be
1
Po k
xeq = xo , (6.257)
P∞
veq = 0. (6.258)
Note
• The equilibrium state is one of rest, veq = 0. This is consistent with a mechanical equilibrium.

• xeq rises if xo increases, Po increases, or P∞ decreases.

• The equilibrium state is independent of the mass of the projectile, and the cross-sectional area.
Now, let us examine the local dynamics in the neighborhood of equilibrium. Consider Taylor series
expansions about xeq and veq of Eqs. (6.247, 6.248):
dv ∂f ∂f
= f (veq , xeq ) + (v − veq ) + (x − xeq ) + . . . , (6.259)
dt | {z } ∂v eq ∂x eq
=0
dx ∂g ∂g
= g(veq , xeq ) + (v − veq ) + (x − xeq ) + . . . . (6.260)
dt | {z } ∂v eq ∂x eq
=0
We have constructed f and g to have values of zero at equilibrium. Now, because veq and xeq are
constants with derivatives of zero, we can rewrite Eqs. (6.259, 6.260) as
d ∂f ∂f
(v − veq ) = (v − veq ) + (x − xeq ) + . . . , (6.261)
dt ∂v eq ∂x eq
d ∂g ∂g
(x − xeq ) = (v − veq ) + (x − xeq ) + . . . . (6.262)
dt ∂v eq ∂x eq
Now, with

P∞ A Po xo k C
f (v, x) = − 1 − v3 , (6.263)
m P∞ x m
We have
∂f C
= −3 v2 . (6.264)
∂v m
At equilibrium, where v = 0, this gives
∂f
= 0. (6.265)
∂v eq
We also get
∂f Po A xo k
= −k . (6.266)
∂x mx x
Evaluating ∂f /∂x at the equilibrium point, we find
1+k
∂f Po A P∞ k
= −k . (6.267)
∂x eq mxo Po
By inspection, we see that ∂g/∂v = 1 and ∂g/∂x = 0. Thus, Eqs. (6.261, 6.262) become
1+k
d Po A P∞ k
(v − veq ) = −k (x − xeq ) + . . . , (6.268)
dt mxo Po
d
(x − xeq ) = (v − veq ) + . . . . (6.269)
dt
Let us define two new variables:
x̂(t) = x(t) − xeq , (6.270)
v̂(t) = v(t) − veq . (6.271)

With the definitions of Eqs. (6.270, 6.271), Eqs. (6.268, 6.269) reduce to
1+k
dv̂ Po A P∞ k
= −k x̂, (6.272)
dt mxo Po
dx̂
= v̂. (6.273)
dt
Take now the time derivative of Eq. (6.273) and substitute it into Eq. (6.272) to get the second order
linear differential equation:
1+k
d2 x̂ Po A P∞ k
2
+k x̂ = 0. (6.274)
dt mxo Po
We can rewrite this in the form of a mass-spring type problem as
d2 x̂
m + ks x̂ = 0, (6.275)
dt2
1+k
Po A P∞ k
ks = k . (6.276)
xo Po
The solution of this equation is well known to be
r ! r !
ks ks
x(t) = C1 sin t + C2 cos t . (6.277)
m m
The solution is oscillatory and has time constant

s
r k+1
m mxo Po k
τ= = . (6.278)
ks kPo A P∞
At early time one can also do a similar linear analysis in the neighborhood of x ∼ xo and v ∼ 0.
One can define new variables x̃ = x − xo , ṽ = v, linearize the equations, and solve at early time.
6.7.2 Calorically imperfect

Consider now an isentropic relation for a CIIG. Recall from Eq. (6.98) that there is no simple
way to write the algebraic form of the entropy, that relies on an integral, available in tabular
form. Consideration of CIIGs usually relies on an unusual combination of table look-up and
equations.
We can use Eq. (6.98) to write the entropy at states 1 and 2 as
P1
s1 = soT1 − R ln , (6.279)
Po
P2
s2 = soT2 − R ln . (6.280)
Po

Recall o denotes the reference pressure, and soT is a temperature-dependent function, that is
available in tabular form. So the entropy difference, s2 − s1 , is

o o P2 P1
s2 − s1 = sT2 − sT1 − R ln − ln , (6.281)
Po Po

o o P2 Po
= sT2 − sT1 − R ln , (6.282)
Po P1

o o P2
= sT2 − sT1 − R ln . (6.283)
P1
If the process is isentropic, we have s2 = s1 , and so
P2
0 = soT2 − soT1 − R ln , (6.284)
P1
P2
soT2 − soT1 = R ln . (6.285)
P1
Example 6.12
Air is isentropically compressed from T1 = 300 K, P1 = 100 kPa to T2 = 2000 K. Find the final
pressure with a CIIG and CPIG model. Find the work per unit mass of compression for CIIG and
CPIG models.
First consider the first law, u2 − u1 = 1q2 − 1w2 . Because the process is isentropic, 1q2 = 0, so
1w2 = u1 − u2 , (6.286)
for both the CIIG and CPIG.
• CIIG analysis: For the CIIG, we find from Table A.7.1 of BS that
kJ kJ
u1 = 214.36 , u2 = 1677.52 . (6.287)
kg kg
So
kJ
1w2 = u1 − u2 = −1463.16 . (6.288)
kg
The work is negative because the fluid is being worked upon.
For this system, Table A.7.1 gives
kJ kJ
soT1 = 6.86926 , soT2 = 8.96611 . (6.289)
kg K kg K
Now, apply Eq. (6.285) for the isentropic process:
P2
soT2 − soT1 = R ln , (6.290)
P1
o
s − soT1
P2 = P1 exp T2 , (6.291)
R
 
kJ kJ
8.96611 kg K − 6.86926 kg K
= (100 kPa) exp  kJ
, (6.292)
0.287 kg K
= 148935 kPa. (6.293)

6.8. TWO CYCLES 277
• CPIG analysis: For the CPIG, with cv = 0.7175 kJ/kg/K, we have

kJ kJ
1w2 = u1 − u2 = cv (T1 − T2 ) = 0.7175 ((300 K) − (2000 K)) = −1219.75 . (6.294)
kg K kg
And we have from Eq. (6.169) for an isentropic CPIG, with k = 7/5
k
k−1 7/5
7/5−1
T2 2000 K
P2 = P1 = (100 kPa) = 76503.4 kPa. (6.295)
T1 300 K
Note the pressure from the CPIG model is half that from the CIIG model! This illustrates the dangers
of using a CPIG model for high temperature applications.
If we had chosen T2 = 400 K, the CIIG model would have given P2 = 274.7 kPa, and the CPIG
model would have given P2 = 273.7 kPa. This emphasizes at low temperatures, the CPIG is an accurate
model.
6.8 Two cycles

In this section, we describe two important thermodynamic cycles. The most scientifically
important is the
• Carnot cycle: thermodynamic cycle with ideal efficiiency,
and we spend some effort in its exposition. We will contrast this with the
• Otto cycle: cycle used to model gasoline engine thermodynamics.
6.8.1 Carnot
Motivated by a practical desire to improve French industry in the aftermath of military de-
feats of earlier decades, Nicolas Lèonard Sadi Carnot (1796-1832), (son of the mathematician
and architect of the military success of the early French revolution, Lazare Carnot) devel-
oped an impractical engine with great theoretical importance. Though the so-called Carnot
engine has never been built, it represents the best heat engine that could be built, and im-
poses useful restrictions for practical engineers with visions of unrealizable efficiencies. Most
importantly, the analysis of Carnot demonstrates how perpetual motion machines of the first
and second kind cannot exist. Those of the first kind violate the first law of thermodynamics;
those of the second kind violate the second law of thermodynamics.
Let us use a piston-cylinder arrangement to illustrate a Carnot cycle. See the sketch of
Fig. 6.24. A sketch of the process in both the P − v and T − s planes for a CPIG is given
in Fig. 6.25.

4 4
3 3
1 1
2 2
TH TL
QH QL
1 to 2 2 to 3 3 to 4 4 to 1
isentropic isothermal isentropic isothermal
compression expansion expansion compression
Figure 6.24: Sketch of a Carnot cycle piston-cylinder device.
P T
2
isot
isotherm 3
2
her isentro
m
isentrope
isentrope
T
H
3
pe
iso
the
ise
1 rm ro 1 isotherm 4
nt
TL p e
4
v s
Figure 6.25: Sketch of a Carnot cycle for a CPIG represented in the P − v and T − s planes.

6.8. TWO CYCLES 279
The Carnot cycle is defined in four stages. Here, we use a different identification of the
states 1, 2, 3, and 4 than do BS to be consistent with more common notation that will be
used later for other engines by BS and most other texts. The four stages are
• 1 → 2: adiabatic reversible (isentropic) compression (Q = 0) from TL to TH ,
• 2 → 3: isothermal reversible expansion at TH ,
• 3 → 4: adiabatic reversible (isentropic) expansion from TH to TL , and
• 4 → 1: isothermal compression at TL .
We always assume the same fluid is present within the Carnot engine, and ignore the effects
of valves. We also ignore the effect of mixtures of combustible gases.
The Carnot cycle for a CPIG is such a foundational idealization for mechanical engineers
that it is centrally incorporated in the logo of the international mechanical engineering
academic honor society, Pi Tau Sigma (ΠTΣ). The logo is reproduced in Fig. 6.26.
Figure 6.26: Logo of the international mechanical engineering honor society, Pi Tau Sigma
(ΠTΣ) featuring the Carnot cycle for a CPIG in the P − v plane as displayed on the campus
of the University of Notre Dame.
Example 6.13
Given k, R, TL = T1 = T4 , TH = T2 = T3 , v1 and v4 , demonstrate for a CPIG undergoing a Carnot
cycle, that the earlier derived Eq. (5.23), η = 1 − TL /TH , is true. Also find the net work and heat
transfer.
First note that cv = R/(k − 1).

• For process 1 → 2, we have the first law
u2 − u1 = 1q2 − 1w2 , (6.296)

| {z } |{z} |{z}
R2
=cv (T2 −T1 ) =0 = 1 P dv
Z 2
cv (T2 − T1 ) = − P dv, (6.297)
1
Z 2
cv (TH − TL ) = − P dv, (6.298)
1
Z 2
R
(TH − TL ) = − P dv. (6.299)
k−1 1
For the isentropic process P v k = P1 v1k , and we know from the per mass version of Eq. (3.92) for work
for a polytropic process that when n = k, the work must be
P2 v2 − P1 v1
1w2 = , (6.300)
1−k
RT2 − RT1
= , (6.301)
1−k
R
= − (T2 − T1 ), (6.302)
k−1
R
= − (TH − TL ). (6.303)
k−1
This is consistent with the change in internal energy being the negative of the work done.
• For process 2 → 3, the first law gives
u3 − u2 = 2q3 − 2w3 , (6.304)

Z 3
cv (T3 − T2 ) = q
2 3 − P dv, (6.305)
| {z } 2
=0
Z 3
2q3 = P dv, (6.306)
2
Z 3
dv
= RTH , (6.307)
2 v
v3
= RTH ln . (6.308)
v2
Now, we have from the isentropic relations that

1
k−1 1
k−1 1
k−1
T1 TL TL
v2 = v1 = v1 , v3 = v4 . (6.309)
T2 TH TH
So
 1
k−1 
TL
v
 4 TH  v4
2q3 = RTH ln  1
k−1  = RTH ln . (6.310)
v1
v1 TTHL

6.8. TWO CYCLES 281
u4 − u3 = 3q4 −3w4 , (6.311)

|{z}
=0
cv (T4 − T3 ) = −3w4 , (6.312)

R
(TL − TH ) = −3w4 . (6.313)
k−1
u1 − u4 = 4q1 − 4w1 , (6.314)

Z 1
cv (T1 − T4 ) = 4q1 − P dv, (6.315)
| {z } 4
=0
Z 1
4q1 = P dv, (6.316)
4
Z 4
dv
= RTL , (6.317)
1 v
v1
= RTL ln . (6.318)
v4
Now, we know the thermal efficiency is from Eq. (5.10)
QL
η =1− . (6.319)
QH
For us
v1 v4
QL ∼ |4q1 | = RTL ln = RTL ln , (6.320)
v4 v1
v4 v4
QH ∼ |2q3 | = RTH ln = RTH ln . (6.321)
v1 v1
So
RTL ln vv14
η =1− , (6.322)
RTH ln vv41
TL
η =1− , Q.E.D.10 (6.323)
TH
Also, we easily get the net work and heat transfer via
v4 v1 v4
wcycle = qcycle = 2q3 + 4q1 = RTH ln + RTL ln = R(TH − TL ) ln . (6.324)
v1 v4 v1
We summarize the first law statements in Table 6.2.
10
abbreviation of the Latin phrase, quod erat demonstrandum, “that which was to have been demon-
,′ ,′
strated.” The original Greek version, used by Euclid and Archimedes, was o πǫρ ǫ δǫι δǫι̂ξαι, translated as
“precisely what was required to be proved,” (https://en.wikipedia.org/wiki/Q.E.D.).

Process ∆u q w
R R
1→2 k−1 (TH
− TL ) 0 − k−1 (TH− TL )
2→3 0 RTH ln vv41 RTH ln vv41
R R
3→4 − k−1 (TH − TL ) 0 k−1 (TH − TL )
v1
4→1 0 RTL ln v4 RTL ln vv14
Total 0 R(TH − TL ) ln vv14 R(TH − TL ) ln vv14
Table 6.2: First law analysis summary for our CPIG undergoing a Carnot cycle.
Example 6.14
A mass of m = 1 kg of water executes a Carnot cycle. The high temperature isothermal expansion
is from P2 = 15 bar, x2 = 0.25 to the saturated vapor state. The adiabatic expansion is to P4 = 1 bar.
Analyze the system.
The system is sketched in Fig. 6.27. We have P2 = 1500 kPa, x2 = 0.25. From the tables we find
P T
2 3 isotherm
2 3
isentrope
isentrope
T=
471.4
7K
T=
1 4 372.7 1 4
7K isotherm
v s
Figure 6.27: Sketch of a Carnot cycle for water represented in the P − v and T − s planes.
the following
T2 = 198.32 ◦ C = 471.47 K, (6.325)
m3 m3
vf 2 = 0.001154 , vf g2 = 0.13062 , (6.326)
kg kg
kJ kJ
uf 2 = 843.14 , uf g2 = 1751.3 , (6.327)
kg kg
kJ kJ
sf 2 = 2.3150 , sf g2 = 4.1298 . (6.328)
kg K kg K

6.8. TWO CYCLES 283
So we get the properties of state 2 via

m3
v2 = vf 2 + x2 vf g2 = 0.033809 , (6.329)
kg
kJ
u2 = uf 2 + x2 uf g2 = 1280.97 , (6.330)
kg
kJ
s2 = sf 2 + x2 sf g2 = 3.3475 . (6.331)
kg K
At state 3, we know P3 = P2 = 1500 kPa, and x3 = 1, so the properties can be read from the
saturated water tables for x3 = 1. We find
m3
v3 = 0.13177 , (6.332)
kg
kJ
u3 = 2594.5 , (6.333)
kg
kJ
s3 = 6.4448 . (6.334)
kg K
At state 4, we know P4 = 100 kPa. We also know that s4 = s3 = 6.4448 kJ/kg/K. So the state is fixed.
At state 4, the tables give
T4 = 99.62 ◦ C = 372.77 K, (6.335)
m3 m3
vf 4 = 0.001043 , vf g4 = 1.69296 , (6.336)
kg kg
kJ kJ
uf 4 = 417.33 , uf g4 = 2088.72 , (6.337)
kg kg
kJ kJ
sf 4 = 1.3025 , sf g4 = 6.0568 . (6.338)
kg K kg K
Knowing s4 , we can find the quality x4 via

s4 − sf 4 6.4448 kgkJK − 1.3025 kgkJK
x4 = = = 0.849013. (6.339)
sf g4 6.0568 kgkJK
We then find
m3
v4 = vf 4 + x4 vf g4 = 1.43838 , (6.340)
kg
kJ
u4 = uf 4 + x4 uf g4 = 2190.68 . (6.341)
kg
Now, we know that for state 1, s1 = s2 = 3.34745 kJ/kg/K, and sf 1 = sf 4 , sf g1 = sf g4 . So
s1 − sf 1
x1 = = 0.337629. (6.342)
sf g1
We also know that T1 = T4 = 372.77 K. We also know that vf 1 = vf 4 , vf g1 = vf g4 , uf 1 = uf 4 ,
uf g1 = uf g4 . So
m3
v1 = vf 1 + x1 vf g1 = 0.572634 , (6.343)
kg
kJ
u1 = uf 1 + x1 uf g1 = 1122.54 . (6.344)
kg
Now, let us consider the first law for each process.

• 1 → 2: The first law gives

u2 − u1 = 1q2 −1w2 , (6.345)
|{z}
=0
1w2 = u1 − u2 , (6.346)

kJ kJ
1w2 = 1122.54 − 1280.97 , (6.347)
kg kg
kJ
= −158.423 . (6.348)
kg
u3 − u2 = 2q3 − 2w3 . (6.349)
Now,
Z 3
2w3 = P dv, (6.350)
2
= P3 (v3 − v2 ), (6.351)

m3 m3
= (1500 kPa) 0.13177 − 0.033809 , (6.352)
kg kg
kJ
= 146.942 . (6.353)
kg
So
2q3 = (u3 − u2 ) + 2w3 , (6.354)

kJ kJ kJ
= 2594.5 − 1280.97 + 146.942 , (6.355)
kg kg kg
kJ
= 1460.48 . (6.356)
kg
u4 − u3 = 3q4 −3w4 . (6.357)
|{z}
=0
So
3w4 = u3 − u4 , (6.358)

kJ kJ
= 2594.5 − 2190.68 , (6.359)
kg kg
kJ
= 403.82 . (6.360)
kg
u1 − u4 = 4q1 − 4w1 . (6.361)
Now
Z 1
4w1 = P dv, (6.362)
4
= P1 (v1 − v4 ), (6.363)

m3 m3
= (100 kPa) 0.572634 − 1.43838 , (6.364)
kg kg
kJ
= −86.5751 . (6.365)
kg

6.8. TWO CYCLES 285
So
4q1 = (u1 − u4 ) + 4w1 , (6.366)

kJ kJ kJ
= 1122.54 − 2190.68 + −86.5751 , (6.367)
kg kg kg
kJ
= −1154.71 . (6.368)
kg
The net work for the cycle is
wcycle = 1w2 + 2w3 + 3w4 + 4w1 , (6.369)

kJ kJ kJ kJ
= −158.423 + 146.942 + 403.82 + −86.5751 , (6.370)
kg kg kg kg
kJ
= 305.764 . (6.371)
kg
The net heat transfer for the cycle is
qcycle = 2q3+ 4q1 , (6.372)

kJ kJ
= 1460.48 + −1154.71 , (6.373)
kg kg
kJ
= 305.764 . (6.374)
kg
Note that, as expected wcycle = qcycle . Now, let us calculate the thermal efficiency.
wcycle
η = , (6.375)
qin
wcycle
= , (6.376)
2q3
kJ
305.764 kg
= kJ
, (6.377)
1460.48 kg
= 0.209359. (6.378)
This should be the same as the Carnot theory’s prediction
TL
η = 1− , (6.379)
TH
372.77 K
= 1− , (6.380)
471.47 K
= 0.209345. (6.381)
The small differences are likely due to interpolation error from using the tables. We summarize the
first law statements in Table 6.3.


kJ kJ kJ
Process ∆u kg q kg w kg
1→2 158.423 0 -158.423
2→3 1313.54 1460.48 146.942
3→4 -403.82 0 403.82
4→1 -1068.14 -1154.71 -86.5751
Total 0 305.764 305.764
Table 6.3: First law analysis summary for our water undergoing a Carnot cycle.
Figure 6.28: Nikolaus August Otto (1832-1891), German developer of the internal combus-
tion engine. Image from https://en.wikipedia.org/wiki/Nicholas Otto.
6.8.2 Otto
The Otto cycle approximates the gasoline engine using what is known as an air standard
approximation. The air standard will be considered in more detail in Ch. 8.2. It is named
for Nikolaus Otto, depicted in Fig. 6.28. Many details are ignored (like inlet and exhaust),
and all material properties are taken to be those of air modeled as a CPIG. It employs a
fixed mass approach. Diagrams for P − v and T − s for the Otto cycle are shown in Fig. 6.29.
One can outline the Otto cycle as follows:
• 1 → 2: isentropic compression in the compression stroke,
• 2 → 3: isochoric heating in the combustion stroke during spark ignition,
• 3 → 4: isentropic expansion in power stroke, and
• 4 → 1: isochoric rejection of heat to the surroundings.

Clearly, the cycle is not a Carnot cycle. The heat transfer during the combustion and
exhaust strokes does not take place at constant temperature. Roughly speaking, we might
expect degradation of the thermal efficiency, relative to an equivalent Carnot engine operating
between the same temperature bounds, because some of the heat transfer of the Otto cycle

6.8. TWO CYCLES 287
P T
3
3
combustion
po
isentrope
power
isochore
we
ion
r
bust
ise
nt com
ro hore
p 2 isoc
compression
e
isen 4
isentrope
trop e
2 4 isochor
exhaust
e
isochore
com st
pres
sion 1 exhau
1
v s
Figure 6.29: P − v and T − s diagrams for the Otto cycle.
occurs at lower temperatures than other parts of the cycle. Recall that for maximum Carnot
efficiency, we would like TH as high as possible. Just past state 2, the heat transferred at T2
is at a lower temperature than the heat transferred at T3 .
Note for isochoric heating, such as 2 → 3, in a fixed mass environment, the first law gives
u3 − u2 = 2q3
− 2w3 , (6.382)
Z v3
= 2q3 − P dv, but v2 = v3 , (6.383)
v2
Z v2
= 2q3 − P dv, (6.384)
v2
| {z }
=0
2q3 = u3 − u2 , (6.385)
= cv (T3 − T2 ), if CPIG. (6.386)
This will give us
QH = mcv (T3 − T2 ). (6.387)
Similarly for the isochoric heat rejection, we have
QL = m|4q1 | = mcv |T1 − T4 | = mcv (T4 − T1 ). (6.388)

The thermal efficiency is found as follows:

Wnet
η = , (6.389)
QH
QH − QL
= , (6.390)
QH
QL
= 1− , (6.391)
QH
mcv (T4 − T1 )
= 1− , (6.392)
mcv (T3 − T2 )
T4 − T1
= 1− , (6.393)
T3 − T2

T1 TT14 − 1
= 1− . (6.394)
T3
T2 T2 − 1
Now, one also has the isentropic relations:

k−1
T2 V1
= , (6.395)
T1 V2
k−1
T3 V4
= . (6.396)
T4 V3
But V4 = V1 and V2 = V3 , so
k−1
T3 V1 T2
= = . (6.397)
T4 V2 T1
Cross multiplying the temperatures, one finds
T3 T4
= . (6.398)
T2 T1
Thus, the thermal efficiency reduces to
T1
η =1− . (6.399)
T2
This looks a lot like the Carnot efficiency. But for a Carnot engine operating between the
same temperature bounds, we would have found η = 1 −T1 /T3 . Because T3 > T2 , the Carnot
engine is more efficient than the ideal Otto engine. This identifies an important
• Thermal engine design principle: To optimize the performance of a thermal

engine, the T − s diagram describing its behavior should be as close to a rectangle as
possible, with the highest possible TH and the lowest possible TL .

6.8. TWO CYCLES 289
One often finds commercial engines characterized by their compression ratios. Modern
gasoline engines may have compression ratios of rv = 10. In terms of the compression ratio
rv = V1 /V2 , one has
1
η = 1 − rv1−k = 1 − k−1 . (6.400)
rv
Note if the compression ratio increases, the thermal efficiency increases, so this is desirable,
in principle. However, high compression ratios introduces a variety of problems including
1) loss of material strength of hot metals in the engine, 2) higher incidence of detonation
or knock in the combustion process, 3) greater tendency to form harmful pollutants such as
NOx .
Some deviations of actual performance from that of the air-standard Otto cycle are as
follows:
• specific heats actually vary with temperature,
• combustion may be incomplete (induces pollution and lowers fuel efficiency),
• work of inlet and exhaust is ignored, and
• losses of heat transfer to engine walls are ignored.
Example 6.15
The temperature prior to compression in an air-standard Otto cycle with a compression ratio of 7
is 500 ◦ R, and the pressure is 1 atm. The maximum temperature is 3500 ◦ R. Find
• temperature and pressure at each stage of the process, and
• thermal efficiency.
For the isentropic compression,

k−1
V1
T2 = T1 , (6.401)
V2
= (500 ◦ R)(7)1.4−1 , (6.402)
= 1088.95 ◦ R. (6.403)
One can use the ideal gas law to get the pressure at state 2:
P2 V2 P1 V1
= , (6.404)
T2 T1
V1 T2
P2 = P1 , (6.405)
V2 T1

1088.95 ◦ R
= (1 atm)(7) , (6.406)
500 ◦ R
= 15.2453 atm. (6.407)

Now, V3 = V2 because the combustion is isochoric. And the maximum temperature is T3 = 3500 ◦ R.
This allows use of the ideal gas law to get P3 :
P3 V3 P2 V2
= , (6.408)
T3 T2
V2 T3
P3 = P2 , (6.409)
V3 T2
|{z}
=1

3500 ◦ R
= (15.2453 atm)(1) , (6.410)
1088.95 ◦ R
= 49 atm. (6.411)
One uses the isentropic relations for state 4:

k−1
T3 V4
= , (6.412)
T4 V3
k−1
T4 V3
= , (6.413)
T3 V4
k−1
V2
T4 = T3 , (6.414)
V1
1−k
V1
= T3 , (6.415)
V2
= (3500 ◦ R)(7)1−1.4 , (6.416)
◦
= 1607.05 R. (6.417)
For the pressure at state 4, use the ideal gas law:

P4 V4 P3 V3
= , (6.418)
T4 T3
V3 T4
P4 = P3 , (6.419)
V4 T3
V2 T4
= P3 , (6.420)
V1 T3

1 1607.05 ◦ R
= (49 atm) , (6.421)
7 3500 ◦ R
= 3.2141 atm. (6.422)
The thermal efficiency is

1
η = 1− , (6.423)
rvk−1
1
= 1− , (6.424)
71.4−1
= 0.54083. (6.425)
A Carnot engine operating between the same upper and lower temperature limits would have had
thermal efficiency η = 1 − (500 ◦ R)/(3500 ◦ R) = 0.857143.

6.9. ENTROPY OF THERMO-MECHANICAL MIXING 291
6.9 Entropy of thermo-mechanical mixing

We believe from experience that mixing is most likely a process that is entropy-generating.
For instance, it is easy to mix water and ink together, but hard to separate them into their
original components. Separation can be accomplished, but it takes energy. Now, we can also
mix the energy of components. We might imagine two gases at different states. This is a
structured arrangement. When the two gases mix, they come to a new equilibrium without
external impetus. There is consequently less structure in the universe, and the entropy
should go up. Moreover, it would require an external action to return the system to its
original state. Let us first demonstrate this with a simple example of thermo-mechanically
mixing of two CPIGs initially at different temperatures and pressures, but with otherwise
identical properties. We will enforce thermal equilibrium via temperature equilibration and
mechanical equilibrium via pressure equilibration.
Example 6.16
Consider a fixed volume chamber that is thermally insulated from its outer surroundings and thus
globally adiabatic. The fixed volume chamber is initially divided into two compartments A and B,
by a thin, thermally insulated barrier. Both compartments contain CPIG air (R = 0.287 kJ/kg/K,
cv = 0.7175 kJ/kg/K). We have PA = 100 kPa, TA = 300 K, mA = 10 kg, PB = 200 kPa, TB = 500 K,
and mB = 2 kg. The thin barrier is removed, and the combined system comes to a new mechanical
and thermal equilibrium at state C. Find the entropy change of the universe.
The process is sketched in Fig. 6.30. Use the ideal gas law to get the various volumes:
A B C
PA =100 kPa PB = 200 kPa PC

TA = 300 K TB = 500 K TC
mA = 10 kg mB = 2 kg mC
before mixing after mixing
Figure 6.30: Schematic for a globally adiabatic thermo-mechanical mixing problem.
RTA RTB
vA = , vB = . (6.426)
PA PB
mA RTA mB RTB
VA = mA vA = , VB = mB vB = . (6.427)
PA PB
Now, mass conservation gives
mC = mA + mB , (6.428)
= (10 kg) + (2 kg), (6.429)
= 12 kg. (6.430)

By geometry, we have

mA RTA mB RTB mA T A mB T B
VC = VA + VB = + =R + . (6.431)
PA PB PA PB
So we can get one intensive property at the final state:

mA TA mB TB
VC R PA + PB
vC = = . (6.432)
mC mA + mB
We can use the first law of thermodynamics to get a second property at the final state:
UC − (UA + UB ) = 1Q2 − 1W2 . (6.433)

|{z} |{z}
=0 =0
Because the combined system is adiabatic and isochoric, there is no global heat transfer or work during
the process, so
UC − (UA + UB ) = 0, (6.434)
UC = UA + UB , (6.435)
mC cv (TC − To ) = mA cv (TA − To ) + mB cv (TB − To ), (6.436)
(mA + mB )(TC − To ) = mA (TA − To ) + mB (TB − To ), (6.437)
(mA + mB )TC = mA T A + mB T B , (6.438)
mA T A + mB T B
TC = . (6.439)
mA + mB
The mixture temperature, TC is the mass-weighted average of the two initial temperatures. Note it
would not matter if we used K or ◦ C to get the mixture temperature. The final temperature for our
system is
(10 kg)(300 K) + (2 kg)(500 K)

TC = = 333.333 K. (6.440)
(10 kg) + (2 kg)
Now, we can use the ideal gas to find the final pressure:
A A m T +m T
B B
RTC mA T A + mB T B
PC = = R mAmTAA+mmBB TB = mA TA . (6.441)
vC R PA + PB P + mPB TB A B
mA +mB
The final pressure is
(10 kg)(300 K) + (2 kg)(500 K)

PC = (10 kg)(300 K) (2 kg)(500 K)
= 114.286 kPa. (6.442)
100 kPa + 200 kPa
Now, there are no interactions with the surroundings, so we need only consider the entropy changes
of A and B. First consider the gas that starts at state A and finishes at state C


TC PC
∆SA = mA cP ln − R ln , (6.443)
TA PA

∆SA cP TC PC
= mA ln − ln , (6.444)
R R TA PA

cP TC PC
= mA ln − ln , (6.445)
cP − cv T A PA

k TC PC
= mA ln − ln , (6.446)
k − 1 TA PA
k−1 k !
∆SA mA TC PC
= ln − ln , (6.447)
R(mA + mB ) mA + mB TA PA
1.4 !
10 kg 333.333 K 1.4−1 114.286 kPa
= ln − ln , (6.448)
(10 kg) + (2 kg) 300 K 100 kPa
= 0.196025. (6.449)
Here, we have generated a dimensionless entropy rise by scaling by RmC . Gas A saw its scaled entropy
rise.
Similarly, for the gas that starts at B and ends at C, we have
k !
∆SB mB TC k−1 PC
= ln − ln , (6.450)
R(mA + mB ) mA + mB TB PB
1.4 !
2 kg 333.333 K 1.4−1 114.286 kPa
= ln − ln , (6.451)
(10 kg) + (2 kg) 500 K 200 kPa
= −0.143252. (6.452)
Gas B saw its scaled entropy fall.
The combined system has
∆SA + ∆SB
= 0.196025 − 0.143252 = 0.0527733. (6.453)
R(mA + mB )
Dimensionally

kJ kJ
∆SA + ∆SB = (0.0527733)RmC = (0.0527733) 0.287 (12 kg) = 0.181751 . (6.454)
kg K K
With some more effort, we could prove that the mixing of arbitrary initial states would result in a global
entropy increase.
We note
• The entropy of the universe increased, and we were able to quantify it.
• The adiabatic mixing process we described is irreversible. That is to say, once mixed, we do not
expect to see a spontaneous return to the initial state.
• The entropy of the universe will increase whenever two systems, initially not in equilibrium, come to
an equilibrium.
• Had the initial states been identical, TA = TB , PA = PB , we would have found TA = TB = TC and
PA = PB = PC and ∆SA + ∆SB = 0. This is a consequence of both A and B being formed of the
same gas that was indistinguishable from chamber A to chamber B.

• Had gas A been distinguishable from gas B, for example A may have been entirely O2 and B may have
been entirely N2 , the analysis would be considerably more nuanced. Such an analysis would require
the development of a mixture theory, not considered here. One would find that even if TA = TB = TC
and PA = PB = PC , that there would be an entropy of mixing, not related to thermo-mechanical
effects. This reflects the fact that once mixed, A and B do not spontaneously return to their original
unmixed state without external energy input. Such a mixture theory is the foundation for the science
of chemistry. The more one is able to distinguish gases, the more one has information about them,
and this is related to the entropy. The relation of entropy to information is foundational and discussed
in the following section.
The previous example showed the second law was satisfied if the system came to equi-
librium because the entropy change was positive for the globally adiabatic problem. Let us
turn this around and demonstrate a stronger result: the equilibrium state is the one that
maximizes entropy in an adiabatic mixing problem.
To do this let us consider a simpler mixing problem: globally adiabatic mixing of the ther-
mal energy of two incompressible, immobile calorically perfect solids, initially with distinct
temperatures and thus not in thermal equilibrium. Rather than assert that the temperatures
equilibrate, we take a different approach. We consider all the possible final states that satisfy
the first law of thermodynamics, and select as the final state the one that maximizes the
final entropy of the mixture.
Example 6.17
Consider a chamber that is thermally insulated from its outer surroundings. The chamber is initially
divided into two compartments A and B, by a thin, thermally insulated barrier. Both compartments
contain the same incompressible solid with identical mass m and identical constant specific heat c.
One mass is initially at T = TA1 , and the other is at T = TB1 . The thin barrier is removed, and
the combined system comes to a new state consistent with energy conservation. Show the state that
conserves energy and maximizes entropy is one of thermal equilibrium at the mean temperature TA2 =
TB2 = (TA1 + TB1 )/2.
The process is sketched in Fig. 6.31. Energy conservation for the mixture requires that
U2 − U1 = 1Q2 − 1W2 , (6.455)

|{z} |{z}
0 0
mc(TA2 − TA1 ) + mc(TB2 − TB1 ) = 0, (6.456)
(TA2 − TA1 ) + (TB2 − TB1 ) = 0, (6.457)
(TA2 + TB2 ) − (TA1 + TB1 ) = 0. (6.458)
Here we have not enforced temperature equilibrium at state 2. So there are an infinite number of ways
to partition the energy at state 2 and still maintain energy conservation. One of those partitions of
energy will result in a maximization of mixture entropy. Let us study that.
The entropy change is
TA2 TB2
S2 − S1 = mc ln + mc ln . (6.459)
TA1 TB1

A B A B
m m m m
before mixing after mixing
Figure 6.31: Schematic for a globally adiabatic thermal mixing problem.
So
TA2 TB2
S2 = S1 + mc ln + mc ln . (6.460)
TA1 TB1
We seek the TA2 and TB2 that both maximize entropy S2 and simultaneously satisfy an energy conser-
vation constraint. This is well suited for the method of Lagrange multipliers. Lagrange is depicted in
Fig. 6.32. Let us define an auxiliary function S∗ as
Figure 6.32: Joseph-Louis Lagrange (1736-1813), Franco-Italian mathematician and me-

chanician; image from https://en.wikipedia.org/wiki/Joseph-Louis Lagrange.
TA2 TB2
S∗ = S1 + mc ln + mc ln −λ ((TA2 + TB2 ) − (TA1 + TB1 )) . (6.461)
TA1 TB1 | {z }
| {z } =0
S2
The term λ is known as the Lagrange multiplier. The term it multiplies is our energy conservation
constraint, that must be zero. So if we maximize S∗ , we also maximize S2 . For S∗ to be maximized,
its partial derivatives with respect to both TA2 and TB2 must be zero. The partial derivatives are
∂S∗ mc
= − λ, (6.462)
∂TA2 TA2
∂S∗ mc
= − λ. (6.463)
∂TB2 TB2

Setting both the partial derivatives to zero, and simultaneously enforcing the energy conservation
constraint, Eq. (6.458), gives three equations for the three unknowns TA2 , TB2 , and λ:
mc
− λ = 0, (6.464)
TA2
mc
− λ = 0, (6.465)
TB2
(TA2 + TB2 ) − (TA1 + TB1 ) = 0. (6.466)
Subtracting the second from the first and dividing by mc gives
1 1
− = 0, (6.467)
TA2 TB2
1 1
= , (6.468)
TA2 TB2
TA2 = TB2 . (6.469)
The final temperatures that maximize entropy must be equal. Let us apply this result to energy
conservation, Eq. (6.458), to see what they are:
(TA2 + TA2 ) − (TA1 + TB1 ) = 0, (6.470)
2TA2 = TA1 + TB1 , (6.471)
TA1 + TB1
TA2 = TB2 = . (6.472)
2
For the system we study, the equilibrium temperature is the mean of the original temperatures. Though
not particularly important, the Lagrange multiplier is
2mc
λ= . (6.473)
TA1 + TB1
We can see the mixture entropy maximized, not minimized, via an alternative related analysis. Let
us first recast Eq. (6.460) so that it gives a dimensionless scaled mixture entropy change Ŝ:

S2 − S1 TA2 TB2
Ŝ = = ln . (6.474)
mc TA1 TB1
Obviously, if S2 is maximized, so is Ŝ. Operate further to get
TA1 TB1 eŜ = TA2 TB2 , (6.475)

TA1 TB1 TA2 TB2
eŜ = , (6.476)
(TA1 + TB1 )2 TA1 + TB1 TA1 + TB1
| {z }
S̃
With S̃ ≡ TA1 TB1 eŜ /(TA1 + TB1 )2 , we see that if we maximize S2 , we maximize Ŝ and S̃. Thus, we
have
TA2 TB2
S̃ = . (6.477)
TA1 + TB1 TA1 + TB1
Next invoke energy conservation, Eq. (6.458) to eliminate TB2 and yield
TB2
z }| {
TA2 (TB1 + TA1 − TA2 )
S̃ = , (6.478)
TA1 + TB1 TA1 + TB1

TA2 TA2
= 1− . (6.479)
TA1 + TB1 TA1 + TB1

With the dimensionless final temperature of A defined as T̂A2 = TA2 /(TA1 + TB1 ), we have
S̃ = T̂A2 (1 − T̂A2 ). (6.480)
It takes on a critical value when its first derivative is zero:
dS̃
= 1 − 2T̂A2 = 0. (6.481)
dT̂A2
This occurs when

1
T̂A2 = , (6.482)
2
TA2 1
= , (6.483)
TA1 + TB1 2
TA1 + TB1
TA2 = . (6.484)
2
This is consistent with our result obtained using the method of Lagrange multipliers, Eq. (6.472. The
second derivative test reveals the critical point is a maximum because it is negative:
d2 S̃
2
= −2. (6.485)
dT̂A2
A plot of the scaled entropy S̃ as a function of the scaled final temperature T̂A2 is given in Fig. 6.33.
maximum scaled entropy when
0.25
0.20
0.15
0.10
0.05
0.2 0.4 0.6 0.8 1.0
Figure 6.33: Scaled entropy as a function of scaled final temperature for a globally adiabatic
thermal mixing problem.

6.10 Probabilistic approach to entropy

One of the more difficult concepts regarding entropy is how it relates to the randomness of
a system. In fact what constitutes randomness vis-à-vis structure may be open to question.
Consider the diagram of Fig. 6.34. Here, we take the level of the grey scale to be propor-
individual blocks at individual blocks at

different temperatures: uniform intermediate
indication of structure, order, temperature: indication
and low entropy of lack of structure,
randomness, and high
entropy
Figure 6.34: Two scenarios for the temperature field with the grey scale proportional to the
temperature.
tional to the local temperature. The blocks on the left are held at a variety of temperatures,
hot, intermediate, and cold. The blocks on the right are held at the same intermediate tem-
perature. Let us restrict attention to the case where the hot and cold temperature blocks
on the left just balance, so that when the net temperature of all the blocks on the left is
calculated, it is precisely the intermediate temperature of the blocks on the right. For such
a case, the total thermal energy of the left and right configurations is equal. Energy being
equal, which configuration has the higher entropy? One is tempted to say that on the left
because it looks to be more random. But in fact it is the configuration on the right, that is
equivalent to that on the left having come to equilibrium, while conserving energy. The con-
figuration on the left has each block at a different temperature. This is properly considered,
in the sense of thermodynamics, to be a structure. Left to itself, the thermal energy would
diffuse, giving rise to the configuration on the right. Now, the grey-level of each block on
the left is representative of that block’s average kinetic energy. Within each block, there will
be a distribution of kinetic energy for each individual molecule. For the blocks on the right,
there is an overall distribution of randomness, the same for each block. That randomness is
not represented by the uniform grey shade, that only captures the average kinetic energy.
It may be possible to better understand the relationship between entropy and randomness,
such as that depicted in Fig. 6.34, by the following discussion. Let us consider a radically

6.10. PROBABILISTIC APPROACH TO ENTROPY 299
different approach to entropy, first advocated by Boltzmann in the late nineteenth century.
Boltzmann, depicted at two disparate stages in his life in Fig. 6.35, had to struggle mightily
Figure 6.35: Ludwig Boltzmann (1844-1906), Austrian physicist whose statistical ap-
proach to thermodynamics laid the foundation for quantum mechanics; images from
https://mathshistory.st-andrews.ac.uk/Biographies/Boltzmann.html.
for his theories to gain acceptance in a time when the atomic theory of matter was not widely
understood. His arguments have become accepted over time relative to those of his many
detractors. Let us define a set of N possible states, each with a probability of pn . By the
nature of probability, we must have
N
X
pn = 1. (6.486)
n=1
Also because of the nature of probability, we will demand that pn ∈ [0, 1]. That is to say
neither negative probability nor probability greater than unity has any meaning. Let us
define the entropy of the system according to what has become known as the Gibbs entropy
formula:
XN
S = −kB pn ln pn . (6.487)
n=1
where we take kB to be the Boltzmann constant. Boltzmann’s tomb has a variant of this
equation cut into its stone, as shown in Fig. 6.35. As an aside, we note that operating on
Eq. (6.487) yields an alternative expression relating S to pn :
N N N
S X X
pn
Y
− = pn ln pn = ln pn = ln ppnn , (6.488)
kB n=1 n=1 n=1
N
S Y
exp − = ppnn . (6.489)
kB n=1

Boltzmann defined kB such that

R
, kB ≡ (6.490)
N
where N is Avogadro’s number, N = 6.02214179 × 1023 molecule/mole. So
J
8.314472 K mole −23 J
kB = molecule
= 1.380650 × 10 . (6.491)
6.02214179 × 1023 mole K molecule
Example 6.18
Consider a CPIG in terms of the Boltzmann constant.
We have
N R
PV = nRT = n RT = nN T, (6.492)
N N
|{z}
=kB
= nN kB T.
|{z} (6.493)
N
We next define the number of molecules N as the product of the number of moles, that we have taken
to be n, and the number of molecules per mole, N:
N = nN. (6.494)
Thus, the ideal gas law becomes
P V = NkB T. (6.495)
Consider next the internal energy. Neglecting the additive constant, we have
u = cv T. (6.496)
Now, scale by the molecular mass to cast this on a per mole basis:
u cv
= T, (6.497)
M M
cv cv 1
u = cv T, = RT = RT = cP RT, (6.498)
R cP − cv cv − 1
1 N R N
= RT = T = kB T, (6.499)
k−1 k−1N k−1
u 1
= kB T. (6.500)
N k−1
Now, u is the energy per mole. And N is the number of molecules per mole. Now, define û as the
energy per molecule:
u
û = . (6.501)
N
Thus,
1
û = kB T, (6.502)
k−1

J J
= [K]. (6.503)
molecule K molecule

For a monatomic ideal gas, k = 5/3; therefore,

3
û = kB T. (6.504)
2
For monatomic ideal gases, there are three degrees of freedom, one for each of the three dimensions
available for motion, x, y, z; thus, each degree of freedom contributes (1/2)kB T to the internal energy
of the molecule.
Example 6.19
Consider a system with four possible states, 1, 2, 3, and 4. Now, consider three configurations A,
B, and C. For each configuration, that may correspond to a bulk energy level, we have a different set
of probabilities for each state. Let us take:
• Energy level A: p1 = 1, p2 = 0, p3 = 0, p4 = 0.
• Energy level B: p1 = 21 , p2 = 41 , p3 = 18 , p4 = 81 .
• Energy level C: p1 = 41 , p2 = 14 , p3 = 41 , p4 = 41 .
P4
Find the entropy of each energy level via S = −kB n=1 pn ln pn .
• A: SA = −kB (1 ln 1 + 0 ln 0 + 0 ln 0 + 0 ln 0) = 0.

• B: SB = −kB 12 ln 21 + 41 ln 41 + 81 ln 81 + 81 ln 81 = 1.21301kB .

• C: SC = −kB 41 ln 41 + 41 ln 41 + 14 ln 41 + 41 ln 41 = 1.38629kB .
Note the least random is at energy level A, where there is certainty that the configuration is state
1. There is no randomness to this, and the entropy is formally zero. Loosely speaking, this might
correspond to a temperature of absolute zero, where the probability of finding a molecule in its ground
state is unity. Note that we have used the mathematical limit limx→0 x ln x = 0, that can be shown
with l’Hôpital’s rule. Configuration B is less random, with some bias towards states 1 and 2. It has
positive entropy. Configuration C has the highest entropy. In this configuration, all states are equally
likely. It is the most random in this sense.
Example 6.20
Let us do the same exercise with a pair of ordinary dice.
We summarize the probability of rolling each number in Table 6.4. The entropy of this configuration
is S = 2.2694kB . Now, if each roll were equally likely, the entropy would be different. We would need a
single die with eleven sides to achieve this! See Table 6.5. The entropy of this configuration is greater,
and in fact, maximum: S = 2.3979kB .

Roll pn −kB pn ln pn
2 1/36 0.0995kB
3 2/36 0.1605kB
4 3/36 0.2071kB
5 4/36 0.2441kB
6 5/36 0.2742kB
7 6/36 0.2986kB
8 5/36 0.2742kB
9 4/36 0.2441kB
10 3/36 0.2071kB
11 2/36 0.1605kB
12 1/36 0.0995kB
P P
Total pn = 1 −kB pn ln pn = 2.2694kB
Table 6.4: Entropy associated with an ordinary pair of six-sided unbiased dice.
2 1/11 0.21799kB
3 1/11 0.21799kB
4 1/11 0.21799kB
5 1/11 0.21799kB
6 1/11 0.21799kB
7 1/11 0.21799kB
8 1/11 0.21799kB
9 1/11 0.21799kB
10 1/11 0.21799kB
11 1/11 0.21799kB
12 1/11 0.21799kB
P P
Table 6.5: Entropy of an equally weighted eleven-sided die.

Example 6.21
Consider the entropy associated with the results of rolling an ordinary unbiased six-sided die.
Compare this with a variety of biased or so-called “loaded” die.
First consider the unbiased or non-“loaded” die. The probability of rolling each number 1 through
6 is pn = 1/6. The entropy associated with each roll is −kB pn ln pn = −kB (1/6) ln(1/6) = 0.298627kB .
See Table 6.6. The total entropy is
6
X
S = −kB pn ln pn = 1.79176kB . (6.505)
n=1
Now, consider the average value of a roll of the die. A given roll will result in a value of n = 1, . . . , 6.
Let us define the average value of n to be <n>. Using the ordinary rules of probability, we expect that
6
X 1 1 1 1 1 1 7
<n> = npn = 1 +2 +3 +4 +5 +6 = . (6.506)
n=1
6 6 6 6 6 6 2
Let us turn our analysis around. Let us say we know <n> = 7/2, but do not know pn . Ther-
modynamically, this is equivalent to saying we know the mean kinetic energy of our molecules, but
do not know the kinetic energy of an individual molecule. Given that there are six possible states of
the system, we would like to know how this energy is distributed among the possible states. Now, we
already know that pn = 1/6 for n = 1, . . . , 6 is a possibility. But it is easy to check that p3 = p4 = 1/2,
p1 = p2 = p5 = pP
P 6 = 0 is another possibility. In fact, there are an infinite set of pn for which
6 6
p
n=1 n = 1 and n=1 npn = 7/2. To fix our selection of the distribution of P pn , let us assert
6
that the entropy is maximized, while simultaneously satisfying the two constraints n=1 pn = 1 and
P6
n=1 npn = 7/2.
This gives rise to a problem in constrained optimization, that can be solved with the method of
Lagrange multipliers. Thus, our problem is to select pn so as to maximize
6 6
! 6
!
∗
X X 7 X
S = −kB pn ln pn + λ1 1− pn + λ2 − npn , (6.507)
n=1 n=1
2 n=1
1 1/6 0.298627kB
2 1/6 0.298627kB
3 1/6 0.298627kB
4 1/6 0.298627kB
5 1/6 0.298627kB
6 1/6 0.298627kB
P P
Table 6.6: Entropy of an equally weighted six-sided die.

subject to the constraints

6
X
pn = 1, (6.508)
n=1
6
X 7
npn = . (6.509)
n=1
2
P6 P6
Here S ∗ = S + λ1 1 − n=1 pn + λ2 7/2 − n=1 npn , and λ1 and λ2 are the Lagrange multipliers.
If we maximize S ∗ , while satisfying the constraints, we also maximize S. To obtain a solution for values
of pn that maximize S ∗ , we differentiate Eq. (6.507) with respect to each pn and set the result to zero.
We thus get six equations:
∂S ∗
= −kB (1 + ln p1 ) − λ1 − λ2 = 0, (6.510)
∂p1
∂S ∗
= −kB (1 + ln p2 ) − λ1 − 2λ2 = 0, (6.511)
∂p2
∂S ∗
= −kB (1 + ln p3 ) − λ1 − 3λ2 = 0, (6.512)
∂p3
∂S ∗
= −kB (1 + ln p4 ) − λ1 − 4λ2 = 0, (6.513)
∂p4
∂S ∗
= −kB (1 + ln p5 ) − λ1 − 5λ2 = 0, (6.514)
∂p5
∂S ∗
= −kB (1 + ln p6 ) − λ1 − 6λ2 = 0. (6.515)
∂p6
These six equations take the general form
−kB (1 + ln pn ) − λ1 − nλ2 = 0, n = 1, . . . , 6, (6.516)
that reduces to
λ1 λ2
ln pn = −1 − − n, (6.517)
k kB
B
λ1 λ2
pn = exp −1 − − n , (6.518)
kB kB

λ1 λ2
pn = exp −1 − exp −n . (6.519)
kB kB
These six equations, embodied in Eq. (6.519), in conjunction with the two constraints, Eqs. (6.508,6.509),
form eight equations for the eight unknowns p1 , . . . , p6 , λ1 /kB , λ2 /kB . They can be solved iteratively
by Newton’s method. We find
7 1
<n> = : p1 = p2 = p3 = p4 = p5 = p6 = . (6.520)
2 6
From here on let us take λ∗1 = λ1 /kB and λ∗2 = λ2 /kB . The values of the Lagrange multipliers are
λ∗1 = 0.79176, λ∗2 = 0, but are not of particular importance. One could use a variety of methods to
show the extreme value for S we found was actually a maximum. Thus, for <n> = 7/2, the probability
distribution function that maximizes entropy is in fact uniform, with pn = 1/6. Thermodynamically,
this implies that if the average energy is 7/2, the six accessible states are equally likely to be populated,

as it is this distribution of probabilities that maximizes the system’s entropy and thus satisfies the
second law of thermodynamics.
Let us decrease the mean die roll to <n> = 2 and ask what probability distribution now maximizes
entropy. That is to say, we are going to employ a “loaded” die. Now, there are again an infinite number
of ways to distribute pn to obtain <n> = 2. One obvious way is to take p2 = 1 and p1 = p3 = p4 = p5 =
p6 = 0. But this P will not maximize entropy; certainly, such a die has no randomness to it. Replacing
6
Eq. (6.509) with n=1 npn = 2 and solving the similar Lagrange multipliers problem to determine the
set of pn that maximizes S, we obtain
<n> = 2 : p1 = 0.47812, (6.521)

p2 = 0.254752, (6.522)
p3 = 0.135737, (6.523)
p4 = 0.0723234, (6.524)
p5 = 0.0385354, (6.525)
p6 = 0.0205324. (6.526)
A die loaded in this fashion will maximize entropy or randomness while simultaneously achieving the
desired average behavior, that can be thought of as conserving energy.
If we choose a mildly loaded die, so that <n> = 4, we find the distribution of pn that maximizes
entropy to be slightly biased:
<n> = 4 : p1 = 0.103065, (6.527)

p2 = 0.122731, (6.528)
p3 = 0.146148, (6.529)
p4 = 0.174034, (6.530)
p5 = 0.207240, (6.531)
p6 = 0.246782. (6.532)
The probability distribution functions pn (n) for a variety of <n> values are shown in Fig. 6.36.
Obviously, if <n> = 1, there is only one possibility: the die must be entirely weighted to always yield
a 1; thus, p1 = 1, and all others are zero. There is no randomness, and the entropy is zero. A similar
condition holds for <n> = 6. For <n> ∈ [1, 6], there is a non-trivial distribution of pn that maximizes
S for the given <n>. Note that when <n> = 7/2, that S is maximized, and the probability distribution
is uniform with pn = 1/6. A plot of S/kB versus <n> is given in Fig. 6.37.

1 1 1
<n> = 1 <n> = 2 <n> = 3
S = 0kB S = 1.36747kB S = 1.74851kB
n n n
0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6
1
<n> = 3.5
S = 1.79176kB
n
0 1 2 3 4 5 6
1 1 1
<n> = 4 <n> = 5 <n> = 6
S = 1.74851kB S = 1.36747kB S = 0kB
n n n
0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6
Figure 6.36: Probability distribution functions for a six-sided die weighted so as to achieve
a variety of mean values <n> = 1, 2, 3, 7/2, 4, 5, 6 while maximizing entropy.
0 1 2 3 4 5 6
<n>
Figure 6.37: Scaled entropy as a function of the mean value of the six-sided die.

P6
As an aside, let us reconsider Eq. (6.519), that itself holds for any value of <n>. Because n=1 pn =
1, we can form
6
X 6
X
pn = 1 = exp (−1 − λ∗1 ) exp (−nλ∗2 ) . (6.533)
n=1 n=1
Now, let us divide Eq. (6.519) by the left and right sides of Eq. (6.533) so as to get
exp (−nλ∗2 )
pn = P6 . (6.534)
∗
n=1 exp (−nλ2 )
We next define the so-called partition function, Z(λ∗2 ), as

6
X
Z(λ∗2 ) = exp (−nλ∗2 ) . (6.535)
n=1
Now, rewriting Eq. (6.517), we have

ln pn = −1 − λ∗1 − nλ∗2 . (6.536)
Now, take the logarithm of Eq. (6.534) and employ the definition of the partition function, Eq. (6.535),
to get
ln pn = −nλ∗2 − ln Z. (6.537)
Combining Eqs. (6.536) and (6.537), we get
1 + λ∗1 = ln Z. (6.538)
This relates the first Lagrange multiplier to the second, that is embodied within Z.
Now, let us uncover a non-obvious, but useful relation through the following sequence of operations.
Let us prove that
d
<n> = − ∗ ln Z. (6.539)
dλ2
We can show this by employing our definitions combined with a series of operations and see if we are
led to an identity. First use the chain rule
dZ d
<n> = − ln Z, (6.540)
dλ∗2 dZ
dZ 1
= − ∗ , (6.541)
dλ2 Z
6
!
d X ∗ 1
= − ∗ exp (−nλ2 ) , (6.542)
dλ2 n=1 Z
6
!
X 1
= n exp (−nλ∗2 ) , (6.543)
n=1
Z
P6
n exp (−nλ∗2 )
= Pn=1
6 ∗
. (6.544)
n=1 exp (−nλ2 )
Now, we also know by definition of the average and the probability pn that
6
X
<n> = npn . (6.545)
n=1

Using Eq. (6.534) to remove pn from Eq. (6.545), we recover exactly Eq. (6.544). Thus our assertion of
Eq. (6.539) is true.
Lastly, we see that given <n>, it is possible to solve Eq. (6.544) to find λ∗2 . With λ∗2 , one can then
use Eq. (6.534) to find each of the pn . In fact, the problem reduces to solving a polynomial equation.
If we take x = exp(−λ∗2 ), Eq. (6.544) reduces to
x + 2x2 + 3x3 + 4x4 + 5x5 + 6x6

<n> = . (6.546)
x + x2 + x3 + x4 + x5 + x6
A factor of x cancels, and after rearrangement, we find a fifth order polynomial in x:
(6 − <n>)x5 + (5 − <n>)x4 + (4 − <n>)x3 + (3 − <n>)x2 + (2 − <n>)x + (1 − <n>) = 0. (6.547)
There are five roots to this equation, and at least one real root. It is likely possible to prove that in
cases of interest there is only one real root. Let us check for the special case in which <n> = 3.5 = 7/2.
In that case, we get
5x5 + 3x4 + x3 − x2 − 3x − 5 = 0, (6.548)

(x − 1)(5x4 + 8x3 + 9x2 + 8x + 5) = 0. (6.549)
There is obviously one real root, x = 1. The other four roots are complex. For x = exp(−λ∗2 ) = 1, we
must have λ∗2 = 0, as we have seen to be the case for the uniform distribution pn = 1/6 associated with
<n> = 7/2.
Unaware of Boltzmann’s theory, electrical engineer and mathematician Claude Shannon

published in 1948 what amounts to precisely the same theory in the context of data com-
munication. He arrived at a similar formula as Gibbs for the so-called information entropy,
Sinf ormation :
N
X
Sinf ormation = − pi ln pi . (6.550)
n=1
This is effectively a dimensionless version of the Gibbs entropy function, Eq. (6.487), differing
only by a factor of kB . When applied in such context, Boltzmann’s theory is known as
information theory. Information theory was constructed to quantify data lost in telephone
line signals. The theory and its author has had a seminal effect on modern computer and
communication technologies.11 12 Shannon as well as scenes from a public park dedicated to
him in his hometown of Gaylord, Michigan, is depicted in Fig. 6.38.
11
C. E. Shannon, 1948, “A mathematical theory of communication,” Bell System Technical Journal, 27(3):
379-423.
12
C. E. Shannon, 1948, “A mathematical theory of communication,” Bell System Technical Journal, 27(4):
623-656.

6.11. SUMMARY STATEMENT OF THERMODYNAMICS 309
Figure 6.38: Claude Elwood Shannon (1916-2001), American electrical engi-

neer and mathematician whose “information entropy” is mathematically iden-
tical to Boltzmann’s statistical definition of entropy; leftmost image from
https://en.wikipedia.org/wiki/Claude Elwood Shannon.
6.11 Summary statement of thermodynamics

We have now covered the major ideas of the natural philosophy that is thermodynamics. We
can summarize for an isolated universe by stating
• The energy of the universe is constant, and its entropy is increasing.
This is a loose translation of the statement of Clausius,13
• Die Energie der Welt ist konstant. Die Entropie der Welt strebt einem Maximum zu.
It is a pessimistic set of principles! Thinking cosmologically, these laws would suggest

that a large concentrated mass-energy complex, initially in a highly structured state, would
ultimately deteriorate into spatially homogeneous space dust at a state of final equilibrium,
the so-called heat death scenario posed initially in the nineteenth century by Kelvin and
others. Present non-equilibrium thermodynamic theory would admit local structures to self-
organize into ordered units such as a solar system or living beings. Such structures could
potentially draw the energy necessary for self-organization from residual energy from the
initial state.
Though some disagree, it is claimed here that the science of thermodynamics is incapable
of definitively answering theological questions that often arise regarding the origin of the
universe, its ultimate fate, and the evolution of structures within it. It simply gives a
framework for what is admissible with a given set of assumptions. In that it can illuminate
some aspects of theology by identifying those parts of it that are in and out of agreement with
13
R. Clausius, 1865, “Ueber verschiedene für die Anwendung bequeme Formen der Hauptgleichungen der
mechanischen Wärmetheorie,” Annalen der Physik und Chemie, 125(7): 353-390.

empirical observations and their consequences, it may be of some use to such disciplines that
are outside its realm. More generally, many scientists follow the train of thought popularized
by the Austrian philosopher of science Karl Popper (1902-1994) who restricted scientific
theories to those that are empirically testable, or more specifically, “falsifiable.” Statements
need not be falsifiable to be true, that thus admits the possibility of theological fact; they
simply are not science.
So, if our universe is formally isolated, we can look forward to “heat death” and the
ultimate equilibrium, first suggested by Thomson.14 If it is not isolated, there is more
uncertainty, and perhaps less reason for pessimism. The so-called laws of thermodynamics
are simply an efficient reflection of present-day empirical data. Science is in that sense
radically pragmatic; if unimpeachable data is found that contradict our present axioms of
thermodynamics, science resorts to new and improved axioms.
14
W. Thomson, 1852, “On a universal tendency in nature to the dissipation of mechanical energy,” Trans-
actions of the Royal Society of Edinburgh, 20(3): 139-142.

Chapter 7
Second law analysis for a control

volume
Read BS, Chapter 7
In this chapter, we will apply notions from control volume analysis to problems that involve
the second law of thermodynamics. Recall that the fundamental description of our axioms
is written for systems. We simply modify these axioms when applying them to control
volumes. We shall omit most of the details of the reduction of the second law to control
volume formulation. It is not unlike that done for mass and energy conservation.
7.1 Irreversible entropy production

First recall an important form of the second law for a system, Eq. (6.31):
Z 2
δQ
S2 − S1 ≥ . (7.1)
1 T
Let us introduce a convenient variable, the
• Irreversible entropy production: a quantity that characterizes that portion of
entropy production that is irreversible.
We note that entropy can be produced by reversible heat transfer as well, that we segregate
and do not consider here. We adopt the common notation of 1σ2 for irreversible entropy
production, with units of kJ/K. Our 1σ2 is equivalent to 1S2 gen of Eq. (7.11) of BS (p. 283),
but is more aligned with the notation of non-equilibrium thermodynamics. We give it the
subscripts to emphasize that it is path-dependent. Mathematically, we recast the second law
for a system by the following two equations:
Z 2
δQ
S2 − S1 = + 1σ2 , (7.2)
1 T
1σ2 ≥ 0. (7.3)
311
312 CHAPTER 7. SECOND LAW ANALYSIS FOR A CONTROL VOLUME
Clearly, this is just a notational convenience that moves the inequality from one equation to
another. On a differential basis, we can say for a system
δQ
dS = + δσ. (7.4)
T
And for time-dependent processes, we say for a system
dS 1 δQ δσ
= + . (7.5)
dt T dt dt
Now, let us expand to an unsteady control volume, that should be similar to that for a
system, with corrections for inlets and exits. We get
dScv X Q̇j X X
= + ṁi si − ṁe se + σ̇cv . (7.6)
dt j
Tj i e
Our irreversible entropy production rate, σ̇cv , with units of kW/K, is equivalent to Ṡgen of
Eq. (7.10) of BS (p. 283).
Let us study this equation in some common limits. First, if the problem is in steady
state, then
X Q̇j X X
0= + ṁi si − ṁe se + σ̇cv , (7.7)
j
Tj i e
steady state limit.
If it is in steady state and there is one entrance and one exit, then mass conservation gives
ṁi = ṁe = ṁ, and
X Q̇j
0= + ṁ(si − se ) + σ̇cv , (7.8)
j
Tj
steady state, one entrance, one exit.
We can rearrange to say
1 X Q̇j σ̇cv
se − si = + . (7.9)
ṁ j Tj ṁ
If there is no heat transfer to the control volume, then

σ̇cv
se − si = , (7.10)
ṁ
no heat transfer to control volume, steady state, one entrance/exit.

7.1. IRREVERSIBLE ENTROPY PRODUCTION 313
Example 7.1
Let us revisit an earlier example problem from p. 193 involving throttling and find the irreversible
entropy production flux. Consider calorically perfect ideal air flowing in a duct at P1 = 100000 Pa,
T1 = 300 K, v1 = 10 m/s. Take cP = 1000 J/kg/K. The air is throttled down by a valve to
P2 = 90000 Pa.
We first repeat some of the earlier analysls. From the ideal gas law we get
P1 100000 Pa kg
ρ1 = = = 1.16144 3 . (7.11)
RT1 287 kgJ K (300 K) m
We write the mass, energy, thermal and caloric state equations as

ρ2 v2 = ρ1 v1 , (7.12)
v2 v12
h2 + 2 = h1 + , (7.13)
2 2
P2 = ρ2 RT2 , (7.14)
h2 − h1 = cP (T2 − T1 ). (7.15)
This simplifies to
ρ2 v2 = ρ1 v1 , (7.16)
v2 v12
cP T 2 + 2 = cP T 1 + , (7.17)
2 2
P2 = ρ2 RT2 . (7.18)
Substituting numbers, we get the system of non-linear algebraic equations

kg m kg
ρ2 v2 = 1.16144 3 10 = 11.6144 2 , (7.19)
m s m s
2
J v22 J 10 ms J
1000 T2 + = 1000 (300 K) + = 300050 , (7.20)
kg K 2 kg K 2 kg

J
90000 Pa = ρ2 287 T2 . (7.21)
kg K
This forms three equations in the three unknowns ρ2 , T2 , v2 . Detailed manipulation can reduce this to
a quadratic equation, with two roots. The first root yields
kg
ρ2 = ,
1.04533 (7.22)
m3
T2 = 299.988 K, (7.23)
m
v2 = 11.1107 . (7.24)
s
This is the physical root. From Eq. (6.87), we have
T2 P2
s2 − s1 = cP ln − R ln , (7.25)
T1 P
1
J 299.988 K J 90000 Pa
= 1000 ln − 287 ln , (7.26)
kg K 300 K kg K 100000 Pa
| {z } | {z }
−0.0390793 J/kg/K −30.2385 J/kg/K
J
= 30.1994 . (7.27)
kg K

We can specialize Eq. (7.10) to say

σ̇cv
= s2 − s1 , (7.28)
ṁ
σ̇cv
= ρ1 v1 (s2 − s1 ), (7.29)
A
kg m J
= 1.16144 3 10 30.1994 , (7.30)
m s kg K
W
= 350.748 2 . (7.31)
m K
As required by the second law, this is a positive quantity. Note that because all the entropy change
is associated with irreversibility, and none with heat transfer, the area under the curve on a T − s
diagram, while non-zero, does not represent any heat transfer. One could say
• If there is reversible heat transfer, there will be area under the curve in a T − s diagram.
• If there is area under a curve in a T − s diagram, it could be due to either irreversibility or heat
transfer.
Example 7.2
A steam turbine has an inlet condition of P1 = 30 bar, T1 = 400 ◦ C, v1 = 160 m/s. Its exhaust
condition is T2 = 100 ◦ C, v2 = 100 m/s, x2 = 1. The work for the turbine is wcv = 540 kJ/kg. Find
σ̇cv /ṁ. If A1 = 0.1 m2 , find ṁ and A2 . The surroundings are at 300 K. See Fig. 7.1.
control volume
steam turbine
Figure 7.1: Steam turbine schematic.
Mass conservation tells us

dmcv
= ṁ1 − ṁ2 , (7.32)
dt }
| {z
=0
ṁ1 = ṁ2 = ṁ. (7.33)


dEcv v12 v22
= Q̇cv − Ẇcv + ṁ h1 + − ṁ h2 + , (7.34)
| dt
{z } 2 2
=0
Q̇cv Ẇcv 1 2
= + (h2 − h1 ) + v2 − v12 , (7.35)
ṁ ṁ 2
1 2
qcv = wcv + (h2 − h1 ) + v2 − v12 . (7.36)
2
We find from the tables that P2 = 101.3 kPa, h2 = 2676.05 kJ/kg, and h1 = 3230.82 kJ/kg. So

kJ kJ kJ
qcv = 540 + 2676.05 − 3230.82
kg kg kg
kJ
1 m 2 m 2 kg
+ 100 − 160 2 , (7.37)
2 s s 1000 ms2
kJ
= −22.6 . (7.38)
kg
This represents a loss of heat to the surroundings.

The second law, Eq. (7.9), tells us
X qcv,j σ̇cv
s2 − s1 = + , (7.39)
Tj ṁ
σ̇cv qcv
= s2 − s1 − . (7.40)
ṁ T
From the tables, we find s1 = 6.9212 kJ/kg/K, s2 = 7.3549 kJ/kg/K. So

kJ
!
σ̇cv kJ kJ −22.6 kg
= 7.3549 − 6.9212 − , (7.41)
ṁ kg K kg K 300 K
kJ
= 0.509033 > 0. (7.42)
kg K
The process is sketched in Fig. 7.2.

The tables tell us that ρ1 = 1/v1 = 1/0.09936 m3 /kg = 10.0644 kg/m3 . Thus

kg m kg
ṁ = ρ1 v1 A1 = 10.0644 3 160 0.1 m2 = 161.031 . (7.43)
m s s
The tables tell us that ρ2 = 1/v2 = 1/1.67290 m3 /kg = 0.597764 kg/m3 . From mass conservation, we
get
ṁ 161.031 kg 2
A2 = = s = 2.69388 m . (7.44)
ρ2 v2 kg
0.597764 m3 100 m
s

T P
1
T=
00
4
o
C
P=3000 kPa
2
T=
100 o
P=101.3 kPa 2 C
s v
Figure 7.2: T − s and P − v diagrams for a steam turbine problem.
control volume
Figure 7.3: Steam diffuser schematic.
Example 7.3
Steam is flowing in a diffuser. At the entrance it has P1 = 0.2 MPa, T1 = 200 ◦ C, v1 = 500 m/s.
At the exhaust it has v2 = 10 m/s. Assume an adiabatic reversible process. Find the final pressure and
temperature. If A1 = 0.1 m2 , find ṁ and A2 . See Fig. 7.3.
Because the process is reversible and adiabatic, the second law simply reduces to
s2 = s1 . (7.45)
We go to the tables and find s1 = 7.5066 kJ/kg/K. So s2 = 7.5066 kJ/kg/K.

Now, the first law tells us

dEcv 1 1
= Q̇cv − Ẇcv +ṁ h1 + v12 + gz1 − ṁ h2 + v22 + gz2 . (7.46)
dt }
| {z |{z} |{z} 2 2
=0 =0
=0
Now, we cannot neglect kinetic energy changes in a diffuser. We can neglect potential energy changes.
We can also neglect unsteady effects as well as control volume work. We were told it is adiabatic, so

heat transfer can be neglected. Thus, we get

1 2 1 2
0 = ṁ h1 + v1 − ṁ h2 + v2 , (7.47)
2 2
1 2
h2 = h1 + v − v22 . (7.48)
2 1
The tables give us h1 = 2870.5 kJ/kg. We thus can get
kJ
kJ 1 m 2 m 2 kg kJ
h2 = 2870.5 + 500 − 10 m 2 = 2995.45 . (7.49)
kg 2 s s 1000 s2 kg
Now, we know two properties, h2 and s2 . To find the final state, we have to double interpolate the
superheated steam tables. Doing so, we find
T2 = 264.415 ◦ C, P2 = 346.235 kPa. (7.50)

See Fig. 7.4 for a diagram of the process. Note the temperature rises in this process. The kinetic
T P
2
2T
= 26 o
4 C
P=346 kPa
1 T=2
P=200 kPa 00 o
C
1
s v
Figure 7.4: Steam diffuser process in T − s and P − v planes.
energy is being converted to thermal energy.

The tables tell us that ρ1 = 1/v1 = 1/1.08034 m3 /kg = 0.925635 kg/m3 . Thus

kg m kg
ṁ = ρ1 v1 A1 = 0.925635 3 500 0.1 m2 = 46.2817 . (7.51)
m s s
A double interpolation of the tables tell us that ρ2 = 1/v2 = 1/0.723216 m3 /kg = 1.38271 kg/m3 .
From mass conservation, we get
ṁ 46.2817 kg 2
A2 = = s = 3.34717 m . (7.52)
ρ2 v2 1.38271 mkg
3 10 m
s
It can be shown that the flow in this nozzle is subsonic throughout. The sound speed at the entrance
is estimated at around 540 m/s using methods to be discussed in Sec. 9.8.

7.2 Bernoulli principle

Let us consider our thermodynamics in appropriate limit to develop the well known
• Bernoulli principle: a useful equation in thermal science, often used beyond its
realm of validity, relating pressure, fluid velocity, density, and fluid height, valid only
in the limit in which mechanical energy is conserved.
The principle was first elucidated, though not without considerable turmoil within his prolific
family, by Daniel Bernoulli,1 depicted in Fig. 7.5.
Figure 7.5: Daniel Bernoulli (1700-1782), Dutch-born mathematician and physicist; image
from https://mathshistory.st-andrews.ac.uk/Biographies/Bernoulli Daniel.html.
For the Bernoulli principle to be formally valid requires some restrictive assumptions.
We shall make them here in the context of thermodynamics. The same assumptions allow
one to equivalently obtain the principle from an analysis of the linear momentum equation
developed in fluid mechanics courses. In such a fluids development, we would need to make
several additional, but roughly equivalent, assumptions. This equivalence is obtained because
we shall develop the equation in the limit that mechanical energy is not dissipated. For our
analysis here, we will make the following assumptions:
• the flow is steady,
• all processes are fully reversible,
• there is one inlet and exit, and
• there is contact with one thermal reservoir in which thermal energy is transferred
reversibly.
1
D. Bernoulli, 1738, Hydrodynamica, sive de Viribus et Motibus Fluidorum Commentarii, J. H. Deckeri,
Strasbourg.

7.2. BERNOULLI PRINCIPLE 319
Though we will not study it, there is another important version of the Bernoulli principle
for unsteady flows.
Our second law, Eq. (7.6), reduces in the limits we study to
dScv Q̇cv
= + ṁs1 − ṁs2 + σ̇cv , (7.53)
dt }
| {z T |{z}
=0
=0
Q̇cv
0 =
+ ṁ(s1 − s2 ), (7.54)
T
Q̇cv
ṁ(s2 − s1 ) = , (7.55)
T
ṁT (s2 − s1 ) = Q̇cv . (7.56)
Now, let us non-rigorously generalize this and allow for differential heat transfer at a variety
of temperatures so as to get Z 2
ṁ T ds = Q̇cv . (7.57)
1
In a more formal analysis from continuum mechanics, this step is much cleaner, but would
require significant development. What we really wanted was a simplification for Q̇cv that we
could use in the relevant energy equation, Eq. (4.182), considered next:

dEcv 1 2 1 2
= Q̇cv −Ẇcv + ṁ h1 + v1 + gz1 − ṁ h2 + v2 + gz2 , (7.58)
dt }
| {z |{z}
R2
2 2
=0 =ṁ 1 T ds
Z 2
Ẇcv 1
= T ds + (h1 − h2 ) + (v12 − v22 ) + g(z1 − z2 ). (7.59)
ṁ
|{z} 1 2
wcv
Now, one form of the Gibbs equation, Eq. (6.65), has T ds = dh − v dP , so

Z 2 Z 2 Z 2
T ds = dh − v dP, (7.60)
1 1 1
Z 2
= h2 − h1 − v dP, (7.61)
1
Z 2 Z 2
T ds + (h1 − h2 ) = − v dP. (7.62)
1 1
Now, substitute Eq. (7.62) into Eq. (7.59) to get

Z 2
1
wcv = T ds + (h1 − h2 ) + (v12 − v22 ) + g(z1 − z2 ), (7.63)
2
|1 {z }
R2
=− 1 v dP
Z 2
1
= − v dP + (v12 − v22 ) + g(z1 − z2 ). (7.64)
1 2

Now, if wcv = 0, we get a rarely used, but valuable generalization of the Bernoulli principle:
Z 2
1
0= v dP + (v22 − v12 ) + g(z2 − z1 ). (7.65)
1 2
7.2.1 Incompressible limit

In the important limit for liquids in which v is approximately constant, we recall that ρ = 1/v
and write Eq. (7.65) as
Z 2
1
0 = v dP + (v22 − v12 ) + g(z2 − z1 ), (7.66)
1 2
1
= v(P2 − P1 ) + (v22 − v12 ) + g(z2 − z1 ), (7.67)
2
P2 − P1 1 2
= + (v2 − v12 ) + g(z2 − z1 ). (7.68)
ρ 2
We can rewrite this as
P 1
+ v2 + gz = constant, (7.69)
ρ 2
Bernoulli principle for an incompressible liquid.
In a different limit, that in which changes in kinetic and potential energy can be neglected,
Eq. (7.64) reduces to
Z 2
wcv = − v dP. (7.70)
1
This integral is not the area under the curve in P − v space. It is the area under the
R 2 curve
in v − P space instead. Contrast this with the system result where we get 1w2 = 1 P dv.
In the important operation of pumping liquids, v is nearly constant, and we can say
wpump = −v(P2 − P1 ), (7.71)

Ẇpump = ṁv(P1 − P2 ). (7.72)
7.2.2 Calorically perfect ideal gas limit

Let us consider the Bernoulli principle for a CPIG undergoing a reversible adiabatic process.
For such a process, we have from Eq. (6.170) that P v k = C. Thus, v = (C/P )1/k . Let us

consider Eq. (7.65) for this case:
Z k1
2
C 1
0 = dP + (v22 − v12 ) + g(z2 − z1 ), (7.73)
1 P 2
k1 ! P2
k C 1
= P + (v22 − v12 ) + g(z2 − z1 ), (7.74)
k−1 P 2
P1
k 1
= (P2 v2 − P1 v1 ) + (v22 − v12 ) + g(z2 − z1 ). (7.75)
k−1 2
Thus, for a CPIG obeying the Bernoulli principle, we can say, taking v = 1/ρ,
k P 1
+ v2 + gz = constant, (7.76)
k−1 ρ 2
Bernoulli principle for a CPIG.
While this is fine and similar to the form given for incompressible flow, it really should
be simplified more because P is directly related to ρ for this compressible flow. Using,
Eq. (6.170), P v k = C = Po vok and ρo = 1/vo , we can rewrite Eq. (7.76) as
k−1
k P k Po 1 2
+ v + gz = constant. (7.77)
k−1 Po ρo 2
Similarly for isentropic pumps or turbines using CPIGs with negligible changes in kinetic
and potential energies, Eq. (7.64) reduces to
Z 2
wcv = − v dP, (7.78)
1
k
= − (P2 v2 − P1 v1 ), (7.79)
k−1
k
= − R(T2 − T1 ), (7.80)
k−1
kRT1 T2
= − −1 , (7.81)
k − 1 T1
k−1 !
P2 k
= −cP T1 −1 . (7.82)
P1

An isothermal pump or compressor using a CPIG has

Z 2
wcv = − v dP, (7.83)
1
Z 2
RT
= − dP, (7.84)
1 P
P2
= −RT1 ln , (7.85)
P1
P2
= −P1 v1 ln . (7.86)
P1
7.2.3 Torricelli’s formula

Let us consider a special case of the Bernoulli principle, known as Torricelli’s formula, de-
veloped by Evangelista Torricelli,2 the inventor of the barometer, and for whom the pressure
unit torr is named (1 torr = 133.322 Pa = 1/760 atm.) Torricelli is sketched in Fig. 7.6.
Figure 7.6: Evangelista Torricelli (1608-1647), Italian physicist and mathematician; image
from https://en.wikipedia.org/wiki/Evangelista Torricelli.
Consider the scenario of Fig. 7.7. Here, a fluid is in an open container. The container has
a small hole near its bottom. The fluid at the top of the container, z = z1 , is at P1 = Patm .
The leaking fluid exhausts at the same pressure P2 = Patm . The fluid leaks at velocity v2
at a hole located at z = z2 . The fluid at the top of the container barely moves; so, it has
negligible velocity, v1 ∼ 0. The fluid exists in a constant gravitational field with gravitational
acceleration g, as sketched. Assume the fluid is incompressible and all of the restrictions of
2
E. Torricelli, 1643, De Motu Gravium Naturaliter Accelerato, Firenze.

Figure 7.7: Fluid container with hole.
the Bernoulli principle are present. Let us apply Eq. (7.69):
P1 1 2 P2 1 2
+ v1 +gz1 = + v2 + gz2 . (7.87)
ρ 2 |{z} ρ 2
∼0
Setting P1 = P2 = Patm and ignoring v1 gives

1 2
gz1 = v + gz2 , (7.88)
2 2
p
v2 = 2g(z1 − z2 ), (7.89)
Torricelli’s formula.
Notice rearranging Torricelli’s formula gives

1 2
v2 = g(z1 − z2 ) . (7.90)
2
|{z} | {z }
potential energy
kinetic energy
It represents a balance of kinetic and potential energy of the fluid, and thus is concerned
only with mechanical energy.
Example 7.4
Let us design a liquid water fountain by cutting a hole in a high pressure water pipe. See Fig. 7.8.
We desire the final height of the water jet to be 30 m. The jet rises against a gravitational field with

g = 9.81 m/s2
30 m
Figure 7.8: Sketch of a simple water fountain.
g = 9.81 m/s2 . The atmospheric pressure is 100 kPa. Water has density ρ = 997 kg/m3 . Find the
necessary pipe gauge pressure P1 and jet exit velocity v2 .
Let us apply the Bernoulli principle between states 1 and 3, the pipe interior and the peak of the
height of the fountain. We will estimate the velocity of the water in the pipe to be small, v1 ∼ 0 m/s.
We will also estimate the velocity at the apex of the motion to be negligible, v3 = 0 m/s.
Let us apply Eq. (7.69):
P1 1 2 P3 1 2
+ v1 +gz1 = + v3 +gz3 , (7.91)
ρ 2 |{z} ρ 2 |{z}
∼0 =0
P1 = Patm + ρg(z3 − z1 ), (7.92)
P1 − Patm = ρg(z3 − z1 ). (7.93)
| {z }
=Pgauge

kg m
Pgauge = 997 3 9.81 2 (30 m) = 2.934 × 105 Pa = 293.4 kPa. (7.94)
m s
We can use the same principle to estimate the exit velocity, v2 . Here, we take z1 ∼ z2 .
P1 1 2 P2 1
+ v1 +gz1 = + v22 + g z2 , (7.95)
ρ 2 |{z} ρ 2 |{z}
∼0 ∼z1
P1 Patm 1
= + v22 , (7.96)
ρ ρ 2
s
2(P1 − Patm )
v2 = , (7.97)
ρ
s
2Pgauge
= . (7.98)
ρ


s
2(2.934 × 105 Pa) m
v2 = kg
= 24.26 . (7.99)
997 m3
s
One could use a similar analysis to estimate the necessary pressure to generate the jet of the University
of Notre Dame’s War Memorial Fountain, depicted in Fig. 7.9.
Figure 7.9: The University of Notre Dame’s War Memorial Fountain, 4 June 2010.
Example 7.5
Perform a similar calculation for the problem sketched in Fig. 7.8, but account for mass conservation.
Take the cross-sectional area of the pipe to be A1 = A4 = 1 m2 , and that of the hole to be A2 = 0.01 m2 .
We measure v1 = 1 m/s. See Fig. 7.10. Assume we have the same Pgauge = 293.4 kPa as calculated
Figure 7.10: Schematic of mass balance for a water fountain problem.

earlier. Find the new height of the fountain, z3 , and the new exit velocity v2 .
The mass balance gives us

dmcv
= ṁ1 − ṁ2 − ṁ4 , (7.100)
dt }
| {z
=0
0 = ṁ1 − ṁ2 − ṁ4 , (7.101)
ṁ4 = ṁ1 − ṁ2 . (7.102)

We recall that ṁ = ρvA, so
ρ4 v4 A4 = ρ1 v1 A1 − ρ2 v2 A2 . (7.103)
We assume incompressible flow, so ρ1 = ρ2 = ρ4 = ρ, and we have A1 = A4 , so
ρv4 A1 = ρv1 A1 − ρv2 A2 , (7.104)

A2
v4 = v1 − v2 . (7.105)
A1
This is nice, but not that useful. It simply predicts a lessening of velocity downstream of the hole.
The Bernoulli principle, Eq. (7.69), applied between 1 and 2 gives
P1 1 P2 1
+ v12 + gz1 = + v22 + g z2 , (7.106)
ρ 2 ρ 2 |{z}
∼z1
P1 1 Patm 1
+ v12 = + v22 , (7.107)
ρ 2 ρ 2
s
2(P1 − Patm )
v2 = + v12 , (7.108)
ρ
s
2Pgauge
= + v12 . (7.109)
ρ
With numbers, we get

s
2(2.934 × 105 Pa) m 2 m
v2 = kg
+ 1 = 24.28 . (7.110)
997 m3
s s
The exit velocity is barely changed from our earlier analysis.

Now, determine the new height. Let us again apply Eq. (7.69):
P1 1 P3 1 2
+ v12 + gz1 = + v3 +gz3 , (7.111)
ρ 2 ρ 2 |{z}
=0
P1 1 Patm
+ v12 + gz1 = + gz3 , (7.112)
ρ 2 ρ
P1 − Patm 1
+ v12 = g(z3 − z1 ), (7.113)
ρ 2
Pgauge 1
z3 − z1 = + v12 . (7.114)
ρg 2g
Substituting numbers, we get
2.934 × 105 Pa 1 m 2
z3 − z1 = + m
1 = 30.05 m. (7.115)
kg
997 m3 9.81 s m 2 9.81 s2
s
The extra boost in height comes from accounting for the initial kinetic energy of the water.

7.3. COMPONENT EFFICIENCY 327
7.3 Component efficiency

Recall for cycles, as shown in the analysis surrounding Eq. (4.326), we defined a thermal
efficiency as what you want/what you pay for, that for a power cycle is η = Wnet /QH . We
can further define efficiencies for components. For a component, we will take an efficiency
for some devices to be
what we get
ηcomponent = . (7.116)
the best we could get
For other devices, we will take an efficiency to be
the least required
ηcomponent = . (7.117)
what is actually required
The optimal is generally an isentropic device. So, for example, for a turbine, we say
actual work w
ηturbine = = . (7.118)
work done by an isentropic turbine ws
Here, the subscript “s” denotes isentropic.
For a nozzle, we would like to maximize the kinetic energy of the working fluid, so we
say
v2
ηnozzle = 2 . (7.119)
vs
However for pumps and compressors, the isentropic pump requires the least work input. So
we take instead
ws
ηpump,compressor = . (7.120)
w
Example 7.6
N2 is adiabatically compressed from T1 = 300 K, P1 = 100 kPa to P2 = 1000 kPa. The compressor
efficiency is ηc = 0.9. Find the final state and the compression work per unit mass. Assume N2 is a
CIIG.
The first law for adiabatic compression gives
w = h2 − h1 . (7.121)
Table A.8 from BS tells us that at T1 = 300 K, h1 = 311.67 kJ/kg, soT1 = 6.8463 kJ/kg/K. But we are
not sure what state 2 is, and the first law does not help yet, as we do not know either w or h2 .
Let us calculate state 2 assuming an isentropic process, and then use our knowledge of compressor
efficiency to correct for real effects. We first note for N2 that
kJ
R 8.31451 kmole K kJ
R= = kg
= 0.2968 . (7.122)
M 28.013 kmole kg K
From Eq. (6.98), we can conclude that for an isentropic process in which s2 = s1 that

P2
s2 − s1 = 0 = soT2 − soT1 − R ln , (7.123)
P1
P2
soT2 = soT1 + R ln , (7.124)
P1

kJ kJ 1000 kPa
= 6.8463 + 0.2968 ln , (7.125)
kg K kg K 100 kPa
kJ
= 7.52971 . (7.126)
kg K
Knowing soT2 , we next interpolate Table A.8 from BS to get
kJ
T2s = 576.133 K, h2s = 601.712 . (7.127)
kg
(Note the CPIG assumption would have yielded T2s = (300 K)(10)0.286 = 579.209 K). So the work for
the isentropic compressor is

kJ kJ kJ
ws = h2s − h1 = 601.712 − 311.67 = 290.042 . (7.128)
kg kg kg
Now, consider the compressor efficiency:

ws
ηc = , (7.129)
w
ws
w = , (7.130)
ηc
kJ
290.042 kg
= , (7.131)
0.9
kJ
= 322.269 . (7.132)
kg
Thus, we have the actual work per unit mass. So the actual enthalpy at state 2 can be derived from
the first law:

kJ kJ kJ
h2 = h1 + w = 311.67 + 322.269 = 633.939 . (7.133)
kg kg kg
Now, knowing h2 , we can again interpolate Table A.8 of BS to find the final temperature T2 to be
T2 = 606.263 K. (7.134)
We had to add more energy to achieve the non-isentropic compression relative to the isentropic com-
pression.

Chapter 8
Cycles
Read BS, Chapters 9, 10
In this chapter, we will delve more deeply into some thermodynamic cycles.
8.1 Rankine
Large electric power plants typically utilize a vapor power cycle. Regardless of the heat
source, be it nuclear or combustion of coal, oil, natural gas, wood chips, etc., the remaining
details of these plants are similar. Typically a pure working fluid, usually water, is circulated
through a cycle, and that fluid trades heat and work with its surroundings. We sketch a
typical power plant cycle for electricity generation in Fig. 8.1. The ideal
• Rankine cycle: thermodynamic cycle for power generation with liquid and vaporous
water as the working fluid,
was first described in 1859 by William John Macquorn Rankine, long after the steam engine
was in wide usage. Expanding on our earlier discussion of Ch. 4.5, the cycle has the following
steps:
• 1 → 2: isentropic compression in a pump,
• 2 → 3: isobaric heating in a boiler,
• 3 → 4: isentropic expansion in a turbine, and
• 4 → 1: isobaric cooling in a condenser.
Two variants of the T − s diagram are given in Fig. 8.2. The first is more efficient as it has
the appearance of a Carnot cycle. However, it is impractical, as it induces liquid water in
the turbine, that can damage its blades. So the second is more common.
329
330 CHAPTER 8. CYCLES
combustion
exhaust
..
+
3 generator
turbine _
4
fuel
cold water
air
boiler
condenser hot water
cooling
pump tower
2 1
work
in
Figure 8.1: Rankine cycle schematic.
The thermal efficiency, consistent with what was first introduced in Eq. (4.327), is
Ẇnet Ẇturbine + Ẇpump

η= = . (8.1)
Q̇H Q̇boiler
This reduces to
ṁ ((h3 − h4 ) + (h1 − h2 ))
η = , (8.2)
ṁ (h3 − h2 )
h4 − h1
= 1− , (8.3)
h3 − h2
qout,condenser
= 1− . (8.4)
qin,boiler
This is consistent with the earlier Eq. (5.8). Note that because the Rankine cycle is not a
Carnot cycle, we have qout,condenser /qin,boiler 6= T1 /T3 .
Power plants are sometimes characterized by their
• back work ratio: bwr, the ratio of pump work to turbine work.
Here,
|pump work| h2 − h1
bwr = = . (8.5)
|turbine work| h3 − h4

8.1. RANKINE 331
T T
3
2 2
4
1 4 1
s s
Figure 8.2: T − s diagrams for two Rankine cycles.
We model the pump work as an isentropic process. Recall our analysis for isentropic pumps
that generated Eq. (7.71). The Gibbs equation, Eq. (6.66), gives dh = T ds + v dP . If
ds = 0, we have
dh = v dP. (8.6)
Thus, for the pump
h2 − h1 = v(P2 − P1 ), (8.7)
because v is nearly constant, so the integration is simple.

It might be tempting to make the Rankine cycle into a Carnot cycle as sketched in
Fig. 8.3. However, it is practically difficult to build a pump to handle two-phase mixtures.
The gas phase can seriously damage the pump. Some features that could be desirable for a
Rankine cycle include
• high power output: One can enhance this by raising the fluid to a high temperature
during the combustion process or by pumping the fluid to a high pressure. Both
strategies soon run into material limits; turbine blades melt and pipes burst. Another
strategy is to lower the condenser pressure. That means that one must maintain a
vacuum, and this can be difficult.
• high thermal efficiency: The key design strategy here lies in 1) increasing component
efficiencies, and 2) rendering the overall cycle as much like a Carnot cycle as is feasible.
Modern power plants have had revolutionary increases in overall thermal efficiency
because of enhancements that make the process more Carnot-like.
There are some important loss mechanisms in the Rankine cycle that inhibit efficiency. They
include

2 3
1 4
Figure 8.3: Impractical Rankine-Carnot cycle.
• Turbine losses: These are the major losses. To avoid these losses requires detailed
consideration of fluid mechanics, material science, and heat transfer and is beyond
the scope of classical thermodynamics. Thermodynamics develops broad measures of
turbine efficiency such as ηturbine = (h3 − h4 )/(h3 − h4s ).
• Pump losses: Again, fluid mechanics, machine design, and material science are required
to analyze how to actually avoid these losses. Thermodynamics characterizes them by
pump efficiency, ηpump = (h2s − h1 )/(h2 − h1 ).
• Heat transfer losses from components.
• Pressure drop in pipes.
• Incomplete fuel combustion.
• Pollution removal devices.
• Loss of heat to surroundings in the condenser.
One simple design strategy to make the system more Carnot-like is to use
• Reheat: a design strategy in which steam is extracted from the turbine before it is
fully expanded, then sent to the boiler again, and re-expanded through the remainder
of the turbine.
This has the effect of making the system more like a Carnot cycle. A schematic and T − s
diagram for the Rankine cycle with reheat is given in Fig. 8.4.

8.1. RANKINE 333
3
T
turbine 3 5
boiler
4
5 6
2 4
2
1
6
pump condenser
1
s
Figure 8.4: Rankine cycle with reheat schematic and T − s diagram.
Example 8.1
Consider water in a Rankine power cycle with reheat. The first turbine has water enter at P3 =
8000 kPa, T3 = 500 ◦ C. The water expands to 600 kPa, undergoes reheat, and then expands again to
10 kPa. The mass flow rate is ṁ = 2.63 × 105 kg/hr. We have ηt = 0.88 for each turbine, and ηp = 0.80
for the pump. Find the net power generated, η, and the heat transfer to the condenser.
Let us consider the big picture first. The net specific power will be the positive effect of the two
turbines and the negative effect of the pump:
wnet = (h3 − h4 ) + (h5 − h6 ) + (h1 − h2 ) . (8.8)
| {z } | {z } | {z }
turbine 1 turbine 2 pump
Now, the heat input for the reheat is in two stages:

qin = (h3 − h2 ) + (h5 − h4 ). (8.9)
Lastly, the heat rejection in the condenser is
qout = h6 − h1 . (8.10)
Let us start at the entrance of the first turbine, at 3. We are given P3 and T3 , so we consult the
tables and find
kJ kJ
h3 = 3398.27 , s3 = 6.7239 . (8.11)
kg kg K
We are given P4 = 600 kPa. Now, let us get the ideal behavior of the turbine: s4s = s3 =
6.7239 kJ/kg/K. At this condition, we find state 4 is a two-phase mixture. At 600 kPa, we find
sf = 1.9311 kJ/kg/K, sg = 6.7600 kJ/kg/K. So

kJ kJ
s4s − sf 6.7239 kg K − 1.9311 kg K
x4s = = = 0.992524. (8.12)
sg − sf 6.7600 kJ
− 1.9311 kJ
kg K kg K
We can thus get h4s by consulting the tables to find

kJ kJ kJ
h4s = hf + x4s hf g = 670.54 + (0.992524) 2086.26 = 2741.2 . (8.13)
kg kg kg

Now, ηt = (h3 − h4 )/(h3 − h4s ), so

h4 = h3 − ηt (h3 − h4s ), (8.14)

kJ kJ kJ
= 3398.27 − (0.88) 3398.27 − 2741.2 , (8.15)
kg kg kg
kJ
= 2820.05 . (8.16)
kg
Now, state 5 is after the reheat, that was isobaric at P4 = P5 = 600 kPa, and the reheating returns
the temperature to T5 = 500 ◦ C. From the superheat tables, we find h5 = 3482.75 kJ/kg, s5 =
8.0020 kJ/kg/K. After expansion in the second turbine, we have s6s = s5 = 8.0020 kJ/kg/K. And we
were given P6 = 10 kPa. We consult the saturation tables to find at this pressure sf = 0.6492 kJ/kg/K,
sg = 8.1501 kJ/kg/K. Thus,

s6s − sf 8.0020 kgkJK − 0.6492 kgkJK
x6s = = = 0.980256. (8.17)
sg − sf 8.1501 kgkJK − 0.6492 kgkJK
The tables then give the necessary information to compute h6s :

kJ kJ kJ
h6s = hf + x6s hf g = 191.81 + (0.980256) 2392.82 = 2537.39 . (8.18)
kg kg kg
Now, the actual h6 is found via
h6 = h5 − ηt (h5 − h6s ), (8.19)

kJ kJ kJ
= 3482.75 − (0.88) 3482.75 − 2537.39 , (8.20)
kg kg kg
kJ
= 2650.83 . (8.21)
kg
Now, the tables give us
kJ m3
h1 = hf = 191.81 , v1 = vf = 0.001010 . (8.22)
kg kg
For the pump, we have P1 = P6 = 10 kPa and P2 = P3 = 8000 kPa. So
ws
ηp = , (8.23)
wp
ws
wp = , (8.24)
ηp
v1 (P2 − P1 )
= , (8.25)
ηp
3

0.001010 m kg ((8000 kPa) − (10 kPa))
= , (8.26)
0.8
kJ
= 10.0874 . (8.27)
kg
So
h2 = h1 + wp , (8.28)

kJ kJ
= 191.81 + 10.0874 , (8.29)
kg kg
kJ
= 201.897 . (8.30)
kg

8.1. RANKINE 335
Now, substitute all these values into Eq. (8.8) and get

kJ kJ
wnet = 3398.27 − 2820.05
kg kg
| {z }
turbine 1

kJ kJ
+ 3482.75 − 2650.83
kg kg
| {z }
turbine 2

kJ kJ
+ 191.81 − 201.897 , (8.31)
kg kg
| {z }
pump
kJ
= 1400.05 . (8.32)
kg
On a mass basis, we have

kg hr kJ
Ẇ = ṁwnet = 2.63 × 105 1400.05 = 1.02282 × 105 kW. (8.33)
hr 3600 s kg
From Eq. (8.9), the heat added is

kJ kJ
qin = 3398.27 − 201.897
kg kg
| {z }
first boiling

kJ kJ
+ 3482.75 − 2820.05 , (8.34)
kg kg
| {z }
second boiling
kJ
= 3859.07 . (8.35)
kg
So the cycle’s thermal efficiency is
kJ
wnet 1400.05 kg
η= = kJ
= 0.362795. (8.36)
qin 3859.07 kg
The heat per unit mass rejected in the condenser is from Eq. (8.10):

kJ kJ kJ
qout = 2650.83 − 191.81 = 2459.02 . (8.37)
kg kg kg
So the power rejected as heat is

5 kg hr kJ
Q̇L = ṁqout = 2.63 × 10 2459.02 = 1.79645 × 105 kW. (8.38)
hr 3600 s kg
Example 8.2
Repeat the previous analysis without reheat. Find the thermal efficiency.

In this case state 4 would be taken down to 10 kPa. We would have

kJ kJ
s4s − sf 6.7239 kg K − 0.6492 kg K
x4s = = = 0.809863. (8.39)
sg − sf 8.1501 kJ
− 0.6492 kJ
kg K kg K
We can thus get h4s by consulting the tables to find

kJ kJ kJ
h4s = hf + x4s hf g = 191.81 + (0.809863) 2392.82 = 2129.67 . (8.40)
kg kg kg
Now
h4 = h3 − ηt (h3 − h4s ), (8.41)

kJ kJ kJ
= 3398.27 − (0.88) 3398.27 − 2129.67 , (8.42)
kg kg kg
kJ
= 2281.90 . (8.43)
kg
We then get

kJ kJ kJ kJ
wnet = h3 − h4 − wp = 3398.27 − 2281.90 − 10.0874 = 1106.28 . (8.44)
kg kg kg kg
We also get
kJ kJ kJ
qin = h3 − h2 = 3398.27 − 201.897 = 3196.37 . (8.45)
kg kg kg
So
kJ
wnet 1106.28 kg
η= = kJ
= 0.346106. (8.46)
qin 3196.37 kg
The thermal efficiency without reheat (0.346106) is less than that with reheat (0.362795). The reheat
altered the topology of the T − s diagram to make it more Carnot-like, and thus generated a more
efficient use of resources.
Example 8.3
A Rankine power cycle with water as the working fluid has ηt = ηp = 0.88. The turbine inlet
pressure and temperature are at P3 = 1200 psia, T3 = 1000 ◦ F. The condenser pressure is at P4 = 1 psia.
The steam generator provides Q̇H = 2×109 Btu/hr. In the condenser the cooling water enters at 60 ◦ F,
and we wish to keep the exit cooling water temperature at 80 ◦ F. Find the net power, the thermal
efficiency, and the mass flow rate of cooling water, ṁcw .
We interpolate the steam tables to find

Btu Btu
h3 = 1499.6 , s3 = 1.6297 . (8.47)
lbm lbm ◦ R

8.1. RANKINE 337
For an isentropic turbine, we have s4s = s3 = 1.6297 Btu/lbm/◦ R. At P4 = 1 psia, we find state 4s is
a two-phase mixture:
Btu
Btu

s4 − sf 1.6297 lbm ◦R − 0.1327 lbm ◦R
x4s = = Btu
Btu
= 0.811. (8.48)
sg − sf 1.9779 lbm ◦R − 0.1327 lbm ◦R
Thus,

Btu Btu Btu
h4s = hf + x4s hf g = 69.74 + (0.811) 1036 = 909.9 . (8.49)
lbm lbm lbm
Now, for the actual turbine, we get
h4 = h3 − ηt (h3 − h4s ), (8.50)

Btu Btu Btu
= 1499.7 − (0.88) 1499.7 − 909.9 , (8.51)
lbm lbm lbm
Btu
= 979.9 . (8.52)
lbm
Now, after the condenser, we take x1 = 0, so h1 = hf , s1 = sf , and v1 = vf , all at P1 = 1 psia. These
yield
Btu Btu ft3
h1 = 69.74 , s1 = 0.1327 , v1 = 0.01614 . (8.53)
lbm lbm ◦ R lbm
Now,
ws
wp = , (8.54)
ηp
v(P4 − P3 )
= , (8.55)
ηp

ft3 lbf lbf
0.01614 lbm 1200 in2 − 1 in2 144 in2 Btu
= , (8.56)
0.88 ft2 778 ft lbf
Btu
= 4.07 . (8.57)
lbm
Now,
Btu Btu Btu
h2 = h1 + wp = 69.74 + 4.07 = 73.81 . (8.58)
lbm lbm lbm
In the boiler, we have
Q̇H = ṁ(h3 − h2 ), (8.59)

Q̇H
ṁ = , (8.60)
h3 − h2
2 × 109 Btu
hr hr
= , (8.61)
1499.7 Btu
lbm − 73.81 Btu
lbm
3600 s
lbm
= 390 . (8.62)
s
We also note

9 Btu hr Btu
Q̇H = 2 × 10 = 5.556 × 105 . (8.63)
hr 3600 s s

Now, the net power is the sum of the turbine and pump work:
Ẇnet = ṁ ((h3 − h4 ) + (h1 − h2 )) , (8.64)

lbm Btu Btu Btu Btu
= 390 1499.6 − 979.9 + 69.74 − 73.81 ,(8.65)
s lbm lbm lbm lbm
Btu
= 2.01 × 105 . (8.66)
s
The thermal efficiency is thus
Ẇnet 2.01 × 105 Btu

s
η= = = 0.3618. (8.67)
Q̇H 5.556 × 105 Btu
s
The cooling water and the water in the Rankine cycle exchange heat in the condenser. This is
sketched in Fig. 8.5. The first law for the heat exchanger is
Rankine cycle water
cooling water
Figure 8.5: Rankine cycle condenser/heat exchanger.
dEcv
= Q̇cv − Ẇcv +ṁ(h4 − h1 ) + ṁcw cP (Tcold − Thot ), (8.68)
dt }
| {z |{z} |{z}
=0 =0
=0
0 = ṁ(h4 − h1 ) + ṁcw cP (Tcold − Thot ), (8.69)
ṁ(h4 − h1 )
ṁcw = , (8.70)
cP (Thot − Tcold )

390 lbm
s 979 Btulbm − 69.74 lbm
Btu
= Btu

◦ F) − (60 ◦ F))
, (8.71)
1.00 lbm ◦ R ((80
lbm
= 17730 . (8.72)
s
8.2 Brayton
Gas turbine power plants, both stationary and those for jet engines operate on the

8.2. BRAYTON 339
Figure 8.6: George Brayton (1830-1892), American mechanical engineer from Exeter, New
Hampshire; image from https://en.wikipedia.org/wiki/George Brayton.
• Brayton cycle:. idealized power cycle for gas turbines.
The cycle is named after George Brayton, an American mechanical engineer. Brayton is
depicted in Fig. 8.6. It has many similarities to the Rankine cycle. A schematic and T − s
and P − v diagrams for the Brayton cycle for a power plant are illustrated in Fig. 8.7.
fuel P T
2 2 isobar 3 3
combustion
chamber 3
isentrope
ope r
isent
w ar
compressor turbine ise
nt isob
ro 2 ar 4
isentrope
p e isob
isobar
products 1 4
air 1
1 4
environmental exhaust return v s
Figure 8.7: Schematic of a Brayton cycle along with P − v and T − s diagrams.
The Brayton cycle is outlined as follows:
• 1 → 2: isentropic compression (W added),
• 2 → 3: isobaric heat addition (Q added),
• 3 → 4: isentropic expansion (W extracted), and
• 4 → 1: isobaric heat rejection (passive exhaust).

Note, the work extracted is greater than the work added, i.e.
|h3 − h4 | > |h2 − h1 |. (8.73)
Often we will be dealing with a CPIG, in which case ∆h = cP ∆T . If so, then we can say
|T3 − T4 | > |T2 − T1 |. (8.74)
The reason for this is that
• isobars diverge in T − s space as s increases.
This is easy to understand when we recall the Gibbs equation, Eq. (6.65): T ds = dh − v dP.
On an isobar, we have dP = 0, so
T ds = dh, on isobar, (8.75)
= cP dT, if IG, (8.76)
∂T T
= . (8.77)
∂s P cP
Because at a given s, a high T isobar sits above a low T isobar, and the slope of the isobar
is proportional to T , it is easily seen how they must diverge. This is illustrated in Fig. 8.8.
r
ba pe
si o slo
T e r
ur he
e ss hig
pr ,
g h er T
hi igh
h
ar
re isob pe
su lo
pres er s
low r T, low
lowe
Figure 8.8: Sketch of diverging isobars for a CPIG in the T − s plane.
There are other classes of Brayton cycle plants. Schematics are shown next.
• Turbojet. In the turbojet, the kinetic energy of the fluid becomes important at two
points in the cycle. In the compression, the freestream fluid, entering the compressor
at the flight speed, has its pressure increased by the so-called “ram effect” where the
fluid decelerates. Second, the point of the turbojet is to produce thrust, that requires
a significant exit velocity. The turbine work is used solely to power the compressor.
See Fig. 8.9.

8.2. BRAYTON 341
C T
Figure 8.9: Turbojet schematic and the associated T − s plane.
• Turbojet with afterburners. We are limited in an ordinary turbojet by how much heat
can be added to the flow in combustion because such flow typically must pass through
the turbine blades, that suffer material degradation if the fluid is too hot. However,
we can add heat after the turbine in so-called afterburners. This releases chemical
energy, turns it into fluid potential energy in the form of high P/ρ, and then converts
to kinetic energy in the nozzle. This can enhance the thrust, though it can be shown
it is not particularly efficient. A sketch is given in Fig. 8.10.
C T
Figure 8.10: Turbojet with afterburners schematic and the associated T − s plane.
• Ramjet. A ramjet is much simpler. The compressor and turbine are removed. We
rely on the ram compression effect alone for compression and convert as much of the
thermal energy as possible into mechanical energy used to generate thrust force. A
sketch is given in Fig. 8.11.
Let us consider an

Figure 8.11: Ramjet schematic and the associated T − s plane.
• Air standard analysis: a common set of assumptions used for idealized cyclic
devices.
The air standard make many compromises in order to admit some simple analysis tools to
be used to make simple estimates for the performance of a variety of devices. Actual design
calculations would have to remedy the many shortcomings. But it is useful for a framework
of understanding. We take the air standard to entail
• The working fluid is air. This ignores any effect of the properties of the fuel or any
other fluid that is mixed with the air.
• The working fluid is an ideal gas. We will often assume it is a CPIG, but sometimes
not.
• We will ignore all details of the combustion process and treat it as a simple heat
addition.
Often in cycle analysis, the formal sign convention is ignored. We take the following
• Turbine work: wt = h3 − h4 . Here, the sign convention is maintained.
• Compressor work: wc = h2 − h1 . Here, the sign convention is ignored.
• Heat addition: qin = h3 − h2 . Here, the sign convention is maintained.
• Heat rejection: qout = h4 − h1 . Here, the sign convention is ignored.
The cycle efficiency for the Brayton power cycle is
(h3 − h4 ) − (h2 − h1 )
| {z } | {z }
wnet turbine compressor
η= = . (8.78)
qin h −h
| 3 {z }2
combustor

8.2. BRAYTON 343
Rearranging Eq. (8.78), we can also say

h −h
| 4 {z }1
heat rejected
η =1− . (8.79)
h −h
| 3 {z }2
heat added
For jet propulsion Brayton cycles, there are additional nuances in defining the efficiency that
we will not consider.
The back work ratio, bwr, is the ratio of compressor to turbine work:
wc h2 − h1
bwr = = . (8.80)
wt h3 − h4
Note the back work ratio will be seen to be much larger for gas phase power cycles than it
was for vapor cycles. For Brayton cycles, we may see bwr ∼ 0.4. For Rankine cycles, we
usually see bwr ∼ 0.01. R
Now, if we have a CPIG, we get ∆h = cP dT to reduce to ∆h = cP ∆T . So Eq. (8.78)
reduces to
cP (T3 − T4 ) − cP (T2 − T1 )
η = , (8.81)
cP (T3 − T2 )
T3 − T4 − T2 + T1
= , (8.82)
T3 − T2
T4 − T1
= 1− , (8.83)
T3 − T2
T
!
T1 T41 − 1
= 1− . (8.84)
T2 TT32 − 1
Now, 1 → 2 is isentropic. Recall from Eq. (6.169) for a CPIG that is isentropic that
T2 /T1 = (P2 /P1 )(k−1)/k . We also have 3 → 4 to be isentropic, so T3 /T4 = (P3 /P4 )(k−1)/k .
But P2 = P3 and P1 = P4 . So
T2 T3
= , (8.85)
T1 T4
T4 T3
= . (8.86)
T1 T2
So
T1
η =1− . (8.87)
T2
With the temperature ratio θ across the compressor defined as
T2
θ= , (8.88)
T1

we have
1
η =1− . (8.89)
θ
In terms of the pressure ratio, the efficiency is
1
η = 1 − k−1 . (8.90)
k P2
P1
A plot of η versus the temperature ratio θ is plotted in Fig. 8.12. As the temperature
1.0
0.8
0.6
0.4
0.2
0.0
0 2 4 6 8 10
Figure 8.12: Thermal efficiency versus temperature ratio across the compressor for an air
standard Brayton cycle.
ratio rises, the thermal efficiency increases for the Brayton cycle. For practical application,
materials have temperature limits. Note the efficiency looks like that for a Carnot cycle,
but it is not. The highest temperature in the Brayton cycle is T3 , so the equivalent Carnot
efficiency would be 1 − T1 /T3 .
Example 8.4
Consider a CPIG air standard Brayton cycle with fixed inlet conditions P1 and T1 . We also fix the
maximum temperature as the metallurgical limit of the turbine blades, Tmax . Find the temperature
ratio θ that maximizes the net work. Then find the temperature ratio that maximizes the thermal
efficiency.
We have
k−1 k−1
P2 k
P4 k
T2 = T1 θ, T2 = T1 , T3 = Tmax , T4 = T3 . (8.91)
P1 P3

8.2. BRAYTON 345
We also have P4 = P1 and P2 = P3 . So

k−1 1−k
P1 k
P2 k
Tmax
T4 = Tmax = Tmax = . (8.92)
P2 P1 θ
Now, the net work is
wnet = (h3 − h4 ) − (h2 − h1 ), (8.93)

= cP (T3 − T4 − T2 + T1 ), (8.94)
= cP (Tmax − Tmax θ−1 − T1 θ + T1 ), (8.95)

Tmax Tmax −1
= cP T 1 − θ −θ+1 . (8.96)
T1 T1
Note wnet = 0 when θ = 1 and when θ = Tmax /T1 .

To find the maximum wnet we take dwnet /dθ and set to zero:

dwnet Tmax −2
= cP T 1 θ −1 , (8.97)
dθ T1

Tmax −2
0 = cP T 1 θ −1 , (8.98)
T1
r
Tmax
θ = ± . (8.99)
T1
We take the positive root, because a negative temperature ratio does not make sense:
r
Tmax
θ= . (8.100)
T1
The second derivative tells us whether our critical point is a maximum or a minimum.
d2 wnet
= −2cP Tmax θ−3 . (8.101)
dθ2
When θ > 0, d2 wnet /dθ2 < 0, so we have found a maximum of wnet . The maximum value is
−1/2 1/2 !
Tmax Tmax
wnet |max = cP Tmax − Tmax − T1 + T1 , (8.102)
T1 T1
−1/2 1/2 !
Tmax Tmax Tmax Tmax
= cP T 1 − − +1 , (8.103)
T1 T1 T1 T1
1/2 !
Tmax Tmax
= cP T 1 −2 +1 , (8.104)
T1 T1
1/2 !2
Tmax
= cP T 1 −1 . (8.105)
T1
Now, at what value of θ is the thermal efficiency maximum? Consider
η = 1 − θ−1 , (8.106)
dη 1
= . (8.107)
dθ θ2

At a maximum, we must have dη/dθ = 0. So we must have θ → ∞ in order to have η reach a maximum.
But we are limited to θ ≤ Tmax /T1 . So the achievable θ that maximizes η is
Tmax
θ= . (8.108)
T1
So the efficiency at our highest allowable θ is

1 T1
η =1− Tmax
=1− . (8.109)
T1
Tmax
But at the value of peak efficiency, the net work is approaching zero! So while this is highly efficient,
it is not highly useful!
Lastly, what is the efficiency at the point where we maximize work?
r
T1
η =1− . (8.110)
Tmax
A plot of scaled net work, wnet /cP /T1 versus
√ temperature ratio is given for Tmax /T1 = 10 in p
Fig. 8.13.
Here, the θ that maximizes wnet is θ = 10 = 3.162. At that value of θ, we find η = 1 − 1/10 =
5
4.675
4
2 4 6 8 10
3.162
Figure 8.13: Scaled net work versus temperature ratio for a Brayton cycle with Tmax /T1 = 10.
√
0.683772 and wnet /cP /T1 = ( 10 − 1)2 = 4.675.
Example 8.5
Consider the Brayton power cycle for a spacecraft sketched in Fig. 8.14. The working fluid is argon,
that is well modeled as a CPIG over a wide range of T and P . We take the pressure in the heating
process to be isobaric, P2 = P3 = 140 kPa, and the pressure in the cooling process to be isobaric,
P4 = P1 = 35 kPa. We are given that T1 = 280 K, T3 = 1100 K. The compressor and turbine both
have component efficiencies of ηt = ηc = 0.8. We are to find the net work, the thermal efficiency, and
a plot of the process on a T − s diagram.

8.2. BRAYTON 347
heat source
heater
2 3
compressor turbine
1 4
cooler
low temperature
reservoir
Figure 8.14: Schematic of a Brayton power cycle for spacecraft.
For argon, we have

kJ kJ 5
R = 0.20813 , cP = 0.5203 , k= ∼ 1.667. (8.111)
kg K kg K 3
Note that cP = kR/(k − 1).
Let us start at state 1. We first assume an isentropic compressor. We will quickly relax this to
account for the compressor efficiency. But for an isentropic compressor, we have for the CPIG
k−1
P2 k T2s
= . (8.112)
P1 T1
Here, T2s is the temperature that would be realized if the process were isentropic. We find
k−1 5/3−1
P2 k 140 kPa 5/3
T2s = T1 = (280 K) = 487.5 K. (8.113)
P1 35 kPa
Now, ηc = ws /wcompressor , so

kJ
ws h2s − h1 cP (T2s − T1 ) 0.5203 kg K (487.5 K − 280 K) kJ
wcompressor = = = = = 135.0 . (8.114)
ηc ηc ηc 0.8 kg
Now wcompressor = h2 − h1 = cP (T2 − T1 ), so
wcompressor 135.0 kJ
kg
T2 = T1 + = (280 K) + = 539.5 K. (8.115)
cP 0.5203 kgkJK

Notice that T2 > T2s . The inefficiency (like friction) is manifested in more work being required to
achieve the final pressure than that which would have been required had the process been ideal.
In the heater, we have

kJ kJ
qH = h3 − h2 = cP (T3 − T2 ) = 0.5203 ((1100 K) − (539.5 K)) = 291.6 . (8.116)
kg K kg
Now, consider an ideal turbine:
k−1
T4s P4 k
= , (8.117)
T3 P3
k−1
P4 k
T4s = T3 , (8.118)
P3
5/3−1
35 kPa 5/3
= (1100 K) , (8.119)
140 kPa
= 631.7 K. (8.120)
But for the real turbine,
wturbine
ηt = , (8.121)
ws
wturbine = ηt ws , (8.122)
= ηt (h3 − h4s ), (8.123)
= ηt cP (T3 − T4s ), (8.124)

kJ kJ
= (0.8) 0.5203 ((1100 K) − (631.7 K)) = 194.9 . (8.125)
kg K kg
Thus, because wturbine = h3 − h4 = cP (T3 − T4 ), we get
wturbine 194.9 kJ
kg
T4 = T3 − = (1100 K) − = 725.4 K. (8.126)
cP 0.5203 kgkJK
Note that T4 is higher than would be for an isentropic process. This indicates that we did not get all
the possible work out of the turbine. Note also that some of the turbine work was used to drive the
compressor, and the rest wnet is available for other uses. We find

kJ kJ kJ
wnet = wturbine − wcompressor = 194.9 − 135.0 = 59.9 . (8.127)
kg kg kg
Now, for the cooler,

kJ kJ
qL = h4 − h1 = cP (T4 − T1 ) = 0.5203 ((725.4 K) − (280 K)) = 231.7 . (8.128)
kg K kg

8.2. BRAYTON 349
We are now in a position to calculate the thermal efficiency for the cycle.
wnet
η = , (8.129)
qH
wturbine − wcompressor
= , (8.130)
qH
cP ((T3 − T4 ) − (T2 − T1 ))
= , (8.131)
cP (T3 − T2 )
(T3 − T4 ) − (T2 − T1 )
= , (8.132)
T3 − T2
((1100 K) − (725.4 K)) − ((539.5 K) − (280 K))
= , (8.133)
(1100 K) − (539.6 K)
= 0.205. (8.134)
If we had been able to employ a Carnot cycle operating between the same temperature bounds, we
would have found the Carnot efficiency to be
T1 280 K
ηCarnot = 1 − =1− = 0.745 > 0.205. (8.135)
T3 1100 K
A plot of the T − s diagram for this Brayton cycle is shown in Fig. 8.15. Note that from 1 to 2 (as
3
Pa
14 0k
2 P=
2s
4
Pa
= 35 k 4s
P
Figure 8.15: T − s diagram of a Brayton power cycle for a spacecraft with turbine and
compressor inefficiencies.
well as 3 to 4) there is area under the curve in the T − s diagram. But the process is adiabatic! Recall
that isentropic processes are both adiabatic and reversible. The 1-2 process is an example of a process
that is adiabatic but irreversible. So the entropy change is not due to heat addition effects but instead
is due to other effects.

Example 8.6
We are given a turbojet flying with a flight speed of 300 m/s. The pressure ratio of the compressor is
7. The ambient air is at Ta = 300 K, Pa = 100 kPa. The turbine inlet temperature is 1500 K. The mass
flow rate is ṁ = 10 kg/s. All of the turbine work is used to drive the compressor. Find the exit velocity
and the thrust force generated. Assume an air standard with a CPIG; k = 1.4, cP = 1.0045 kJ/kg/K.
A plot of the T − s diagram for this Brayton cycle is shown in Fig. 8.16. We first calculate the ram
T 3
1 5
Figure 8.16: T − s diagram of a Brayton cycle in a turbojet engine.
compression effect:
1 2 1
h1 + v1 = ha + va2 . (8.136)
2
|{z} 2
∼0
We typically neglect the kinetic energy of the flow once it has been brought to near rest within the
engine. So we get
1 2
h1 − ha = v , (8.137)
2 a
1 2
cP (T1 − Ta ) = v , (8.138)
2 a
va2
T1 = Ta + , (8.139)
2cP
2
300 ms kJ
= (300 K) + m2
, (8.140)
kJ
2 1.0045 kg K 1000 s2
= 344.8 K. (8.141)

8.2. BRAYTON 351
Now, consider the isentropic compression in the compressor. For this, we have
  k−1
k
T2  P2 
=   , (8.142)
T1  P1 
|{z}
=7
1.4−1
T2 = (344.8 K)(7) 1.4 , (8.143)
= 601.27 K. (8.144)
Let us calculate P2 /Pa , that we will need later. From the isentropic relations,
k
k−1
P2 T2
= , (8.145)
Pa Ta
1.4
1.4−1
601.27 K
= , (8.146)
300 K
= 11.3977. (8.147)
We were given the turbine inlet temperature, T3 = 1500 K. Now, the compressor work must equal
the turbine work. This amounts to, ignoring the sign convention,
wc = wt , (8.148)
h2 − h1 = h3 − h4 , (8.149)
cP (T2 − T1 ) = cP (T3 − T4 ), (8.150)
T2 − T1 = T3 − T4 , (8.151)
T4 = T3 − T2 + T1 , (8.152)
= (1500 K) − (601.27 K) + (344.8 K), (8.153)
= 1243.5 K. (8.154)
Now, we use the isentropic relations to get T5 . Process 3 to 5 is isentropic with P5 /P3 = Pa /P2 =
1/11.3977 , so we have
k
k−1
P5 T5
= , (8.155)
P3 T3
k−1
P5 k
T5 = T3 , (8.156)
P3
1.4−1
1 1.4
= (1500 K) , (8.157)
11.3977
= 748.4 K. (8.158)

Now, we need to calculate the exhaust velocity. Take an energy balance through the nozzle to get
1 2 1
h4 + v4 = h5 + v52 , (8.159)
2 |{z} 2
∼0
1
h4 = h5 + v52 , (8.160)
p 2
v5 = 2(h4 − h5 ), (8.161)
p
= 2cP (T4 − T5 ), (8.162)
v
u 2
t2 1.0045 kJ
u 1000 ms2
= ((1243.5 K) − (748.4 K)) kJ
, (8.163)
kg K kg
m
= 997.3 . (8.164)
s
Now, Newton’s second law for a control volume can be derived by applying the methods of Sec. 4.1.
Leaving out details, it can be shown for one dimensional flow with one inlet and exit to be
d
(ρv) = Fcv + ṁvi − ṁve . (8.165)
dt
It says the time rate of change of momentum in the control volume is the net force acting on the control
volume plus the momentum brought in minus the momentum that leaves. Often terms of the form ṁv
are known as momentum flux.. We will take the problem to be steady and take the force to be the
thrust force. So
Fcv = ṁ(ve − vi ), (8.166)
= ṁ(v5 − va ), (8.167)

kg m m
= 10 997.3 − 300 , (8.168)
s s s
= 6973 N. (8.169)
8.3 Refrigeration
A simple way to think of a refrigerator is a cyclic heat engine operating in reverse. Rather
than extracting work from heat transfer from a high temperature reservoir and rejecting heat
to a low temperature reservoir, the refrigerator takes a work input to move heat from a low
temperature reservoir to a high temperature reservoir.
A common refrigerator is based on a
• vapor-compression cycle: cycle used for refrigeration and air conditioning.
This is a Rankine cycle in reverse. While one could employ a turbine to extract some work,
it is often impractical. Instead the high pressure gas is simply irreversibly throttled down to
low pressure.
One can outline the vapor-compression refrigeration cycle as follows:

8.3. REFRIGERATION 353
• 1 → 2: isentropic compression
• 2 → 3: isobaric heat transfer to high temperature reservoir in condenser,
• 3 → 4: adiabatic expansion in throttling valve, and
• 4 → 1: isobaric (and often isothermal) heat transfer from a low temperature reservoir
to an evaporator.
A schematic and associated T − s diagram for the vapor-compression refrigeration cycle is

shown in Fig. 8.17. One goal in design of refrigerators is low work input. There are two
condenser 2
3
Compressor
expansion compressor
valve
4 evaporator
T
1 2
3 condenser
compressor temperature of
surroundings
valv
e
evaporator temperature inside

refrigerator
4 1
refrigerator
Figure 8.17: Schematic and T − s diagrams for the vapor-compression refrigeration cycle.
main strategies in this:
• Design the best refrigerator to minimize Q̇in . This really means reducing the conductive
heat flux through the refrigerator walls. One can use a highly insulating material. One
can also use thick walls. Thick walls will reduce available space for storage however.
This is an example of a design trade-off.
• For a given Q̇in , design the optimal thermodynamic cycle to minimize the work nec-
essary to achieve the goal. In practice, this means making the topology of the cycle
as much as possible resemble that of a Carnot refrigerator. Our vapor compression
refrigeration cycle is actually close to a Carnot cycle.

The efficiency does not make sense for a refrigerator as 0 ≤ η ≤ 1. Instead, much as our
earlier analysis for Carnot refrigerators, a coefficient of performance, β, is defined as
what one wants
β = , (8.170)
what one pays for
qL
= . (8.171)
wc
Note that a heat pump is effectively the same as a refrigerator, except one desires qH
rather than qL . So for a heat pump, the coefficient of performance, β ′ , is defined as
qH
β′ = . (8.172)
wc
Example 8.7
R-134a, a common refrigerant, enters a compressor at x1 = 1, T1 = −15 ◦ C. At the compressor
inlet, the volume flow rate is 1 m3 /min. The R-134a leaves the condenser at T3 = 35 ◦ C, P3 = 1000 kPa.
Analyze the system.
We have the state at 1, knowing x1 and T1 . The tables then give

kJ kJ m3
h1 = 389.20 , s1 = 1.7354 , v1 = 0.12007 . (8.173)
kg kg K kg
The process 2 to 3 is along an isobar. We know P3 = 1000 kPa, so P2 = 1000 kPa. We assume an
isentropic compression to state 2, where P2 = 1000 kPa. We have s2 = s1 = 1.7354 kJ/kg/K. We
interpolate the superheat tables to get
kJ
h2 = 426.771 . (8.174)
kg
State 3 is a subcooled liquid, and we have no tables for it. Let us approximate h3 as hf at T3 = 35 ◦ C,
that is
kJ
h3 ∼ 249.10 . (8.175)
kg
In the expansion valve, we have
kJ
h4 = h3 = 249.10 . (8.176)
kg
Now, !
m3
Av 1 min min kg
ṁ = ρAv = = m3
= 0.138808 . (8.177)
v1 0.12007 kg
60 s s
Note the term Av has units m3 /s and is a volume flow rate. Now, the compressor power is

kg kJ kJ
Ẇ = ṁ(h2 − h1 ) = 0.138808 426.771 − 389.20 = 5.2152 kW. (8.178)
s kg kg
The refrigerator capacity is

kg kJ kJ
Q̇in = ṁ(h1 − h4 ) = 0.138808 389.20 − 249.10 = 19.447 kW. (8.179)
s kg kg

8.3. REFRIGERATION 355
With a 5.2152 kW input, we will move 19.447 kW out of the refrigerator.

At what rate does heat exit the back side?

kg kJ kJ
Q̇H = ṁ(h2 − h3 ) = 0.138808 426.771 − 249.10 = 24.6622 kW. (8.180)
s kg kg
Note that
Q̇H = Q̇in + Ẇ , (8.181)

24.6622 kW = (19.447 kW) + (5.2152 kW). (8.182)
The coefficient of performance is
Q̇in 19.447 kW
β= = = 3.72891. (8.183)
Ẇ 5.2152 kW
We could also say
Q̇in
β = , (8.184)
Q̇H − Q̇in
1
= . (8.185)
Q̇H
Q̇
− 1
in
Because we do not have a Carnot refrigerator for this problem, we realize that Q̇H /Q̇in 6= T3 /T1 .
The University of Notre Dame Power Plant also serves as a generator of chilled water for
air conditioning campus buildings. This is effectively a refrigerator on a grand scale, though
we omit details of the actual system here. A photograph of one of the campus chillers is
shown in Fig. 8.18.

Figure 8.18: Chiller in the University of Notre Dame power plant, 14 June 2010.

Chapter 9
Mathematical foundations
Read BS, Chapters 12, 15
This chapter will serve as an introduction to some of the mathematical underpinnings of

thermodynamics. Though the practicality is not immediately obvious to all, this analysis
is a necessary precursor for building many useful and standard theories. Important among
those are theories to describe chemical reactions, that have widespread application in a
variety of engineering scenarios, including combustion, materials processing, and pollution
control.
9.1 Maxwell relations

We begin with a discussion of the so-called Maxwell1 relations, named after the great nine-
teenth century physicist, James Clerk Maxwell, shown in Fig. 9.1.
Figure 9.1: James Clerk Maxwell (1831-1879) Scottish physicist; image from
https://mathshistory.st-andrews.ac.uk/Biographies/Maxwell.html.
Recall that if w = w(x, y), we have Eq. (3.3):

1
J. C. Maxwell, 1871, Theory of Heat, reprinted 2001, Dover, Mineola, New York, p. 169.
357
358 CHAPTER 9. MATHEMATICAL FOUNDATIONS
∂w ∂w
dw = dx + dy. (9.1)
∂x y ∂y x
And also recall if dw = M(x, y) dx + N(x, y) dy, the requirement for an exact differential is
∂w ∂w
= M, = N, (9.2)
∂x y ∂y x
∂2w ∂M ∂2w ∂N
= , = . (9.3)
∂y∂x ∂y x ∂x∂y ∂x y
These equations are the same as Eqs. (3.7, 3.8). Because order of differentiation does not
matter for functions that are continuous and differentiable, we must have for exact differen-
tials, Eq. (3.9):
∂N ∂M
= . (9.4)
∂x y ∂y x
Compare the Gibbs equation, Eq. (6.59), to our equation for dw:
du = −P dv + T ds, (9.5)
dw = M dx + N dy. (9.6)
We see the equivalences
w → u, x → v, y → s, M → −P, N → T, (9.7)
and just as one expects w = w(x, y), one then expects the natural, or canonical form of
u = u(v, s). (9.8)
Application of Eq. (9.4) to the Gibbs equation, Eq. (6.59), gives then
∂T ∂P
=− . (9.9)
∂v s ∂s v
Equation (9.9) is known as a
• Maxwell relation: a relation between thermodynamic variables induced by sym-

metry of second partial derivatives.
Moreover, specialization of Eq. (9.2) to the Gibbs equation gives
∂u ∂u
= −P, = T. (9.10)
∂v s ∂s v

9.2. FUNCTIONS OF TWO INDEPENDENT VARIABLES 359
9.2 Functions of two independent variables

Consider a general implicit function linking three variables, x, y, z:
f (x, y, z) = 0. (9.11)
In x − y − z space, this will represent a surface. If the function can be inverted, it will be
possible to write the explicit forms
x = x(y, z), y = y(x, z), z = z(x, y). (9.12)
Differentiating the first two of Eqs. (9.12) gives
∂x ∂x
dx = dy + dz, (9.13)
∂y z ∂z y
∂y ∂y
dy = dx + dz. (9.14)
∂x z ∂z x
Now, use Eq. (9.14) to eliminate dy in Eq. (9.13):

∂x ∂y ∂y ∂x
dx = dx + dz + dz, (9.15)
∂y z ∂x z ∂z x ∂z y
| {z }
=dy
!
∂x ∂y ∂x ∂y ∂x
1− dx = + dz, (9.16)
∂y z ∂x z ∂y z ∂z x ∂z y
!
∂x ∂y ∂x ∂y ∂x
0 dx + 0 dz = − 1 dx + + dz. (9.17)
∂y z ∂x z ∂y z ∂z x ∂z y
| {z } | {z }
=0 =0
Because x and z are independent, so are dx and dz, and the coefficients on each in Eq. (9.17)
must be zero. Therefore, from the coefficient on dx in Eq. (9.17), we have
∂x ∂y
− 1 = 0, (9.18)
∂y z ∂x z
∂x ∂y
= 1, (9.19)
∂y z ∂x z
∂x 1
= ∂y , (9.20)
∂y z ∂x z

and also from the coefficient on dz in Eq. (9.17), we have
∂x ∂y ∂x
+ = 0, (9.21)
∂y z ∂z x ∂z y
∂x ∂x ∂y
= − , (9.22)
∂z y ∂y z ∂z x
∂x ∂y ∂z
= −1. (9.23)
∂z y ∂x z ∂y x
If one now divides Eq. (9.13) by a fourth differential, dw, one gets
dx ∂x dy ∂x dz
= + . (9.24)
dw ∂y z dw ∂z y dw
Demanding that z be held constant in Eq. (9.24) gives
∂x ∂x ∂y
= , (9.25)
∂w z ∂y z ∂w z
∂x
∂w z ∂x
∂y
= , (9.26)
∂w z
∂y z
∂x ∂w ∂x
= . (9.27)
∂w z ∂y z ∂y z
If x = x(y, w), one then gets
∂x ∂x
dx = dy + dw. (9.28)
∂y w ∂w y
Divide now by dy while holding z constant so
∂x ∂x ∂x ∂w
= + . (9.29)
∂y z ∂y w ∂w y ∂y z
These general operations can be applied to a wide variety of thermodynamic operations.
9.3 Legendre transformations

Here we introduce the
• Legendre transformation, a transformation that for thermodynamics allows one

to exchange one set of independent thermodynamic variables to another.

9.3. LEGENDRE TRANSFORMATIONS 361
The transformation and its general background are described in more detail in the Appendix.
For further details, the interested student can consult Zia, et al.2 or Abbott and van Ness.3
The transformation is named after Adrien-Marie Legendre, whose work was not motivated
by thermodynamic concerns, but has found application in thermodynamics. The only known
image of Legendre is shown in Fig. 9.2.
Figure 9.2: Adrien-Marie Legendre (1752-1833) French mathematician; image from

https://mathshistory.st-andrews.ac.uk/Biographies/Legendre.html.
The Gibbs equation, Eq. (6.59): du = −P dv + T ds, is the fundamental equation

of classical thermodynamics. It is a canonical form that suggests the most natural set of
variables in which to express internal energy u are v and s:
u = u(v, s). (9.30)
However, v and s may not be convenient for a particular problem. There may be other
combinations of variables whose canonical form gives a more convenient set of independent
variables for a particular problem. An example is the enthalpy, Eq. (3.218):
h = u + P v. (9.31)
Differentiating the enthalpy gives
dh = du + P dv + v dP. (9.32)
We repeat the analysis used to obtain Eq. (6.66) earlier. Use Eq. (9.32) to eliminate du in
the Gibbs equation, Eq. (6.59), to give
dh {z − v dP} = −P dv + T ds,
| − P dv (9.33)
=du
dh = T ds + v dP. (9.34)
2
R. K. P. Zia, E. F. Redish, and S. R. McKay, 2009, “Making sense of the Legendre transform,” Ameri-
can Journal of Physics, 77(7): 614-622.
3
M. M. Abbott and H. C. van Ness, 1972, Thermodynamics, Schaum’s Outline Series in Engineering,
McGraw-Hill, New York.

So the canonical variables for h are s and P . One then expects
h = h(s, P ). (9.35)
This exercise can be systematized with the Legendre transformation. The basic outline
of the Legendre transformation is as follows. The form du = −P dv + T ds, suggests u is
the fundamental dependent variable, v and s are the canonical independent variables, with
−P and T serving as so-called conjugate variables. We seek transformations that can render
conjugate variables to be canonical variables. We can achieve this by defining new dependent
variables as the difference between the original dependent variable and simple second order
combinations of the canonical and conjugate variables. For the Gibbs equation, there are
only three combinations, −P v, T s, and −P v + T s, that are dimensionally consistent with
u. We subtract each of these from u to define new dependent variables as follows: They are
h = h(P, s) = u − (−P v) = u + P v, enthalpy, (9.36)

a = a(v, T ) = u − (T s) = u − T s, Helmholtz free energy, (9.37)
g = g(P, T ) = u − (−P v + T s) = u + P v − T s, Gibbs free energy. (9.38)
The Helmholtz free energy was developed by Helmholtz.4 It is symbolized by a in recognition

of the German word arbeit, or “work.” An image of the original appearance of the notion
from Helmholtz’s 1882 work is shown in Fig. 9.3. The notation F is our Helmholtz free energy
a; U is our u; J is our mechanical equivalent of heat J; ϑ is our temperature T ; and S is
our entropy s. The Gibbs free energy was introduced by Gibbs.5 An image of a roundabout
appearance of the Gibbs free energy from Gibbs’ 1873 work is shown in Fig. 9.4. Here, ǫ is
our u, E is our U, η is our s, and H is our S.
It has already been shown for the enthalpy that dh = T ds + v dP , so that the canonical
variables are s and P . One then also has
∂h ∂h
dh = ds + dP, (9.39)
∂s P ∂P s
from which one deduces that

∂h ∂h
T = , v= . (9.40)
∂s P ∂P s
From Eq. (9.40), a second Maxwell relation can be deduced by differentiation of the first
with respect to P and the second with respect to s:
∂T ∂v
= . (9.41)
∂P s ∂s P
4
H. Helmholtz, 1882, “Die Thermodynamik chemischer Vorgänge,” Sitzungsberichte der Königlich
Preuβischen Akademie der Wissenschaften zu Berlin, 1: 22-39.
5
J. W. Gibbs, 1873, “A method of geometrical representation of the thermodynamic properties of sub-
stances by means of surfaces,” Transactions of the Connecticut Academy of Arts and Sciences, 2: 382-404.

9.3. LEGENDRE TRANSFORMATIONS 363
Figure 9.3: Image of the original 1882 appearance of the Helmholtz free energy.
The relations for Helmholtz and Gibbs free energies each supply additional useful relations
including two new Maxwell relations. First consider the Helmholtz free energy
a = u − T s, (9.42)
da = du − T ds − s dT, (9.43)
= (−P dv + T ds) −T ds − s dT, (9.44)
| {z }
du
= −P dv − s dT. (9.45)
So the canonical variables for a are v and T . The conjugate variables are −P and −s. Thus
∂a ∂a
da = dv + dT. (9.46)
∂v T ∂T v
So one gets
∂a ∂a
−P = , −s = . (9.47)
∂v T ∂T v
and the consequent Maxwell relation
∂P ∂s
= . (9.48)
∂T v ∂v T

Figure 9.4: Image of the original 1873 appearance of a combination of terms that is now
known as the Gibbs free energy.
For the Gibbs free energy

g = |u +{zP v} −T s, (9.49)
=h
= h − T s, (9.50)
dg = dh − T ds − s dT, (9.51)
= (T ds + v dP ) −T ds − s dT, (9.52)
| {z }
=dh
= v dP − s dT. (9.53)
Many find some of these equations to have sufficient appeal to cast them in concrete. The
extensive version of Eq. (9.51), unfortunately restricted to the isothermal limit, is depicted
in the floor of University of Notre Dame’s Jordan Hall of Science atrium, see Fig. 9.5.
So for Gibbs free energy, the canonical variables are P and T , while the conjugate vari-
ables are v and −s. One then has g = g(P, T ), that gives
∂g ∂g
dg = dP + dT. (9.54)
∂P T ∂T P
So one finds
∂g ∂g
v= , −s = . (9.55)
∂P T ∂T P

Figure 9.5: Figure cast in the atrium floor of the University of Notre Dame’s Jordan Hall of
Science containing an isothermal extensive version of Eq. (9.51), among other things.
u=u h = u + Pv a = u − Ts g = u + Pv − Ts
du = −P dv + T ds dh = T ds + v dP da = −P dv − s dT dg = v dP − s dT
u = u(v, s) h = h(s, P ) a = a(v, T ) g = g(P, T )
− ∂P
∂s v
= ∂T
∂v s
∂T
∂P s
= ∂v
∂s P
∂P
∂T v
= ∂s
∂v T
∂v
∂T P
∂s
= − ∂P T
Table 9.1: Summary of Maxwell relations and their generators.
The resulting Maxwell relation is then

∂v ∂s
=− . (9.56)
∂T P ∂P T
Table 9.1 gives a summary of the Maxwell relations and their generators. An image
showing the first published appearance of the Maxwell relations is given in Fig. 9.6. In
Fig. 9.6, the “thermodynamic function” φ is our s, and θ is our T . Note that typography
for partial derivatives was non-existent in most texts of the nineteenth century.
9.4 Specific heat capacity

Recall from Eqs. (3.229, 3.232) that specific heat capacities are defined as
∂u
cv = , (9.57)
∂T v
∂h
cP = . (9.58)
∂T P

Figure 9.6: Maxwell’s relations as first written by Maxwell, 1871.
Then perform operations on the Gibbs equation, Eq. (6.59):
du = T ds − P dv, (9.59)
∂u ∂s
= T , (9.60)
∂T v ∂T v
∂s
cv = T . (9.61)
∂T v
Likewise, operating on Eq. (6.66), we get,
dh = T ds + v dP, (9.62)
∂h ∂s
= T , (9.63)
∂T P ∂T P
∂s
cP = T . (9.64)
∂T P
One finds further useful relations by operating on the Gibbs equation, Eq. (6.59):
du = T ds − P dv, (9.65)
∂u ∂s
= T − P, (9.66)
∂v T ∂v T
∂P
= T − P. (9.67)
∂T v

So one can then say
u = u(T, v), (9.68)

∂u ∂u
du = dT + dv, (9.69)
∂T v ∂v T

∂P
= cv dT + T − P dv. (9.70)
∂T v
| {z }
∂u
|
∂v T
For an ideal gas, one has
∂u ∂P
= T − P, (9.71)
∂v T ∂T v

R RT
= T − , (9.72)
v v
= 0. (9.73)
Consequently, we have proved what was asserted in Sec. 3.8.1: u is not a function of v for an
ideal gas, so u = u(T ) alone. Because h = u + P v, h for an ideal gas reduces to h = u + RT .
Thus,
h = u(T ) + RT = h(T ). (9.74)
Now, return to general equations of state. With s = s(T, v) or s = s(T, P ), one gets
∂s ∂s
ds = dT + dv, (9.75)
∂T v ∂v T
∂s ∂s
ds = dT + dP. (9.76)
∂T P ∂P T
Now, using Eqs. (9.41, 9.56, 9.61, 9.64) one gets
cv ∂P
ds = dT + dv, (9.77)
T ∂T v
cP ∂v
ds = dT − dP. (9.78)
T ∂T P
Subtracting Eq. (9.78) from Eq. (9.77), one finds
cv − cP ∂P ∂v
0 = dT + dv + dP, (9.79)
T ∂T v ∂T P
∂P ∂v
(cP − cv ) dT = T dv + T dP. (9.80)
∂T v ∂T P

Now, divide both sides by dT and hold either P or v constant. In either case, one gets
∂P ∂v
cP − cv = T . (9.81)
∂T v ∂T P
Also, because ∂P/∂T |v = −(∂P/∂v|T )(∂v/∂T |P ), Eq. (9.81) can be rewritten as
2
∂v ∂P
cP − cv = −T . (9.82)
∂T P ∂v T
Now, because T > 0, (∂v/∂T |P )2 > 0, and for all known materials ∂P/∂v|T < 0, we must
have
cP > cv . (9.83)
Example 9.1
For a CIIG, prove Mayer’s relation, Eq. (3.248), cP (T ) − cv (T ) = R.
For the ideal gas, P v = RT , one has

∂P R ∂v R
= , = . (9.84)
∂T v v ∂T P P
So, substituting these into Eq. (9.81), we get
RR
cP − cv = T , (9.85)
vP
R2
= T , (9.86)
RT
= R. (9.87)
This holds even if the ideal gas is calorically imperfect. That is,
cP (T ) − cv (T ) = R. Q.E.D. (9.88)
For the ratio of specific heats for a general material, one can use Eqs. (9.61) and (9.64)
to get
∂s
cP T ∂T P
k= = ∂s
, then apply Eq. (9.20) to get (9.89)
cv T ∂T v
∂s ∂T
= , then apply Eq. (9.22) to get (9.90)
∂T P ∂s v

∂s ∂P ∂T ∂v
= − − , (9.91)
∂P T ∂T s ∂v s ∂s T

∂v ∂s ∂P ∂T
= , (9.92)
∂s T ∂P T ∂T s ∂v s
∂v ∂P
= . (9.93)
∂P T ∂v s

9.5. THE FIRST LAW AND COORDINATE TRANSFORMATIONS 369
The first term can be obtained from P − v − T data. The second term is related to the
isentropic sound speed of the material, that is also a measurable quantity.
Let us see how one can use a general thermal equation of state, P = P (v, T ) to deduce
an expression for entropy.. Because we can expect s = s(T, P ), we can say
∂s ∂s
ds = dT + dP. (9.94)
∂T P ∂P T
Using Eq. (9.64) and the Maxwell relation of Eq. (9.56), we can rewrite as
cP ∂v
ds = dT − dP. (9.95)
T ∂T P
One can integrate to find s(T, P ).

Similarly for s = s(T, v), we can say
∂s ∂s
ds = dT + dv. (9.96)
∂T v ∂v T
Using Eq. (9.61) and the Maxwell relation of Eq. (9.48), we can rewrite as
cv ∂P
ds = dT + dv. (9.97)
T ∂T v
One can integrate to find s(T, v).
9.5 The first law and coordinate transformations

One can apply standard notions from the mathematics of coordinate transformations to
the
H firstH law of thermodynamics. Recall the primitive form of the first law Eq. (3.167):
δQ = δW . In intensive form, this becomes
I I
δq = δw. (9.98)
We also know that δq = T ds and δw = P dv, so that

I I
T ds = P dv. (9.99)
Geometrically, one could say that an area in the T − s plane has the same value in the P − v
plane. Moreover, because the cyclic integral is direction-dependent, one must insist that an
area in the T −s plane maintain its orientation in the P −v plane. As an example, a rotation
of a two-dimensional geometric entity preserves area and orientation, while a reflection of
the same entity preserves area, but not orientation.

Now we can consider equations of state to be coordinate mappings; for example, consider
the general equations of state
T = T (P, v), (9.100)

s = s(P, v). (9.101)
These are mappings that take points in the P −v plane into the T −s plane. The differentials
of Eqs. (9.100, 9.101) are
∂T ∂T
dT = dP + dv, (9.102)
∂P v ∂v P
∂s ∂s
ds = dP + dv. (9.103)
∂P v ∂v P
In matrix form, we could say

∂T ∂T

dT dP
= ∂P
∂s
v ∂v
∂s
P . (9.104)
ds dv
| ∂P v {z ∂v P
}
=J
Here, we have defined the Jacobian matrix of the mapping from the standard mathematics
of coordinate transformations: ∂T
∂T
J = ∂P ∂s
v ∂v P
∂s . (9.105)
∂P v ∂v P
In a standard result from mathematics, for a coordinate transformation to be area- and
orientation-preserving, its Jacobian determinant, J must have a value of unity:
∂T ∂T
J ≡ det J = ∂P v ∂v P = 1. (9.106)
∂s ∂s
∂P v ∂v P
Expanding the Jacobian determinant, we require

∂T ∂s ∂T ∂s
J= − = 1. (9.107)
∂P v ∂v P ∂v P ∂P v
For general mathematical background of Jacobian matrices and coordinate transformations,

the interested reader can consult a variety of sources, for example, Kaplan,6 or Powers and
Sen.7
Example 9.2
Show a CPIG has a mapping from the T − s plane to the P − v plane that is area- and orientation-
preserving.
6
W. Kaplan, 2003, Advanced Calculus, Fifth Edition, Addison-Wesley, Boston, pp. 90-95, pp. 331-336.
7
J. M. Powers and M. Sen, 2015, Mathematical Methods in Engineering, Cambridge, New York, pp. 31-42.

9.6. THE VAN DER WAALS GAS 371
For a CPIG, it is easily shown, see Eq. (6.91) for s(P, v), that
Pv
T (P, v) = , (9.108)
R
v P
s(P, v) = so + cP ln + cv ln . (9.109)
vo Po
Calculating J from Eq. (9.107), we find

v c P c c − c R
P v P v
J= − = = = 1. (9.110)
R v R P R R
9.6 The van der Waals gas

A van der Waals gas is a common model for a non-ideal gas, introduced earlier in Ch. 2.4.2.1.
It can capture some of the behavior of a gas as it approaches the vapor dome. Its form was
first presented in Eqs. (2.46), (2.47):
RT a
P (T, v) = − 2, (9.111)
v−b v
where b accounts for the finite volume of the molecules, and a accounts for intermolecular
forces.
If we select
27 R2 Tc2 1 RTc
a= , b= , (9.112)
64 Pc 8 Pc
where Tc and Pc are the critical point temperature and pressure, respectively, we approximate
some physical behavior well, namely
• at the critical point ∂P/∂v|T = 0; that is an isotherm has a zero slope in the P − v
plane at the critical point, and
• at the critical point ∂ 2 P/∂v 2 |T = 0; that is an isotherm has a point of inflection in the
P − v plane at the critical point.
It is also easy to show that at the critical point, we have
3 RTc
vc = 3b = . (9.113)
8 Pc

Example 9.3
Show an isotherm has a slope of zero in the P − v plane at the critical point for a van der Waals
model.
Taking a partial derivative of Eq. (9.111), we see that

∂P RT 2a
=− + 3. (9.114)
∂v T (v − b)2 v
At the critical point, this has value
∂P RTc 2a
=− + 3. (9.115)
∂v T critical point (vc − b)2 vc
Now, substitute values for a, b, and vc for the van der Waals gas to get

R2 Tc2
∂P RTc 2 2764 Pc
= − 2 + 3 , (9.116)
∂v T critical point 3 RTc RTc 3 RTc
8 Pc − 8Pc 8 Pc

RT
RTc RTc 27 32 Pc
c
= − 2 + 3 , (9.117)
1 RTc 3 RTc
4 Pc 8 Pc
27
!
RTc 1 32
= 2 − 2 + , (9.118)
RTc 1 3 3
Pc 4 8

Pc2 27 512
= −16 + = 0. (9.119)
RTc 32 27
One could similarly take the second derivative and show the critical point is a point of inflection.
Example 9.4
Consider the van der Waals equation for water.
For water, we have Tc = 374.15 ◦ C = 647.3 K and Pc = 22120 kPa. We also have M =
18.015 kg/kmole. So
kJ
R 8.314 kmole K kJ
R= = kg
= 0.4615 . (9.120)
M 18.015 kmole kg K
Thus, our constants are
2
kJ
27 0.4615 kg K (647.3 K)2 kPa m6
a = = 1.70201 , (9.121)
64 (22120 kPa) kg2

0.4615 kgkJK (647.3 K) m3
b = = 0.00168813 . (9.122)
8(22120 kPa) kg

So our van der Waals equation of state for water is

m6
0.4615 kgkJK T 1.70201 kPa
kg 2
P = − 2
. (9.123)
v − 0.00168813 kg m3 v
Some isotherms for water, as predicted by the van der Waals equation, are given along with the actual
data for the vapor dome in Fig. 9.7. Obviously, the van der Waals predictions are not valid inside the
P (kPa)
K
20000
T
0
62
T
=
Tc
=
T
70
K
0
64
600
K
7.
3
T=
K
=
580
T=
10000
vapor dome
0
0.02 0.04 v (m3/kg)
Figure 9.7: Isotherms for water as predicted by the van der Waals equation along with actual
data for the vapor dome.
vapor dome, where isotherms must be isobars.
The critical volume is predicted by the van der Waals model, Eq. (9.113), to be

kJ
3 RTc 3 0.4615 kg K (647.3 K) m3
vc = 3b = = = 0.00506439 . (9.124)
8 Pc 8 22120 kPa kg
The actual data from Table B.1.2 of BS gives vc = 0.003155 m3 /kg, so clearly the van der Waals equation
has some inaccuracy, even near the critical point. This inaccuracy is clearly seen in Fig. 9.7 as the
isotherm corresponding to T = Tc = 647.3 K has its zero-slope inflection point at P = Pc = 22120 kPa,
slightly displaced from the measured value of vc .
Example 9.5
Find a general expression for the caloric equation of state for u(T, v) if we have a van der Waals

gas thermal equation of state.

RT a
P (T, v) = − . (9.125)
v − b v2
Proceed as before: First because u = u(T, v) we have Eq. (9.70):

∂u ∂u
du = dT + dv. (9.126)
∂T v ∂v T
Recalling Eqs. (3.228), (9.67), we also have that

∂u ∂u ∂P
= cv , =T − P. (9.127)
∂T v ∂v T ∂T v
Now, for the van der Waals gas, we find

∂P R
= , (9.128)
∂T v v−b
∂P RT
T −P = − P, (9.129)
∂T v v−b

RT RT a
= − − 2 , (9.130)
v−b v−b v
| {z }
P
a
= . (9.131)
v2
So we have
∂u a
= , (9.132)
∂v T v2
a
u(T, v) = − + f (T ). (9.133)
v
Here, f (T ) is some as-of-yet arbitrary function of T . To evaluate f (T ), take the derivative with respect
to T holding v constant:
∂u df
= = cv . (9.134)
∂T v dT
Because f is a function of T at most, here cv can be a function of T at most, so we allow cv = cv (T ).
Integrating, we find f (T ) as
Z T
f (T ) = C + cv (T̂ ) dT̂ , (9.135)
To
where C is an integration constant. Thus, u is

Z T
a
u(T, v) = C + cv (T̂ ) dT̂ − . (9.136)
To v
Taking C = uo + a/vo , we get
Z T
1 1
u(T, v) = uo + cv (T̂ ) dT̂ + a − . (9.137)
To vo v

We also find
Z T
1 1
h = u + P v = uo + cv (T̂ ) dT̂ + a − + P v, (9.138)
To vo v
Z T
1 1 RT v a
h(T, v) = uo + cv (T̂ ) dT̂ + a − + − . (9.139)
To vo v v−b v
Example 9.6
Consider an isothermal compression of water at T = 400 ◦ C from P1 = 8000 kPa to P2 = 10000 kPa.
Analyze this using a van der Waals gas model. Compare some results to those from an ideal gas model
and those from a steam table model.
At state 1, we have T1 = 400 ◦ C = 673.15 K. Note that T1 > Tc . Now, we also have P1 = 8000 kPa.
Note that P1 < Pc . Let us find an estimate for v1 from the van der Waals equation, Eq. (9.123):

m6
0.4615 kgkJK (673.15 K) 1.70201 kPa
kg 2
P1 = 8000 kPa = − 2 . (9.140)
v1 − 0.00168813 kgm3 v1
Omitting units, Eq. (9.140) expands to −3.59151 × 10−7 + 0.000212751v1 − 0.0405208v12 + v13 = 0. This
cubic equation has three roots:
m3
v1 = 0.00291758 ± 0.00135727i , non-physical, (9.141)
kg
m3
v1 = 0.0346857 , physical. (9.142)
kg
The two non-physical roots are a complex conjugate pair. In other cases, they might appear to be
physical, but will still be nothing more than mathematical relics with no physical meaning. At this
state, the ideal gas law, v1 = RT1 /P1 predicts a value of v1,ideal gas = 0.0388327 m3 /kg. Data from the
steam tables show that v1,steam tables = 0.03432 m3 /kg. So the van der Waals equation gives a better
prediction of v1 than does an ideal gas assumption.
At state 2, we have T2 = 400 ◦ C = 673.15 K and P2 = 10000 kPa. Again, we find an estimate for
v2 from the van der Waals equation, Eq. (9.123):

m6
0.4615 kgkJK (673.15 K) 1.70201 kPa
kg2
P2 = 10000 kPa = − . (9.143)
v2 − 0.00168813 m
3 v22
kg
Again, we find three roots:
m3
v2 = 0.0029749 ± 0.00136715i , non-physical, (9.144)
kg
m3
v2 = 0.0268045 , physical. (9.145)
kg

At this state, the ideal gas law, v2 = RT2 /P2 predicts a value of v2,ideal gas = 0.0310662 m3 /kg. Data
from the steam tables show that v2,steam tables = 0.02641 m3 /kg. So again, the van der Waals equation
gives a better prediction of v2 than does an ideal gas assumption.
The first law of thermodynamics tells us
u2 − u1 = 1q2 − 1w2 . (9.146)
Because we can compute u2 − u1 and 1w2 , the first law lets us get the heat transfer 1q2 . Let us first
compute the work:
Z 2
1w2 = P dv, (9.147)
1
Z 2
RT a
= − dv, (9.148)
1 v − b v2

v2 − b 1 1
= RT1 ln +a − , (9.149)
v1 − b v2 v1

m3 m 3

kJ
0.0268045 kg − 0.00168813 kg
= 0.461504 (673.15 K) ln
kg K 0.0346857 m
3
− 0.00168813 m3
kg kg
!
kPa m6 1 1
+ 1.70201 3 − 3 , (9.150)
kg2 0.0268045 m kg 0.0346857 mkg
kJ
= −70.3563 . (9.151)
kg
The work in compression is negative. From the steam tables, we could estimate the work via numerical
integration along an isotherm. We only have two data points from the steam tables, so the estimate is
simple 1w2 ∼ Pave (v2 − v1 ). This gives (9000 kPa)(0.02641 m3 /kg − 0.03432 m3 /kg) = −71.19 kJ/kg.
Now, let us find u2 − u1 . Note if we had an ideal gas, this would be zero, because it is an
isothermal process. However, the van der Waals gas has u(T, v), and because v changes, so does u.
From Eq. (9.137), we can deduce that
Z T2
1 1
u2 − u1 = cv (T ) dT −a − , (9.152)
T1 v2 v1
| {z }
=0
!
kPa m6 1 1
= − 1.70201 m3
− m3
, (9.153)
kg2 0.0268045 kg 0.0346857 kg
kJ
= −14.4277 . (9.154)
kg
We can compare this with u2 −u1 from the steam tables, that gives (2832.38 kJ/kg)−(2863.75 kJ/kg) =
−31.37 kJ/kg.
So the heat transfer is
1q2 = u2 − u1 + 1w2 , (9.155)

kJ kJ
= −14.4277 + −70.3563 , (9.156)
kg kg
kJ
= −84.7839 . (9.157)
kg

This compares with an estimate based on the steam tables of 1q2 = −102.56 kJ/kg.
Now, let us get the entropy change. We start with the Gibbs equation, Eq. (6.59): du = T ds−P dv.
Rearranging, we get
T ds = du + P dv. (9.158)
We can differentiate Eq. (9.137) to get

a
du = cv (T ) dT + dv. (9.159)
v2
Substitute Eq. (9.159) into Eq. (9.158) to get
a
T ds = cv (T ) dT + P + 2 dv. (9.160)
v
Now, we know from the van der Waals gas equation, Eq. (9.111), that P + a/v 2 = (RT )/(v − b), so
Eq. (9.160) reduces to
RT
T ds = cv (T ) dT + dv, (9.161)
v−b
cv (T ) R
ds = dT + dv, (9.162)
T v−b
Z 2
cv (T ) v2 − b
s2 − s1 = dT + R ln . (9.163)
1 T v1 − b
For the isothermal changes of this problem, we have simply
v2 − b
s2 − s1 = R ln , (9.164)
v1 − b

m3 m3

kJ
0.0268045 kg − 0.00168813 kg
= 0.461504 ln , (9.165)
kg K 0.0346857 m3
− 0.00168813 m3
kg kg
kJ
= −0.12591 . (9.166)
kg K
The estimate of the entropy change from the steam tables is s2 −s1 = 6.2119 kJ/kg/K−6.3633 kJ/kg/K =
−0.1514 kJ/kg/K.
Note that for this isothermal problem
Z 2 kJ
δq 1q2 −84.7839 kg kJ
s2 − s1 = = = = −0.125951 . (9.167)
1 T T 673.15 K kg K
Thermal energy left the system, and its entropy went down. This thermal energy entered the sur-
roundings, that needed to have Tsurr ≤ 673.15 K for this process to occur. Had Tsurr = 673.15 K, the
heat transfer process would have been slow, and the entropy of the surroundings would have risen by
precisely enough to balance the loss in the system, keeping the entropy of the universe constant. The
lower the surroundings’ temperature, the higher the total entropy change of the universe would have
been.

Example 9.7
Use Eq. (9.97) as an alternative way to find the entropy of a van der Waals gas.
We start with Eq. (9.97):

cv ∂P
ds = dT + dv. (9.168)
T ∂T v
From Eq. (9.134) we know for a van der Waals gas that at most cv = cv (T ). And for this material, we
also have ∂P/∂T |v = R/(v − b). So we get
cv (T ) R
ds = dT + dv. (9.169)
T v−b
Integrating, we get
Z T
cv (T̂ ) v−b
s(T, v) = so + + R ln . (9.170)
To T̂ vo − b
This is equivalent to what we found in Eq. (9.163).
Example 9.8
Find cP − cv for a van der Waals gas.
We start with Eq. (9.81):

∂P ∂v
cP − cv = T . (9.171)
∂T v ∂T P
Now, ∂v/∂T |P is difficult to compute for a van der Waals gas. But its reciprocal is not. Therefore let
us seek
∂P
∂T v
cP − cv = T ∂T
. (9.172)
∂v P
Now, for the van der Waals gas, Eq. (2.46), we have
RT a
P = − , (9.173)
v − b v2
and we easily get
∂P R
= . (9.174)
∂T v v−b
We also solve for T to get
(v − b)(a + P v 2 )
T = . (9.175)
Rv 2
Leaving out details, one can show that
∂T P v 3 − av + 2ab
= . (9.176)
∂v P Rv 3

9.7. THE REDLICH-KWONG GAS 379
Eliminating P , one can then show, leaving out details, that
∂T RT v 3 − 2ab2 + 4abv − 2av 2

= . (9.177)
∂v P Rv 3 (v − b)
Substituting for cP − cv , it is seen, omitting many algebraic details, that

!
1
cP − cv = R 2a(v−b)2
. (9.178)
1− RT v 3
Thus, Mayer’s relation, Eq. (3.248), does not hold for the van der Waals gas. Note that when a = 0,
Mayer’s relation again holds, even for b 6= 0.
9.7 The Redlich-Kwong gas

For the Redlich-Kwong equation of state, Eq. (2.48),
RT a
P = − √ , (9.179)
v − b v(v + b) T
one can show using similar methods that


v
√To /T 
3a v+b
u(T, v) = uo + cvo (T − To ) + √ ln  vo
, (9.180)
2b To vo +b
v

3a ln b+v
cv (T, v) = cvo − , (9.181)
4bT 3/2
 ( TTo ) 3/2 
v
T v−b a v+b
s(T, v) = so + cvo ln + R ln + 3/2
ln  vo
.
To vo − b 2bTo vo +b
(9.182)
9.8 Adiabatic sound speed

With help from the mass, linear momentum, and energy equations, along with validation
from experiment, it can be shown that the speed of sound waves, c, is given by the formula
s
∂P
c= . (9.183)
∂ρ s

As the entropy is constant for such a calculation, this is sometimes called the adiabatic sound
speed.
Let us calculate c. From the Gibbs equation, Eq. (6.60), we have
T ds = du + P dv. (9.184)
Now, because v = 1/ρ, we get dv = −(1/ρ2 ) dρ, and Eq. (9.184) can be rewritten as
P
T ds = du − dρ. (9.185)
ρ2
Now, for simple compressible substances, we can always form u = u(P, ρ). Thus, we also
have
∂u ∂u
du = dP + dρ. (9.186)
∂P ρ ∂ρ P
Now, use Eq. (9.186) to eliminate du in Eq. (9.185) so to get
∂u ∂u P
T ds = dP + dρ − 2 dρ, (9.187)
∂P ρ ∂ρ P ρ
| {z }
=du

∂u ∂u P
= dP + − dρ. (9.188)
∂P ρ ∂ρ P ρ2
p
Now, to find c = ∂P/∂ρ|s , take ds = 0, divide both sides by dρ, and solve for ∂P/∂ρ|s in
Eq. (9.188) so as to get
∂u P
∂P − ∂ρ
+ ρ2
P
= ∂u
. (9.189)
∂ρ s ∂P ρ
Now, Eq. (9.189) is valid for a general equation of state. Let us specialize it for a CPIG. For
the CPIG, we have
u = cv T + constant, (9.190)

P
= cv + constant, (9.191)
ρR
cv P
= + constant, (9.192)
cP − cv ρ
1 P
= cP + constant, (9.193)
cv
−1 ρ
1 P
= + constant. (9.194)
k−1 ρ

9.8. ADIABATIC SOUND SPEED 381
Thus, we have for the CPIG

∂u 1 1
= , (9.195)
∂P ρ k−1ρ
∂u −1 P
= . (9.196)
∂ρ P k − 1 ρ2
Now, substitute Eqs. (9.195, 9.196) into Eq. (9.189) to get
1 P
∂P k−1 ρ2
+ ρP2
= 1 1 , (9.197)
∂ρ s k−1 ρ
P P
= + (k − 1) , (9.198)
ρ ρ
P
= k , (9.199)
ρ
= kRT. (9.200)
Thus,
∂P
c2 = = kRT, (9.201)
∂ρ s
s
√ P
c= kRT = k . (9.202)
ρ
Compare this to the isothermal sound speed:

s
∂P √
cT = = RT . (9.203)
∂ρ T
By use of the ideal gas law, one can also say

s
P
cT = . (9.204)
ρ
This is the form Newton used in 1687 to estimate the sound speed; however, he probably used
an approach different from assuming Boyle’s law and taking derivatives. Newton’s approach
was corrected by Laplace in 1816 who generated what amounts to our adiabatic prediction,
long before notions of thermodynamics were settled. Laplace is depicted in Fig. 9.8. Laplace’s
notions rested on an uncertain theoretical foundation; he in fact adjusted his theory often,
and it was not until thermodynamics was well established several decades later that our
understanding of sound waves clarified. The interested reader can consult Finn.8
8
B. S. Finn, 1964, “Laplace and the speed of sound,” Isis, 55(1): 7-19.

Figure 9.8: Pierre-Simon Laplace (1749-1827), French mathematician and

physicist who improved Newton’s sound speed estimates; image from
https://mathshistory.st-andrews.ac.uk/Biographies/Laplace.html.
Example 9.9
At T = 300 K, estimate the adiabatic sound speed and compare it to the isothermal sound speed.
The adiabatic sound speed is

s
√ 7 J m
c= kRT = 287 (300 K) = 347 . (9.205)
5 kg K s
The isothermal sound speed is

s
√ J m
cT = RT = 287 (300 K) = 293 . (9.206)
kg K s
Newton’s published estimate in the first edition of his Principia was 968 ft/s = 295 m/s. In his
second edition, he adjusted his estimate to 979 ft/s = 298 m/s. However, it was well known at that
time that the measured speed of sound in air was roughly 1100 ft/s = 335 m/s. Newton put forth
some speculations to try to come to agreement with experiment, but these did not stand the test of
time. Careful experiment with the local temperature carefully monitored shows conclusively that the
adiabatic sound speed better predicts the data observed in nature than the isothermal sound speed.
Lastly, we recall Eq. (9.93), valid for general materials, k = (∂v/∂P |T )(∂P/∂v|s ). This actually
allows us to relate adiabatic and isothermal sound speeds for general materials. Slightly rearranging
Eq. (9.93) and using ρ = 1/v, we get
dρ ∂P
∂P
dv ∂ρ s − v12 ∂P
∂ρ s
∂P
∂ρ s

c
2
∂v s
k= ∂P
= = = = . (9.207)
∂v T
dρ ∂P
− v12 ∂P ∂P cT
dv ∂ρ T ∂ρ T ∂ρ T
That is to say, the ratio of specific heats k is also the ratio of the square of the ratio of the adiabatic
and isothermal sound speeds.

Example 9.10
Estimate the adiabatic and isothermal sound speeds of H2 O at P1 = 100 kPa, T1 = 200 ◦ C.
We will need to use the steam tables to get appropriate finite difference approximations to the two
relevant derivatives. The tables do not have ρ, but do have v, so let us recast the adiabatic sound speed
in terms of v:
∂P
c2 = , (9.208)
∂ρ s
dv ∂P
= , (9.209)
dρ ∂v s
1 ∂P
= dρ ∂v
, (9.210)
dv s
1 ∂P
= , (9.211)
−1/v 2 ∂v s
∂P
= −v 2 , (9.212)
∂v s
s
∂P
c = v − . (9.213)
∂v s
Similarly it is easy to show that

s
∂P
cT = v − . (9.214)
∂v T
Now the steam tables tell us at the given state that

m3 kJ
v1 = 2.17226 , s1 = 7.8342 . (9.215)
kg kg K
Let us now perturb the pressure, holding entropy constant, to a nearby pressure for which data is
available. Let us take then P2 = 200 kPa. We then take s2 = s1 = 7.8342 kJ/kg/K and interpolate to
get
m3
v2 = 1.27893 . (9.216)
kg
We then estimate the adiabatic sound speed with a finite difference, taking care to convert the pressures
to Pa,
r
∆P
c ≈ v1 − , (9.217)
∆v
r
P2 − P1
≈ v1 − , (9.218)
v2 − v1
v
m3 u u (200000 Pa) − (100000 Pa)
≈ 2.17226 t− , (9.219)
kg m3
1.27893 kg − 2.17226 m
3
kg
m
≈ 726.785 . (9.220)
s

The isothermal sound speed requires no interpolation. For P2 = 200 kPa, T2 = T1 = 200 ◦ C, the tables
tell us
m3
v2 = 1.08034 . (9.221)
kg
Then we get
r
∆P
cT ≈ v1 − , (9.222)
∆v
r
P2 − P1
≈ v1 − , (9.223)
v2 − v1
v
m3 u u (200000 Pa) − (100000 Pa)
≈ 2.17226 t− , (9.224)
kg m3
1.08034 kg − 2.17226 m
3
kg
m
≈ 657.38 . (9.225)
s
We can compare these to over-simplistic estimates from approximating steam as an ideal gas with
R = 461.5 J/kg/K and k = 1.327 that gives us
s
√ kJ m
c ≈ kRT = 1.327 461.5 ((200 + 273.15) K) = 538.295 , (9.226)
kg K s
s
√ kJ m
cT ≈ RT = 461.5 ((200 + 273.15) K) = 467.289 . (9.227)
kg K s
The estimates from ideal gas theory are low relative to those from the tables.
We close this section with the relevant passage from a translation of Newton’s Principia.9
The title page and a portion of the original 1687 Latin text are depicted in Fig. 9.9. Note
the 1687 Latin text employs a slightly different numbering system for the “Problem” than
does the translation.
Proposition L. Problem XII
To find the distances of the pulses
Let the number of the vibrations of the body, by whose tremor the pulses are
produced, be found to any given time. By that number divide the space which a
pulse can go over in the same time, and the part found will be the breadth of one
pulse. Q.E.I.
9
I. Newton, 1934, Principia, Cajori’s revised translation of Motte’s 1729 translation, U. California Press,
Berkeley. Link to the 1726 edition in the original Latin.

Figure 9.9: Images from the original 1687 edition of Newton’s Principia.
Scholium
The last Propositions respect the motions of light and sounds; for since light
is propagated in right lines, it is certain that it cannot consist in action alone
(by Prop. XLI and XLII). As to sounds, since they arise from tremulous bodies,
they can be nothing else but pulses of the air propagated through it (by Prop.
XLIII); and this is confirmed by the tremors which sounds, if they be loud and
deep, excite in the bodies near them, as we experience in the sound of drums;
for quick and short tremors are less easily excited. But it is well known that any
sounds, falling upon strings in unison with the sonorous bodies, excite tremors
in those strings. This is also confirmed from the velocity of sounds; for since the
specific gravities of rain-water and quicksilver are to one another as about 1 to 13
2/3, and when the mercury in the barometer is at the height of 30 inches of our
measure, the specific gravities of the air and of rain-water are to one another as
about 1 to 870, therefore the specific gravities of air and quicksilver are to each
other as 1 to 11890. Therefore when the height of the quicksilver is at 30 inches,
a height of uniform air, whose weight would be sufficient to compress our air to
the density we find it to be of, must be equal to 356700 inches, or 29725 feet of
our measure; and this is that very height of the medium, which I have called A in
the construction of the forgoing Proposition. A circle whose radius is 29725 feet
is 186768 feet in circumference. And since a pendulum 39 1/5 inches in length
completes one oscillation, composed of its going and return, in two seconds of
time, as is commonly known, it follows that a pendulum 29725 feet, or 256700
inches in length will perform a like oscillation in 190 3/4 seconds. Therefore in

that time a sound will go right onwards 186768 feet, and therefore in one second
979 feet.
But in this computation we have made no allowance for the crassitude of the solid
particles of the air, by which the sound is propagated instantaneously. Because
the weight of air is to the weight of water as 1 to 870, and because salts are almost
twice as dense as water; if the particles of air are supposed to be of about the same
density as those of water or salt, and the rarity of the air arises from the intervals
of the particles; the diameter of one particle of air will be to the interval between
the centres of the particles as 1 to about 9 or 10, and to the interval between the
particles themselves as 1 to 8 or 9. Therefore to 979 feet, which according to the
above calculation, a sound will advance forwards in one second of time, we may
add 979
9
, or about 109 feet, to compensate for the crassitude of the particles of
air: and then a sound will go forwards about 1088 feet in one second of time.
Moreover, the vapors floating in the air being of another spring, and a different
tone, will hardly, if at all, partake of the motion of the true air in which the
sounds are propagated. Now, if these vapors remain unmoved, that motion will
be propagated the swifter through the true air alone, and that as the square root
of the defect of the matter. So if the atmosphere consist of ten parts of true air
and one part of vapors, the motion of sounds will be swifter as the square root of
the ratio of 11 to 10, or very nearly in the entire ratio of 21 to 20, than if it were
propagated through eleven parts of true air: and therefore the motion of sounds
above discovered must be increased in that ratio. By this means the sound will
pass through 1142 feet in one second of time.
These things will be found true in spring and autumn, when the air is rarefied by
the gentle warmth of those seasons, and by that means its elastic force becomes
somewhat more intense. But in winter, when the air is condensed by the cold,
and its elastic force is somewhat more remitted, the motion of sounds will be
slower as the square root of the density; and, on the other hand, swifter in the
summer.
Now, by experiments it actually appears that sounds do really advance in one
second of time about 1142 feet of English measure, or 1070 feet of French measure.
The velocity of sounds being known, the intervals of the pulses are known also.
For M. Sauveur, by some experiments that he made, found that an open pipe
about five Paris feet in length gives a sound of the same tone with a viol sting
that vibrates a hundred times in one second. therefore there are near 100 pulses
in a space of 1070 Paris feet, which a sound runs over in a second of time; and
therefore one pulse fills up a space of about 10 7/10 Paris feet, that is, about twice
the length of the pipe. From this it is probably that the breadths of the pulses, in
all sounds made in open pipes, are equal to twice the length of the pipes.
Moreover, from the Corollary of Prop. XLVII appears the reason why the sounds

9.9. INTRODUCTION TO COMPRESSIBLE FLOW 387
immediately cease with the motion of the sonorous body, and why they are heard
no longer when we are at a great distance from the sonorous bodies that when we
are very near them. And besides, from the foregoing principles, it plainly appears
how it comes to pass that sounds are so mightily increased in speaking-trumpets;
for all reciprocal motion tends to be increased by the generating cause at each
return. And in tubes hindering the dilatation of the sounds, the motion decays
more slowly, and recurs more forcibly; and therefore is the more increased by the
new motion impressed at each return. And these are the principal phenomena of
sounds.
9.9 Introduction to compressible flow
We close these course notes with an opening to later coursework in which thermodynamics
and the adiabatic sound speed plays a critical role: compressible fluid mechanics. We only
sketch a few critical results here and leave the details for another semester.
To see the importance of the sound speed for compressible flows, let us consider briefly
the equations of motion for a one-dimensional flow in a duct with area change. We ignore
effects of momentum and energy diffusion as embodied in viscosity and heat conduction.
The conservation laws of mass, linear momentum, and energy can be shown to be
∂ ∂
(ρA) + (ρvA) = 0, (9.228)
∂t ∂x
∂v ∂v ∂P
ρ +v = − , (9.229)
∂t ∂x ∂x

∂u ∂u ∂v ∂v
+v = −P +v . (9.230)
∂t ∂x ∂t ∂x
Note, we have not specified any equation of state. It can be shown that viscosity and
heat diffusion, that we have neglected, are the only mechanisms to generate entropy in a
flow without shock waves. Because we have neglected these mechanisms, our equations are
isentropic as long as there are no shock waves. Note that Eq. (9.230) can be rewritten as
du/dt = −P dv/dt when we define the material derivative as d/dt = ∂/∂t + v∂/∂x. Thus,
Eq. (9.230) also says du = −P dv. Comparing this to the Gibbs equation, Eq. (6.59),
du = T ds − P dv, we see that our energy equation, Eq. (9.230), is isentropic, ds = 0.
We can thus replace Eq. (9.230) by ds/dt = ∂s/∂t + v∂s/∂x = 0. We also take a general
equation of state P = P (ρ, s). So our governing equations, Eqs. (9.228-9.230) supplemented
by the general equation of state become

∂ ∂
(ρA) + (ρvA) = 0, (9.231)
∂t ∂x
∂v ∂v ∂P
ρ +v = − , (9.232)
∂t ∂x ∂x
∂s ∂s
+v = 0, (9.233)
∂t ∂x
P = P (ρ, s). (9.234)
9.9.1 Acoustics
Let us first explore the acoustic limit in which disturbances to an otherwise stationary
material are small but non-zero. We restrict attention to purely isentropic flows, so s =
constant, and all its derivatives are zero. We first consider the state equation, Eq. (9.234)
so as to remove P from our analysis.
∂P ∂P
dP = dρ + ds, (9.235)
∂ρ s ∂s ρ
∂P ∂P ∂ρ ∂P ∂s
= + , (9.236)
∂x ∂ρ ∂x ∂s ∂x
| {z }s ρ |{z}
=0
=c2
∂ρ
= c2 . (9.237)
∂x
We next consider Eq. (9.231) in the limit where A is a constant and Eq. (9.232) where ∂P/∂x
is replaced in favor of ∂ρ/∂x via Eq. (9.237):
∂ρ ∂
+ (ρv) = 0, (9.238)
∂t
∂x
∂v ∂v ∂ρ
ρ +v = −c2 . (9.239)
∂t ∂x ∂x
We next assume that the state variables ρ and v are the sum of a constant state and a
small perturbation:
ρ = ρo + ρ̃, (9.240)
v = 0 + ṽ. (9.241)
The velocity is assumed to be perturbed about zero, the stationary state. We substitute
Eqs. (9.240-9.241) into Eqs. (9.238-9.239) to get
∂ ∂
(ρo + ρ̃) + ((ρo + ρ̃) ṽ) = 0, (9.242)
∂t ∂x
∂ṽ ∂ṽ ∂
(ρo + ρ̃) + ṽ = −c2 (ρo + ρ̃) . (9.243)
∂t ∂x ∂x

We expand to get
∂ρo ∂ ρ̃ ∂ṽ ∂
+ + ρo + (ρ̃ṽ) = 0, (9.244)
∂t
|{z} ∂t ∂x ∂x
| {z }
 =0  ∼0

 ∂ṽ ∂ṽ  ∂ṽ ∂ṽ ∂ρo ∂ ρ̃
ρo  + ṽ  + ρ̃ + ṽ = −c2 −c2 . (9.245)
∂t |{z}∂x ∂t ∂x ∂x
|{z} ∂x
| {z }
∼0 ∼0 =0
Neglecting terms involving the product of small terms, we are left with
∂ ρ̃ ∂ṽ
+ ρo = 0, (9.246)
∂t ∂x
∂ṽ ∂ ρ̃
ρ0 = −c2 . (9.247)
∂t ∂x
Now, take the time derivative of Eq. (9.246) and the space derivative of Eq. (9.247) and get
∂ 2 ρ̃ ∂ 2 ṽ
+ ρo = 0, (9.248)
∂t2 ∂t∂x
∂ 2 ṽ ∂ 2 ρ̃
ρo = −c2 2 . (9.249)
∂x∂t ∂x
Next, realizing the order of the mixed second partial derivatives does not matter for functions
that are continuous and differentiable, we eliminate ∂ 2 ṽ/∂t∂x and get
∂ 2 ρ̃ 2
2 ∂ ρ̃
= c . (9.250)
∂t2 ∂x2
Taking P = Po + P̃ , and using Eq. (9.202) for c2 , we have
P
c2 = k , (9.251)
ρ
Po + P̃
= k , (9.252)
ρo + ρ̃

Po 1 + PP̃o
= k , (9.253)
ρo 1 + ρρ̃o
!
Po P̃ ρ̃
= k 1+ 1− + ... , (9.254)
ρo Po ρo
!
Po P̃ ρ̃
= k 1+ − + ... . (9.255)
ρo Po ρo

We retain only the most important term and take then

Po
c2 = c2o + . . . , with c2o = k . (9.256)
ρo
So we get
∂ 2 ρ̃ 2
2 ∂ ρ̃
= c o . (9.257)
∂t2 ∂x2
This is the well known wave equation that is satisfied by the well known d’Alembert solution:
ρ̃(x, t) = f (x + co t) + g(x − co t). (9.258)
It is named after Jean le Rond d’Alembert, see Fig. 9.10, Here, f and g are arbitrary
Figure 9.10: Jean le Rond d’Alembert (1717-1783), French mathematician, physicist, and
music theorist; image from https://en.wikipedia.org/wiki/Jean le Rond d’Alembert.
functions. In a physical problem, they are determined by the actual initial and boundary
conditions that are appropriate for the particular problem. The so-called “phase” φ of f is
φ = x + co t. We can find the speed of a point with constant phase by considering φ to be a
constant, and taking appropriate derivatives:
φ = constant = x + co t, (9.259)
dφ dx
=0 = + co , (9.260)
dt dt
dx
= −co . (9.261)
dt
Thus, waves described by ρ̃(x, t) = f (x + co t) are traveling to the left (negative x direction)
with speed co . Similarly the waves given by g(x − co t) are traveling to the right (positive x
direction) with speed co .

Example 9.11
Let us verify that ρ̃(x, t) = f (x + co t) satisfies the wave equation, Eq. (9.257). The proof for
g(x − co t) is similar.
We simply need to calculate derivatives and then substitute into the original equation. The appro-
priate derivatives are
∂ ρ̃
= co f ′ (x + co t), (9.262)
∂t
∂ 2 ρ̃
= c2o f ′′ (x + co t), (9.263)
∂t2
∂ ρ̃
= f ′ (x + co t), (9.264)
∂x
∂ 2 ρ̃
= f ′′ (x + co t). (9.265)
∂x2
Substituting into the wave equation, Eq. (9.257) gives
c2o f ′′ (x + co t) = c2o f ′′ (x + co t). (9.266)
The wave equation is satisfied.
9.9.2 Compressible Bernoulli principle

We can employ the Bernoulli principle for a compressible CPIG to arrive at some common
notions in compressible flow. This is best illustrated by examining the Bernoulli principle
for an isentropic CPIG, Eq. (7.76) for a so-called stagnation flow, also neglecting gravity
forces. A stagnation point is a point in the flow where the velocity is zero.
We take the common notation of “o” for a stagnation state. Then neglecting gravity
forces, the Bernoulli principle, Eq. (7.76), can be recast as
k P 1 k Po 1
+ v2 = + v2 , (9.267)
k−1 ρ 2 k − 1 ρo |{z}
2
0
k P 1 k Po
+ v2 = . (9.268)
k−1 ρ 2 k − 1 ρo
We now use Eq. (9.202) to employ the square of the isentropic sound speed, taking c2 =

kP/ρ = kRT and c2o = kPo /ρo = kRTo . This gives
c2 1 c2o
+ v2 = , (9.269)
k−1 2 k−1
k − 1 v2 c2o
1+ = , (9.270)
2 c2 c2
k − 1 v2 kRTo
1+ = , (9.271)
2 c2 kRT
k − 1 v2 To
1+ = . (9.272)
2 c2 T
Now, we define the Mach number, M, as the ratio of the fluid velocity to the sound speed,
so that
v2
M2 = . (9.273)
c2
Mach is pictured in Fig. 9.11. Thus
Figure 9.11: Ernst Waldfried Josef Wenzel Mach (1838-1916). Austrian physicist and
philosopher. Image from https://en.wikipedia.org/wiki/Ernst Mach.
To k−1 2
=1+ M. (9.274)
T 2
So we see that the stagnation temperature To rises with Mach number. This represents
a conversion of kinetic energy to thermal energy as the fluid decelerates to the stagnation
condition. Then, we can use our isentropic relations for a CPIG, Eq. (6.169), to say
k−1
To Po k k−1 2
= =1+ M. (9.275)
T P 2

Rearranging, we get
k
Po k − 1 2 k−1
= 1+ M . (9.276)
P 2
Taylor series of this equation about M = 0 yields
Po k k
= 1 + M2 + M4 + . . . , (9.277)
P 2 8
k 2 1 2
= 1 + M 1 + M + ... , (9.278)
2 4
2

k ρv 1 2
= 1+ 1 + M + ... , (9.279)
2 kP 4
2

ρv 1 2
= 1+ 1 + M + ... , (9.280)
2P 4
2

ρv 1 2
Po = P + 1 + M + ... . (9.281)
2 4
For M = 0, we recover the incompressible Bernoulli principle, see Eq. (7.69), and the cor-
rection at small finite Mach number to the incompressible Bernoulli principle estimate for
stagnation pressure is evident.
Note, we can also achieve Eq. (9.274) by simple modification to our first law equation for
an isentropic nozzle, Eq. (4.269), to get the simple.
v2
ho = h + , (9.282)
2
v2
ho − h = , (9.283)
2
v2
cP (To − T ) = , (9.284)
2
v2
To = T+ , (9.285)
2cP
To v2
= 1+ , (9.286)
T 2cP T
kR v2
= 1+ , (9.287)
2cP kRT
cP cP − cv 2
= 1+ M, (9.288)
cv 2cP
k−1 2
= 1+ M. (9.289)
2
Example 9.12
A high Mach number vehicle travels at M = 5 into air at T = 300 K, P = 100 kPa. Find the
stagnation temperature and pressure at the leading edge of the vehicle. Neglect all effects of shock

waves and non-ideal gas effects and assume isentropic CPIG air. Compare the stagnation pressure to
predictions of various versions of the Bernoulli principle.
We first apply Eq. (9.275) and find

k−1 2
To = T 1+ M , (9.290)
2
7
5 −1 2
= (300 K) 1 + 5 , (9.291)
2
= 1800 K. (9.292)
The temperature increased significantly as the kinetic energy was converted to thermal energy. As a
consequence, the vehicle is subjected an extreme thermal environment.
For the stagnation pressure, Eq. (9.276) gives
k
k − 1 2 k−1
Po = P 1+ M , (9.293)
2
7/5
7 7/5−1
5 −1 2
= (100 kPa) 1 + 5 , (9.294)
2
= 52909 kPa. (9.295)
The pressure rose by a factor of 529, so the mechanical stress on the vehicle is high. These predictions
are really crude estimates. Both shock waves and non-ideal effects would play a significant role for a
real vehicle, and this would need to be considered in any realistic design.
Now let us examine the predictions of less accurate versions of the Bernoulli principle. For this we
need the ambient density, that by the ideal gas law is
P 105 Pa kg
ρ= = = 1.16144 3 . (9.296)
RT 287 kgJ K (300 K) m
We also need the ambient sound speed, that for the ideal gas is
s
√ 7 J m
c = kRT = 287 (300 K) = 347.189 (9.297)
5 kg K s
Thus, the flight velocity is

m m
v = Mc = 5 347.189 = 1735.94 . (9.298)
s s
In the incompressible limit, we artificially take M = 0 and reduce Eq. (9.281) to the incompressible
Bernoulli principle:
ρv2
Po ≈ P+ , (9.299)
2
kg

m 2
1.16144 m3 1735.94 s
5
= 10 Pa + , (9.300)
2
= 1.85 × 106 Pa, (9.301)
= 1850 kPa. (9.302)

The prediction of the incompressible Bernoulli principle has significant error! It under-predicts the
pressure by a factor of 28.6. Let us see how the corrected approximation does. Applying Eq. (9.281),
we get the approximation

ρv2 1 2
Po ≈ P + 1+ M , (9.303)
2 4
2
kg
1.16144 m 3 1735.94 ms 52

5
= 10 Pa + 1+ , (9.304)
2 4
= 1.27875 × 107 Pa, (9.305)
= 12787.5 kPa. (9.306)
This approximation is better, but still differs by a factor of 4.13 from the actual value. One notes that
the approximations were developed from a Taylor series in the limit of small Mach number, and the
Mach number here is not small. So there is no surprise that the approximate formulæ do not work
well. The approximations are both valid and accurate for small Mach finite M ≪ 1. For example, if
M = 0.3, the exact value is Po = 106.44303 kPa. The incompressible Bernoulli limit approximation
gives Po ∼ 106.30000 kPa, and the correction for small finite Mach number gives the more accurate
Po ∼ 106.44175 kPa.
9.9.3 Steady flow with area change

Let us now return to the full equations, Eqs. (9.231-9.234). In particular, we will now consider
potentially large fluid velocities, v; more specifically, the kinetic energy changes of the flow
may be as important as the internal energy changes. Let us also consider only steady flows;
thus, ∂/∂t = 0. Our governing equations, Eqs. (9.231-9.234), reduce to
d
(ρvA) = 0, (9.307)
dx
dv dP
ρv = − , (9.308)
dx dx
ds
= 0, (9.309)
dx
P = P (ρ, s). (9.310)
Specializing Eq. (9.237) for steady flows, we have
dP dρ
= c2 . (9.311)
dx dx
Using Eq. (9.311), in the linear momentum equation, Eq. (9.308), and expanding the
mass equation, Eq. (9.307), our mass and linear momentum equations become
dA dv dρ
ρv + ρA + vA = 0, (9.312)
dx dx dx
dv dρ
ρv = −c2 . (9.313)
dx dx

We next use Eq. (9.313) to eliminate dρ/dx in the mass equation, Eq. (9.312), to get

dA dv ρv dv
ρv + ρA + vA − 2 = 0, (9.314)
dx dx c dx
1 dA 1 dv v dv
+ − 2 = 0, (9.315)
A dx v dx
c 2dx
1 dv v 1 dA
1− 2 = − , (9.316)
v dx c A dx
v dA
dv A dx
= v2
. (9.317)
dx c2
−1
Recalling the Mach number, Eq. (9.273), M ≡ v/c, we can restate Eq. (9.317) as
v dA
dv
= A2 dx . (9.318)
dx M −1
Notice when the Mach number is unity, there is a potential singularity in dv/dx. This caused
great concern in the design of early supersonic vehicles. The only way to prevent the singular
behavior is to require at a sonic point, where M = 1, for dA/dx to be simultaneously zero.
Remarkably, this is precisely how nature behaves and is the reason why supersonic nozzles
are first converging, then diverging. At the point where dA/dx = 0, the flow becomes locally
sonic and can undergo a transition from subsonic to supersonic.
In terms of differentials, we can restate Eq. (9.318) as
dv 1 dA
= 2 . (9.319)
v M −1 A
Note, if the flow is subsonic, M < 1, with v > 0 and area increasing dA > 0, then dv < 0: area
increase in a subsonic nozzle generates velocity decrease. For supersonic flow, the opposite
is true: area increase in a supersonic nozzle generates velocity increase. One can see a
converging-diverging nozzle as well as its use in generating supersonic flow at its exit plane
in an image of a 2010 space shuttle launch depicted in Fig. 9.12.

a) b)
Figure 9.12: a) diverging section of a nozzle for the space shuttle main engine; image
from https://en.wikipedia.org/wiki/Space Shuttle main engine, b) Launch of Space
Shuttle Atlantis, STS-132, 14 May 2010, with a crew including astronaut Michael T. Good,
BSAE 1984, MSAE 1986, University of Notre Dame; image from https://www.nasa.gov.


Appendix
This appendix will consider two peripheral topics: Legendre transformations, first introduced
in Sec. 9.3, and the method of least squares.
Legendre transformations
Here we will draw upon the work of Zia, et al.10 to better understand Legendre transforma-
tions in general and how they relate to thermodynamics. The method is general and is often
applied to other problems in physics, especially in dynamics.
Consider a function F (x). Let us insist that F be such that
d2 F
≥ 0.
dx2
That is to say, its slope increases or does not change as x increases. Such a function may
possess a minimum value, and can be considered to be convex. Formally, one can say that
F is convex when the region above it, known as the epigraph, forms a so-called convex set.
A set is convex if special linear combinations of any of its elements also reside within the
set. Examples of convex and non-convex functions are shown in Fig. 9.13. For the convex
function F (x) = x2 , shown in Fig. 9.13a, we have d2 F/dx2 = 2 > 0. We also see that
special linear combinations of any points within the epigraph will lie within the epigraph.
Mathematically, this is expressed as follows: for s ∈ [0, 1] we must have F (sx1 + (1 − s)x2 ) ≤
sF (x1 ) + (1 − s)F (x2 ). This is illustrated by a sample line whose interior points all lie within
the epigraph. For the non-convex function F (x) = x3 − x2 − x, shown in Fig. 9.13b, we have
d2 F/dx2 = 6x − 2. Obviously, this is not positive for all x, and so the function is non-convex.
And we see that lines exist connecting points within the epigraph that contain points outside
of the epigraph. Hence it is non-convex.
Let us define the slope of F as w. Mathematically, we can say
dF
= w(x).
dx
Because F is a function of x, its derivative also is a function of x. Now because we have
insisted that w is increasing as x increases, we can always find a unique inverse such that
10
R. K. P. Zia, E. F. Redish, and S. R. McKay, 2009, “Making sense of the Legendre transform,” Ameri-
can Journal of Physics, 77(7): 614-622.
399
F(x) F(x)
2.0 2
0.990m
e
%
ativ
°
60
//KKgDr
Rel
n
atio
50%
epigraph epigraph
tur
yﬁr
1.5
rSa
1
llo
tBu
We
sob
1.0
0
0.5
-1
0.0
convex function non-convex
30 35 function 40
- 0.5 -2
- 1.0 - 0.5 0.0 0.5 1.0 - 1.0 - 0.5 0.0 0.5 1.0 1.5 2.0
x x
a) b)
Figure 9.13: Plots illustrating a) the convex function F (x) = x2 and b) the non-convex
function F (x) = x3 − x2 − x.
x(w) exists. That is to say, just as the slope is a function of x, x can be identified as a
function of slope. Now let us define the Legendre transformation G as
G(w) = wx(w) − F (x(w)).
Given F (x), and convexity, it is always possible to compute G. We can write this in a form
that will be more useful for thermodynamics:
G(w) + F (x) = wx.
Here we see a symmetry in the relationship. Now differentiate G with respect to w. We get
dG dx dF dx
=w + x(w) − .
dw dw dx dw
Rearrange to get
dG dx dF
= w− +x(w)
dw dw dx
| {z }
0
Now because dF/dx = w, we get dG/dw = x. Thus we get a set of symmetric relations
dF dG
= w, = x.
dx dw
We could also say then that
dG dF
G(w) + F (x) =
.
dw dx
It can remarkably be shown that the Legendre transformation of G returns us to the original
function F .
400
We also have
d2 F dw d2 G dx
2
= , 2
= .
dx dx dx dw
So
d2 F d2 G dw dx
2 2
= = 1.
dx dw dx dw
Example 9.1
Find the Legendre transformation of F (x) = ex .
First, we find that

dF d x
w= = (e ) = ex .
dx dx
We also see that
d2 F
= ex > 0,
dx2
so F is convex. We also see that w(x) exists, as does its inverse x(w):
w(x) = ex , x(w) = ln w.
For the inverse to be real valued, we must require that w > 0, that is the case for w = ex . So
G(w) = wx(w) − F (x(w)),

= w ln w − exp(ln w),
= w ln w − w,
= w(ln w − 1).
Example 9.2
Find the Legendre transformation of F (x) = x(ln x − 1). Note the function F has the same form
as the Legendre transformation of ex studied in the previous example.
We easily see that

dF
= ln x,
dx
so
w(x) = ln x.
We also see that
d2 F 1
2
= ,
dx x
so F is convex as long as x > 0. We then get that
x(w) = ew .
401
The Legendre transformation is
G(w) = wx(w) − F (x(w)),

= wew − ew (ln ew − 1),
= wew − ew (w − 1),
= ew .
The Legendre transformation of the Legendre transformation returns us to the original function. Thus,
the transformation is its own inverse.
Example 9.3
Find the Legendre transformation of F (x) = αx2 /2, where α is a positive scalar constant α > 0.
First, we find that

dF d αx2
w= = = αx.
dx dx 2
We also see that
d2 F
= α.
dx2
Because α > 0, F is convex. We also see that w(x) exists, as does its inverse x(w):
w
w(x) = αx, x(w) = .
α
So
G(w) = wx(w) − F (x(w)),

w α w 2
= w − ,
α 2 α
w2
= .
2α
Example 9.4
Find the Legendre transformation of F (x) = x2 /(2α), with α > 0. Note the function F has the
same form as the Legendre transformation of αx2 /2 studied in the previous example.
We easily see that

dF x
= ,
dx α
so
x
w(x) = .
α
402
We also see that
d2 F 1
= ,
dx2 α
so F is convex. We then get that
x(w) = αw.
The Legendre transformation is
G(w) = wx(w) − F (x(w)),

1
= w(αw) − (αw)2 ,
2α
α
= αw2 − w2 ,
2
α 2
= w .
2
The Legendre transformation of the Legendre transformation returns us to the original function. Thus,
the transform is its own inverse.
Example 9.5
Apply the formalism of Legendre transformations to a linear mass spring system.
We know the potential energy of such a system is

1 2
U (x) = kx ,
2
where k is the spring constant, and x is the displacement of the mass from its equilibrium position of
x = 0. We think of U (x) as standing in for F (x). Let us let then let f stand in for w and take
dU
f= = kx.
dx
Thus, we associate f with the traditional spring force kx. We see
d2 U
= k,
dx2
and because k > 0, U (x) is convex.
Now we also see
f
x(f ) = .
k
Let us let V stand in for G and define the Legendre transformation
V (f ) = xf − U (x).
So 2
f2 1 f
V (f ) = − k .
k 2 k
403
Thus,
1 f2
V (f ) = .
2 k
Note that
dV f
= = x(f ).
df k
Our Legendre transformation then gives
U (x) + V (f ) = f x,
1 2 1 f2
kx + = f x,
2 2 k
(kx)2 + f 2 − 2kf x = 0,
(f − kx)2 = 0.
This yields
f = kx,
as it must. We can think of U (x) as a potential associated with the position. We can think of V (f ) as
a potential associated with the control parameter f , the spring force.
Example 9.6
Consider the Legendre transformation in terms of traditional Lagrangian and Hamiltonian dynamics
for a particle of mass m with position q, velocity v = q̇, momentum p = mv = mq̇, Lagrangian L(q̇, q, t),
and Hamiltonian H(p, q, t). The particle moves in a potential field with the potential a known function
of position, V (q). We consider the so-called “coordinates” to be p and q̇
We typically think of Hamiltonian systems as non-dissipative; for particle dynamics, this implies
that the sum of the kinetic energy T = mv 2 /2 and potential energy V = mgz is a constant independent
of time:
H = T + V = constant.
Note the kinetic and potential energy can change with time, but their sum does not for the non-
dissipative system. Here z is a traditional position coordinate that we take to be z = q. And we simply
require the potential V to be a general function of position only: V (q). The kinetic energy in terms of
momentum p is
1 m2 v 2 p2
T = mv 2 = = .
2 2m 2m
So
p2
H(p, q, t) = + V (q).
2m
For the Legendre transformation formalism, we wish to consider the variables p and q̇. Let us then
define the “slope”
∂H p
= = q̇.
∂p q,t m
404
With the definition of momentum p = mv, we see that
q̇ = v.
We see that H is convex by computing

∂2H 1
= > 0,
∂p2 q,t m
because m > 0.
The Legendre transformation allows us to define the Lagrangian L in terms of the Hamiltonian H:
L(q̇, q, t) = pq̇ − H(p(q̇), q, t).
So
(mq̇)2
L(q̇, q, t) = pq̇ − + V (q) .
2m
Simplifying, we get
mq̇ 2
L(q̇, q, t) = mq̇ 2 − + V (q) ,
2
or
q̇ 2 mv 2
L(q̇, q, t) = m − V (q) = − V (q) = T − V.
2 2
So we have
H = T + V, L = T − V.
We could then say
L(q̇, q, t) + H(p, q, t) = pq̇.
Note that
∂L ∂H p
= mq̇ = mv = p, = = v = q̇.
∂ q̇ q,t ∂p q,t m
Now with L = L(q, q̇, t), we have the total derivative
∂L ∂L ∂L
dL = dq + dq̇ + dt.
∂q ∂ q̇ ∂t
Now p = ∂L/∂ q̇, so
∂L ∂L
dL = dq + p dq̇ + dt.
∂q ∂t
With the product rule, we say
∂L ∂L
dL = dq + d(pq̇) − q̇ dp + dt.
∂q ∂t
Rearranging gives
∂L ∂L
d (L − pq̇) = dq − q̇ dp + dt.
∂q ∂t
∂L ∂L
d (pq̇ − L) = − dq + q̇ dp − dt.
∂q ∂t
Recalling pq̇ − L = H, we get
∂L ∂L
dH = − dq + q̇ dp − dt.
∂q ∂t
405
Scale by dt to get
dH ∂L dp ∂L
=− q̇ + q̇ − .
dt ∂q dt ∂t
dH ∂L d ∂L ∂L
=− q̇ + q̇ − .
dt ∂q dt ∂ q̇ ∂t
Now because H = pq̇ − L, we have
∂H ∂L
=− .
∂t p,q̇ ∂t p,q̇
And because for non-dissipative systems H is constant, it does not vary with time, so we also have
∂L/∂t = 0, along with dH/dt = 0. Thus, after scaling by q̇, we get the well-known Euler-Lagrange
equation:
∂L d ∂L
− = 0.
∂q dt ∂ q̇
We can get the so-called Hamilton’s equations for a Hamiltonian system by the following. For a
constant Hamiltonian, we have H = H(p, q). So
∂H ∂H
dH = dp + dq = 0.
∂p q ∂q p
Scale by dt to get
dH ∂H ∂H
= ṗ + q̇ = 0.
dt ∂p q ∂q p
Because we already have q̇ = ∂H/∂p, we get

∂H ∂H ∂H
ṗ + = 0.
∂p q ∂q p ∂p q̇
For this to hold, we must require ṗ = −∂H/∂q. Thus, we have the well-known Hamilton’s equations:
∂H ∂H
= q̇, = −ṗ.
∂p q ∂q p
Let us see how this fits with thermodynamics. We will need some unusual notation to
identify the similarities. Let us take
P̃ ≡ −P, h̃ ≡ −h, ã ≡ −a, g̃ ≡ −g.
Our definition of enthalpy, h = u + P v, can be rewritten as
u − h = −P v.
Now recall from Eq. (9.8) that the canonical form for u is u(v, s), and that from Eq. (9.35),
the canonical form for h is h(P, s). Also invoking Eq. (9.10) to eliminate P and Eq. (9.40)
to eliminate v, we get
∂u ∂h
u(v, s) − h(P, s) = .
∂v s ∂P s
406
In terms of P̃ and h̃, we could say
∂u ∂ h̃
u(v, s) + h̃(P̃ , s) = .
∂v s ∂ P̃ s
Recognizing that s is a frozen variable here, we see that this is precisely the form of the
Legendre transformation G(w) + F (x) = wx, with dF/dx = w, dG/dw = x. We could also
say
u(v, s) = v P̃ (v, s) − h̃(P̃ (v, s), s),
to be consistent with the original form G(w) = wx(w) − F (x(w)). So u is playing the role
of G, v is playing the role of w, P̃ is playing the role of x, and h̃ is playing the role of F .
Rearranging gives
(−h̃(P̃ (v, s), s)) = u(v, s) + (−P̃ (v, s))v,
h(P (v, s), s)) = u(v, s) + P (v, s)v.
For convexity, we need d2 F/dx2 ≥ 0. This extends to
∂ 2 h̃
≥ 0.
∂ P̃ 2 s
In terms of h and P , this becomes

∂2h
≤ 0.
∂P 2 s
This is a requirement for convexity for a general equation of state. For example, we can
consider a CPIG, that has h = cP To (P/Po )(k−1)/k exp(s/cP ) + (ho − cP To ). For this special
case, we easily compute
s
− k+1
k
P
∂2h (k − 1)cP To e cP
Po
=− ≤ 0.
∂P 2 s k 2 Po2
As P > 0, Po > 0, To > 0, k > 1, cP > 0, the convexity condition is guaranteed. It is also
easy to show that
∂2h ∂2u
= −1.
∂P 2 s ∂v 2 s
Bethe11 , Zel’dovich12 , Thompson13 , and many others have explored consequences of non-
convexity of equations of state in what are now known as BZT fluids. As they explain,
11
Bethe, H. A., 1942, “The theory of shock waves for an arbitrary equation of state,” Technical Report
No. 545, Office of Scientific Research and Development.
12
Zel’dovich, Y. B., 1946, “On the possibility of rarefaction shock waves,” Zhurnal Eksperimental’noi i
Teoreticheskoi Fiziki, 16(4): 363-364, (English translation: Joint Publications Research Service 7320, Ar-
lington, VA., 1960).
13
Thompson, P. A., 1971, “A fundamental derivative in gasdynamics,” Physics of Fluids, 14(9): 1843-1849.
407
non-convexity can be introduced by non-ideal effects such as reflected in a van der Waals
equation of state.
Let us repeat this for a and g. First for Helmholtz free energy from Eq. (9.42) we get
u − a = T s.
Invoking Eqs. (9.10) and (9.47), we get

∂u ∂a
u−a= − .
∂s v ∂T v
In terms of ã, we get

∂u ∂ã
u(s, v) + ã(T, v) = .
∂s v ∂T v
Recognizing that v is a frozen variable here, we once again see that this is precisely the form
of the Legendre transformation G(w) + F (x) = wx, with dF/dx = w, dG/dw = x.
For Gibbs free energy, from Eq. (9.50), we can say
h − g = T s.
Invoking Eqs. (9.40) and (9.55), we can say

∂h ∂g
h−g = − .
∂s P ∂T P
Invoking g̃, we can say

∂h ∂g̃
h(s, P ) + g̃(T, P ) = .
∂s P ∂T P
Recognizing that P is a frozen variable here, we once again see that this is precisely the form
of the Legendre transformation G(w) + F (x) = wx, with dF/dx = w, dG/dw = x.
Method of least squares
One important application of data analysis is the method of least squares. This method is
often used to fit data to a given functional form. The form is most often in terms of polyno-
mials, but there is absolutely no restriction; trigonometric functions, logarithmic functions,
Bessel functions can all serve as well. Here, we will restrict ourselves to strictly scalar
functions of the form
x = f (t; aj ), j = 1, . . . , M,
where x is a dependent variable, t is an independent variable, f is an assumed functional
form, and aj is a set of M constant parameters in the functional form. The analysis can
408
easily be extended for functions of many variables. General mathematical background is
given by Strang.14
Mathematically, the fundamental problem is, given
• a set of N discrete data points, xi , ti , i = 1, . . . , N, and
• an assumed functional form for the curve fit f (t; aj ) that has M parameters aj , j =
1, . . . , M,
find the best set of parameter values aj so as to minimize the least squares error between the
curve fit and the actual data points. That is, the problem is to find aj , j = 1, . . . , M, such
that v
u N
uX
ℓ2 = ||xi − f (ti ; aj )||2 ≡ t (xi − f (ti ; aj ))2 ,
i=1
is minimized. Here, ℓ2 represents a total error of the approximation. It is sometimes called a

“norm” of the approximation or an “L-two norm.” The notation || · ||2 represents the L-two
norm of a vector represented by “·.” In that it is a square root of the sum of squares, it can
be thought of as an unusual distance, as motivated by Pythagoras’15 theorem.
In the least squares method, one
• examines the data,
• makes a non-unique judgment of what the functional form might be,
• substitutes each data point into the assumed form so as to form an over-constrained
system of equations,
• uses straightforward techniques from linear algebra to solve for the coefficients that
best represent the given data if the problem is linear in the coefficients aj ,
• uses techniques from optimization theory to solve for the coefficients that best represent
the given data if the problem is non-linear in aj .
The most general problem, in which the dependency aj is non-linear, is difficult, and
sometimes impossible. For cases in which the functional form is linear in the coefficients aj or
can be rendered linear via simple transformation, it is possible to get a unique representation
of the best set of parameters aj . This is often the case for common curve fits such as straight
line, polynomial, or logarithmic fits.
Let us first consider polynomial curve fits. Now, if one has say, ten data points, one can
in principle, find a ninth order polynomial that will pass through all the data points. Often
times, especially when there is much experimental error in the data, such a function may
be subject to wild oscillations, that are unwarranted by the underlying physics, and thus is
14
G. Strang, 1988, Linear Algebra and its Application, Harcourt Brace Jovanovich, Orlando, Florida.
15
Pythagoras of Samos, c. 570 B.C.-495 B.C., Ionian Greek philosopher and mathematician.
409
not useful as a predictive tool. In such cases, it may be more useful to choose a lower order
curve that does not exactly pass through all experimental points, but that does minimize
the error.
Unweighted least squares
This is the most common method used when one has equal confidence in all the data.
Example 9.7
Find the best straight line to approximate the measured data relating x to t.
t x
0 5
1 7
2 10
3 12
6 15
A straight line fit will have the form
x = a1 + a2 t,
where a1 and a2 are the terms to be determined. Substituting each data point to the assumed form,
we get five equations in two unknowns:
5 = a1 + 0a2 ,
7 = a1 + 1a2 ,
10 = a1 + 2a2 ,
12 = a1 + 3a2 ,
15 = a1 + 6a2 .
This is an over-constrained problem, and there is no unique solution that satisfies all of the equations!
If a unique solution existed, then the curve fit would be perfect. However, there does exist a solution
that minimizes the error, as is often proved in linear algebra textbooks (and will not be proved here).
The procedure is straightforward. Rearranging, we get
   
1 0 5
1 1 7
  a1  
1 2
 a2 = 10 .
 

1 3 12
1 6 15
This is of the form A · a = b. We then find
AT · A · a = AT · b,
−1
a = AT · A · AT · b.
410
Substituting, we find that
  −1  
1 0 5
 1 1   7 
a1  1 1 1 1 1 1
 1 1 1 1 1 
10 = 5.7925 .

= 2

a2  0 1 2 3 6 1
 0 1 2 3 6   1.6698
 3 12
1 6 15
So the best fit estimate is

x = 5.7925 + 1.6698 t.
The least squares error is ||A · a − b||2 = 1.9206. This represents what is known as the ℓ2 error norm
of the prediction. In MATLAB, this is found by the command norm(A ∗ a − b) where A, a, and b are the
coefficient matrix A, the solution a and the input vector b, respectively. If the curve fit were perfect,
the error norm would be zero. A plot of the raw data and the best fit straight line is shown in Fig. 9.14.
20
16
Data Points
12
x
x = 5.7925 + 1.6698 t
4
2 4 6
t
Figure 9.14: Plot of x − t data and best least squares straight line fit.
Weighted least squares
If one has more confidence in some data points than others, one can define a weighting
function to give more priority to those particular data points.
411
Example 9.8
Find the best straight line fit for the data in the previous example. Now however, assume that we
have five times the confidence in the accuracy of the final two data points, relative to the other points.
Define a square weighting matrix W:
 
1 0 0 0 0
0 1 0 0 0 
 
W= 0 0 1 0 0  .

0 0 0 5 0 
0 0 0 0 5
Now, we perform the following operations:
A·a = b,
W·A·a = W · b,
T T
(W · A) · W · A · a = (W · A) · W · b,
−1
T T
a = (W · A) · W · A (W · A) · W · b.
With the above values of W, direct substitution leads to

a1 8.0008
a= = .
a2 1.1972
So the best weighted least squares fit is
x = 8.0008 + 1.1972 t.
A plot of the raw data and the best fit straight line is shown in Fig. 9.15.
When the measurements are independent and equally reliable, W is the identity matrix.
If the measurements are independent but not equally reliable, W is at most diagonal. If the
measurements are not independent, then non-zero terms can appear off the diagonal in W.
It is often advantageous, for instance in problems in which one wants to control a process in
real time, to give priority to recent data estimates over old data estimates and to continually
employ a least squares technique to estimate future system behavior. The previous example
does just that. A famous fast algorithm for such problems is known as a Kalman16 Filter.
Power law/logarithmic curve fits
It is common and useful at times to fit data to a power law form, especially when the
data range over wide orders of magnitude. For clean units, it is advisable to scale both x
16
Rudolf Emil Kálmán, 1930-2016, Hungarian-American electrical engineer.
412
20
weighted data points

16
12
x
8
x = 8.0008 + 1.1972 t
2 4 6
t
Figure 9.15: Plot of x − t data and best weighted least squares straight line fit.
and t by characteristic values. Sometimes this is obvious, and sometimes it is not. Whatever
the case, the following form can usually be found
a2
x(t) t
= a1 .
xc tc
Here, x is a dependent variable, t is an independent variable, xc is a characteristic value of x
(perhaps its maximum), tc is a characteristic value of t (perhaps its maximum), and a1 and
a2 are curve fit parameters. This fit is not linear in the coefficients, but can be rendered so
by taking the logarithm of both sides to get
a2
x(t) t t
ln = ln a1 = ln(a1 ) + a2 ln .
xc tc tc
Often times one must not include values at t = 0 because of the logarithmic singularity
there.
Example 9.9
An experiment yields some data, shown next.
t(s) x(nm)
0.0 0.0
1 × 10−3 1 × 100
1 × 10−2 5 × 101
1 × 100 3 × 105
1 × 101 7 × 109
1 × 102 8 × 1010
413
Analyze.
A plot of the raw data is shown in Fig. 9.16. Notice that the linear plot obscures the data at small
x 1010
8 1012
a) b)
6
108
x (nm)
x (nm)
4
104
2
100
20 40 60 80 100 10-4 10-2 100 102
t (s) t (s)
Figure 9.16: Plot of x − t data in a) linear and b) log-log plots.
time, while the log-log plot makes the trends more clear. Now, to get a curve fit for the log-log plot, we
assume a power law form. We first eliminate the point at the origin, then scale the data, in this case
by the maximum values of t and x, and take appropriate logarithms to get to following values.

t x
t (s) x (nm) t/tmax x/xmax ln tmax ln xmax
1 × 10−3 1 × 100 1 × 10−5 1.25 × 10−11 −11.5129 −25.1053
1 × 10−2 5 × 101 1 × 10−4 6.25 × 10−10 −9.2013 −21.1933
1 × 100 3 × 105 1 × 10−2 3.75 × 10−6 −4.6052 −12.4938
1 × 101 7 × 109 1 × 10−1 8.75 × 10−2 −2.3026 −2.4361
1 × 102 8 × 1010 1 × 100 1 × 100 0.0000 0.0000
Now, we prepare the system of linear equations to solve

x t
ln = ln a1 + a2 ln ,
xmax tmax
−25.1053 = ln a1 + a2 (−11.5129),
−21.1933 = ln a1 + a2 (−9.2013),
−12.4938 = ln a1 + a2 (−4.6052),
−2.4361 = ln a1 + a2 (−2.3026),
0.0000 = ln a1 + a2 (0.0000).
414
In matrix form, this becomes
   
1 −11.5129 −25.1053
1 −9.2013  −21.1933
  ln a1  
1 −4.6052  −12.4938 .
= 
  a2
1 −2.3026   −2.4361 
1 0.0000 0.0000
This is of the form
A · a = b.
T
As before, we multiply both sides by A and then solve for a, we get
−1
a = AT · A · AT · b.
Solving, we find
0.4206
a= .
2.2920
So that
ln a1 = 0.4206, a2 = 2.2920,
or
a1 = 1.5228.
So the power law curve fit is
2.2920
x(t) t
= 1.5228 ,
8.000 × 1010 nm 100 s
or 2.2920
11
t
x(t) = 1.2183 × 10 nm .
100 s
A plot of the raw data and curve fit is shown in Fig. 9.17.
Higher order curve fits
As long as the assumed form for the curve fit is linear in the coefficients, it is straight-
forward to extend to high order curve fits as demonstrated in the following example.
Example 9.10
An experiment yields the data that follows.
t x
0.0 1.0
0.7 1.6
0.9 1.8
1.5 2.0
2.6 1.5
3.0 1.1
Find the least squares best fit coefficients a1 , a2 , and a3 if the assumed functional form is
415
1012
x (nm) 108
104
100
10-4
10-4 10-2 100 102
t (s)
Figure 9.17: Plot of raw x − t data and power law curve fit to the data: x(t) =
(1.2183 × 1011 nm) (t/100 s)2.2920 .
1. x = a1 + a2 t + a3 t2 ,

2. x = a1 + a2 sin 6t + a3 sin t
3 .
Plot on a single graph the data points and the two best fit estimates. Which best fit estimate has the
smallest least squares error?
• x = a 1 + a 2 t + a 3 t2 :
We substitute each data point into the assumed form and get the following set of linear equations
1.0 = a1 + a2 (0.0) + a3 (0.0)2 ,

1.6 = a1 + a2 (0.7) + a3 (0.7)2 ,
1.8 = a1 + a2 (0.9) + a3 (0.9)2 ,
2.0 = a1 + a2 (1.5) + a3 (1.5)2 ,
1.5 = a1 + a2 (2.6) + a3 (2.6)2 ,
1.1 = a1 + a2 (3.0) + a3 (3.0)2 .
These can be rewritten as    

1 0.0 0.0 1.0
1 0.7 0.49   1.6
  a1  
1 0.9 0.81 a2  = 1.8 .
 

1 1.5 2.25 2.0

1 a 3
 
2.6 6.76 1.5
1 3.0 9.00 1.1
416
This is of the form
A · a = b.
As before, we multiply both sides by AT and then solve for a to get
−1
a = AT · A · AT · b.
Solving, we find  
0.9778
a =  1.2679  .
−0.4090
So the best quadratic curve fit to the data is
x(t) ∼ 0.9778 + 1.2679t − 0.4090t2.
The least squares error norm is

||A · a − x||2 = 0.0812.
t

• x = a1 + a2 sin 6 + a3 sin 3t :
This form has applied a bit of intuition. The curve looks like a sine wave of wavelength 6 that has
been transposed. So we suppose it is of such a form. The term a1 is the transposition; the term on
a2 is the fundamental frequency, also known as the first harmonic, that fits in the domain; the term
on a3 is the second harmonic; we have thrown that in for good measure.
We substitute each data point into the assumed form and get the following set of linear equations

0.0 0.0
1.0 = a1 + a2 sin + a3 sin ,
6 3

0.7 0.7
1.6 = a1 + a2 sin + a3 sin ,
6 3

0.9 0.9
1.8 = a1 + a2 sin + a3 sin ,
6 3

1.5 1.5
2.0 = a1 + a2 sin + a3 sin ,
6 3

2.6 2.6
1.5 = a1 + a2 sin + a3 sin ,
6 3

3.0 3.0
1.1 = a1 + a2 sin + a3 sin .
6 3
This can be rewritten as    
1 0.0 0.0 1.0
1 0.1164 0.2312   1.6
  a1  
1 0.1494 0.2955
 a2  = 1.8 .
 

1 0.2474 0.4794 2.0
  a3  
1 0.4199 0.7622 1.5
1 0.4794 0.8415 1.1
This is of the form
A · a = b.
T
As before, we multiply both sides by A and then solve for a, we get
−1
a = AT · A · AT · b.
417
Solving, we find  
1.0296
a = −37.1423 .
21.1848
So the best curve fit for this form is

t t
x(t) ∼ 1.0296 − 37.1423 sin + 21.1848 sin .
6 3
The least squares error norm is

||A · a − x||2 = 0.1165.
Because the error norm for the quadratic curve fit is less than that for the sinusoidal curve fit, the
quadratic curve fit is better in this case. A plot of the raw data and the two best fit curves is shown
in Fig. 9.18.
two-term sinusoidal
2 curve fit
quadratic
x
polynomial
curve fit
1 2 3
t
Figure 9.18: Plot of x − t data and two least squares curve fits x(t) ∼ 0.9778 + 1.2679t −
0.4090t2 , and x(t) ∼ 1.0296 − 37.1423 sin (t/6) + 21.1848 sin (t/3).
418
Bibliography
M. M. Abbott and H. C. van Ness, 1989, Thermodynamics with Chemical Applications, Second Edition,
Schaum’s Outline Series in Engineering, McGraw-Hill, New York.
First published in 1972, this is written in the style of all the Schaum’s series. It has extensive solved
problems and a crisp rigorous style that is readable by undergraduate engineers. It has a chemical
engineering emphasis, but is also useful for all engineers.
J. D. Anderson, 2019, Hypersonic and High-Temperature Gas Dynamics, Third Edition, AIAA, Reston,
Virginia.
First published in 1989, this readable text includes details of how equilibrium and non-equilibrium
thermodynamics influences hypersonic fluid mechanics.
J. D. Anderson, 2020, Modern Compressible Flow with Historical Perspective, Fourth Edition, McGraw-Hill,
New York.
First published in 1982, this readable text includes essential thermodynamics for high Mach number
fluid mechanics.
P. Atkins, 2010, The Laws of Thermodynamics: a Very Short Introduction, Oxford University Press, Oxford.
This is a short book by an eminent chemist summarizing the foundations of thermodynamics for an
interested general reader.
A. Bejan, 2016, Advanced Engineering Thermodynamics, Fourth Edition, John Wiley, Hoboken, New Jersey.
This is an advanced undergraduate text first published in 1988. It gives a modern treatment of
the science of classical thermodynamics. It does not confine its attention to traditional engineering
problems, and considers applications across biology and earth sciences as well; some readers will
find parts of the discussion to be provocative, as received wisdom is occasionally challenged. The
thermodynamics of irreversible processes are discussed in detail.
T. L. Bergman, A. S. Lavine, F. P. Incropera, and D. P. DeWitt, 2020, Fundamentals of Heat and Mass
Transfer, Eighth Edition, John Wiley, New York.
This has evolved since its introduction in 1981 into the standard undergraduate text in heat transfer.
The interested thermodynamics student will find the subject of heat transfer builds in many ways
on a classical thermodynamics foundation. Especially relevant to thermodynamics are chapters on
boiling and condensation, heat exchangers, as well as extensive tables of thermal properties of real
materials. The final two authors were the authors of the first edition; their names do not appear on
more recent editions.
R. S. Berry, S. A. Rice, and J. Ross, 2000, Physical Chemistry, Second Edition, Oxford University Press,
Oxford.
This is a rigorous general text in physical chemistry at a senior or first year graduate level. First
appearing in 1980, it has a full treatment of classical and statistical thermodynamics as well as
quantum mechanics.
419
C. Bohren and B. Albrecht, 2023, Atmospheric Thermodynamics, Second Edition, Oxford University Press,
Oxford.
This is readable advanced undergraduate text on thermodynamics as it applies to the atmosphere,
especially weather. First appearing in 1998, it adopts an irreverent style to treat some serious and
important subjects with rigor.
L. Boltzmann, 1995, Lectures on Gas Theory, Dover, New York.
This is a detailed monograph by the founding father of statistical thermodynamics. This edition is a
translation of the original German version, Gastheorie, from 1896-1898.
C. Borgnakke and R. E. Sonntag, 2017, Fundamentals of Thermodynamics, Ninth Edition, John Wiley,
New York.
This classic and popular undergraduate mechanical engineering text, that in earlier editions was
authored by J. G. van Wylen and Sonntag, has stood the test of time and has a full treatment of
most classical problems. Its first edition appeared in 1965. More recent editions exist, but are missing
important material such as Maxwell relations.
M. Born, 1949, Natural Philosophy of Cause and Chance, Clarendon Press, Oxford.
This monograph is a summary of the Waynflete Lectures delivered at Oxford University in 1948 by
the author, the winner of the 1954 Nobel Prize in physics. The lectures consider various topics, and
include an important chapter on thermodynamics with the author’s earlier defense of the approach of
Carathèodory playing a prominent role.
R. Boyle, 2003, The Sceptical Chymist, Dover, New York.
This is a reprint of the original 1661 work of the famous early figure of the scientific revolution.
P. W. Bridgman, 1943, The Nature of Thermodynamics, Harvard, Cambridge.
This short monograph by the winner of the 1946 Nobel Prize in physics gives a prosaic introduction to
thermodynamics that is directed at a scientifically literate audience who are not interested in detailed
mathematical exposition.
H. A. Buchdahl, 1966, The Concepts of Classical Thermodynamics, Cambridge, Cambridge.
This short text outlines the basic principles of thermodynamics.
H. B. Callen, 1985, Thermodynamics and an Introduction to Thermostatistics, Second Edition, John Wiley,
New York.
This advanced undergraduate text, an update of the 1960 original, has an emphasis on classical physics
applied to thermodynamics, with a few chapters devoted to quantum and statistical foundations.
S. Carnot, 2005, Reflections on the Motive Power of Fire, Dover, New York.
This is a translation of the author’s foundational 1824 work on the heat engines, Réflexions sur la
Puissance Motrice du Feu et sur les Machines propres à développer cette Puissance (“Reflections on
the Motive Power of Fire and on Machines Fitted to Develop that Power”). Also included is a paper
of Clausius.
Y. A. Çengel and M. A. Boles, 2023, Thermodynamics: An Engineering Approach, Tenth Edition, McGraw-
Hill, Boston.
This popular undergraduate mechanical engineering text that first appeared in 1989 has most of the
features expected in a modern book intended for a large and varied audience.
S. Chandrasekhar, 2010, An Introduction to the Study of Stellar Structure, Dover, New York.
This monograph, first published in 1939, on astrophysics is by the winner of the 1983 Nobel Prize
in physics. It has large sections devoted to rigorous axiomatic classical thermodynamics in the style
of Carathèodory, highly accessible to engineering students, that show how thermodynamics plays a
critical role in understanding the physics of the heavens.
420
R. J. E. Clausius, 2008, The Mechanical Theory of Heat, Kessinger, Whitefish, Montana.
This is a reprint of the 1879 translation of the 1850 German publication of the great German scientist
who in many ways founded classical thermodynamics.
S. R. de Groot and P. Mazur, 1984, Non-Equilibrium Thermodynamics, Dover, New York.
This is an influential monograph, first published in 1962, that summarizes much of the work of the
famous Belgian school of thermodynamics. It is written at a graduate level and has a strong link to
fluid mechanics and chemical reactions.
E. Fermi, 1936, Thermodynamics, Dover, New York.
This short 160 page classic clearly and efficiently summarizes the fundamentals of thermodynamics.
It is based on a series of lectures given by this winner of the 1938 Nobel Prize in physics. The book
is highly recommended, and the reader can benefit from multiple readings.
R. P. Feynman, R. B. Leighton, and M. Sands, 1963, The Feynman Lectures on Physics, Volume 1, Addison-
Wesley, Reading, Massachusetts.
This famous series documents the introductory undergraduate physics lectures given at the California
Institute of Technology by the lead author, the 1965 Nobel laureate in physics. Famous for their clarity,
depth, and notable forays into challenging material, they treat a wide range of topics from classical
to modern physics. Volume 1 contains chapters relevant to classical and modern thermodynamics.
J. B. J. Fourier, 2009, The Analytical Theory of Heat, Cambridge, Cambridge.
This reprint of the 1878 English translation of the 1822 French original, Théorie Analytique de la
Chaleur is a tour de force of science, engineering, and mathematics. It predates Carnot and the
development of the first and second laws of thermodynamics, but nevertheless successfully develops
a theory of non-equilibrium thermodynamics fully consistent with the first and second laws. In so
doing, the author makes key advances in the formulation of partial differential equations and their
solution by what are now known as Fourier series.
J. W. Gibbs, 1957, The Collected Works of J. Willard Gibbs, Yale U. Press, New Haven.
This compendium gives a complete reproduction of the published work of the monumental American
engineer and scientist of the nineteenth century, including his seminal work on classical and statistical
thermodynamics.
W. T. Grandy, 2008, Entropy and the Time Evolution of Macroscopic Systems, Oxford, Oxford.
This monograph gives an enlightening description, fully informed by both historical and modern
interpretations, of entropy and its evolution in the context of both continuum and statistical theories.
E. A. Guggenheim, 1933, Modern Thermodynamics by the Methods of Willard Gibbs, Methuen, London.
This graduate-level monograph was important in bringing the work of Gibbs to a wider audience.
E. P. Gyftopoulos and G. P. Beretta, 2010, Thermodynamics: Foundations and Applications, Dover, Mine-
ola, New York.
First published in 1991, this beginning graduate text has a rigorous development of classical thermo-
dynamics.
C. S. Helrich, 2009, Modern Thermodynamics with Statistical Mechanics, Springer, Berlin.
This is an advanced undergraduate text aimed mainly at the physics community. The author includes
a full treatment of classical thermodynamics and moves easily into statistical mechanics. The author’s
undergraduate training in engineering is evident in some of the style of the text that should be readable
by most undergraduate engineers after a first class in thermodynamics.
421
J. O. Hirschfelder, C. F. Curtis, and R. B. Bird, 1954, Molecular Theory of Gases and Liquids, Wiley, New
York.
This comprehensive tome is a valuable addition to any library of thermal science. Its wide ranging
text covers equations of state, molecular collision theory, reactive hydrodynamics, reaction kinetics,
and many other topics all from the point of view of careful physical chemistry. Much of the work
remains original.
J. R. Howell and R. O. Buckius, 1992, Fundamentals of Engineering Thermodynamics, Second Edition,
McGraw-Hill, New York.
This is a good undergraduate text for mechanical engineers; it first appeared in 1987.
W. F. Hughes and J. A. Brighton, 1999, Fluid Dynamics, Third Edition, Schaum’s Outline Series, McGraw-
Hill, New York.
Written in the standard student-friendly style of the Schaum’s series, this discussion of fluid mechanics
includes two chapters on compressible flow that bring together fluid mechanics and thermodynamics.
It first appeared in 1967.
R. J. Kee, M. E. Coltrin, and P. Glarborg, 2018, Chemically Reacting Flow: Theory and Practice, Second
Edition, John Wiley, New York.
This comprehensive text gives an introductory graduate level discussion of fluid mechanics, thermo-
chemistry, and finite rate chemical kinetics. The focus is on low Mach number reacting flows, and
there is significant discussion of how to achieve computational solutions.
J. H. Keenan, F. G. Keyes, P. G. Hill, and J. G. Moore, 1985, Steam Tables: Thermodynamic Properties
of Water Including Vapor, Liquid, and Solid Phases, John Wiley, New York.
This is a version of the original 1936 data set that is the standard for water’s thermodynamic prop-
erties.
J. Kestin, 1966, A Course in Thermodynamics, Blaisdell, Waltham, Massachusetts.
This is a foundational textbook that can be read on many levels. All first principles are reported in a
readable fashion. In addition the author makes a great effort to expose the underlying mathematical
foundations of thermodynamics.
C. Kittel and H. Kroemer, 1980, Thermal Physics, Second Edition, Freeman, San Francisco.
This is a classic undergraduate text, first introduced in 1970; it is mainly aimed at physics students.
It has a good introduction to statistical thermodynamics and a short effective chapter on classical
thermodynamics. Engineers seeking to broaden their skill set for new technologies relying on micro-
scale thermal phenomena can use this text as a starting point.
S. Klein and G. Nellis, 2012, Thermodynamics, Cambridge, Cambridge.
This is an undergraduate textbook for engineers treating a standard set of topics.
D. Kondepudi and I. Prigogine, 2015, Modern Thermodynamics: From Heat Engines to Dissipative Struc-
tures, Second Edition, John Wiley, New York.
This is a detailed modern exposition that exploits the authors’ unique vision of thermodynamics with
both a science and engineering flavor. The authors, the second of whom is one of the few engineers
who was awarded the Nobel Prize (chemistry 1977, for the work summarized in this text), often
challenge the standard approach to teaching thermodynamics, and make the case that the approach
they advocate, with an emphasis on non-equilibrium thermodynamics, is better suited to describe
natural phenomena and practical devices than the present approach, that is generally restricted to
equilibrium states.
422
K. K. Kuo, 2005, Principles of Combustion, Second Edition, John Wiley, New York.
This is a readable graduate level engineering text for combustion fundamentals. First published in
1986, it includes a full treatment of reacting thermodynamics as well as discussion of links to fluid
mechanics.
K. J. Laidler, 1987, Chemical Kinetics, Third Edition, Prentice-Hall, Upper Saddle River, NJ.
This is a standard advanced undergraduate chemistry text on the dynamics of chemical reactions. It
first appeared in 1965.
L. D. Landau and E. M. Lifshitz, 2000, Statistical Physics, Part 1, Volume 5 of the Course of Theoretical
Physics, Third Edition, Butterworth-Heinemann, Oxford.
This book, part of the monumental series of graduate level Russian physics texts, first published in
English in 1951 from Statisticheskaya fizika, gives a fine introduction to classical thermodynamics as
a prelude to its main topic, statistical thermodynamics in the spirit of Gibbs.
B. H. Lavenda, 1978, Thermodynamics of Irreversible Processes, John Wiley, New York.
This is a lively and opinionated monograph describing and commenting on irreversible thermodynam-
ics. The author is especially critical of the Prigogine school of thought on entropy production rate
minimization.
A. Lavoisier, 1984, Elements of Chemistry, Dover, New York.
This is the classic treatise by the man known as the father of modern chemistry, translated from the
1789 Traité Élementaire de Chimie, that gives the first explicit statement of mass conservation in
chemical reactions.
G. N. Lewis and M. Randall, 1961, Thermodynamics and the Free Energy of Chemical Substances, Second
Edition, McGraw-Hill, New York.
This book, first published in 1923, was for many years a standard reference text of physical chemistry.
H. W. Liepmann and A. Roshko, 2002, Elements of Gasdynamics, Dover, New York.
This is an influential text in compressible aerodynamics that is appropriate for seniors or beginning
graduate students. First published in 1957, it has a strong treatment of the physics and thermody-
namics of compressible flow along with elegant and efficient text. Its treatment of both experiment
and the underlying theory is outstanding, and in many ways is representative of the approach to
engineering sciences fostered at the California Institute of Technology, the authors’ home institution.
J. C. Maxwell, 2001, Theory of Heat, Dover, New York.
This is a short readable book by the nineteenth century master, first published in 1871. Here, the
mathematics is minimized in favor of more words of explanation.
M. J. Moran, H. N. Shapiro, D. D. Boettner, and M. B. Bailey, 2018, Fundamentals of Engineering Ther-
modynamics, Ninth Edition, John Wiley, New York.
This is a standard undergraduate engineering thermodynamics text, and one of the more popular.
First published in 1988, it has much to recommend it including good example problems, attention to
detail, good graphics, and a level of rigor appropriate for good undergraduate students.
P. M. Morse, 1969, Thermal Physics, Second Edition, Benjamin, New York.
This is a good undergraduate book on thermodynamics from a physics perspective. It covers classical
theory well in its first sections, then goes on to treat kinetic theory and statistical mechanics. It first
appeared in 1964.
I. Müller and T. Ruggeri, 1998, Rational Extended Thermodynamics, Springer-Verlag, New York.
This modern, erudite monograph gives a rigorous treatment of some of the key issues at the frontier
of modern continuum thermodynamics.
423
I. Müller and W. Weiss, 2005, Entropy and Energy, Springer-Verlag, New York.
This is a unique treatise on fundamental concepts in thermodynamics. The author provide mathe-
matical rigor, historical perspective, and examples from a diverse set of scientific fields.
I. Müller, 2007, A History of Thermodynamics: the Doctrine of Energy and Entropy, Springer-Verlag,
Berlin.
The author gives a readable text at an advanced undergraduate level that highlights some of the many
controversies of thermodynamics, both ancient and modern.
I. Müller and W. H. Müller, 2009, Fundamentals of Thermodynamics and Applications: with Historical
Annotations and Many Citations from Avogadro to Zermelo, Springer, Berlin.
The author presents an eclectic view of classical thermodynamics with much discussion of its history.
The text is aimed at a curious undergraduate who is unsatisfied with industrial-strength yet narrow
and intellectually vapid textbooks.
W. Nernst, 1969, The New Heat Theorem: Its Foundation in Theory and Experiment, Dover, New York.
This monograph gives the author’s exposition of the development of the third law of thermodynamics.
It first appeared in English translation in 1917 and was originally published in German.
S. Paolucci, 2016, Continuum Mechanics and Thermodynamics of Matter, Cambridge, New York.
This graduate level monograph has an extensive discussion of how thermodynamics fits within the
broader structure of continuum mechanics.
S. Paolucci, 2019, Undergraduate Lectures on Thermodynamics, BreviLiber, South Bend.
This short book gives summaries of traditional thermodynamics lectures with insightful example
problems.
S. Paolucci, 2019, Undergraduate Lectures on Intermediate Thermodynamics, BreviLiber, South Bend.
This short book gives summaries of traditional intermediate thermodynamics lectures with insightful
example problems.
W. Pauli, 2000, Thermodynamics and the Kinetic Theory of Gases, Dover, New York.
This is a monograph on statistical thermodynamics by the winner of the 1945 Nobel Prize in physics.
It is actually derived from his lecture course notes given at ETH Zurich, as compiled by a student in
his class, E. Jucker, published in 1952.
M. Planck, 1990, Treatise on Thermodynamics, Dover, New York.
This brief book, that originally appeared in German in 1897, gives many unique insights from the
great scientist who was the winner of the 1918 Nobel Prize in physics. It is rigorous, but readable by
an interested undergraduate student.
H. Poincaré, 1892, Thermodynamique: Leçons Professèes Pendant le Premier Semestre 1888-89, Georges
Carré, Paris.
This text of classical undergraduate thermodynamics has been prepared by one of the premier math-
ematicians of the nineteenth century.
M. Potter and C. Somerton, 2009, Thermodynamics for Engineers, Second Edition, Schaum’s Outline Series
in Engineering, McGraw-Hill, New York.
First published in 1993, this is a standard contribution in the Schaum format of many solved example
problems.
J. M. Powers, 2016, Combustion Thermodynamics and Dynamics, Cambridge, New York.
This graduate level monograph focusing on combustion has a detailed discussion of advanced under-
graduate chemical thermodynamics.
424
J. M. Powers, 2024, Mechanics of Fluids, Cambridge, Cambridge.
This graduate level textbook has an extensive discussion of thermodynamics and its coupling with
fluid mechanics.
I. Prigogine, 1967, Introduction to Thermodynamics of Irreversible Processes, Third Edition, Interscience,
New York.
This is a famous book that summarizes the essence of the work of the Belgian school for which the
author was awarded the 1977 Nobel Prize in chemistry. This book first appeared in 1955.
W. J. M. Rankine, 1908, A Manual of the Steam Engine and Other Prime Movers, Seventeenth Edition,
Griffin, London.
This in an accessible undergraduate text for mechanical engineers of the nineteenth century. It first
appeared in 1859. It contains much practical information on a variety of devices, including steam
engines.
L. E. Reichl, 2016, A Modern Course in Statistical Physics, Fourth Edition, John Wiley, New York.
This full service graduate text has a good summary of key concepts of classical thermodynamics and
a strong development of modern statistical thermodynamics.
O. Reynolds, 1903, Papers on Mechanical and Physical Subjects, Volume III, The Sub-Mechanics of the
Universe, Cambridge, Cambridge.
This volume compiles various otherwise unpublished notes of Reynolds and includes his detailed deriva-
tions of general equations of conservation of mass, momentum, and energy employing his transport
theorem.
W. C. Reynolds, 1968, Thermodynamics, Second Edition, McGraw-Hill, New York.
This is an unusually good undergraduate text written for mechanical engineers. The author has
wonderful qualitative problems in addition to the usual topics in such texts. A good introduction
to statistical mechanics is included as well. This particular edition is highly recommended; the first
edition appeared in 1965.
W. C. Reynolds and P. Colonna, 2018, Thermodynamics: Fundamentals and Engineering Applications,
Cambridge, New York.
This is a significant modernization of Reynolds’ 1968 test. It maintains the rigor and style of the
original and treats new and important topics.
S. I. Sandler, 2017, Chemical, Biochemical, and Engineering Thermodynamics, Fifth Edition, John Wiley,
New York.
This is an advanced undergraduate text in thermodynamics from a chemical engineering perspective
with a good mathematical treatment. It first appeared in 1977.
E. Schrödinger, 1989, Statistical Thermodynamics, Dover, New York.
This is a short monograph written by the one of the pioneers of quantum physics, the co-winner of
the 1933 Nobel Prize in Physics. It is based on a set of lectures delivered to the Dublin Institute for
Advanced Studies in 1944, and was first published in 1946.
A. H. Shapiro, 1953, The Dynamics and Thermodynamics of Compressible Fluid Flow, Vols. I and II, John
Wiley, New York.
This classic two volume set has a comprehensive treatment of the subject of its title. It has numerous
worked example problems, and is written from a careful engineer’s perspective.
J. M. Smith, H. C. van Ness, and M. Abbott, 2017, Introduction to Chemical Engineering Thermodynamics,
Eighth Edition, McGraw-Hill, New York.
This is probably the most common undergraduate text in thermodynamics in chemical engineering.
It first appeared in 1959. It is rigorous and has went through many revisions.
425
A. Sommerfeld, 1956, Thermodynamics and Statistical Mechanics, Lectures on Theoretical Physics, Vol. V,
Academic Press, New York.
This is a compilation of the author’s lecture notes on this subject. The book reflects the author’s
stature of a leader of theoretical physics of the twentieth century who trained a generation of students
(e.g. Nobel laureates Heisenberg, Pauli, Debye and Bethe). The book gives a fine description of
classical thermodynamics with a seamless transition into quantum and statistical mechanics.
J. W. Tester and M. Modell, 1997, Thermodynamics and Its Applications, Third Edition, Prentice Hall,
Upper Saddle River, New Jersey.
First appearing in 1974, this entry level graduate text in thermodynamics is written from a chemical
engineer’s perspective. It has a strong mathematical development for both classical and statistical
thermodynamics.
K. S. Thorne and R. D. Blandford, 2017, Modern Classical Physics: Optics, Fluids, Plasmas, Elasticity,
and Statistical Physics, Princeton University Press, Princeton, New Jersey.
This comprehensive text is accessible to advanced undergraduates who appreciate a physics perspec-
tive. It treats a wide range of topics that include and transcend thermodynamics.
C. A. Truesdell, 1980, The Tragicomic History of Thermodynamics, 1822-1854, Springer Verlag, New York.
This idiosyncratic monograph has a lucid description of the history of nineteenth century thermal
science. It is written in an erudite fashion, and the reader who is willing to dive into a difficult subject
will be rewarded for diligence by gain of many new insights.
C. A. Truesdell, 1984, Rational Thermodynamics, Second Edition, Springer-Verlag, New York.
This is an update on the evolution of classical thermodynamics in the twentieth century. The book
itself first appeared in 1969. The second edition includes additional contributions by some contempo-
raneous leaders of the field.
S. R. Turns and D. C. Haworth, 2020, An Introduction to Combustion: Concepts and Applications, Fourth
Edition, McGraw-Hill, Boston.
This is a senior-level undergraduate text on combustion that uses many notions from thermodynamics
of mixtures. It first appeared in 1996.
S. R. Turns and L. L. Pauley, 2020, Thermodynamics: Concepts and Applications, Second Edition, Cam-
bridge, Cambridge.
This is detailed undergraduate book on engineering thermodynamics.
H. C. van Ness, 1983, Understanding Thermodynamics, Dover, New York.
This is a short readable monograph from a chemical engineering perspective. It first appeared in 1969.
W. G. Vincenti and C. H. Kruger, 1976, Introduction to Physical Gas Dynamics, Krieger, Malabar, Florida.
This graduate text on high speed non-equilibrium flows contains a good description of the interplay
of classical and statistical mechanics. There is an emphasis on aerospace science and fundamental
engineering applications. It first appeared in 1965.
F. M. White, 2020, Fluid Mechanics, Ninth Edition, McGraw-Hill, Boston.
This standard undergraduate fluid text draws on thermodynamics in its presentation of the first law
and in its treatment of compressible flows. It first appeared in 1979.
B. Woodcraft, 1851, The Pneumatics of Hero of Alexandria, London, Taylor Walton and Maberly.
This is a translation and compilation from the ancient Greek of the work of Hero (10 A.D.-70 A.D.).
The discussion contains descriptions of the engineering of a variety of technological devices including
a primitive steam engine. Other devices that convert heat into work are described as well.
426
L. C. Woods, 1975, The Thermodynamics of Fluid Systems, Clarendon, Oxford.
This graduate text gives a good, detailed survey of the thermodynamics of irreversible processes,
especially related to fluid systems in which convection and diffusion play important roles.
427
Index
absolute pressure, 39 base load, 204

absolute temperature, 35, 37, 52, 60, 129, 168, Bernoulli principle, 318–324, 326, 391, 393, 394
219, 229 Bernoulli, D., 318
absolute zero, 15, 36, 242, 301 biology, 17
acceleration, 24, 37, 38, 41, 77, 120, 121, 171, Biot number, 147
264, 270, 322 Biot, M., 147
accelertion, 61 black swan, 17
acoustics, 388 Black, J., 15, 125
adiabatic, 110, 111, 150, 185, 188, 197, 200, 201, boiler, 204–206, 208, 329, 332, 337
205, 206, 222, 242, 247, 248, 266, 269, boiling point, 36, 45, 46
279, 282, 291–294, 297, 316, 320, 327, Boltzmann constant, 35, 299
349, 353, 380–382, 387 Boltzmann, L., 15, 299
sound speed, 380, 383 Boulton and Watt steam engine, 21
æolipile, 18 Boyle’s law, 16, 50
afterburner, 341 Boyle, R., 14
air, 134, 286, 342, 344, 350 Brayton cycle, 339, 340, 343, 344, 346, 349, 350
air standard, 286, 342, 344, 350 Brayton, G., 339
Amagat, É, 68 Bridgman, P., 16
ammonia, 73, 247, 248
Amontons, G., 50 calculus, 31, 87, 159–161, 165, 252
angular momenta principle, 168 vector, 158
Aquinas, T., 16 caloric equation of state, 127, 128, 139–141, 143,
argon, 346, 347 145, 146, 190, 201, 211, 215, 251
Aristotle, 12 caloric theory, 112, 125
Arrhenius, 152 calorically imperfect ideal gas, 127, 131, 132,
Arrhenius kinetic rate, 149 138, 255, 368
astronomy, 159 calorically perfect ideal gas, 127, 128, 254
Avogadro’s law, 50 calorie, 112
Avogadro’s number, 32, 52, 300 thermochemical, 113
axiom, 15–17, 36, 155, 157, 158, 167, 168, 211, calorimetry, 15, 125
218, 242, 310, 311 canonical form, 358, 361
angular momenta, 168 canonical variables, 362, 363
conservation, 168 Carathéodory, C., 15, 221
energy conservation, 168, 181 carbon dioxide, 73, 219
entropy evolution, 168 Carnot cycle, 230, 277–279, 282, 286, 329–332,
linear momenta, 168 344, 349, 353
mass conservation, 168–170 Carnot heat engine, 225
Newton’s second, 119–121 Carnot, S., 15, 225, 277
axiomatic approach, 36, 221 Celsius
degrees, 36
back work ratio, 330, 343 Celsius scale, 36
bacon, 109, 110 cesium, 35
428
change of state, 34 convection, 110, 145, 146, 148, 154
Charles’ law, 50 converging-diverging nozzle, 396
Charles, J., 50, 52 convex, 399, 401, 402
chemical energy, 118, 150, 341 convex set, 399
chemical equilibrium, 34 coordinate
chemical kinetics, 152 mapping, 370
chemical potential, 34 transformation, 370
chemistry, 68, 218, 294 Coriolis, G., 92
chiller, 355 critical point, 46, 48, 297, 371, 372
Clapeyron, É., 52 critical pressure, 46, 247
classical thermodynamics, 11–13, 17, 31, 77, 116, critical temperature, 46, 67, 247
144, 332 critical volume, 373
Clausius, R., 15, 17, 115, 220, 237 crystal, 48
Clermont, 21 crystallography, 159
coefficient of performance, 231, 354, 355 curl, 87, 89, 90
collision theory, 31 curve fitting
combustion, 101, 151, 205, 207, 286, 289, 290, higher order, 415
329, 331, 332, 341, 342, 357 linear, 410
Comet, 21 power law, 412
component efficiency, 327 cycle, 22, 34, 112, 221
compressed liquid, 46, 71, 176, 208 Brayton, 339, 340, 343, 344, 346, 349, 350
compressibility, 192, 193 Carnot, 230, 277–279, 282, 286, 329–332,
chart, 64, 65 344, 349, 353
factor, 63 Otto, 277, 286, 287, 289
compressible flow, 391 Rankine, 203–206, 209, 329–333, 338, 339,
compressible flow with area change, 395 343, 352
compressible fluid mechanics, 387 refrigeration, 352, 353
compression, 126 steam power, 22, 204, 206
compression ratio, 289 thermal efficiency, 208
compressor, 180, 198, 199, 322, 327, 328, 340, thermodynamic, 203, 208, 220, 239, 277,
341, 343, 346–351, 354 329, 353
condenser, 203, 204, 208, 210, 329, 331–333, vapor compression, 352
335–338, 353, 354 cyclic, 239
conduction, 109 engine, 227, 228, 237
conjugate variables, 362, 363 integral, 112, 116, 239, 240, 369
conservation axiom, 168
conservation of energy, 112, 118, 158, 168, 181, d’Alembert solution, 390
196 d’Alembert, J., 390
conservation of mass, 158, 168, 169, 196 Dalton, J., 50, 68
conservation of mass-energy, 156 degrees
conservation of momentum, 196 Celsius, 36
continuous, 83 Fahrenheit, 37
continuum, 31, 211 Kelvin, 35
mechanics, 209, 319 Rankine, 36, 37
contour, 86, 91 density, 12, 32, 35, 39, 40, 50, 62, 64, 145, 151,
closed, 87, 91, 112 165, 169–172, 185, 186, 194, 211, 212,
integral, 87 246, 264, 318, 324, 385, 386
control surface, 30, 187 derivative, 82, 159, 161, 163, 170, 182, 274, 374,
control volume, 30, 123, 157, 158, 167, 173–177, 381, 383, 388, 390, 391, 399
183–187, 189–192, 196, 208, 311, 312, first, 297
316, 352 material, 170–172, 213, 387
429
partial, 82, 83, 128, 163, 295, 296, 365, 372, 313, 316–320, 323, 326–328, 340, 341,
389 350, 352, 377, 379, 387, 395
second, 297, 345, 372 chemical, 150
space, 389 chemical, 118, 149, 153, 341
time, 145, 161, 180, 184, 275, 389 conservation of, 15, 112, 118, 294
total, 82, 128, 179 electric, 118
Descartes, R., 81 evolution of, 118
deviatoric stress tensor, 211 internal, 71, 118–120, 122, 130, 136, 138,
diatomic gas, 132, 134–136, 201 139, 144, 150, 187, 249, 260, 280, 300,
Diesel engine, 22 301, 361, 395
differential, 41, 172, 312, 319 kinetic, 12, 14, 24–26, 28, 81, 117–119, 133,
exact, 81, 83, 86, 358 134, 187, 193, 196, 197, 199–201, 203,
inexact, 84, 111 221, 229, 250, 298, 303, 316, 317, 320,
total, 82, 171 323, 326, 327, 340, 341, 350, 392, 394,
differential equation, 30 395
first order, 146, 151 magnetic, 118
non-linear, 151, 271 mechanical, 12, 16, 18, 24, 25, 27, 28, 81,
ordinary, 24, 146, 151, 271 117, 141, 197, 198, 225, 250, 318, 323,
partial, 209 341
second order, 24 potential, 24–26, 28, 81, 117, 118, 120, 134,
diffuser, 196, 197, 316, 317 141, 179, 187, 192, 196, 197, 199–201,
diffusivity 203, 221, 250, 316, 320, 323, 341, 403
thermal, 215 rotational, 133
disorder, 219 specific internal, 122
dissociation, 62, 134 thermal, 12, 16, 18, 28, 81, 109, 111, 112,
distance, 31, 39, 92–94, 102, 108, 112, 119, 120, 117–119, 139, 148–150, 199, 205, 206,
180, 272, 387, 409 208, 218, 225, 228, 231, 232, 245, 250,
district heating and cooling, 204 257, 298, 317, 318, 341, 377, 392, 394
divergence, 89, 90, 170, 172 total, 81, 118, 122, 145, 155, 179, 184, 187,
divergence theorem, 159, 160, 169, 173, 182, 184 250
dot product, 93, 180 translational, 133
double interpolation, 76 vibrational, 133, 135, 202
drag, 24, 26–28, 81, 270, 272, 273 energy conservation principle, 168, 181
dynamics, 399, 404 engine
automobile, 237
earth, 110 Diesel, 22
efficiency gasoline, 22
component, 327 jet, 157, 237
thermal, 208, 226 steam, 93
Einstein, A., 16, 17, 155 enthalpy, 71, 123, 124, 127, 136–138, 182, 187,
elasticity 190, 192, 193, 252, 328, 361, 362, 406
modulus, 212 specific, 123
electric energy, 118 total, 123, 184, 185, 187, 188
electron, 134 entropy, 15, 71, 111, 168, 218–220, 237, 238,
energy, 12–16, 18, 24–28, 39, 61, 71, 81, 93, 108– 240–246, 249–253, 255–258, 260, 269,
112, 117–120, 122, 127, 133–135, 139, 270, 275, 276, 291–299, 301–306, 309,
141, 148, 155, 157, 158, 168, 179, 187, 311, 314, 362, 369, 377, 380, 383, 387
189, 190, 192, 193, 196–203, 205–209, ideal gas, 253
211–213, 217–221, 225, 228, 229, 231– information, 309
235, 237, 245, 250, 256, 257, 291, 294– irreversible production, 311
296, 298, 300, 301, 303, 304, 309, 311, mixture, 294
430
randomness, 298 277, 280, 281, 283–285, 287, 292, 294,
thermo-mechanical mixing, 291 316, 327, 328, 369, 376
entropy evolution principle, 168 flow, 31, 157, 225, 318, 387, 391, 396
epidemiology, 152, 154 compressible, 321, 387, 391
epigraph, 399 fluid, 30
equation of state, 50, 66, 71, 81, 205, 209, 367, incompressible, 321, 326
370, 380, 387 inviscid, 224
caloric, 127, 128, 139–141, 143, 145, 146, isentropic, 388
211, 251, 373 one-dimensional, 171, 387
non-ideal, 62 stagnation, 391
Redlich-Kwong, 63, 379 steady, 395
tabular, 65 supersonic, 396
thermal, 50, 62, 65, 127, 142, 211, 369, 374 unsteady, 319
van der Waals, 62, 371–379 fluid, 30, 40–42, 45, 67, 73, 100, 110, 112, 125,
equilibrium, 33–35, 43, 71, 94, 140, 147, 178, 157, 158, 166, 167, 169–172, 176, 177,
190, 219, 222, 267, 271–274, 291, 293, 179–181, 186, 187, 197, 199, 205, 206,
294, 298, 309, 310, 403 212, 224, 263, 264, 276, 279, 318, 322,
chemical, 34 323, 327, 329, 331, 332, 336, 340–342,
mechanical, 33, 41, 102, 264, 266, 270, 273, 346, 387, 392, 395
291 fluid mechanics, 157, 166, 212, 332
pressure, 267 compressible, 212, 387
thermal, 33, 35, 153, 291, 294 force, 12, 24, 26–28, 30, 31, 37–39, 41, 48, 61,
Euclid, 16, 17, 36 62, 64, 77, 78, 81, 92–94, 102, 106, 108,
Euler method, 152 119–121, 134, 166, 168, 180, 212, 250,
Euler-Lagrange equation, 406 263, 264, 270, 272, 273, 341, 350, 352,
evolution of energy, 118 371, 386, 403
exact differential, 81, 83, 86, 179, 358 body, 37
exothermic, 149, 150 Fourier’s law, 109, 211, 213
explicit Euler method, 271 Fourier, J., 109, 213
extensive property, 33, 117, 119, 165, 168, 179 Fowler, R., 36
extrapolation, 74 free body diagram, 60, 102, 106
free energy
Fahrenheit Gibbs, 362, 364, 408
degrees, 37 Helmholtz, 362, 363, 408
Fermi, E., 16 freezing point, 36
Feynman, R., 134 frequency, 417
field, 82, 87 friction, 119, 206, 225, 348
electric, 108 Fulton, R., 21
gravitational, 26, 27, 112, 179, 322, 323 fundamental theorem for line integrals, 87
scalar, 82, 88 fundamental theorem of calculus, 86, 159, 160,
temperature, 298 165
vector, 89, 90 furnace, 157, 237
velocity, 170, 172
finite difference, 271, 383 Galileo, G., 14
first harmonic, 417 gas, 45, 108
first law of thermodynamics, 15, 36, 81, 112, diatomic, 132, 134–136, 201
115–121, 123–126, 129, 130, 137, 140, monatomic, 128, 129, 132, 136
141, 143–146, 155, 168, 179, 181, 185, noble, 128, 129
187, 188, 201, 202, 206, 208, 209, 213, turbine, 22
214, 218, 221, 225, 227, 231, 232, 234, gauge pressure, 39, 40, 324
237, 239, 247, 250, 256, 257, 269, 276, Gauss’s theorem, 159
431
Gay-Lussac, J., 52 isothermal, 353
gedankenexperiment, 43, 218 radiative, 110
general transport theorem, 166 reversible, 243, 250, 311
Giauque, W., 16 Helmholtz free energy, 362, 363, 408
Gibbs entropy formula, 299 Helmholtz, H., 15, 115
Gibbs equation, 250–253, 258, 259, 261, 319, Hero of Alexandria, 14, 18
331, 340, 358, 361, 362, 366, 377, 380, Hess, G., 15
387 holonomic system, 92
first, 251
second, 252 ideal gas, 14, 50, 52–54, 57–62, 64–66, 71, 97,
Gibbs free energy, 362, 364, 408 98, 127, 128, 130, 131, 136–138, 193,
Gibbs, J., 15, 250 194, 201, 211, 251, 253–256, 259–262,
governing equations of continuum mechanics, 209 267, 272, 289–292, 300, 301, 313, 342,
gradient, 87, 88, 109, 160 367, 368, 375, 376, 381, 384
pressure, 264 calorically imperfect, 131, 132, 138, 255,
temperature, 109 368
gravity, 212, 391 calorically perfect, 128, 254
entropy of, 253
Hamilton’s equations, 406 isentropic relations, 261
Hamiltonian, 404–406 ideal gas tables, 137
heat, 11, 12, 15, 29, 45, 108, 110, 112, 118, 124, incompressibility, 213
155, 168, 218, 233, 239, 329, 332, 341, incompressible, 126, 150, 170, 212, 258, 294,
342, 355 321, 322, 326, 393, 394
conduction, 212 incompressible liquid, 258
latent, 124, 147, 248 incompressible solid, 258
sensible, 124, 249 industrial revolution, 22
sign convention, 110 inertia, 61, 77, 212, 273
heat capacity inexact differential, 84, 111
specific, 125, 365 inflection point, 373
heat death, 310 information, 17
heat diffusion, 109, 387 information entropy, 309
heat engine, 15 information theory, 308
Carnot, 225 integral, 83, 84, 87, 91, 105, 117, 159, 161, 170,
heat equation, 213, 215 173, 181–184, 186, 201, 241, 244, 320
heat exchanger, 157, 199 cyclic, 112, 116, 239, 240, 369
counterflow, 200 surface, 160, 163, 165, 167, 169, 182, 184
heat fluix, 214 volume, 159, 160, 165, 167, 170, 184
heat flux vector, 109, 211 intensive property, 33, 49, 52, 77, 78
heat pump, 232, 354 internal energy, 71, 118–120, 122, 130, 136, 139,
Carnot, 231 144, 150, 187, 214, 249, 260, 280, 300,
heat transfer, 31, 108, 110, 111, 117, 122–124, 301, 361, 395
129, 130, 137, 139, 144, 147, 149, 153, interpolation, 73–75, 79, 208, 247, 285, 384
155, 166, 168, 179, 181–183, 185, 187, double, 74, 76, 206
188, 192, 196, 197, 199, 203–206, 208, linear, 73–76, 247
213, 228, 230, 233, 241–245, 248, 260, single, 74
279, 281, 285–287, 289, 292, 312, 314, ion, 134
317, 319, 332, 333, 352, 376, 377 ionization, 134
coefficient, 145 irreversibility, 314
convective, 145, 152 irreversible, 224
irreversible, 243 entropy production, 311
isobaric, 353 process, 224, 243, 244
432
work, 227 Leibniz’s rule, 161
isentrope, 242, 248, 260, 262, 263 Leibniz, G, 81, 161
isentropic, 247, 248, 261–264, 266, 268, 271, length, 12, 31, 32, 35, 41, 108, 168, 272, 385,
275–277, 279, 280, 286, 288–290, 321, 386
327–329, 331, 337, 339, 343, 347–349, linear interpolation, 73–76, 247
351, 353, 354, 369, 387, 388, 391, 392, linear momenta principle, 168
394 linearized analysis of differential equations, 273
isobar, 57, 71, 260, 340, 354, 373 liquid, 45, 46, 251
isobaric, 34, 44, 45, 48, 50, 97, 98, 100, 105, 124, compressed, 46, 71, 176, 208
126, 137, 248, 249, 329, 334, 339, 346, incompressible, 258
353 saturated, 44, 46, 66, 71, 74, 122, 208, 252
isochore, 259, 260 subcooled, 354
isochoric, 34, 61, 97, 100, 105, 123, 125, 143,
144, 256, 257, 269, 286, 287, 290, 292 Müller, I., 14
isolated system, 29 Mach number, 392, 393, 395, 396
isotherm, 79, 252, 262, 263, 371–373 Mach, E., 392
isothermal, 14, 34, 45, 50, 51, 97, 98, 248, 249, magnetic energy, 118
262, 279, 282, 322, 353, 364, 365, 375– manometer, 40, 41
377, 381, 382 manometry, 40
isothermal sound speed, 381 Mariotte’s law, 50
Mariotte, E., 50
Jacobian determinant, 370 mass, 12, 24, 29, 30, 33, 35, 38, 43, 45, 47, 52,
Jacobian matrix, 370 53, 61, 67–70, 81, 93, 112, 119, 120,
Joule, J., 15, 112 122, 124, 125, 130, 137, 139–141, 143,
147, 151, 155–158, 165, 168–171, 173–
Kalman filter, 412 177, 183, 185–187, 189, 190, 193, 194,
Kelvin 196, 198–200, 204, 205, 209, 211–213,
degrees, 35 217–219, 234, 241, 247, 250, 256, 267,
Kelvin scale, 36 268, 270, 274–276, 280, 282, 287, 291,
kinetic energy, 12, 14, 24–26, 28, 81, 117–119, 292, 294, 309, 311–313, 325, 327, 328,
133, 134, 187, 196, 197, 199–201, 203, 333, 335, 336, 350, 379, 387, 395, 396,
221, 229, 250, 298, 303, 316, 317, 320, 403
323, 326, 327, 340, 341, 350, 392, 395 flow rate, 174, 176, 186, 205, 333, 336, 350
kinetic rate mass conservation principle, 168–170
Arrhenius, 149 material
kinetics derivative, 170–172, 213
chemical, 152 region, 168, 169, 171
surface, 167
Lagrange multiplier, 295–297, 303–305, 307 volume, 167, 169, 172, 179, 180
Lagrange, J., 159, 295 mathematics, 86, 157–159, 166, 222, 369, 370
Lagrangian, 404, 405 matter, 157
Laplace, S., 381 maximum, 17, 152, 153, 287, 289, 290, 297, 344–
lasagna, 109 346, 413, 414
latent heat, 124, 147, 248 Maxwell relations, 15, 357, 358, 362, 365
law of mass action, 149 Maxwell, J., 15, 135, 357
least squares, 408, 409, 418 Mayer’s relation, 128, 139, 368, 379
unweighted, 410 Mayer, J., 15, 112
weighted, 411 mean free path, 12, 31, 32
Legendre transformation, 360, 362, 403, 405, mean molecular speed, 32
407, 408 mechanical energy, 12, 16, 18, 24, 25, 27, 28, 81,
Legendre, A., 361 117, 141, 197, 198, 225, 250, 318, 323,
433
341 nozzle, 196, 197, 327, 341, 352, 396, 397
mechanical equilibrium, 33, 41, 102, 264, 266, converging-diverging, 396
270, 273, 291 subsonic, 396
mechanical equivalent of heat, 15, 112 supersonic, 396
mechanics, 14, 16, 115, 155, 159, 168
continuum, 209, 319 Onnes, H., 16, 123
fluid, 157, 166, 212, 332, 387 Onsager reciprocity, 15
Newtonian, 16, 28, 31, 92, 119 Onsager, L., 15, 16
quantum, 135, 299 optics, 159
solid, 136 Otto cycle, 277, 286, 287, 289
statistical, 15, 135 Otto, N., 286
methane, 73, 219 oven, 110
method of least squares, 408 microwave, 110
Michelson-Morley experiment, 17
minimum, 17, 81, 232, 399 pan, 110
mixing, 294 pandemic, 154
entropy of, 291, 294 Papin, D., 15, 18
mixture, 294 partial derivative, 82, 83, 128, 295, 296, 365,
Amagat, 68 372, 389
Dalton, 68 partial differential equation, 209
mixture entropy, 296 partial pressure, 68
mixture theory, 68, 294 partial volume, 68
modulus of elasticity, 136 partition function, 307
molecular diameter, 32 Pascal, B., 40
molecular mass, 32, 53, 54, 149, 300 path-dependent, 93, 95, 111, 117, 125, 206, 311
molecule, 134 path-independent, 83, 86, 87, 91, 95, 117, 138,
momenta, 211 240, 241
momentum, 31, 196, 318, 352, 379, 387, 395, 404 phase, 33, 331, 343
momentum flux, 352 gas, 331, 343
monatomic gas, 128, 129, 132, 136 solid, 48
phase boundaries, 33
nano-scale heat transfer, 31 physics, 17, 24, 95, 133, 135, 157, 159, 169, 178,
Nernst, W., 16, 242 218, 219, 266, 273, 399, 409
Newcomen, T., 15 Planck, M., 16
Newton’s law of cooling, 110, 145 point of inflection, 372
Newton’s method, 304 Poisson’s ratio, 212
Newton’s second law, 12, 24, 37, 41, 77, 155, polytropic exponent, 96
168, 250, 263, 270, 352 polytropic process, 96–98, 105, 130, 262, 280
Newton, I., 12, 14, 15, 17, 81, 145, 161, 381, 382 Popper, K., 310
Newtonian fluid, 211 position, 12
Newtonian mechanics, 12, 119 potato, 29, 145–154
nitrogen, 48, 73 potential, 404
noble gas, 128, 129 potential energy, 24–26, 28, 81, 117, 118, 120,
non-equilibrium thermodynamics, 15, 34, 144, 134, 141, 179, 187, 192, 196, 197, 199–
309 201, 203, 221, 250, 316, 320, 323, 341,
non-holonomic system, 92 403
non-ideal gas, 53, 97, 251, 371 power, 11, 12, 18, 21, 22, 93, 202–204, 206, 208–
non-linear, 151, 409 210, 232, 233, 237, 286, 329, 331, 333,
algebra, 194, 313 335, 336, 338–340, 343, 346, 347, 349,
ordinary differential equations, 271 354, 356
norm, 409, 418 electric, 208
434
power law curve fitting, 412 quantum effects, 11
Power, H., 50 quantum mechanics, 135
pressure, 18, 30, 31, 33–36, 39–46, 48, 50, 57, 59, quasi-equilibrium, 209
61, 64–66, 68–71, 74, 76–79, 94, 100,
102, 105, 106, 108, 126, 128, 139, 142, R-134a refrigerant, 73, 354
144, 180, 181, 191–193, 205, 207, 250, R-410a refrigerant, 73
255, 264, 266, 267, 270, 272, 273, 276, radiation, 35, 110
277, 289–292, 316, 318, 322–325, 331, ramjet, 341
334, 336, 340, 344, 346, 348, 352, 371, random, 81, 119, 298, 301
383, 394 randomness, 219, 246, 298, 301, 305
absolute, 39 Rankine
critical, 46, 247 degrees, 36
gauge, 39, 40, 324 Rankine cycle, 203–206, 209, 329–333, 338, 339,
partial, 68 343, 352
saturation, 46 reheat, 332
stagnation, 393, 394 Rankine, W., 110, 329
pressure ratio, 350 rarefied gas dynamics, 31
Prigogine, I., 16 ratio, 132, 168, 386, 392
probability, 17, 135, 299, 301 back work, 330, 343
process, 17, 34, 45, 54, 61, 77, 79, 94, 96–100, compression, 289
102, 105, 106, 110, 117, 118, 123, 130, of specific heats, 128, 368, 382
140, 141, 143, 144, 193, 221, 224, 294, Poisson’s, 212
315, 317, 320, 348, 377 pressure, 350
adiabatic, 110, 316, 349 quality, 47
combustion, 331, 342 surface to volume, 147
cooling, 346 temperature, 343–346
cyclic, 225 thermal efficiency, 208
expansion, 101 reaction, 149–154
heat transfer, 108, 149, 377 chemical, 149
heating, 126, 346 dynamics, 149
irreversible, 224, 243, 244 reaction progress, 149
isentropic, 327, 331, 348 reciprocity
isobaric, 44, 97, 124, 137 Onsager, 15
isochoric, 61, 97, 100, 123, 143, 144 Redlich-Kwong equation of state, 63, 142, 379
isothermal, 50, 51, 97, 376 reduced pressure, 64
non-equilibrium, 108 reduced temperature, 64
polytropic, 96–98, 105, 130, 262, 280 refrigeration cycle, 22, 352, 353
quasi-equilibrium, 209 refrigerator, 232, 352, 355
reversible, 224, 240, 242–244, 246, 251, 316 capacity, 354
reversible, adiabatic, 247, 266, 320 Carnot, 231, 355
sublimation, 71 Regnault, H., 52
two-step, 98 reheat, 332–336
property, 33, 34, 49, 208 reversible, 237
extensive, 33, 117, 119, 165, 168, 179 process, 224, 240, 242–244, 246, 251, 316
intensive, 33, 49, 52, 77, 78, 241, 292 reversible work, 227
pump, 14, 18, 157, 180, 198, 199, 203, 205, 208, Reynolds transport theorem, 166, 167, 169, 173,
237, 321, 322, 327, 329–334, 338, 354 182
pure substance, 43 Reynolds, O., 166
Rocket, 21
quality, 47, 49, 66, 67, 70, 71, 74, 76, 102, 122, rocket, 22
143, 283 rotational energy, 133, 197
435
saturated liquid, 44, 46, 66, 71, 74, 122, 208, 252 at constant volume, 125
saturated vapor, 45, 46, 66, 122, 142, 144, 252, specific heat capacity, 125
282 specific internal energy, 122
saturation pressure, 46, 48 specific volume, 39, 41, 45, 46, 53, 66, 68, 69,
saturation temperature, 44–46, 48, 71 71, 120, 268
Savery, T., 15 specific work, 93
scalar, 93, 160, 163, 181, 213, 408 speed of light, 155
field, 87 spring, 76, 77, 79, 102, 403
scalar field, 82, 88 stagnation, 391, 392
Schrödinger, E., 16 flow, 391
second harmonic, 417 point, 391
second law, 314 pressure, 393, 394
second law of thermodynamics, 15, 16, 36, 168, temperature, 392, 393
208, 218–226, 228, 231–233, 237, 239, state, 33, 34, 36, 44, 45, 49, 50, 54, 59, 60,
240, 243–246, 257, 277, 305, 311, 315, 66, 76–79, 84, 92–96, 98–100, 102–105,
316, 319 107, 108, 111, 117, 120, 121, 127, 130,
Carathéodory statement, 222 139, 141–145, 176, 190, 193, 199, 206,
Clausius statement, 220 208, 220, 224, 237, 242, 246–248, 253,
entropy-based statement, 219 254, 267, 268, 282, 283, 287, 289–294,
Kelvin-Planck statement, 220, 239 299, 301, 303, 304, 309, 317, 324, 327,
sensible heat, 124, 249 328, 333, 334, 336, 337, 347, 354, 375,
shaft work, 180, 185, 188, 189, 206 376, 383, 388
Shannon, C., 308 equilibrium, 33, 178, 222, 273, 274
shock wave, 195, 387, 394 reference, 122
SI (Système International), 35 statics, 12
simple compressible substance, 48, 49, 250 statistical mechanics, 15, 135
Simpson’s rule, 101 statistical thermodynamics, 15
single interpolation, 74 statistics, 17
singularity, 396, 413 steady state, 174, 175, 178, 199, 205, 312
slope, 57, 59, 172, 259, 260, 262, 263, 265, 340, steam digester, 15
371–373, 399, 400 steam engine, 15, 18, 21, 93, 218, 230
Snow, C., 218 Boulton and Watt, 21
social distance, 154 steam power cycle, 204, 206
social distancing, 152 steam tables, 66, 142, 193, 202, 205, 206, 246,
social network, 154 317, 336, 375–377, 383
solid, 211, 213, 251, 294 Stefan-Boltzmann constant, 110
incompressible, 258 Stephenson, R., 21
solid mechanics, 136 Stokes’ assumption, 211
Sommerfeld, A., 36 Stokes’ theorem, 87
sonic, 396 stove, 109
sound speed, 192, 369, 379, 381, 382, 387, 392 strain, 136, 211, 212
adiabatic, 379, 380, 382, 383, 387 strain rate, 211
isentropic, 369, 391 stress, 136, 180, 211, 212, 394
isothermal, 381, 382, 384 deviatoric, 211
space, 48, 50, 78, 79, 167, 209, 359, 384, 389 sublimation, 48, 71, 246
space shuttle, 396, 397 subsonic, 317, 396
spacecraft, 195 subsonic nozzle, 396
species, 149 superheated vapor, 46, 49, 71, 79
specific heat, 124–127, 139, 141, 149, 214, 289, supersonic, 396
294, 365 supersonic nozzle, 396
at constant pressure, 126 surface, 167
436
surroundings, 29, 93, 108, 119, 120, 145, 153, thermodynamics, 12, 15–17, 22, 28, 30, 31, 33,
155, 206, 208, 220, 224, 232, 246, 256– 34, 36, 37, 39, 53, 81, 93, 96, 109, 111,
258, 266, 270, 286, 291, 292, 294, 329, 112, 117, 118, 122, 127, 133, 145, 157,
332, 377 168, 179, 181, 187, 208, 209, 212, 213,
system, 12, 17, 29, 30, 33, 34, 49, 81, 93, 94, 99, 217–224, 226, 228, 231, 237, 239, 240,
108, 110–112, 117–121, 130, 147, 150, 242–244, 247, 249, 250, 257, 269, 277,
152, 155, 157, 175, 190, 205, 209, 217, 292, 298, 299, 305, 309–311, 318, 357,
219, 220, 222, 224, 225, 228–231, 294, 361, 369, 376, 381, 387, 399, 406
311, 312, 320, 332, 354, 355, 377 classical, 11–13, 17, 31, 77, 116, 144, 332
boundary, 29, 108, 111, 141, 155 equilibrium, 34, 109, 209
combined, 141, 228 non-classical, 17
holonomic, 92 non-equilibrium, 15, 34, 109, 144, 309, 311
isolated, 29, 155, 156, 219 statistical, 15
non-holonomic, 92 thermostatics, 12
thermodynamic, 29 third law of thermodynamics, 242
Thompson, B. (Count Rumford), 15, 112
tabular equation of state, 65 Thomson, W. (Lord Kelvin), 11, 15, 310
Taylor series, 195, 274, 393, 395 throttling device, 192, 196
temperature, 15, 16, 30, 31, 33–37, 44–46, 48, throttling valve, 353
50, 52, 59–62, 64–67, 71–74, 77, 81, 99, thrust, 340, 341, 350, 352
108, 109, 119, 124–126, 128–137, 139– time, 12, 24–27, 31–35, 77, 81, 93, 94, 109, 144,
142, 144, 145, 147–150, 152–154, 168, 145, 147, 148, 151, 159, 161, 163, 168,
181, 189–191, 193–195, 201, 202, 215, 170, 172, 178–180, 184, 186, 187, 190,
219, 229–231, 233, 234, 237, 242, 245, 191, 207, 209, 217, 219, 244, 271–273,
246, 248, 249, 255–257, 265, 276, 277, 275, 299, 352, 382, 384–386, 389
282, 286–292, 294, 296, 298, 301, 316, time constant, 147, 178, 190, 191
317, 319, 328, 331, 334, 336, 344–347, time-dependency, 144
349–353, 362, 371, 377, 382, 394 Torricelli’s formula, 322
absolute, 35, 37, 52, 60, 129, 168, 219 Torricelli, E., 322
critical, 46, 67, 247 total derivative, 82, 128, 179
saturation, 44–46, 71 total differential, 82, 171
stagnation, 392, 393 total energy, 81, 118, 122, 145, 155, 179, 184,
tensor, 211, 212 187, 250
stress, 212 total enthalpy, 123, 184, 185, 187, 188
theology, 16 Towneley, R., 50
thermal conductivity, 109, 148, 213, 215 transformation, 36
thermal diffusivity, 215 coordinate, 369
thermal efficiency, 208, 226 Legendre, 360, 362, 399
thermal energy, 12, 16, 18, 28, 81, 109, 111, 112, translational kinetic energy, 133
117–119, 139, 148, 150, 199, 205, 206, transport theorem
208, 218, 225, 228, 231, 232, 245, 250, general, 166
257, 298, 317, 318, 341, 377, 392 Reynolds, 166, 167, 169, 182
thermal equation of state, 50, 62, 65, 127, 142, transpose, 86
211 trapezoidal method, 102
thermal equilibrium, 33, 35, 153, 291 triple point, 35, 36, 48, 67, 71, 230
thermal radiation, 110 Truesdell, C., 14, 115
thermochemical calorie, 113 turbine, 157, 180, 197, 198, 200–204, 206–208,
thermochemistry, 149 314, 316, 321, 327, 329–334, 336, 337,
thermodynamic cycle, 203, 208, 220, 239, 277, 340, 341, 343, 348–350
329, 353 steam, 202, 314, 316
thermodynamic system, 29 turbojet, 340, 341, 350
437
two-phase mixture, 33, 46, 47, 49, 66, 69, 70, vibrational energy, 133, 135, 202
74, 76, 99, 100, 122, 143, 242, 251, 331, vis viva, 81
333, 337 viscosity, 212, 225, 387
volume, 29, 30, 33, 34, 39, 41, 44, 45, 48, 50,
universal gas constant, 52, 132 55, 57, 60–62, 67–70, 95, 100, 106, 122,
universe, 29, 245, 246, 256–258, 270, 291, 293, 142, 145, 147, 159–168, 170, 173, 184,
309, 310, 377 185, 189, 234, 250, 266, 291, 354, 371
isolated, 309 critical, 373
University of Notre Dame, 22, 23, 113, 115, 205, flow rate, 354
207, 209, 210, 279, 355, 356, 364, 397 partial, 68
power plant, 23, 205, 355 specific, 39, 41, 45, 46, 53, 66, 68, 69, 71,
120, 268
vacuum, 108 von Guericke, O., 14
vacuum pump, 14
valve, 188, 192, 193, 313 water, 14, 33, 35, 36, 43–45, 47–49, 66, 69–72,
expansion, 354 74–78, 80, 99–101, 112, 122, 140–144,
van der Waals equation of state, 62 147, 148, 175, 176, 204, 205, 208, 209,
van der Waals gas, 141, 371–379 217–219, 230, 234, 246, 248, 249, 252,
van der Waals, J., 16, 63 260, 266–268, 282, 283, 286, 291, 323–
van’t Hoff, J., 16 326, 329, 333, 336, 338, 355, 372, 373,
vapor, 45 375, 385, 386
compression, 22 Watt, J., 15, 18, 125
saturated, 45, 46, 66, 122, 142, 144, 252, wave equation, 390, 391
282 wavelength, 417
superheated, 46, 71, 79 welding, 134
water, 43 William of Occam, 16
vapor compression cycle, 352 Wilson, K., 16
vapor dome, 62, 63, 65, 78, 79, 129, 247, 248, work, 12, 15, 22, 29, 39, 49, 50, 82, 90, 92–99,
252, 371, 373 102, 105, 107, 108, 110, 112, 115, 117,
vapor pressure curve, 46 118, 121, 122, 126, 130, 139, 141, 143,
variables 150, 155, 166, 168, 179–181, 183, 185,
canonical, 362, 363 187, 188, 192, 197, 200–202, 205, 207,
conjugate, 362, 363 208, 213, 218–222, 225, 228, 230, 231,
vector, 86, 87, 91, 93, 109, 156, 158, 160, 162, 233, 237, 239, 244, 246, 247, 250, 260,
180, 181, 186, 211, 212, 409, 411 269, 271, 276, 279–281, 285, 289, 292,
column, 86, 90 314, 327–331, 340, 343, 345, 346, 348,
distance, 93 350–353, 361, 362, 376
field, 87, 89, 90, 170 chemical, 49
force, 12, 93 electric, 49, 90, 108
gradient, 88 irreversible, 227
heat flux, 109 magnetic, 49
position, 12 reversible, 227
row, 86, 90 shaft, 180, 185, 188, 189, 206
velocity, 156, 166, 170 sign convention, 93
velocity, 14, 25, 26, 30, 120, 121, 155, 156, 162, surface force, 108
166, 169–172, 180, 186, 192, 196, 211, tensile force, 108
212, 270, 272, 318, 322, 324–326, 340,
Young, T., 12
350, 352, 385, 386, 388, 391, 392, 394,
396, 404 zeroth law of thermodynamics, 36
field, 170, 171
gradient, 172
438

Notes

Uploaded by

Copyright:

Available Formats

Notes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Notes

Uploaded by

Copyright:

Available Formats

LECTURE NOTES ON

Department of Aerospace and Mechanical Engineering

CC BY-NC-ND. 03 July 2024, J. M. Powers.

2 Properties of a pure substance 43

2.4.4.6 Tables for other materials . . . . . . . . . . . . . . . . . . . 73

3 The first law of thermodynamics 81

4 First law analysis for a control volume 157

CC BY-NC-ND. 03 July 2024, J. M. Powers.

4.1.2.2.2 Work term . . . . . . . . . . . . . . . . . . . . . . 180

5 The second law of thermodynamics 217

CC BY-NC-ND. 03 July 2024, J. M. Powers.

6.5 Entropy for an incompressible solid or liquid . . . . . . . . . . . . . . . . . . 258

7 Second law analysis for a control volume 311

9 Mathematical foundations 357

CC BY-NC-ND. 03 July 2024, J. M. Powers.

CC BY-NC-ND. 03 July 2024, J. M. Powers.

CC BY-NC-ND. 03 July 2024, J. M. Powers.

CC BY-NC-ND. 03 July 2024, J. M. Powers.

Read BS, Chapter 1

1.1 Some semantics

The modifier “classical” is used to connote a description in which quantum mechanical

CC BY-NC-ND. 03 July 2024, J. M. Powers.

CC BY-NC-ND. 03 July 2024, J. M. Powers.

Figure 1.4: Gabrielle Émilie Le Tonnelier de Breteuil, marquise du Châtelet (1706-1749),

1.2 Historical milestones

• 1593: Galileo develops a water thermometer.

CC BY-NC-ND. 03 July 2024, J. M. Powers.

• 1687: Isaac Newton publishes PhilosophiæNaturalis Principia Mathematica.

• 1698: Thomas Savery patents an early steam engine.

• 1710: Thomas Newcomen creates a more practical steam engine.

• 1760s: Joseph Black develops calorimetry.

• 1780s: James Watt improves the steam engine.

• 1798: Benjamin Thompson (Count Rumford) considers the mechanical equivalent of

• 1824: Nicolas Lèonard Sadi Carnot discusses idealized heat engines.

• 1840s: Julius Robert von Mayer relates heat and work.

• 1840s: James Prescott Joule relates heat and work.

• 1847: Hermann von Helmholtz publishes his theory of energy conservation.

• 1848: William Thomson (Lord Kelvin) postulates an absolute zero of temperature.

• 1865: Clausius introduces the concept of entropy.

• 1871: James Clerk Maxwell develops the Maxwell relations.

• 1870s: Josiah Willard Gibbs further formalizes mathematical thermodynamics.

• 1870s: Maxwell and Ludwig Boltzmann develop statistical thermodynamics.

• 1909: Constantin Carathéodory publishes a rigorous axiomatic formulation of the laws

• 1931: Lars Onsager advances non-equilibrium thermodynamics with his reciprocity

CC BY-NC-ND. 03 July 2024, J. M. Powers.

• Jacobus Henricus van’t Hoff (1901),

• Johannes van der Waals (1910),

• Heike Kamerlingh Onnes (1913),

• Max Planck (1918),

• Walther Nernst (1920),

• Albert Einstein (1921),

• Erwin Schrödinger (1933),

• Enrico Fermi (1938),

• Percy Bridgman (1946),

• William Giauque (1949),

• Lars Onsager (1968),

• Ilya Prigogine (1977), and

• Kenneth Wilson (1982).

1.3 Philosophy of science note

CC BY-NC-ND. 03 July 2024, J. M. Powers.