Fluid Mechanics
Fluid Mechanics
Continuous Matter
Second Edition
B Lautrup
The Niels Bohr International Academy, The Niels Bohr Institute
Copenhagen, Denmark
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2011 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and
information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission
to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or
retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact
the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides
licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation
without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents
Preface xi
1 Continuous matter 1
1.1 Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 The continuum approximation . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Newtonian mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Reference frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5 Cartesian coordinate systems . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6 Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
I Fluids at rest 19
2 Pressure 21
2.1 What is pressure? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2 The pressure field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 Hydrostatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4 Equation of state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 Bulk modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.6 Application: Earth’s homentropic atmosphere . . . . . . . . . . . . . . . . . 34
2.7 Application: The Sun’s convective envelope . . . . . . . . . . . . . . . . . . 38
4 Hydrostatic shapes 57
4.1 Fluid interfaces in hydrostatic equilibrium . . . . . . . . . . . . . . . . . . . 57
4.2 The centrifugal force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.3 The figure of Earth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.4 The Earth, the Moon, and the tides . . . . . . . . . . . . . . . . . . . . . . . 62
4.5 Application: The tides of Io . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5 Surface tension 69
5.1 Basic physics of surface tension . . . . . . . . . . . . . . . . . . . . . . . . 69
5.2 Soap bubbles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.3 Pressure discontinuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.4 The Rayleigh–Plateau instability . . . . . . . . . . . . . . . . . . . . . . . . 78
5.5 Contact angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
vi PHYSICS OF CONTINUOUS MATTER
II Solids at rest 95
6 Stress 97
6.1 Friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.2 Stress fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.3 The nine components of stress . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.4 Mechanical equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.5 Asymmetric stress tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7 Strain 109
7.1 Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
7.2 The displacement field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.3 Geometrical meaning of the strain tensor . . . . . . . . . . . . . . . . . . . 116
7.4 Work and energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
7.5 Large deformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
15 Viscosity 243
15.1 Shear viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
15.2 Velocity-driven planar flow . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
15.3 Dynamics of incompressible Newtonian fluids . . . . . . . . . . . . . . . . . 250
15.4 Classification of flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
15.5 Dynamics of compressible Newtonian fluids . . . . . . . . . . . . . . . . . . 255
15.6 Application: Viscous attenuation of sound . . . . . . . . . . . . . . . . . . . 257
22 Energy 371
22.1 First Law of Thermodynamics . . . . . . . . . . . . . . . . . . . . . . . . . 371
22.2 Incompressible fluid at rest . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
22.3 Incompressible fluid in motion . . . . . . . . . . . . . . . . . . . . . . . . . 380
22.4 General homogeneous isotropic fluids . . . . . . . . . . . . . . . . . . . . . 384
23 Entropy 393
23.1 Entropy in classical thermodynamics . . . . . . . . . . . . . . . . . . . . . . 393
23.2 Entropy balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
23.3 Fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
30 Convection 547
30.1 Heat-driven convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
30.2 Convective instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
30.3 Linear stability analysis of convection . . . . . . . . . . . . . . . . . . . . . 555
30.4 Application: Rayleigh–Bénard convection . . . . . . . . . . . . . . . . . . . 557
31 Turbulence 565
31.1 Scaling in fully developed turbulence . . . . . . . . . . . . . . . . . . . . . 565
31.2 Mean flow and fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . 571
31.3 Universal inner layer near a smooth wall . . . . . . . . . . . . . . . . . . . . 574
31.4 The outer layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 579
31.5 Application: Turbulent channel flow . . . . . . . . . . . . . . . . . . . . . . 581
31.6 Application: Turbulent pipe flow . . . . . . . . . . . . . . . . . . . . . . . . 582
31.7 Turbulence modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
VI Appendices 587
A Newtonian mechanics 589
A.1 Dynamic equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
A.2 Force and momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 590
A.3 Moment of force and angular momentum . . . . . . . . . . . . . . . . . . . 591
x PHYSICS OF CONTINUOUS MATTER
References 665
Index 673
Preface
This second edition of Physics of Continuous Matter is, as the first, primarily an introduction
to the basic ideas of continuum physics and their application to the wealth of macroscopic
phenomena. The equations of continuum mechanics were developed between 1750 and 1850,
and are so simple that they can be derived from Newtonian particle mechanics in a couple of
pages. Unfortunately, they are generally unsolvable by analytic methods, although they have
during the last half century yielded to direct numerical computation. Over the years many
approximate methods have been developed that give insight into the rich physics hidden in
the basic equations, and these methods take, as earlier, center stage in this book.
In this edition the main “story line” has been kept largely intact. Everywhere the physics
arguments and their mathematical presentation have been rethought and to some extent rewrit-
ten to improve clarity and consistency. Some structural changes could not be avoided. An
originally introductory chapter on Cartesian vector and tensor algebra has been moved into an
appendix, while two chapters on gravity and stars have been dropped. The original chapter on
ideal flow has been split into two chapters on, respectively, incompressible and compressible
inviscid fluids. Conversely, the former chapter on lubrication is now joined with the chapter
on creeping flow. A new chapter has been added on elasticity of slender rods, as well as two
new chapters on, respectively, energy and entropy. After a single introductory chapter, this
second edition proceeds immediately to the basics of continuum physics.
The book is aimed squarely at third-year undergraduate or older students. Although I
originally taught second-year students from the first edition, much space and time were wasted
on mathematical concepts that the students naturally would encounter in a second-year course
on electromagnetism. Now these concepts are merely recapitulated in appendices that may
be sampled as the need arises. The necessary mathematical tools are developed along with
the physics on a “need-to-know” basis in order to avoid lengthy and boring mathematical
preliminaries seemingly without purpose. Mathematical rigor is, as before, only used when it
is necessary for a clear understanding of the physics.
The important thing to learn from this book is how to reason about physics, both qualita-
tively and—especially—quantitatively. Numerical simulations may be fine for obtaining solu-
tions to practical problems, but are of very little aid in obtaining real understanding. Physicists
must learn to think in terms of fundamental principles and generic methods. Solving one prob-
lem after another of a similar kind seems unnecessary and wasteful. This does not mean that
the physicist should not be able to reach a practical result through calculation, but the physical
principles behind equations and the conditions underlying approximations must never be lost
from sight.
A structured text is, in my opinion, an important prerequisite for learning new material.
Beyond parts, chapters, sections, subsections, and paragraphs, further structuring has been
introduced in this edition by marking certain section and subsection headings with “Appli-
cation” or “Case”. These categories may be seen as “worked examples” and can mostly be
included or left out according to the reader’s preferences. Headers marked with an asterisk
indicate that the subject matter is harder or falls outside the main line of presentation, and
may be left out in a first reading without breaking the continuity of the study. The book is
xii PHYSICS OF CONTINUOUS MATTER
divided into six parts: Fluids at rest, Solids at rest, Fluids in motion, Balance and Conserva-
tion, Selected topics, and Appendices. The part headings now include a short description of
the chapters contained in the part with some comments on what a minimal curriculum could
contain.
Each chapter ends with a collection of problems that has been somewhat expanded in this
edition. Most problems are meant to illustrate and further develop the physical and mathemat-
ical concepts introduced in the chapter text. Problems marked with an asterisk are generally
harder than the unmarked ones. Answers to all odd-numbered problems are found in the back
of the book. On request, a complete collection of solutions to all problems is available to
course teachers, although my experience shows that such collections rapidly tend to leak into
the hands of the students. Secrets are, as demonstrated many times, hard to keep in the age
of the Internet. Anyway, students are strongly urged to try their hands on the problems before
turning to the answers.
Illustrations have, as before, been included everywhere. Not only are there now more
margin drawings to aid the understanding of the text, but also more photographs to lighten the
presentation. Images resulting from simulations have been redone and improved. As in the
first edition, micro-biographies of the major players in the historic development of continuum
physics have been provided in the margin, and some have now been illustrated with portraits.
The reader should again (and again) be warned that the history of continuum physics is much
more complicated than can be learned here. Nevertheless, I believe that it is important for
the understanding of this subject at least to be able to place the major contributors properly in
time, space, and physics context.
The field of continuum physics is so highly developed that it cannot be given justice in
a single textbook. Nearly every chapter in this book represents a separate subfield, each
covered by a number of specialized textbooks and monographs, some of which are referred
to in the list of references. There are numerous, more advanced topics that could—or perhaps
should—have earned a place in this book, for example viscoelastic, anisotropic, and artificial
materials; elasticity of membranes; non-linear elasticity; nonlinear surface waves; magneto-
and plasma-hydrodynamics; turbulence modeling; finite volume numerics; and so on. What
is and is not included reflects the intended level, my personal predilections and, of course, the
limitations imposed by the size of the book1 . But I hope that students of this book will have
acquired the basic tools necessary for entering into topics not covered here.
My colleagues through many years John Renner Hansen, Poul Olesen, Poul Henrik Dam-
gaard, and Mogens Høgh Jensen are thanked for their kind and generous support. Besides the
persons mentioned in the first edition, special thanks go to Anders Andersen, Hassan Aref,
Tomas Bohr, Predrag Cvitanovic, Joachim Mathiesen, and Niels Kjær Nielsen, and to all the
students around the world who have sweated over it. Finally, I thank my wife Birthe Østerlund
for her unwavering support during the ardors of the preparation of the second edition over the
past two years.
The book is written for adults with a serious intention to learn physics. I have selected for
readers what I think are the interesting topics in continuum physics, and presented these as
pedagogically as I can without trying to cover everything. I sincerely hope that my own joy
in understanding and explaining the physics shines through everywhere.
Benny Lautrup
Copenhagen, Denmark
1 Extra material may be found at the book’s homepage, which also contains the programs used for working out the
H2 H2 O
C O2 ! Figure 1.1. How continuous matter really looks at the atomic scale. Noise-filtered image of freshly
H2 H2 O
cleaved Mica obtained by atomic force microscopy, approximately 225 Angstrom on a side. This gran-
ularity of matter is ignored in continuum physics. (Source: Mark J. Waner, PhD dissertation, Michigan
State University, 1998. With permission.)
The quantitative meaning of a
chemical formula. The boxes
represent fixed amounts, for
example moles. 1.1 Molecules
Chemical reactions such as 2H2 C O2 ! 2H2 O are characterized by simple integer coeffi-
cients. Two measures of hydrogen plus one measure of oxygen yield two measures of water
without anything left over of the original ingredients. What are these measures? For gases at
the same temperature and pressure, a measure is simply a fixed volume, for example a liter, so
that two liters of hydrogen plus one liter of oxygen yield two liters of water vapor, assuming
that the water vapor without condensing can be brought to the same temperature and pressure
as the gases had before the reaction. In 1811, Count Avogadro of Italy proposed that the sim-
ple integer coefficients in chemical reactions between gases could be explained by the rule
that equal volumes of gases contain equal numbers of molecules (at the same temperature and
pressure).
The various measures do not weigh the same. A liter of oxygen is roughly 16 times heavier
Lorenzo Romano Amadeo Car-
lo Avogadro (1776–1856). Ital- than a liter of hydrogen at the same temperature and pressure. The mass of any amount of
ian philosopher, lawyer, chemist, water vapor must—of course—be the sum of the masses of its ingredients, hydrogen and
and physicist. Count of Quaregna oxygen. The reaction formula tells us that two liters of water vapor weigh roughly .2 1/ C
and Cerratto. Formulated that .1 16/ D 18 times a liter of hydrogen. Such considerations led to the introduction of the
equal volumes of gas contain concept of relative molecular mass (or weight) in the ratio 1:16:9 for molecular hydrogen,
equal numbers of molecules. molecular oxygen, and water. Today, most people would prefer to write these proportions
Also argued that simple gases as 2:32:18, reflecting the familiar molecular masses of H2 , O2 , and H2 O, respectively. In
consist of diatomic molecules. practice, relative molecular masses deviate slightly from integer values, but for the sake of
(Source: Wikimedia Commons.) argument we shall disregard that here.
1. CONTINUOUS MATTER 3
Unit of mass: The definition of Avogadro’s number depends on the definition of the unit of
mass, the kilogram, which is (still) defined by a prototype from 1889 stored by the International
Bureau of Weights and Measures near Paris, France. Copies of this prototype and balances for
weighing them can be made to a precision of one part in 109 . Maybe already in 2011 a new
definition of the kilogram will replace this ancient one [MMQ&06], for example by defining the
kilogram to be the total mass of an exact number of 12 C atoms. Avogadro’s number will then also
become an exact number without error.
where D M=V is the mass density. This molecular separation length sets the scale at which
the molecular granularity of matter dominates the physics, and any conceivable continuum
description of bulk matter must utterly fail.
For liquids and solids where the molecules touch each other, this length is roughly the size
of a molecule. For solid iron we get Lmol 0:23 nm, and for liquid water Lmol 0:31 nm.
Since by Avogadro’s hypothesis equal gas volumes contain an equal number of molecules, the
molecular separation length for any (ideal) gas at normal temperature and pressure (p D 1 atm
and T D 20ı C) becomes Lmol 3:4 nm. There is a lot of vacuum in a volume of gas, in fact
about 1000 times the true volume of the molecules at normal temperature and pressure.
4 PHYSICS OF CONTINUOUS MATTER
1
* Mixtures : The above expression for Lmol may also be used for a mixture of pure substances,
provided Mmol is taken to be a suitable average over the molar masses Mimol of the i-th pure
Pof mass M containing
component (consisting of only one kind of molecules). For a mixture sample
the mass Mi of each component, the total mass becomes the sum M D i Mi . The number of
molesPof the i-th component is ni D Mi =Mimol and the total number of moles in the sample is
n D i ni . Characterizing the composition of the mixture by the molar fraction Xi D ni =n of
each component, the average molar mass, Mmol D M=n, becomes
X
Mmol D Xi Mimol ; (1.2)
i
P
where i Xi D 1. If we instead describe the composition by the mass fraction Yi D Mi =M D
Xi Mimol =Mmol of each component, and use that i Xi D 1, the average molar mass is determined
P
by the reciprocal sum,
1 X Yi
D ; (1.3)
Mmol
i
Mimol
P
where i Yi D 1.
Dry air is a molar mixture of 78.08% nitrogen, 20.95% oxygen and 0.93% argon with an
average molar mass of Mmol D 28:95 g mol 1 . By mass the mixture is 75.56% nitrogen, 23.15%
oxygen, and 1.29% argon, and has of course the same average molar mass.
Solids: In solid matter the binding is so strong that thermal motion cannot overcome it. The
molecules remain bound to each other by largely elastic forces, and constantly undergo small-
amplitude thermal motion around their equilibrium positions. If increasing external forces are
applied, solids will begin to deform elastically, until they eventually become plastic or even
fracture. A solid body retains its shape independently of the shape of a container large enough
to hold it, apart from small deformations, for example due to gravity.
Liquids: In liquid matter the binding is weaker than in solid matter, although it is still hard
for a molecule on its own to leave the company of the others through an open liquid surface.
The molecules stay in contact but are not locked to their neighbors. Molecular conglomerates
may form and stay loosely connected for a while, as for example chains of water molecules.
1 The asterisk indicates that this part of the text can be skipped in a first reading.
1. CONTINUOUS MATTER 5
V
1.0
0.5 gas
r
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
fluid
-0.5
solid
-1.0
Figure 1.2. Sketch of the intermolecular potential energy V .r/ between two roughly spherical neutral
molecules as a function of the distance r between their centers. It is attractive at moderate range and
strongly repulsive at close distance. The equilibrium between attraction and repulsion is found at the
minimum of the potential. Here the units are arbitrarily chosen such that the equilibrium distance be-
comes r D 1 and the minimum potential V .1/ D 1. The horizontal dashed lines suggest the thermal
energy levels for solid, liquid, and gaseous matter.
Under the influence of external forces, for example gravity, a liquid will undergo bulk motion,
called flow, a process that may be viewed as a kind of continual fracturing. A liquid will not
expand to fill an empty container completely, but will under the influence of external forces
eventually adapt to its shape wherever it touches it.
Gases: In gaseous matter the molecules are bound so weakly that the thermal motion easily
overcomes it, and they essentially move around freely between collisions. A gas will always
expand to fill a closed empty container completely. Under the influence of external forces, for
example a piston pushed into the container, a gas will quickly flow to adapt to the changed
container shape.
Granular matter
Molecules and atoms represent without discussion the basic granularity of all macroscopic
matter, but that does not mean that all macroscopic matter can be viewed as continuous.
Barley grain, quartz sand, living beings, buildings, and numerous materials used in industry
are examples of matter having a discrete, granular substructure that influences the material
properties even at considerably larger length scales. The main difference between granular
matter and molecular matter lies in the friction that exists between the granular elements.
Although grain may flow like water in great quantities, the friction between the grains can
make them stick to each other, forming a kind of solid that may block narrow passages. A
heap of grain does not flatten out like a puddle of water, and neither do the wind-blown
ripples of dry sand in the desert. You may also build a castle of wet sand, the main part of
which remains standing even when dried out.
In this book we shall not consider granular matter as such, although at sufficiently large
length scales, granular matter may in many respects behave like continuous matter, whether
the grains are tiny as the quartz crystals in sand or enormous as the galaxies in the universe.
6 PHYSICS OF CONTINUOUS MATTER
Density fluctuations
.......................
.........
...... ...... Consider a fixed small volume V of a pure gas with molecules of mass m. If at a given time t
..
...... .....
....
.. .
. ... the number of molecules in this volume is N , the mass density at this time becomes
A...
..... A .....
*
.
.... A .. Nm
... .-.. D : (1.4)
.... A ..
.... .... V
..... U A ........
....... H..... Due to rapid random motion of the gas molecules, the number N will be different at a later
..............H
.................Y
H
time t C t . Provided the time interval t is much larger than the time interval between
molecular collisions, the molecules in the volume V will at the later time be an essentially
In a gas the molecules move random sample of molecules taken from a much larger region around V . The probability that
rapidly in and out of a small any particular molecule from this larger region ends up in the volume V will be tiny. From
volume with typical velocities general statistical considerations it follows thatp the root-mean-square size of the fluctuations
comparable to the speed of in the number of molecules is given by N x D N (see Figure 1.3, and Problem 1.1). Since
sound. the density is linear in N , the relative fluctuation in density becomes
x
x
N 1
D Dp : (1.5)
N N
In classical macroscopic thermodynamics where typically N NA , the relative fluctuation
becomes of magnitude 10 12 and can safely be ignored.
In continuum physics this is not so. If we want the relative density fluctuation to be smaller
x
than a given value = . , we must require N & 2 . The smallest acceptable number of
molecules, , occupies a volume 2 L3mol , where Lmol is the molecular separation length
2
2=3
Lmicro D Lmol : (1.6)
Thus, to secure a relative precision D 10 3 for the density, the microscopic cell should
contain 106 molecules and have side length Lmicro D 100Lmol . For an ideal gas under normal
conditions we find Lmicro 0:34 m.
The micro scale diverges for ! 0, substantiating the claim that it is impossible to
maintain a continuum description to arbitrarily high precision.
1. CONTINUOUS MATTER 7
Ρ
2.0
1.5
1.0
0.5
0.0 V
0 100 200 300 400 500
Macroscopic smoothness
As a criterion for a smooth continuum description we demand that the relative variation in
density between neighboring cells should be less than the desired precision . The change
in density between the centers of neighboring cells along some direction x is of magnitude
Lmicro j@=@xj. Demanding the relative variation to be smaller than the precision,
= . , we obtain a constraint on the magnitude of the density derivative,
ˇ ˇ
ˇ @ ˇ
ˇ ˇ. ; (1.7)
ˇ @x ˇ L
macro
where
1
Lmacro D Lmicro : (1.8)
As long as the above condition is fulfilled, the density may be considered to vary smoothly,
because the density changes over the micro-scale are imperceptible. Any significant change
in density must take place over distances larger than Lmacro . With D 10 3 we find Lmacro
1000Lmicro 0:34 mm for an ideal gas under normal conditions.
The thickness of interfaces between macroscopic bodies is typically on the order of Lmol
and thus much smaller than Lmacro . Consequently, these regions of space fall outside the
smooth continuum description. In continuum physics, interfaces appear instead as surface
discontinuities in the otherwise smooth macroscopic description of matter.
Velocity fluctuations
Everyday gas speeds are small compared to the molecular velocities—unless one is traveling
by jet aircraft or cracking a whip. What we normally mean by wind is the bulk drift of
air, not the rapid molecular motions. So even if the individual molecules move very fast in
random directions, the center of mass of a collection of N molecules in a small volume V will
normally move with much slower velocity, which for large N approximates the drift speed v.
8 PHYSICS OF CONTINUOUS MATTER
Gases consist mostly of vacuum, and apart from an overall drift, the individual molecules
move in all possible directions with average root-mean-square speed (see page 24),
s
3Rmol T
vmol D ; (1.9)
Mmol
where Rmol D 8:31447 J K 1 mol 1 is the universal molar gas constant and T the absolute
temperature. For air at normal temperature one finds vmol 500 m s 1 .
Under very general p assumptions the root-mean-square fluctuation in the center-of-mass
x D vmol = N (see Problem 1.3). Since the fluctuation scale is set by the molec-
speed is v
ular velocity, it takes much larger numbers of molecules to be able ignore the fluctuations in
everyday gas velocities. To maintain a relative precision in the velocity fluctuations we must
x
require v=v . , implying that the linear size of a gas volume must be larger than
v 2=3
mol
L0micro D Lmicro : (1.10)
v
The velocity fluctuations of a gentle steady wind, say v 0:5 m s 1 , can be ignored with
precision 10 3 for volumes of linear size larger than L0micro D 100Lmicro 34 m.
The smoothness scale should similarly be L0macro D 1 L0micro , which in this case becomes
L0macro 34 mm. A hurricane wind, v 50 m s 1 , only requires volumes of linear size
L0micro D 4:6Lmicro 1:6 m to yield the desired precision, but in this case fluctuations due
to turbulence will anyway completely swamp the molecular fluctuations.
Since =Lmol 2=3 , the mean free path will in sufficiently dilute gases always become
larger than the micro scale and should instead be used to define the smallest linear scale for
the continuum description.
Air consists mainly of nitrogen and argon molecules with an average molar mass Mmol
A sphere of diameter d will col-
29 g mol 1 and average diameter d 0:37 nm [2]. At normal temperature and pressure the
lide with any other sphere of the
same diameter with its center
mean free path becomes 65 nm, which is five times smaller than the microscopic length
inside a cylinder of diameter 2d . scale for the density, Lmicro 340 nm (for D 10 3 ). The mean collision time may be
estimated as =vmol , and becomes for air 0:13 ns.
1. CONTINUOUS MATTER 9
Figure 1.4. Newton’s own copy of Principia with handwritten corrections for the second edition.
(Source: Andrew Dunn (2004). Wikimedia Commons.)
1. There exist (inertial) reference frames in which a particle moves with constant velocity
along a straight line when it is not acted upon by any forces.
2. The mass times the acceleration of a particle equals the sum of all forces acting on it.
Sir Isaac Newton (1643–1727).
3. If a particle acts on another with a certain force, the other particle acts back on the first English physicist and mathemati-
with an equal and opposite force. cian. Founded classical me-
chanics on three famous laws in
In Appendix A you will find a little refresher course in Newtonian mechanics. his books Philosophiae Naturalis
Newton’s Second Law is the fundamental equation of motion. Mathematically, this law Principia Mathematica (1687).
is expressed as a second-order differential equation in time. Since the force acting on any Newton developed calculus to
given particle can depend on the positions and velocities of the particle itself and of other solve the equations of motion,
and formulated theories of optics
particles as well as on external parameters, the dynamics of a collection of particles becomes
and of chemistry. He still stands
a web of coupled ordinary second-order differential equations in time. Even if macroscopic as perhaps the greatest scientific
bodies are huge collections of atoms and molecules, it is completely out of the question to genius of all time. (Source: Wiki-
solve the resulting web of differential equations. In addition, there is the problem that molec- media Commons.)
ular interactions are quantum mechanical in nature, so that Newtonian mechanics does not
apply at the atomic level. This knowledge is, however, relatively new and has as mentioned
earlier some difficulty in making itself apparent at the macroscopic level. So although quan-
tum mechanics definitely rules the world of atoms, its special character is rarely amplified
to macroscopic proportions, except in low-temperature phenomena such as superconductivity
and superfluidity.
10 PHYSICS OF CONTINUOUS MATTER
Laws of balance
In Newtonian particle mechanics, a “body” is taken to be a fixed collection of point particles
with unchangeable masses, each obeying the Second Law. For any body one may define
various global mechanical quantities, which like the total mass are calculated as sums over
contributions from each and every particle in the body. Some of the global quantities are
kinematic: momentum, angular momentum, and kinetic energy. Others are dynamic: force,
moment of force, and power (rate of work of all forces).
Newton’s Second Law for particles leads to three simple laws of balance—often called
conservation laws—relating the kinematic and dynamic quantities. They are (in addition to
the trivial statement that the mass of a Newtonian body never changes):
Even if these laws are insufficient to determine the dynamics of a multi-particle body, they
represent seven individual constraints (two vectors and one scalar) on the motion of any col-
lection of point particles, regardless of how complex it is.
In continuum mechanics any volume of matter may be considered to be a body. As the
dynamics unfolds, matter is allowed to be exchanged between the environment and the body,
but we shall see that the laws of balance can be directly carried over to continuum mechanics
when exchange of matter between a body and its environment is properly taken into account.
Combined with simplifying assumptions about the macroscopic behavior, for example sym-
metry, the laws of balance for continuous matter also turn out to be quite useful for obtaining
quick solutions to a variety of problems.
Material particles
In continuum physics we generally speak about material particles as the elementary con-
stituents of continuous matter, obeying Newton’s equations. A material particle will always
contain a large number of molecules but may in the continuum description be thought of as in-
finitesimal or point-like. From the preceding analysis we know that material particles cannot
be truly infinitesimal, but represent the smallest bodies that may consistently be considered
part of the continuum description within the required precision. Continuum physics does not
“on its own” go below the level of the material particles. Even if the mass density may be
determined by adding together the masses of all the molecules in a material particle and di-
viding the sum by the volume of the particle, this procedure falls, strictly speaking, outside
continuum physics.
Although we normally think of material particles as being identical in different types of
matter, it is sometimes necessary to go beyond the continuum approximation and look at their
differences. In solids, we may with some reservation think of solid particles as containing a
fixed collection of molecules, whereas in liquids and especially in gases we should not forget
that the molecules making up a fluid particle at a given instant will shortly be replaced by
other molecules. If the molecular composition of the matter in the environment of a material
particle has a slow spatial variation, this incessant molecular game of “musical chairs” may
slowly change the composition of the material inside the particle. Such diffusion processes
driven by spatial variations in material properties lie at the very root of fluid mechanics. Even
a spatial velocity variation will drive momentum diffusion, causing internal (viscous) friction
in the fluid.
1. CONTINUOUS MATTER 11
Time
Time is the number you read on your clock. There is no better definition. Clocks are phys-
ical objects that may be shared, compared, copied, and synchronized to create an objective
meaning of time. Most clocks, whether they are grandfather clocks with a swinging pendulum
or oscillating quartz crystals, are based on periodic physical systems that return to the same
state again and again. Time intervals are simply measured by counting periods. There are
also aperiodic clocks, for example hour glasses, and clocks based on radioactive elements. It
is especially the latter that allow time to be measured on geological time scales. On extreme
cosmological time scales the very concept of time becomes increasingly more theory laden;
see for example [RZ09].
Like all macroscopic physical systems, clocks are subject to small fluctuations in the way
they run. Some clocks are considered better than others because they keep time more stably
with respect to copies of themselves as well as with clocks built on other principles. Grand-
father clocks are much less stable than mechanical maritime chronometers that in turn are
less stable than modern quartz clocks. The international frame of reference for time is always
based on the most stable clocks currently available.
Unit of time: The unit of time, the second, was formerly defined as 1/86,400 of a mean solar
day. But the Earth’s rotation is not that stable, and since 1966 the second has been defined by
international agreement as the duration of 9,192,631,770 oscillations of the microwave radiation
absorbed in a certain hyperfine transition in cesium-133, a metal that can be found anywhere on
Earth [1]. A beam of cesium-133 atoms is used to stabilize a quartz oscillator at the right frequency
by a resonance method, so what we call an atomic clock is really an atomically stabilized quartz
clock. The intrinsic precision in this time standard has been continually improving and is now
about 4 10 16 corresponding to about 1 second in 80 million years [LHJ07].
In the extreme mathematical limit, time may be taken to be a real number, and in Newto-
nian physics its value is assumed to be universally knowable.
12 PHYSICS OF CONTINUOUS MATTER
Figure 1.5. Independently of how different their reference frames, two observers who agree on the
unique reality of any point in space can in principle determine the coordinate transformation relating
them by listing their respective coordinates for each and every point in space. (Source: Fragment of
“The Creation of Adam”, by Michelangelo Buonarroti (1511). Wikimedia Commons.)
Space
It is a mysterious and so far unexplained fact that physical space has three dimensions, which
ra means that it takes exactly three real numbers to locate a point in space. These numbers
are called the coordinates of the point, and the reference frame for coordinates is called the
rb
rx coordinate system. It must contain all the operational specifications for locating a point given
the coordinates, and conversely obtaining the coordinates given the location. In this way we
have relegated all philosophical questions regarding the real nature of points and of space to
Points may be visualized as dots the operational procedures contained in the reference frame.
on a piece of paper. Each point is
labeled by its position in the cho- Earth coordinates: On Earth everybody navigates by means of a geographical coordinate
sen coordinate system (not visu- system agreed upon by international convention, in which a point is characterized by latitude ı,
alized here). longitude , and elevation h. Latitude and longitude are angles fixed by the Earth’s rotation axis
and the position of the former Royal Observatory in Greenwich near London (UK). Elevation is
defined as the signed height above the average sea level. The modern Global Positioning System
(GPS) uses instead “fixed points in the sky” in the form of at least 24 satellites, and the geographical
coordinates of any point on Earth as well as the absolute time in this point is determined from radio
signals received from four or more of these satellites.
The triplet of coordinates that locates a point in a particular coordinate system is called
ra $ a 0 its position in that coordinate system, and usually marked with a single symbol printed in
boldface2 , for example x D .x1 ; x2 ; x3 /. Given the position x D .x1 ; x2 ; x3 / of a point in
r b $ b0
rx $ x 0 one coordinate system, we now require that the position x 0 D .x10 ; x20 ; x30 / of the exact same
point in another coordinate system must be calculable from the first:
In different coordinate systems x10 D f1 .x1 ; x2 ; x3 /; x20 D f2 .x1 ; x2 ; x3 /; x30 D f3 .x1 ; x2 ; x3 /: (1.12)
the same points have different co-
ordinates, connected by a trans-
More compactly this may be written x 0 D f .x/.
formation x 0 D f .x/. The postulate that there should exist a unique, bijective transformation connecting any
given pair of coordinate systems reflects that physical reality is unique (see Figure 1.5), and
2 The printed boldface notation is hard to reproduce in calculations with pencil on paper, so other graphical means
are commonly used, for example a bar (x), an arrow (x), E or underlining (x).
1. CONTINUOUS MATTER 13
that different coordinate systems are just different ways of representing the same physical
space in terms of real numbers.
What’s in a symbol?: There is nothing sacred about the symbols used for the coordinates.
Mostly the coordinates are given suggestive or conventional symbolic names in particular coor-
dinate systems, for example ı, , and h for the Earth coordinates above. In physics the familiar
Cartesian coordinates are often denoted x, y, and z, while in more formal arguments one may
retain the index notation, x1 , x2 , and x3 . Cylindrical coordinates are denoted r, , and z, and
spherical coordinates r, , and . These three coordinate systems are in fact the only ones to be
used in this book.
s
s sP
Pt b
Length road s
s s
From the earliest times humans have measured the length of a road between two points, say s s shortest
a and b, by counting the number of steps it takes to walk along this road. This definition of
length depends, however, on how you are built. In order to communicate to others the length t
a
of a road, the count of steps must be accompanied by a clear definition of the length of a step
in terms of an agreed-upon unit of length. The length of the road between
the positions a and b is measured
Unit of length: Originally the units of length—inch, foot, span, and fathom—were directly re- by counting steps along the road.
lated to the human body, but increasing precision in technology demanded better-defined units. In Different roads have different
1793 the meter was introduced as a ten millionth of the distance from equator to pole on Earth, and lengths, but normally there is a
until far into the twentieth century a unique “normal meter” was stored in Paris, France. In 1960 unique shortest road.
the meter became defined as a certain number of wavelengths of a certain spectral line in krypton-
........................
86, an isotope of a noble gas that can be found anywhere on Earth. Since 1983 the meter has been .... ....
... ....
... ...
defined by international convention to be the distance traveled by light in exactly 1/299,792,458 .. .. ...
... ...
of a second [1]. The problem of measuring lengths has thus been transferred to the problem of
...
... sb
...
.
...
path ...
measuring time, which makes sense because the precision of the time current standard is at least a ...
..
...
thousand times better than any proper length standard. ...
.
.... shortest
..
.....
s
...
This method for determining the length of a path may be refined to any desired practical ...
precision by using very short steps. In the extreme mathematical limit, the steps become a
infinitesimally small, and the road becomes a continuous path.
In the mathematical limit the
shortest continuous path con-
Distance necting a and b is called the
The shortest path between two points is called a geodesic and represents the “straightest line” geodesic. Normally, there is only
between the points. In Euclidean space, a geodesic is indeed what we would intuitively call one geodesic between any two
points.
a straight line. On the spherical surface of the Earth geodesics are great circles, and airplanes
and ships travel along them for good reason. The distance between two points is defined to
be the length of a geodesic connecting them. Since the points are completely defined by their
coordinates a and b in the chosen coordinate system, the distance must be a real positive
function d.a; b/ of the two sets of coordinates.
It is clear that the distance between two points must be the same in all coordinate systems,
b0 $
r b
because it can, in principle, be determined by laying out rulers or counting steps between two
points without any reference to coordinate systems. Distance is a property of space rather
than a property of the coordinate system. The actual distance function d 0 .a0 ; b0 / in a new d 0 .a0 ; b0 / D d.a; b/
coordinate system may be different from the old, d.a; b/, but the numerical values have to be
the same,
r
d 0 .a0 ; b0 / D d.a; b/; (1.13) a0 $ a
where a0 D f .a/ and b0 D f .b/ are calculated by the coordinate transformation (1.12). The distance is invariant under
Knowing the distance function d.a; b/ in one coordinate system, it may be calculated in any coordinate transformations.
other coordinate system by means of the appropriate coordinate transformation.
14 PHYSICS OF CONTINUOUS MATTER
imal coordinate differences in the infinitesimal neighborhood of any point (the local Euclidean tangent space).
1. CONTINUOUS MATTER 15
1.6 Fields
In the extreme mathematical limit, material particles are taken to be truly infinitesimal and English field
German feld
all physical properties of the particles as well as the forces acting on them are described by
Dutch veld
smooth—or at least piecewise smooth—functions of space and time. Continuum physics Danish felt
is therefore a theory of fields. Mathematically, a field f is simply a real-valued function, Swedish fält
f .x; t / D f .x1 ; x2 ; x3 ; t / or f .x; t / D f .x; y; z; t /, of the spatial coordinates x and time t , French champ
representing the value of a physical quantity in this point of space at this time5 . Italian campo
Spanish campo
The fields of continuum physics Russian polje
The use of the word “field” in
We have already met the mass density field .x; t /. Knowing this field, the mass dM of a physics to denote a function of
material particle occupying the volume d V near the point x at time t can be calculated as the spacetime coordinates has
an unclear origin. The orig-
dM D .x; t / d V: (1.15)
inal meaning of phrases such
We shall mostly suppress the explicit space and time variables and just write dM D d V . as “gravitational field”, “elec-
Sometimes a collection of functions is also called a field and the individual real-valued tric field”, or “magnetic field”
was presumably to denote re-
members are called its components. The most fundamental field of fluid mechanics, the ve-
gions of gravitational, electric, or
locity field v D .v1 ; v2 ; v3 / or v D .vx ; vy ; vz /, has three components, one for each of the
magnetic influences in the other-
Cartesian coordinate directions. The velocity field v.x; t / determines the momentum, wise empty space around a body.
d P D v.x; t / dM; (1.16) The meaning was later shifted to
the mathematical representation
of a material particle of mass dM near x at time t. The velocity field will be of major of the strength (and direction) of
importance in formulating the dynamics of continuous systems. such influences in every point of
Besides fields characterizing the state of the material, such as mass density and velocity, it space.
is convenient to define fields that characterize the forces acting on and within the material. The
gravitational acceleration field g.x; t / penetrates all bodies from afar and acts on a material
particle of mass dM with a force
d F D g.x; t / dM: (1.17)
Using that dM D d V we may also write gravity as a body force (or volume force) of the
form, d F D f d V , with a density of force f D g. It has infinite range, and the same is
true for electromagnetism, but that also ends the list. No other forces in nature seem to have
infinite range.
Some force fields are only meaningful for regions of space where matter is actually
present, as for example the density or the pressure field p, which acts across the imagined
contact surfaces that separate neighboring volumes of a fluid at rest. Pressure is, however,
not the only contact force. Fluids in motion, solids and more general materials, have more
complicated contact forces that can only be fully described by the nine-component stress
field, D fij g, which is a (3 3) matrix field with rows and columns labeled by Cartesian
coordinates: i; j D 1; 2; 3 or i; j D x; y; z.
Mass density, velocity, gravity, pressure, and stress are the usual fields of continuum me-
chanics and will all be properly introduced in the chapters to come. Other fields are thermo-
dynamic, like the temperature T , the specific internal energy U , or the specific entropy S .
Some describe different states of matter, for example the electric charge density e and cur-
rent density je together with the electric and magnetic field strengths, E and B. Like gravity
g, these force fields are thought to exist in regions of space completely devoid of matter.
5 In mathematics the tendency is to use coordinates labeled by integers. In physics we shall—as mentioned
before—mostly label the coordinate axes x, y, and z, and use these labels as vector indices u D .ux ; uy ; uz /.
The general position is denoted x D .x; y; z/ and becomes admittedly a bit inconsistent because of the strange re-
lations, xx D x, xy D y and xz D z. In many physics texts the general position is instead denoted r D .x; y; z/,
but that comes with its own esthetic problems in more formal analysis. There seems to be no easy way to get the best
of both worlds.
16 PHYSICS OF CONTINUOUS MATTER
There are also fields that refer to material properties, for example the coefficient of shear
elasticity of a solid and the coefficient of shear viscosity of a fluid. Such fields are often
nearly constant within homogeneous bodies, that is, independent of space and time, and are
mostly treated as material constants rather than true fields.
Field equations
Like all physical variables, fields evolve with time according to dynamical laws, called field
equations. In continuum mechanics, the central equation of motion descends directly from
Newton’s Second Law applied to every material particle. Mass conservation, which is all but
trivial and most often tacitly incorporated in particle mechanics, turns in continuum theory
into an equation of motion for the mass density. Still other field equations such as Maxwell’s
equations for the electromagnetic fields have completely different and non-mechanical ori-
gins, although they do couple to the mechanical equations of motion via the Lorentz force.
Mathematically, field equations are partial differential equations in both space and time.
This makes continuum mechanics considerably more difficult than particle mechanics where
the equations of motion are ordinary differential equations in time. On the other hand, this
greater degree of mathematical complexity also leads to a plethora of new and sometimes quite
unexpected phenomena. Mathematically, field equations in three-dimensional space would be
quite cumbersome to deal with, were it not for an efficient extension of vector methods to
what is now called called vector calculus or field calculus. It is introduced along the way in
the chapters to come and presented with some rigor in Appendix C.
In some field theories, for example Maxwell’s electromagnetism, the field equations are
linear in the fields, but that is not the case in continuum mechanics. The non-linearity of the
field equations of continuum mechanics is caused by the velocity field, which behaves like
a “wind” that carries other fields (and itself) along in the motion. This adds a further layer
of mathematical difficulty to this subject, making it very different from linear theories—and
much richer. The non-linearity leads to dynamic instabilities and gives rise to the chaotic and
as yet not fully understood phenomenon of turbulence, well known from our daily dealings
with water and air.
Mathematically, the local and global equations are equivalent, and must both be presented
in any textbook, including this one. The local equations allow us to find exact solutions, either
analytically or numerically, while the global equations are well suited for getting approxima-
tive solutions and making estimates.
Problems
1.1 Consider a small volume V of a much larger volume of gas, such that the probability for any
molecule to be found in V is exceedingly small. If the average number of molecules in V is known to
be N , the probability of finding precisely n of the molecules in V is given by the Poisson distribution,
Nn N
Pr.njN / D e :
nŠ
Show that
(a) It is normalized.
(b) The mean value is hni D˝ N . p
(c) The variance is N 2 D .n N /2 D N , that is N D N .
˛
1.2 The Lennard–Jones potential is often used to describe the interaction energy between two neutral
atoms. It is given by the conventional formula
!
12 6
V .r/ D 4 ;
r 12 r6
where r is the distance between centers of the atoms. The parameters and have dimensions of energy
and length.
(a) Determine the equilibrium distance r D a where the potential is minimal, and its minimal value.
(b) Determine the leading behavior of the potential around minimum.
(c) Calculate the frequency of radial harmonic vibrations around equilibrium with one atom held fixed.
(d) As an example, take argon, which has molar mass Mmol D 40 g mol 1 , molar energy D
1 kJ mol 1 , and equilibrium distance a D 2:87 Å.
1.3 Consider a collection of N identical molecules (a “material particle”) taken from a large volume
of gas. Let the instantaneous molecular velocities be vn for n D 1; 2 ; N . Collisions with other
molecules in the gas at large will randomly change the velocity of each of the selected molecules, but
if there is no overall drift in the gas, the velocity of individual molecules should average out to zero,
hvn i D 0, the velocities of different molecules should be uncorrelated, hvn˝vm i˛ D 0 for n ¤ m, and the
average of the square of the velocity should be the same for all molecules, vn2 D v02 .
(a) p
Show that the root-mean-square average of the center-of-mass velocity of the collection equals
v0 = N .
Show that the Cartesian distance function (1.14) satisfies these axioms.
Part I
Fluids at rest
20
Fluids at rest
In the following chapters we only consider fluids without any macroscopic motion. Micro-
scopically, matter is never truly at rest because of random molecular motion (heat).
The essential chapters for a minimal curriculum are Chapter 2 and part of Chapter 3.
Chapters 4 and 5 may be included in part or whole depending on the level and length of the
course.
List of chapters
2. Pressure: The intuitive concept of pressure in static fluids is brought on a formal foot-
ing by means of the pressure field. The basic equations of hydrostatics are developed
and applied to the sea, the atmosphere, and the outermost convective layer of the Sun.
3. Buoyancy and stability: Archimedes’ principle is derived from the balance of forces
and expanded to include balance of moments. The stability of floating objects is ana-
lyzed.
4. Hydrostatic shapes: The flattening of the Earth as well as the tides of Earth are ana-
lyzed. The tides that Jupiter generates on its moon Io are calculated.
5. Surface tension: The physical origin and formal definition of surface tension are pre-
sented. Soap bubbles are described and the importance of Marangoni forces is em-
phasized. The Young–Laplace law is derived and the Rayleigh–Plateau instability ana-
lyzed. Capillary effects, menisci, and drop shapes are discussed.
2
Pressure
If the Sun did not shine, if no heat were generated inside the Earth, and no energy radiated
into space, all the winds in the air and the currents in the sea would die away, and the air and
water on the planet would in the end come to rest in equilibrium with gravity. In the absence
of external driving forces or time-dependent boundary conditions, and in the presence of
dissipative contact forces, any fluid must eventually reach a state of hydrostatic equilibrium,
where nothing moves anymore and all fields become constant in time. This state must be the
first approximation to the sea, the atmosphere, the interior of a planet, or a star.
In mechanical equilibrium of a continuous system there is everywhere a balance between Tension Force
short-range contact forces and long-range body forces, such as gravity. Contact interactions 6
between material bodies or even between parts of the same body take place across contact .
.
surfaces. A contact force acting on a tiny patch of a surface can in principle take any direction " ."... ..
" .. ... .
."
. .-
.
relative to the surface, and may be resolved into a normal and a tangential component. If the "" " .
. .. ..
" Shear
.".... .
.. ...
normal component has the same orientation as the surface normal, it is called a tension force, "" ... ..
.. .. ... .. .. ... .. .. ... ..
"
.. .. .. .. .. .. .. .. ..
and if opposite a pressure force. The component acting tangentially to the surface is called a
The force acting on the material
shear force or a traction force. Fluids in motion, and solids at rest or in motion, are able to
underneath a small patch of a sur-
sustain shear forces, whereas fluids at rest cannot. Should shear forces arise in a fluid at rest,
face can always be resolved into a
it will begin to flow until it again reaches mechanical equilibrium without shear forces. perpendicular pressure force and
In this chapter we shall first investigate the basic properties of pressure, and afterward a tangential shear force. The
develop the mathematical formalism that permits us to analyze hydrostatic equilibrium in the pressure is positive if the force
sea and the atmosphere. Along the way we shall recapitulate some basic rules of thermody- is directed toward the patch, and
namics. In the following chapters we shall continue to study the implications of hydrostatic negative if it (as here) is directed
equilibrium for balloons and ships, and the shapes of large fluid bodies subject only to gravity away from it.
and small fluid bodies subject mainly to surface tension.
atmosphere p D p0 0 g0 z; (2.3)
-y where p0 is the surface pressure (equal to the atmospheric pressure). The pressure field is
independent of the horizontal coordinates, x and y, because the flat-Earth sea has the same
sea properties everywhere at the same depth.
x With 0 103 kg m 3 and g0 10 m s 2 , the scale of the pressure increase per unit of
depth in the sea becomes 0 g0 104 Pa m 1 D 1 bar=10 m. Thus, at the deepest point in
The flat-Earth coordinate system. the sea—the Challenger Deep of the Mariana Trench in the Western Pacific Ocean—with z
11 km, the pressure is a about 1,100 bar. The assumption of constant density is reasonably
well justified even to this depth, because the density of water is only about 5% higher there
than at the surface (see page 34).
z Example 2.1 [Sluice gate force]: Water is stemmed up to height h behind a sluice gate of
6 width L. On the water surface and on the outer side of the gate there is atmospheric pressure p0 .
What is the total horizontal force acting on the gate when h D 20 m and L D 30 m?
p0
0 Using that the pressure is the same in all directions in a fluid at rest (see the following section),
we find by integrating the pressure excess, p.z/ p0 , on the inside of the gate,
p.z/ - p0 Z 0 Z 0
1
FD .p.z/ p0 / Ldz D 0 g0 L z dz D 0 g0 Lh2 : (2.4)
h h h 2
This result should have been anticipated, because the pressure rises linearly with depth, such that
Water stemmed up behind a
the total force is simply the product of the area of the sluice gate Lh with the average pressure
sluice gate. The pressure varies
excess hp p0 i D 21 0 g0 h acting on the gate. With the given numbers the force becomes F
linearly with depth z.
6 107 N, corresponding to the weight of 6,000 metric tons of water.
Figure 2.1. The mean air pressure at sea level is not constant, but varies by a few percent around
the standard atmosphere: 1 atm D 101; 325 Pa D 1:01325 bar. This picture shows the world-wide
sea level air pressure averaged over a 22-year period (1979–2001). The contours are isobars and the
numbers are millibars. (Source: European Center for Medium-Range Weather Forecasts (ECMWF).
With permission.)
With surface pressure p0 D 1 atm and air density 0 D 1:18 kg m 3 at 25ı C we get h0 D
8:72 km, which is a tiny bit lower than the height of Mount Everest (8:848 km). This is, of
course, meaningless since climbers have reached the summit of that mountain without oxygen
masks. Nevertheless, this height sets the correct scale for major changes in the atmospheric
properties. As seen in Figure 2.2 on page 36, the linear decrease of air pressure with height is
quite a decent approximation for the first 2 km.
z
6
Microscopic origin of pressure in gases
In a liquid the molecules directly touch each other and the container walls, and liquid pressure .vx ; vy ; vz / .vx ; vy ; vz /
may be seen as a result of these contacts. A gas consists mostly of vacuum with the molecules @
moving freely along straight lines between collisions with each other, and the gas pressure on @ -y
a solid wall arises from the incessant random molecular bombardment of the wall. Rr
@
A
We shall under very general and reasonable assumptions obtain the pressure from the
average rate of molecular momentum transfer to a small flat area A of the wall orthogonal x
to the z-axis. Let us begin with the (evidently unreasonable) assumption that all molecules A molecule of mass m that is
have the same velocity vector v D .vx ; vy ; vz / with vz < 0, such that they all eventually reflected from the wall changes
will strike the wall at z D 0. The molecules that collectively hit the area A in a small time the sign of its z-component but
interval dt must all lie within a “striking distance” dz D vz dt from the wall, so that the leaves the other components un-
total mass of these molecules becomes dM D Adz D A. vz /dt , where is the mass changed, such that the momen-
density of the gas. Next we assume that all molecules are reflected perfectly from the wall, tum transferred to the wall is
thereby changing the z-component of the velocity from vz to vz while leaving the other 2mvz .
24 PHYSICS OF CONTINUOUS MATTER
components unchanged. The momentum transferred to the area A in the time interval dt
becomes d Pz D 2vz dM D 2vz2 Adt , and the corresponding force
dPz
Fz D D 2vz2 A: (2.6)
dt
Notice the sign, which shows that the force is always directed into the wall, as it should.
Finally, we take into account that the many molecular collisions create a mixture of all
possible velocities. We do not need to know this mixture, except that it ought to have the same
probability for all velocity directions. Denoting the average over the mixture of velocities by
a bracket hi, we get by symmetry hvx2 i D hvy2 i D hvz2 i D 13 hvx2 C vy2 C vz2 i D 13 hv2 i. As only
half the molecules, those with vz < 0, can hit the wall, the pressure defined as the average
force acting on a unit of area becomes
1 h Fz i 1 ˝ ˛
D vz2 D v2 ;
˝ ˛
p (2.7)
2 A 3
and solving for the root-mean-square molecular velocity, we obtain
s
p 3p
vmol hv2 i D : (2.8)
We shall see later that this velocity—as one might suspect—is of the same order of magnitude
as the velocity of sound. At a pressure of p D 1 atm and temperature T D 18 ı C D 291K,
the density of air is 1:2 kg m 3 , and the molecular velocity calculated from the above
expression becomes v 500 m s 1 .
The total pressure force acting on any oriented surface S is obtained by adding (that is, ........
integrating) all the little vector contributions from each surface element, A ................................ ................ S
nO K
....... ... ... ...
... A P ... .
......... ...
......... Ar
P ... ...
.
. .. dS.. ..
Z
... . ..
.. .. P ..
...
FD p dS: (2.11) ...... ...
P .. .
. . .
. ..
.. . d V ... ....
S ..
... .. .
.... ...
. ..
...
...
. .... .
..
.. . ....
...
Here we have again suppressed the explicit dependence on x. This is the force that acts on the
cork in the champagne bottle, moves the pistons in the cylinders of your car engine, breaks a
.....
...
...
... V
...
... ....
... ....
... .....
.... .. ...
.
. .
dam, and fires a bullet from a gun. It is also this force that lifts fish, ships, and balloons. ........... ...... . ..... ........................
..............................
Gauss’ theorem
The total pressure force on a body of volume V can thus be obtained in two ways: either by
adding (integrating) all the pressure forces that act on its surface S , or by adding (integrating)
all the forces on its constituent material particles. Reversing the sign on both sides, we must Johann Carl Friedrich Gauss
in other words have (1777–1855). German math-
ematician of great genius.
I Z Contributed to number theory, al-
p dS D r p d V: (2.13) gebra, non-Euclidean geometry,
S V and complex analysis. In physics
he developed the magnetometer.
The circle through the surface integral is only there to remind us that the surface is closed. The older (cgs) unit of magnetic
This is one version of Gauss’ theorem, a purely mathematical relation between the integral of strength is named after him.
an arbitrary function p.x/ over a closed surface S and the integral of its gradient r p.x/ over (Source: Wikimedia Commons.)
the enclosed volume V .
26 PHYSICS OF CONTINUOUS MATTER
Usually Gauss’ theorem is formulated as a relation between the surface integral over a
vector field U .x/ and the volume integral over its divergence r U D rx Ux C ry Uy C
rz Uz D @Ux =@x C @Uy =@y C @Uz =@z,
I Z
U dS D r U d V: (2.14)
S V
2.3 Hydrostatics
Two classes of forces are at play in standard continuum physics. The first class consists of
body forces that act throughout the volume of a body. The only known body forces are gravity
and electromagnetism, but in standard hydrostatics we shall only discuss gravity. The second
class consists of contact forces acting only on the surface of a body. Pressure is the only
contact force that is possible in fluids at rest, but there are others in fluids in motion and in
solids. The full set of possible contact forces will be introduced in Chapter 6.
...
? ? ? ? ..
..
section with a force d F D p d S on every surface element of the body. Explicitly, these ..
.....
..
..
.
...
... ...
two contributions to the total force on the fluid body are ..
.... ? ? ? ? ? ..
....
....... .
.....
......... ..
............ .......
................... ........
.................................................................
Z I
@
I
@
FG D g d V; FB D p dS: (2.17) 6
V S
Z I
F D F G C FB D g d V p dS D 0 (2.18)
V S
for each and every volume V of fluid. This is the equation of global hydrostatic equilibrium.
It states that the weight of the fluid in the volume must exactly balance its buoyancy. In Section
2.1 we used intuitively that the weight of a vertical box of sea water must be in balance with
the pressure forces on the bottom and the top of the box, and could in this way derive the
pressure in the sea. If the balance is not perfect, as for example if a small volume of fluid is
heated or cooled relative to its surroundings, the fluid must start to move, either upward if the
buoyancy force wins over the weight or downward if it loses.
The problem with the global equilibrium equation is that we have to know the force den-
sity g and the pressure p in order to calculate the integrals. Sometimes symmetry consid-
erations can get us a long way (see the example below), but in general we need a local form
of the equation of hydrostatic equilibrium valid at each point x in any geometry, to be able to
determine the hydrostatic pressure anywhere in a fluid.
Example 2.3 [The incompressible sea]: In constant gravity and using flat-Earth coordi-
nates, we have g.x/ D g0 D .0; 0; g0 /. For symmetry reasons the sea on the flat Earth ought
to have the same properties for all x and y, such that the pressure can only depend on z. Taking a
rectangular vertical box with height h and cross-sectional area A, we find
FG D 0 g0 AheOz ; FB D p.0/A C p. h/A eOz ; (2.19)
because the pressure contributions from the sides of the box cancel. Global hydrostatic balance
immediately leads back to Equation (2.2).
28 PHYSICS OF CONTINUOUS MATTER
d F D d F G C d F B D .g r p/ d V: (2.20)
f D g r p: (2.21)
It must again be emphasized that the effective force density is the sum of the true long-range
force (gravity) and the resultant of all the short-range pressure forces acting on its surface.
In hydrostatic equilibrium the total force d F D f d V must vanish for every material
particle, implying that the effective force density must vanish everywhere, f D 0, or
r p D g: (2.22)
This is the local equation of hydrostatic equilibrium. It is a differential equation for the
pressure, valid everywhere in a fluid at rest, and it encapsulates in an elegant way all the
physics of hydrostatics. Gauss’ theorem (2.13) allows us to convert the local equation back
z into the global one (2.18), showing that there is complete mathematical equivalence between
6 the local and global formulations of hydrostatic equilibrium.
g0
?
Case: Hydrostatics in constant gravity
-y
In a flat-Earth coordinate system, the constant field of gravity is g.x/ D .0; 0; g0 / for all x.
Written out explicitly in coordinates, the equation of local equilibrium (2.22) becomes three
x individual equations:
@p @p @p
The flat-Earth coordinate system D 0; D 0; D g0 : (2.23)
with constant gravity pointing @x @y @z
against the direction of the z-axis.
The two first equations express that the pressure does not depend on x and y but only on z,
which confirms the previous argument based on symmetry. It also shows that, independently
of the shape of a fluid container, the pressure will always be the same at a given depth (in
constant gravity). For the special case of constant density, .z/ D 0 , the last equation
may immediately be integrated to yield the previous result (2.3) for the pressure field in the
incompressible sea.
Paradoxes of hydrostatics
a It is easily understood why the pressure can only depend on z in an open liquid container,
b
because it must everywhere carry the weight of the liquid column above. But what if the
A schematic “boot”. The pres-
sure at a horizontal bottom un- liquid column does not reach all the way to the surface, as in a “boot” filled with water (see the
der the open shaft of the boot (a) margin figure)? Since there is only a short column of water in the “toe”, it raises the question
has to carry the full weight of the of where the constant pressure comes from. What is it “up against”? The only possibility is
liquid above, but what about the that the material of the boot must supply the necessary downward forces to compensate for the
pressure in the toe of the boot (b) missing weight of the water column. These forces arise as a response to the elastic extension
where the water column is much of the material of the boot caused by the pressure. If the material in the toe cannot deliver a
shorter? sufficient response, it will break and a fountain of water will erupt from the toe.
2. PRESSURE 29
Another “paradox” goes back to Pascal. Wine is often stored for aging in large wooden
casks containing cubic meters of the precious liquid. Such casks may only be able to with-
stand perhaps 30% pressure difference between liquid and atmosphere. Even if it would seem
natural to mount a pipeline to transport the wine to and from the casks, such a system comes
with its own problems. If, for example, the pipe entrance lies 5 m above the casks (that often
are placed in cellars), and if the unwary employee fills the cask and the whole pipe with wine,
half an atmosphere extra pressure will be added to the cask, which may as a consequence
develop a leak or even burst. Paradoxically, this does not depend on the diameter of the pipe.
Finally, what happens if someone presents you with a fluid density .x; y; z/ that actu-
ally depends on the horizontal coordinates? Such a situation could, for example, arise when
you remove a separating wall in a container with petrol on one side and water on the other.
The hydrostatic equations (2.23) cannot be fulfilled because they imply that g0 @=@x D
@2 p=@x@z D 0, contradicting the assumption that varies with x. The conclusion must be
that the liquid cannot be static but must flow until the dependence on the horizontal coordi-
nates has gone. In the end, all the petrol will lie in a horizontal layer on top of the water.
Non-horizontal separating surfaces between immiscible fluids at rest can simply not exist.
Since D M=V D nMmol =V we may write the ideal gas law in the local form,
103 Mmol 10 3 R
H2 2:0 4:157
He 4:0 2:079 Rmol
Ne 20:2 0:412 p D R T; RD : (2.27)
Mmol
N2 28:0 0:297
O2 32:0 0:260
Ar 39:9 0:208 In the margin the specific gas constant R is tabled for a few gases. Various extensions of the
CO2 44:0 0:189 ideal gas law take into account the excluded molecular volume as well as the strong molecular
Air 29:0 0:287 repulsion (see Problem 2.7). The ideal gas law is not only valid for pure gases but also for a
kg J
mol K kg mixture of pure gases provided one uses the average molar mass (see page 4). Real air with
Mmol D 28:95 g mol 1 is quite well described by the ideal gas law at normal temperatures,
The molar mass, Mmol , and
the specific gas constant, R D although in precise calculations it may be necessary to include various corrections (see for
Rmol =Mmol , for a few gases (in example [Kaye and Laby 1995]).
SI units).
* Microscopic origin of the ideal gas law
Even if the derivation of the equation of state belongs to statistical mechanics, it is nevertheless
of interest to see how the ideal gas law arises in the simplest possible molecular picture of a
gas at rest, already discussed on page 23. The gas consists of particles of mass m that are free
except for random collisions with the walls and with each other, so that the velocity vector,
v D .vx ; vy ; vz /, of a typical particle incessantly fluctuates around v D 0.
In statistical mechanics one derives (under certain conditions that we shall not discuss
here) the equipartition theorem, which states that each independent kinematic variable on
average carries the thermal energy 12 kB T , where kB D 1:3806510 23 J K 1 is Boltzmann’s
constant. Accordingly, the average kinetic energy of a molecule of mass m becomes
1 ˝ 2˛ 1 ˝ ˛ 1 ˝ ˛ 1 ˝ ˛ 3
Ludwig Boltzmann (1844– m v D m vx2 C m vy2 C m vz2 D kB T; (2.28)
1906). Austrian theoretical phys- 2 2 2 2 2
icist. Made fundamental contri- and inserting this into (2.7) we obtain
butions to statistical mechanics
and the understanding of the re- 1 ˝ 2˛ kB T
lations between the macroscopic pD v D : (2.29)
3 m
and microscopic descriptions of
nature. Believed firmly in the re- Multiplying the numerator and denominator of the fraction on the right by Avogadro’s number
ality of atoms of molecules, even NA and using that the molar mass is Mmol D NA m, we obtain the ideal gas law with the molar
if most of his contemporaries did gas constant Rmol D NA kB , as it should be.
not. His recurrent depressions
eventually led him to suicide.
(Source: Wikimedia Commons.) Case: Isothermal atmosphere
Everybody knows that the atmospheric temperature varies with height; but if we nevertheless
assume that it has constant temperature, T .x/ D T0 , and combine the equation of hydrostatic
equilibrium (2.23) with the ideal gas law (2.27), we obtain
dp g0 p
D g0 D : (2.30)
dz RT0
This is an ordinary differential equation for the pressure, and using the initial condition p D
p0 for z D 0, we find the solution
z= h0 RT0 p0
p D p0 e ; h0 D D : (2.31)
g0 0 g0
In the last step we have again used the ideal gas law at z D 0 to show that the expression for
h0 is identical to the incompressible atmospheric scale height (2.5).
2. PRESSURE 31
In the isothermal atmosphere the pressure thus decreases exponentially with height on a
characteristic length scale again set by h0 8; 728 m calculated for 1 atm and 25ı C. The
pressure at the top of Mount Everest (z D 8;848 m) is now finite and predicted to be 368 hPa.
As seen in Figure 2.2, the isothermal approximation is better than the incompressible one and
fits the data to a height of about 4 km. In Section 2.6 we shall obtain a third estimate, the
homentropic atmosphere, lying between the two.
which does not depend on the local temperature T .x/. This is in fact not as far-fetched as
it might seem at first. The condition of constant density .x/ D 0 that we used in the
preceding section to calculate the pressure in the sea is a trivial example of such a relationship
in which the density is independent of both pressure and temperature. A less trivial example
is obtained if the walls containing a fluid at rest are held at a fixed temperature T0 . The
omnipresent heat conduction will eventually cause all of the fluid to attain this temperature,
T .x/ D T0 ; and in this state of isothermal equilibrium the equation of state (2.25) simplifies
to F ..x/; p.x/; T0 / D 0, which is indeed a barotropic relationship.
Polytropic relation: In the following we shall often meet a barotropic relation of the so-
called polytropic form,
p D C ; (2.33)
The function ˆ.x/ represents the potential energy per unit of mass, so that a body of mass m
at the point x has a gravitational potential energy mˆ.x/. For constant gravity in flat-Earth
coordinates we have g.x/ D .0; 0; g0 /, and the potential is simply ˆ D g0 z.
32 PHYSICS OF CONTINUOUS MATTER
In terms of the potential, the equilibrium equation (2.22) may now be written as
rp
rˆ C D 0: (2.35)
If the density is constant, .x/ D 0 , it follows that
p
ˆ D ˆ C (2.36)
0
ˆ D ˆ C w.p/: (2.38)
The requirement ˆ .x/ D const is completely equivalent to the equation of hydrostatic equi-
librium for barotropic fluids. There is no agreement in the literature about a name for ˆ ,
but we have chosen to call it the effective potential because it combines the true gravitational
potential with the pressure potential, and because the effective force density (2.21) in general
is given by f D g with g D r ˆ .
Example 2.4 [Isothermal gas]: Under isothermal conditions, the pressure potential of an
ideal gas is calculated by means of the ideal gas law (2.27)
dp
Z
w D RT0 D RT0 log p: (2.39)
p
In flat-Earth gravity, the constancy of ˆ D g0 z C RT0 log p immediately leads to the exponen-
tially decreasing pressure (2.31) in the isothermal atmosphere.
Example 2.5 [Polytropic fluid]: For fluids obeying a polytropic relation (2.33), the pressure
potential becomes
1 d p
Z
1
wD C DC D : (2.40)
1 1
When the fluid is an ideal gas with p D RT , this takes the even simpler form
wD RT D cp T; cp D R; (2.41)
1 1
where cp is the specific heat at constant pressure (see margin table on the preceding page). In
Section 2.6 we shall use this expression to model the atmosphere.
2. PRESSURE 33
dp dp dp
KD D D : (2.42)
d V =V d= d The archetypal thought-experi-
ment in thermodynamics: A
In the second step we have used the constancy of the mass M D V of the fluid in the volume
cylindrical chamber with a move-
to derive that dM D d V C Vd D 0, from which we get d V =V D d=.
able piston.
The above definition makes immediate sense for a barotropic fluid, where p D p./ is
a function of density. For general fluid states it is necessary to specify the conditions under
which the bulk modulus is defined, for example whether the temperature is held constant Liquid T Œı C KŒGPa
(isothermal) or whether there is no heat transfer (adiabatic or isentropic). Thus, the equation
Mercury 20 24:93
of state (2.27) for an ideal gas implies that the isothermal bulk modulus is
Glycerol 0 3:94
Water 20 2:18
@p
KT D D p; (2.43) Benzene 25 1:04
@ T
Ethanol 20 0:89
where the index—as commonly done in thermodynamics—indicates that the temperature T Methanol 20 0:82
is held constant. Similarly, for an isentropic ideal gas obeying the polytropic relation (2.33) Hexane 20 0:60
that the isentropic bulk modulus becomes
Isothermal bulk modulus for
common liquids [Lide 1996].
dp
KS D D p; (2.44)
d S
where the index indicates that the entropy S is held constant. It larger than the isothermal
bulk modulus (2.43) by a factor of > 1, because adiabatic compression also increases the
temperature of the gas, which further increases the pressure.
The definition of the bulk modulus (and the above equation) shows that it is measured in
the same units as pressure, for example pascals, bars, or atmospheres. The bulk modulus is
actually a measure of incompressibility, because the larger it is, the greater is the pressure
increase that is needed to obtain a given fractional increase in density. The inverse bulk
modulus ˇ D 1=K may be taken as a measure of compressibility.
Example 2.6 [Water compression]: For sea water the bulk modulus is K 23:2 kbar D
2:32 GPa. As long as the pressure change is much smaller than the bulk modulus, dp K, we
may calculate the relative change in density from (2.42) to be d= dp=K. In the deepest
abyss of the sea, the pressure is about 1,100 bar, implying that the relative density change is
d= 1=20 5%.
K D K0 C .p p0 /; (2.45)
z 1=. 1/
-60
K0
.z/ D 0 1 ; h1 D : (2.47)
-80
h1 . 1/0 g0
-100
With the given parameters the length scale becomes h1 D 47:3 km. The pressure as a function
Polytropic water as a function of of depth z is obtained by inserting this into the equation of state (2.46). Since the resulting
depth z in units of km. Top: den- exponent, =. 1/ D 1:2, is close to unity, the pressure increases nearly linearly with depth,
sity in units of g cm 3 and Bot- as seen in the margin figures. At the deepest point of the sea, z D 10:924 km, the density
tom: pressure p in units of kbar. 4.3% higher than the surface density. The pressure at the deepest point is increased by about
2.2%, corresponding to 24 bar above the pressure that would have been obtained if the density
had been constant.
Vertical mixing
Let us imagine that we take a small amount of air and exchange it with another amount with z
'$
the same mass, taken from a different height with a different volume and pressure. In order 6
to fill out the correct volume, one air mass would have to contract and the other expand. If
T2
this is done quickly, there will be no time for heat exchange with the surrounding air, and
one air mass will consequently be heated up by compression and the other cooled down by &%
expansion. If the atmosphere initially were in isothermal equilibrium, the temperature of the
swapped air masses would not be the same as the temperature of the surrounding air in the T0
new positions, and the atmosphere would be brought out of equilibrium.
If, however, the surrounding air initially had a temperature distribution, such that the #
swapped air masses after the expansion and compression would arrive at precisely the correct
T1
temperatures of their new surroundings, a kind of “equilibrium” could be established, in which
the omnipresent vertical mixing had essentially no effect. Intuitively, it is reasonable to expect "!
that the end result of fast vertical mixing and slow heat conduction might be precisely such
a state. It should, however, not be forgotten that this state is not a true equilibrium state but Swapping air masses from dif-
rather a dynamically balanced state depending on the incessant small-scale turbulent motion ferent heights. If the air has
in the atmosphere. temperature T0 before the swap,
the compressed air would be
warmer, T1 > T0 , and the
Homentropic equation of state expanded air colder, T2 < T0 .
A process that takes place without exchange of heat between a system and its environment
is said to be adiabatic. If furthermore the process is reversible, it will conserve the entropy
and is called isentropic. In that case it follows from the thermodynamics of ideal gases (see
Appendix E) that an isentropic compression or expansion of a fixed amount M of an ideal gas
that changes the volume from V0 to V and the pressure from p0 to p, will obey the rule
pV D p0 V0 : (2.48)
Here is the adiabatic index, which for a gas like air with diatomic molecules is approxi-
mately 7=5 D 1:4.
Expressed in terms of the density D M=V , an isentropic process that locally changes
density 0 and pressure p0 to and p must obey the local polytropic relation:
p D p0 0 : (2.49)
In the dynamically balanced state of the atmosphere described above, this rule would not
only be valid for a local process, but also apply to the swapping of two air masses from
different heights, for example from the ground z D 0 to any value of z. The lower part of the
atmosphere, the troposphere, is at least approximatively in such a homentropic state in which
this law is valid everywhere, with the right-hand side being a constant expressed in terms of
the ground values of pressure and density.
The pressure potential was calculated in Example 2.5 with the result
w D cp T; cp D R: (2.50)
1
The constant R D Rmol =Mmol is the specific gas constant, and cp is the specific heat at
constant gas pressure (see Appendix E for details). For dry air with R 0:287 J K 1 g 1 and
D 7=5, its value is cp 1 J K 1 g 1 , which is about 4 times smaller than for liquid water.
36 PHYSICS OF CONTINUOUS MATTER
p @hPaD
1000
800
600
400
200
z @kmD
5 10 15 20 25 30
Figure 2.2. The pressure as a function of altitude (for T0 D 25ı C) in three different atmospheric
models analyzed in this chapter: constant density (dotted), isothermal (dashed), and homentropic (dot-
dashed). The solid curve is the Standard Atmosphere (1976) model (see page 37).
Homentropic solution
It was shown before on page 32 that hydrostatic equilibrium implies that ˆ D ˆ C w.p/ D
g0 z C cp T is a constant, independent of x. Defining the value of this constant to be ˆ D
cp T0 , we find the temperature
g0
T .z/ D T0 z; (2.51)
cp
where h0 is the scale height (2.31) for the isothermal atmosphere. For D 7=5 and T0 D
25ı C, we find the isentropic scale height h2 31 km.
The ideal gas law combined with the homentropic law (2.49) implies that T1 and
1
p T are also constant, so that the density and pressure become
1=. 1/ =. 1/
z z
D 0 1 ; p D p0 1 ; (2.53)
h2 h2
where 0 and p0 are the density and pressure at sea level, respectively. Temperature, density,
and pressure all vanish at z D h2 , which is of course unphysical.
In Figure 2.2 the pressure in the various atmospheric models has been plotted together
with the 1976 Standard Atmosphere model (see below). Even if the homentropic model gives
the best fit, it fails at altitudes above 10 km. The real atmosphere is more complicated than
any of the models considered here.
2. PRESSURE 37
Atmospheric stability
In the real atmosphere the temperature lapse rate will generally not be constant throughout
the troposphere. Sunlight will often heat bubbles of air relative to the surrounding atmosphere
because of local topographic variations. Since the pressure must be the same in the heated
bubble and the atmosphere, the ideal gas law tells us that a higher temperature is always ac-
companied by lower mass density. The bubble will thus weigh less than the volume of air it
displaces, and must according to Archimedes principle begin to rise (see the following chap-
ter). As it rises it will always be in pressure balance with the environment, but the temperature
will generally not have time to equilibrate because of the low heat conductivity of air.
If the temperature of the bubble drops faster with falling pressure than that of the atmo-
sphere, it will soon stop rising. If, on the other hand, the bubble temperature drops slower
with pressure than that of its environment, the bubble temperature will increase relative to the
environment, and the bubble will continue to rise faster and faster. This will, for example,
happen if a bubble of warm humid air rises in a relatively dry atmosphere. To begin with, it
rises slowly because of the small initial temperature excess in the bubble; but when the hu-
midity begins to condense, the released heat will make the bubble’s temperature drop much
slower. Strong vertical currents may be generated in this way, for example in thunderstorms.
dp g0 p
D ; (2.54)
dz RT
where R is the specific gas constant for air. Since T is assumed to be known, this equation
38 PHYSICS OF CONTINUOUS MATTER
z z z
80 80 80
60 60 60
40 40 40
20 20 20
dT
Layer z dz T p Ρ
-80 -60 -40 -20 0 20 10-5 10-4 10-3 10-2 10-1 1 10-5 10-4 10-3 10-2 10-1 1
Mesopause 85
High mesosphere 71 -2.0 Figure 2.3. Plots of temperature, pressure, and density in the Standard Atmosphere (1976) for sea-
Low mesosphere 51 -2.8 level temperature T0 D 25ı C. The vertical height is measured in kilometers, the temperature in Celsius,
Stratopause 47 0.0 the pressure in bars, and the density in kilograms per cubic meter.
High stratosphere 32 +2.8
Low stratosphere 20 +1.0
Tropopause 11 0.0
Troposphere 0 -6.5 can be immediately integrated to yield
The temperature gradient in the
z
dz 0
g0
Z
layered model of the US Stan-
p.z/ D p0 exp 0 ; 0 D ; (2.55)
dard Atmosphere (1976). The 0 T .z 0 / R
height z is measured in km and
the gradient d T =dz in K m 1 .
where p0 is the pressure at z D 0. The exponential coefficient 0 D 34:16 K km 1 is called
the atmospheric constant.
In the margin table is shown the temperature gradient for the Standard Atmosphere (1976).
It is given as a piecewise constant function, implying that the temperature itself is piecewise
linear. This way of presenting the model is convenient for analytic or numeric calculations.
The result for the pressure is shown as the fully drawn curve in Figure 2.2. Plots of tempera-
ture, pressure, and density are shown in Figure 2.3 for T0 D 25ı C.
T A106 KE
10
r A103 kmE
100 200 300 400 500 600 700
Figure 2.4. The temperature distribution in the Sun as a function of the distance from the center.
The vertical lines are boundaries between the three major layers of the Sun. The fully drawn curve is
extracted from the Standard Sun model [CDDA&96] and [3]. The dashed curve is the approximative
homentropic solution (2.57) for the convective envelope.
M0
ˆD G ; (2.56)
r
With the Sun’s parameters we find T1 D 5:4 106 K, a value that also sets the scale for the
Sun’s central temperature. The precise surface of the Sun is not too well defined, but normally
chosen to be that spherical surface at radius r D a where the plasma turns into neutral gas and
becomes transparent. The transition to complete transparency actually takes place over a few
hundred kilometers. The surface temperature is about T0 6;000 K, which is about 1,000
times smaller than T1 and can mostly be ignored. At the surface of the Sun the temperature
40 PHYSICS OF CONTINUOUS MATTER
lapse rate becomes d T =dr D T1 =a D 7:8 K km 1 , which accidentally is nearly the same
as in the Earth’s troposphere.
In Figure 2.4 the solution is plotted together with the temperature distribution from the
Standard Sun model. The agreement is nearly perfect in the convective zone. The density
and pressure can as in the preceding section be determined from the adiabatic law. To do
that one needs reliable values for the solar surface density and pressure, in which the mean
molar mass variation due to incomplete ionization near the surface is taken into account. We
shall not go further into this question here but refer the reader to the literature [CDDA&96,
Hansen and Kawaler 1994].
Problems
2.1 The normal human systolic blood pressure is usually quoted as 120 mm mercury (above atmo-
spheric pressure). A clinical sphygomanometer used to measure blood pressure is constructed from a
U-tube half filled with mercury. During measurement, the air pressure in one arm of the manometer is
supplied by an inflated blood-constricting cuff around the upper arm or the wrist, whereas the other arm
of the manometer is exposed to atmospheric pressure. (a) How long should the arms of the manometer
be when it must accommodate a measurement range of ˙100% around normal?
2.2 Consider a canal with a dock gate that is 12 m wide and has water depth 9 m on one side and 6 m
on the other side. Calculate
(a) The pressure in the water on both sides of the gate at a height z over the bottom.
(b) The total force on the gate.
(c) The total moment of force around the bottom of the gate.
(d) The height over the bottom at which the total force acts.
2.3 An underwater lamp is covered by a hemispherical glass with a radius of a D 15 cm and is placed
with its center at a depth of h D 3 m on the side of the pool. (a) Calculate the total horizontal force from
the water on the lamp when there is air at normal pressure inside.
2.4 Using a manometer, the pressure in an open container filled with liquid is found to be 1:6 atm at
a height of 6 m over the bottom, and 2:8 atm at a height of 3 m. (a) Determine the density of the liquid
and (b) the height of the liquid surface.
2.5 An open jar contains two non-mixable liquids with densities 1 > 2 . The heavy layer has
thickness h1 and the light layer on top of it has thickness h2 . (a) An open glass tube is now lowered
vertically into the liquids toward the bottom of the jar. Describe how high the liquids rise in the tube
(disregarding capillary effects). (b) The open tube is already placed in the container with its opening
close to the bottom when the heavy fluid is poured in, followed by the light. How high will the heavy
fluid rise in the tube?
2.6 Show that a mixture of ideal gases (see page 4) also obeys the equation of state (2.27).
2.8 Calculate the pressure and density in the flat-Earth sea, assuming constant bulk modulus. (a) Show
that both quantities are singular at a certain depth and calculate this depth.
* 2.9 Calculate the isentropic scale height for the Mars atmosphere.
3
Buoyancy and stability
Fishes, whales, submarines, balloons, and airships all owe their ability to float to buoyancy, the “Buoy”, pronounced “booe”,
lifting power of water and air. The understanding of the physics of buoyancy goes back as far probably of Germanic origin. A
as antiquity and probably sprung from the interest in ships and shipbuilding in classic Greece. tethered floating object used to
mark a location in the sea.
The basic principle is due to Archimedes. His famous law states that the buoyancy force on a
body is equal and oppositely directed to the weight of the fluid that the body displaces. Before
his time it was thought that the shape of a body determined whether it would sink or float.
The shape of a floating body and its mass distribution do, however, determine whether it
will float stably or capsize. Stability of floating bodies is of vital importance to shipbuilding,
and to anyone who has ever tried to stand up in a small rowboat. Newtonian mechanics not
only allows us to derive Archimedes’ principle for the equilibrium of floating bodies, but also
to characterize the deviations from equilibrium and calculate the restoring forces. Even if a
body floating in or on water is in hydrostatic equilibrium, it will not be in complete mechanical
balance in every orientation, because the center of mass of the body and the center of mass of
the displaced water, also called the center of buoyancy, do not in general coincide.
The mismatch between the centers of mass and buoyancy for a floating body creates a Archimedes of Syracuse (287–
moment of force, which tends to rotate the body toward a stable equilibrium. For submerged 212 BC). Greek mathematician,
bodies, submarines, fishes, and balloons, the stable equilibrium will always have the center of physicist, and engineer. Dis-
gravity situated directly below the center of buoyancy. But for bodies floating stably on the covered the formulae for area
surface, ships, ducks, and dumplings, the center of gravity is mostly found directly above the and volume of cylinders and
spheres, and invented rudimen-
center of buoyancy. It is remarkable that such a configuration can be stable. The explanation
tary infinitesimal calculus. For-
is that when the stably floating body is tilted away from equilibrium, the center of buoyancy mulated the Law of the Lever,
moves instantly to reflect the new volume of displaced water. Provided the center of gravity and wrote two volumes on hy-
does not lie too far above the center of buoyancy, this change in the displaced water creates a drostatics titled On Floating Bod-
moment of force that counteracts the tilt. ies, containing his Law of Buoy-
ancy. Killed by a Roman soldier.
(Source: Photograph of Fields
3.1 Archimedes’ principle Medal courtesy Stefan Zachow.
Wikimedia Commons.)
Mechanical equilibrium takes a slightly different form than global hydrostatic equilibrium
(2.18) on page 27 when a body of another material is immersed in a fluid. If its material is
incompressible, the body retains its shape and displaces an amount of fluid with exactly the
same volume. If the body is compressible, like a rubber ball, the volume of displaced fluid
will be smaller. The body may even take in fluid, like a sponge or the piece of bread you dunk
into your coffee, but we shall disregard this possibility in the following.
42 PHYSICS OF CONTINUOUS MATTER
...............
..........
..... ...
..
..
In the field of gravity, an unrestrained body with mass density body is subject to two forces:
........ ? ? ? ..
... ..
.......... .
.......
..
..
.
its weight,
@ ... .
.
....
@
R ..
... ? ? ? ? ..
... ..
..
..... ....
Z
... ..
...
.... ? ? ? ? ? ..
... FG D body g d V; (3.1)
....
...... ....
........
.......... ......
.... V
............... ........
.............................................................................
I
@
@
6 and the buoyancy due to pressure acting on its surface,
I
Gravity pulls on a body over FB D p dS: (3.2)
its entire volume while pressure S
only acts on the surface.
If the total force F D F G C F B does not vanish, an unrestrained body will accelerate in
the direction of F according to Newton’s Second Law. Therefore, in mechanical equilibrium,
weight and buoyancy must precisely cancel each other at all times to guarantee that the body
will remain in place.
Assuming that the body does not itself contribute to the field of gravity, the local balance
of forces in the fluid, expressed by Equation (2.22) on page 28, will be the same as before the
body was placed in the fluid. In particular the pressure in the fluid cannot depend on whether
...............................................................................
......... ...........
the volume V contains material that is different from the fluid itself. The pressure acting on
..... ........
... .....
...
.... ....
...
the surface of the immersed body must for this reason be identical to the pressure on a body
... ...
...
...
...
... of fluid of the same shape, but then the global equilibrium condition (2.18) on page 27 for any
... ..
...
...
...
..
..
..
..
volume of fluid tells us that F fluid
G C F B D 0, or
... ..
... ...
... ...
... ....
... DISPLACEMENT .
..
Z
... ..
...
...
... ...
..
..
.. FB D F fluid
G D fluid g d V: (3.3)
... .. V
... ...
...
.... ...
.... .. ..
.... ...
.... ....
.....
.......
......... ..........
... This theorem is indeed Archimedes’ principle:
........................... .
A body partially submerged in a The force of buoyancy is equal and opposite to the weight of the displaced fluid.
liquid. The displacement is the
amount of water that has been The total force on the body may now be written
displaced by the body below
the waterline. In equilibrium
Z
the weight of the displaced fluid F D F G C FB D .body fluid /g d V; (3.4)
V
equals the weight of the body.
explicitly confirming that when the body is made from the same fluid as its surroundings—
so that body D fluid —the resultant force vanishes automatically. In general, however, the
distributions of mass in the body and in the displaced fluid will be different.
Karl Friedrich Hieronymus
Freiherr von Münchhausen
Münchhausen effect: Archimedes’ principle is valid even if the gravitational field varies
(1720–1797). German soldier,
across the body, but fails if the body is so large that its own gravitational field cannot be neglected,
hunter, nobleman, and delightful
such as would be the case if an Earth-sized body fell into Jupiter’s atmosphere. The additional
story-teller. In one of his stories,
gravitational compression of the fluid near the surface of the body increases the fluid density and
he lifts himself out of a deep lake
thus size of the buoyancy force. In semblance with Baron von Münchhausen’s adventure, the body
by pulling at his bootstraps.
in effect lifts itself by its bootstraps (see Problems 3.8 and 3.9).
3. BUOYANCY AND STABILITY 43
Since the total force is the sum of these contributions, one might say that buoyancy acts as if
the displacement were filled with fluid of negative mass Mfluid . In effect the buoyancy force
acts as a kind of antigravity.
The total force on an unrestrained object is now
If the body mass is smaller than the mass of the displaced fluid, the total force is directed
upward (i.e., against gravity) and the body will begin to rise, and conversely if the force is
directed downward it will sink. Alternatively, if the body is kept in place, the restraints must
deliver a force F to prevent the object from moving.
In constant gravity, a body can only hover motionlessly inside a fluid (or on its surface) if
its mass equals the mass of the displaced fluid:
A fish achieves this balance by adjusting the amount of water it displaces (Mfluid ) through con-
traction and expansion of its body by means of an internal gas-filled bladder. A submarine, in
contrast, adjusts its mass (Mbody ) by pumping water in and out of ballast tanks while keeping
its shape unchanged.
Bermuda Triangle Mystery: It has been proposed that the mysterious disappearance of ships
near Bermuda could be due to a sudden release of methane from the vast deposits of methane
hydrates known to exist on the continental shelves. What effectively could happen is the same as
when you shake a bottle of soda. Suddenly the water is filled with tiny gas bubbles with a density
near that of air. This lowers the average density of the frothing water to maybe only a fraction of
normal water, such that the mass of the ship’s displacement falls well below the normal value. The
ship is no more in buoyant equilibrium and drops like a stone, until it reaches normal density water
or hits the bottom where it will usually remain forever because it becomes filled with water on the
way down. Even if this sounds like a physically plausible explanation for the sudden disappearance
of surface vessels, there is no consensus that this is what really happened in the Bermuda Triangle,
nor in fact whether there is a mystery at all [4].
The physical phenomenon is real enough. It is, for example, well known to whitewater sailors
that “holes” can form in which highly aerated water decreases the buoyancy, even to the point
where it cannot carry their craft. You can yourself do an experiment in your kitchen using a
partially filled soda bottle with a piece of wood floating on the surface. When you tap the bottle
hard, carbon dioxide bubbles are released, and the “ship” sinks. In this case the “ship” will however
quickly reappear on the surface because it does not take in water.
Example 3.1 [Gulf Stream surface height]: The warm Gulf Stream originates in the Gulf
of Mexico and crosses the Atlantic to Europe after having followed the North American coast to
Newfoundland. It has a width of about 100 km and a depth of about d D 1;000 m. Its temperature
is about T D 10ı C above the surrounding ocean, of course warmer close to its origin and close
to the surface. Since the fractional expansion of seawater (also called the expansivity) is about
˛ D 2 10 4 K 1 , and taking into account that the temperature excess varies through the water
column, we estimate the increase in surface height from half the depth: h D ˛T d=2. Inserting
the numbers we find h 1 m. This agrees quite well with known values and with seasonal
variations in surface height of about h 15 cm [KSH99].
44 PHYSICS OF CONTINUOUS MATTER
where 0 the average density of the gas, is the average density of the displaced air, and V
the volume of the gas at height z. If the left-hand side of this equation is smaller or larger than
the right-hand side, the balloon will rise or fall.
z
Hot-air balloons
6 A hot-air balloon is open at the bottom so that the inside pressure is always the same as the
atmospheric pressure outside. The air in the balloon is warmer (T 0 > T ) than the outside
......
............................................
......... ......
....
temperature and the density is correspondingly lower (0 < ). Using the ideal gas law (2.27)
...
and the equality of the inside and outside pressures, we obtain 0 T 0 D T so that the inside
0 ....
.. .. T ... T
... ...
... ...
...
..
.. 0
...
.. density is 0 D T =T 0 . Up to a height of about 10 km one can use the expressions (2.52) and
... ...
... .... (2.53) for the homentropic temperature and density of the atmosphere.
... ..
.... ...
...... ...
....... .....
... ........
p p Example 3.2 [La Montgolfière]: The first Montgolfier balloon used for human flight on
November 21, 1783, was about 15 m in diameter with an oval shape and had a constant volume
V 60;000 foot3 1700 m3 . It carried two persons, rose to a ceiling of z 1;000 m and flew
A hot-air balloon has higher tem- about 9 km in 25 minutes. The machine is reported to have weighed 1;600 lbs 725 kg, and
perature T 0 > T and lower adding the two passengers and their stuff the total mass to lift could have been M 850 kg. The
density 0 < but essentially the November air being fairly cold and dense, we guess that its density must have been D 1:1 kg m 3
same pressure as the surrounding at 1,000 m above Paris, and from (3.8) we then conclude that the hot air density must have been
atmosphere because it is open 0 D .V M /=V D 0:60 kg m 3 . Taking T D 0ı C D 273 K it follows from the ideal gas law
below. that the hot air temperature was T 0 D T=0 D 501 K D 228ı C. This is uncomfortably close to
the famous ignition temperature of paper, 451ı F D 233ı C. The balloon’s material did actually get
scorched by the burning straw used to heat the air in flight, but the fire was quickly extinguished
with water brought along for this eventuality.
1 On the right-hand side we have left out the tiny buoyancy V D M= due to the displaced air volume V
0 0 0
of the balloon’s solid material with average density 0 D M=V0 . Typically =0 . 10 3 .
3. BUOYANCY AND STABILITY 45
Figure 3.1. Contemporary pictures of the first balloons. The first ascents were witnessed by huge
crowds. Benjamin Franklin, scientist and one of the founding fathers of the United States, also witnessed
the first ascents and became deeply interested in the future possibilities of this invention, but did not live
to see the first American hot-air balloon flight in 1793. Left: : The Montgolfier hot-air balloon. The
first ascent was piloted by Jean-Francois Pilatre de Rozier and the Marquis d’Arlandes. In 1785 Rozier
attempted to cross the English Channel with a combination of a hydrogen and hot-air balloon, but the
balloon suddenly deflated and he died in the crash. (Source: Wikimedia Commons.) Right: The Charles
hydrogen balloon. The first ascent was piloted by Charles himself and Nicolas-Louis Robert. At the end
of the first flight Robert jumped out, and the balloon with Charles rapidly shot up to about 3,000 m.
Charles actually never flew again. (Source: Wikimedia Commons.)
Gas balloons
A modern large hydrogen or helium balloon typically begins its ascent being only partially
filled, assuming an inverted tear-drop shape. During the ascent the gas expands because of
the fall in ambient air pressure, and eventually the balloon becomes nearly spherical and
stops expanding (or bursts) because the “skin” of the balloon cannot stretch further. To avoid
bursting, the balloon can be fitted with a safety valve, or simply be open at the bottom. Since
the density of the surrounding air falls with height, causing the buoyancy to drop, the balloon
will eventually reach a ceiling where it would hover permanently if it did not lose gas. In the
end no balloon stays aloft forever2.
Example 3.3 [La Charlière]: The hydrogen balloon used by Charles for the ascent on De-
cember 1, 1783, carried two passengers to a height of 600 m. The balloon was made from rub-
berized silk and nearly spherical with a diameter of 27 ft, giving it a nearly constant volume of
V D 292 m3 . It was open at the bottom to make the pressure the same inside and outside. Tak-
ing also the temperatures to be the same inside and outside, it follows from the ideal gas law that
the ratio of densities equals the ratio of molar masses, 0 = D Mhydrogen =Mair 1=15, inde-
pendently of height. Assuming a density D 1:1 kg m 3 at the ceiling, we obtain from (3.8)
M D . 0 /V D 300 kg. Assuming furthermore a skin thickness of 1:5 mm and a skin den-
sity 0:4 g cm 3 , the skin mass becomes 128 kg, leaving only 172 kg for the gondola and the two
passengers. Assuming they each weight 60 kg, the mass of the gondola can have been only about
50 kg, subject of course to the uncertainty in the assumptions that we have made.
2 Curiously, no animals seem to have developed balloons for floating in the atmosphere, although both the physics
M D MG C MB ; (3.9)
If the total force vanishes, F D 0, the total moment will be independent of the origin of the
coordinate system, as may be easily shown.
Assuming again that the presence of the body does not change the local hydrostatic bal-
ance in the fluid, the moment of buoyancy will be independent of the nature of the material
inside V . If the actual body is replaced by an identical volume of the ambient fluid, this fluid
volume must be in total mechanical equilibrium, such that both the total force as well as the
total moment acting on it have to vanish. Using that MfluidG C MB D 0, we get
Z
MB D Mfluid
G D x fluid g d V: (3.12)
V
The moment of buoyancy is equal and opposite to the moment of the weight of the
displaced fluid.
This result is a natural corollary to Archimedes’ principle, and of great help in calculating
the buoyancy moment. A formal proof of this theorem, starting from the local equation of
hydrostatic equilibrium, is found in Problem 3.4.
The moment of gravity (3.10) may be expressed in terms of the center of mass xG of the
body, here called the center of gravity:
1
Z
MG D xG M g0 ; xG D xbody d V: (3.13)
M V
1
Z
MB D xB M g0 ; xB D xfluid d V; (3.14)
M V
where xB is the moment of gravity of the displaced fluid, also called the center of buoyancy.
Although each of these moments depends on the choice of origin of the coordinate system,
the total moment,
M D .xG xB / M g0 ; (3.15)
............
.............. ................
will be independent. This is also evident from the appearance of the difference of the two ............
...........
........... ....
...
...
......
. ...
center positions. A shift of the origin of the coordinate system will affect the centers of .......... ...
......
........... ..
.
.......... ..
...... ..
gravity and buoyancy in the same way and therefore cancel out. .. ...
.
......
..... 6 ..
..
.
.
.. .. .
.
.
As long as the total moment is non-vanishing, the unrestrained body is not in complete ..
..
....
rG rB
..
..
..
... ..
mechanical equilibrium, but will start to rotate toward an orientation with vanishing moment. ... ..
.
..
... ...
... ...
Except for the trivial case where the centers of gravity and buoyancy coincide, the above ....
......
....... ...
.
..
....
........ ......
equation tells us that the total moment can only vanish if the centers lie on the same vertical ...........
............... .........
................................................................................
?
line, xG xB / g0 . Evidently, there are two possible orientations satisfying this condition:
one where the center of gravity lies below the center of buoyancy, and another where the Fully submerged body in buoyant
center of gravity is above. At least one of these must be stable, for otherwise the body would equilibrium with non-vanishing
never come to rest. total moment (which here sticks
out of the paper). The moment
will for a fully submerged body
Fully submerged body always tend to rotate it (here anti-
clockwise) such that the center of
In a fully submerged rigid body, for example a submarine, both centers are always in the same gravity is brought below the cen-
place relative to the body, barring possible shifts in the cargo. If the center of gravity does ter of buoyancy.
not lie directly below the center of buoyancy, but is displaced horizontally, for example by
.......................................................................
.......... ............
rotating the body, the direction of the moment will always tend to turn the body so that the ....
...... .......
....
.... ....
...
center of gravity is lowered with respect to the center of buoyancy. The only stable equilib- ...
..
6 ...
...
...
... ...
rium orientation of the body is where the center of gravity lies vertically below the center of ...
...
...
rB
..
... ..
..
buoyancy. Any small perturbation away from this orientation will soon be corrected and the ...
...
...
..
..
..
body brought back to the equilibrium orientation, assuming of course that dissipative forces ...
G r 0 ...
rG
... .. . .
.
... ......... .
.... ..
... ..
(friction) can seep off the energy of the perturbation, for otherwise it will oscillate. A similar ... ..
... ...
... ...
argument shows that the other equilibrium orientation with the center of gravity above the ...
... ...
.... ...
.... ..
.... ...
center of buoyancy is unstable and will flip the body over, if perturbed the tiniest amount. ....
.....
?
...
....
.
..
.......
.......... .......
Now we understand better why the gondola hangs below an airship or balloon. If the ..........................
gondola were on top, its higher average density would raise the center of gravity above the
A fully submerged rigid body (a
center of buoyancy and thereby destabilize the craft. Similarly, a fish goes belly-up when it submarine) in stable equilibrium
dies because various gases fill the swim bladder, which enlarges into the belly and reverses must have the center of gravity
the positions of the centers of gravity and buoyancy. It also loses buoyant equilibrium and situated directly below the center
floats to the surface, and stays there until it becomes completely waterlogged and sinks to the of buoyancy. If G is moved to G 0 ,
bottom. Interestingly, a submarine always has a conning tower on top to serve as a bridge for example by rotating the body,
when sailing on the surface, but its weight will be offset by the weight of heavy machinery a restoring moment is created that
at the bottom, so that the boat remains fully stable when submerged. Surfacing, the center sticks out of the plane of the pa-
of buoyancy of the submarine is obviously lowered, but as we shall see below this need not per, as shown in the upper figure.
destabilize the boat even if it comes to lie below the center of gravity.
48 PHYSICS OF CONTINUOUS MATTER
.........
..........................................................................
...........
Body floating on the surface
...... .......
... ....
.... ....
... ...
... ...
..
... ...
...
At the surface of a liquid, a body such as a ship or an iceberg will according to Archimedes’
... ...
...
...
... 6
..
..
..
principle always arrange itself so that the mass of displaced liquid exactly equals the mass
... ..
...
... rB .
..
..
..
of the body. Here we assume that there is vacuum or a very light fluid such as air above the
rG
... .
.
...
... ..
..
.
liquid. The center of gravity is always in the same place relative to the body if the cargo is
... ..
... ...
... ... fixed (see however Figure 3.2), but the center of buoyancy depends now on the orientation of
...
... ...
...
....
....
....
?
...
.. the body, because the volume of displaced fluid changes place and shape, while keeping its
.... ..
..... .... mass constant, when the body orientation changes.
...... ...
......... ......
.............................
Stability can—as always—only occur when the two centers lie on the same vertical line,
A floating body may have a sta- but there may be more than one stable orientation. A sphere made of homogeneous wood
ble equilibrium with the center of floating on water is stable in all orientations. None of them are in fact truly stable, because
gravity directly below the center it takes no force to move from one to the other (disregarding friction). This is however a
of buoyancy. marginal case.
....................... A floating body may, like a submerged body, possess a stable orientation with the center
...... .....
...... .... of gravity directly below the center of buoyancy. A heavy keel is, for example, used to lower
. ....
.... ...
... the center of gravity of a sailing ship so much that this orientation becomes the only stable
...
... ...
.. .. equilibrium. In that case it becomes virtually impossible to capsize the ship, even in a very
.... rG ...... strong wind. The stable orientation for most floating objects, such as ships, will in general
... ...
. ...
..... ? ...
...
have the center of gravity situated directly above the center of buoyancy. This happens always
... ...
... when an object of constant mass density floats on top of a liquid of constant mass density, for
... rB
6 ...
..
...
.... . example an iceberg on water. Since ice, like water, is homogeneous, the part of the iceberg
...... .
. ..
.............. ...
. that lies below the waterline must have its center of buoyancy in exactly the same place as its
..................................................
center of gravity. The part of the iceberg lying above the water cannot influence the center of
A floating body generally has a buoyancy whereas it always will shift the center of gravity upward.
stable equilibrium with the center How can that situation ever be stable? Will the moment of force not be of the wrong
of gravity directly above the cen- sign if the ship is perturbed? Why don’t ducks and tall ships capsize spontaneously? The
ter of buoyancy. qualitative answer is that when the body is rotated away from such an equilibrium orientation,
the volume of displaced water will change place and shape in such a way as to shift the center
of buoyancy back to the other side of the center of gravity, reversing thereby the direction of
the moment of force to restore the equilibrium. We shall now make this argument quantitative.
z
6
rM
... ..
...
.... rG ...
.... ...
....
.... ... 3.4 Ship stability
..
....
....
....
r ..
...
.....
..... V rB ......
....... .
.
......... .... Sitting comfortably in a small rowboat, it is fairly obvious that the center of gravity lies above
................. .......................
..
the center of buoyancy, and that the situation is stable with respect to small movements of
the body. But many a fisherman has learned that suddenly standing up may compromise the
Ship in an equilibrium orienta- stability and throw him out among the fishes. There is, as we shall see, a strict limit to how
tion with aligned centers of grav- high the center of gravity may be above the center of buoyancy. If this limit is violated, the
ity (G) and buoyancy (B). The boat becomes unstable and capsizes. As a practical aid to the captain, the limit is indicated
metacenter (M ) lies in this case by the position of the so-called metacenter, a fictive point usually placed on the vertical line
above the center of gravity, so through the equilibrium positions of the centers of buoyancy and gravity (the “mast”). The
that the ship is stable against stability condition then requires the center of gravity to lie below the metacenter (see the
small perturbations. The horizon- margin figure).
tal line at z D 0 indicates the
Initially, we shall assume that the ship is in complete mechanical equilibrium with vanish-
surface of the water. In buoy-
ant equilibrium, the ship always
ing total force and vanishing total moment of force. The aim is now to calculate the moment
displaces the volume V , indepen- of force that arises when the ship is brought slightly out of equilibrium. If the moment tends
dently of its orientation. to turn the ship back into equilibrium, the initial orientation will be stable, otherwise it is
unstable.
3. BUOYANCY AND STABILITY 49
Figure 3.2. The Flying Enterprise (1952). A body can float stably in many orientations, depending on
the position of its center of gravity. In this case the list to port was caused by a shift in the cargo which
moved the center of gravity to the port side. The ship and its lonely captain Carlsen became famous
because he stayed on board during the storm that eventually sent it to the bottom. (Source: Politiken,
Denmark. With permission.)
y
Center of roll 6
.......................
........ ....
Most ships are mirror symmetric in a plane, but we shall be more general and consider a
..... ....... ....
....
.....
“ship” of arbitrary shape. In a flat-Earth coordinate system with vertical z-axis, the waterline .
. .... . ...
...
..
is naturally taken to lie at z D 0. In the waterline the ship covers a horizontal region A of ... ...
..... ..
.
arbitrary shape. The geometric center or area centroid of this region is defined by the average .... .... -x
... ..
... .
of the position, ... ..
... A ....
.... ..
1 ....
Z
......
............ .....
.x0 ; y0 / D .x; y/ dA; (3.16) ....................................
A A
where dA D dx dy is the area element. Without loss of generality we may always place the The area A of the ship in the wa-
coordinate system such that x0 D y0 D 0. In a ship that is mirror symmetric in a vertical terline may be of quite arbitrary
plane, the area center will also lie in this plane. shape.
z
To discover the physical significance of the centroid of the waterline area, the ship is tilted
6
(or “heeled” as it would be in maritime language) through a tiny positive angle ˛ around the
x-axis, such that the equilibrium waterline area A comes to lie in the plane z D ˛y. The net
change V in the volume of the displaced water is to lowest order in ˛ given by the difference
in volumes of the two wedge-shaped regions between new and the old waterline. Since the ˛
-y
C
displaced water is removed from the wedge at y > 0 and added to the wedge for y < 0, the
volume change becomes
Z Z
V D z dA D ˛ y dA D 0: (3.17)
A A Tilt around the x-axis. The
In the last step we have used that the origin of the coordinate system coincides with the change in displacement consists
centroid of the waterline area (i.e., y0 D 0). There will be corrections to this result of order ˛ 2 of moving the water from the
due to the actual shape of the hull just above and below the waterline, but they are disregarded wedge to the right into the wedge
here. To leading order the two wedges have the same volume. to the left.
50 PHYSICS OF CONTINUOUS MATTER
Figure 3.3. The Queen Mary 2 set sail on its maiden voyage on January 2, 2004. It was at that time the
world’s largest ocean liner with a length of 345 m, a height of 72 m from keel to funnel, and a width of
41 m. Having a draft of only 10 m, its superstructure rises an impressive 62 m over the waterline. The
low average density of the superstructure, including 2,620 passengers and 1,253 crew, combined with
the high average density of the 117 megawatt engines and other heavy facilities close to the bottom of
the ship nevertheless allow the stability condition (3.28) to be fulfilled. (Source: Wikimedia Commons.)
Since the direction of the x-axis is quite arbitrary, the conclusion is that the ship may be
heeled around any line going through the centroid of the waterline area without any first-order
change in volume of displaced water. This guarantees that the ship will remain in buoyant
equilibrium after the tilt. The centroid of the waterline area may thus be called the ship’s
center of roll.
The metacenter
Taking water to have constant density, the center of buoyancy is simply the geometric average
of the position over the displacement volume V (below the waterline),
1
Z
G0 Gb .xB ; yB ; zB / D .x; y; z/ d V: (3.18)
b V V
˛
In equilibrium, the horizontal positions of the centers of buoyancy and gravity must be equal
r r r xB D xG and yB D yG . The vertical position zB of the center of buoyancy will normally be
B00 B B0 different from the vertical position of the center of gravity zG , which depends on the actual
mass distribution of the ship, determined by its structure and load.
The tilt rotates the center of grav- The tilt around the x-axis changes the positions of the centers of gravity and buoyancy.
ity from G to G0 , and the center of The center of gravity xG D .xG ; yG ; zG / is supposed to be fixed with respect to the ship (see
buoyancy from B to B0 . In addi-
however Figure 3.2) and is to first order in ˛ shifted horizontally by a simple rotation through
tion, the change in displaced wa-
ter shifts the center of buoyancy
the infinitesimal angle ˛,
back to B00 . In stable equilibrium ıyG D ˛zG : (3.19)
this point must for ˛ > 0 lie to
the left of the new center of grav- There will also be a vertical shift, ızG D ˛yG , but that is of no importance to the stability
ity G0 . because gravity is vertical and therefore the shift creates no moment.
3. BUOYANCY AND STABILITY 51
The center of buoyancy is also shifted by the tilt, at first by the same rule as the center of
gravity, but because the displacement also changes there will be another contribution yB to
the total shift, so that we may write
ıyB D ˛zB C yB : (3.20)
As discussed above, the change in the shape of the displacement amounts to moving the water
from the wedge at the right (y > 0) to the wedge at the left (y < 0). The ensuing change
in the horizontal position of the center of buoyancy may according to (3.18) be calculated by
averaging the position change y yB over the volume of the two wedges,
1 ˛ I
Z Z
yB D .y yB / z dA D y 2 dA D ˛ ;
V A V A V
where
Z
I D y 2 dA (3.21)
A
is the second-order moment of the waterline area A around the x-axis. The movement
R of dis-
placed water will also create a shift in the x-direction, xB D ˛J =V where J D A xy dA,
which does not destabilize the ship.
The total horizontal shift in the center of buoyancy may thus be written
I
ıyB D ˛ zB C : (3.22)
V
This shows that the complicated shift in the position of the center of buoyancy can be written
as a simple rotation of a point M that is fixed with respect to the ship with z-coordinate,
I
zM D zB C : (3.23)
V
This point, called the metacenter, is usually placed on the straight line that goes through the
centers of gravity and buoyancy, such that xM D xG D xB and yM D yG D yB . The
calculation shows that when the ship is heeled through a small angle, the center of buoyancy
will always move so that it stays vertically below the metacenter.
The metacenter is a purely geometric quantity (for a liquid with constant density), depend-
ing only on the displacement volume V , the center of buoyancy xB , and the second-order
moment of the shape of the ship in the waterline. The simplest waterline shapes are
Rectangular waterline area: If the ship has a rectangular waterline area with sides 2a and
2b, the roll center coincides with the center of the rectangle, and the second moment around
the x-axis becomes
Z a Z b
4
I D dx dy y 2 D ab 3 : (3.24)
a b 3
If a > b, this is the smallest moment around any tilt axis because ab 3 < a3 b.
Elliptic waterline area: If the ship has an elliptical waterline area with axes 2a and 2b, the
roll center coincides with the center of the ellipse, and the second moment around the x-axis
becomes
Z a Z b p1 x 2 =a2
4 3 1
Z
2 3=2
I D dx p y dy D ab .1 t 2 / dt D ab 3 : (3.25)
a b 1 x 2 =a2 3 0 4
Notice that this is about half the value for the rectangle.
52 PHYSICS OF CONTINUOUS MATTER
Stability condition
The tilt generates a restoring moment around the x-axis, which may be calculated from (3.15),
vM
L Mx D .yG yB /M g0 : (3.26)
L
L Since we have yG D yB in the original mechanical equilibrium, the difference in coordinates
L vG
L after the tilt may be written, yG yB D ıyG ıyB , where ıyG and ıyB are the small
L horizontal shifts of order ˛ in the centers of gravity and buoyancy, calculated above. The
L shift xB D ˛J =V will create a moment My D xB M g0 , which tends to pitch the ship
along x but does not affect its stability.
L
L
L In terms of the height of the metacenter zM , the restoring moment becomes
L
L
Mx D ˛.zG zM /M g0 :
L
v Lv (3.27)
B00 B
vG For the ship to be stable, the restoring moment must counteract the tilt and thus have opposite
L sign of the tilt angle ˛. Consequently, the stability condition becomes
L
zG < zM :
L
L vM (3.28)
L
L Evidently, the ship is only stable when the center of gravity lies below the metacenter. For an
L
L alternative derivation of the stability condition, see Problem 3.13.
L
L
L Example 3.4 [Elliptical rowboat]: An elliptical rowboat with vertical sides has major axis
L 2a D 2 m and minor axis 2b D 1 m. The smallest moment of the elliptical area is I D
L .=4/ab 3 0:1 m4 . If your mass is 75 kg and the boat’s is 50 kg, the displacement will be
B00 v Lv
B V D 0:125 m3 , and the draft d V =4ab D 6:25 cm, ignoring the usually curved shape of the
boat’s hull. The coordinate of the center of buoyancy becomes zB D 3:2 cm and the metacenter
The metacenter M lies always zM D 75 cm. Getting up from your seat may indeed raise the center of gravity so much that it gets
directly above the actual center close to the metacenter and the boat begins to roll violently. Depending on your weight and mass
of buoyancy B00 . Top: The ship distribution, the boat may even become unstable and turn over.
is stable because the metacenter
lies above the center of gravity.
Bottom: The ship is unstable Metacentric height and righting arm
because the metacenter lies
below the center of gravity. The orientation of the coordinate system with respect to the ship’s hull was not specified in the
analysis, which is therefore valid for a tilt around any direction. For a ship to be fully stable,
the stability condition must be fulfilled for all possible tilt axes. Since the displacement V is
the same for all choices of tilt axis, the second moment of the area on the right-hand side of
(3.23) should be chosen to be the smallest one. Often it is quite obvious which moment is the
smallest. Many modern ships are extremely long with the same cross-section along most of
their length and with mirror symmetry through a vertical plane. For such ships the smallest
moment is clearly obtained with the tilt axis parallel to the longitudinal axis of the ship.
The restoring moment (3.27) is proportional to the vertical distance, zM zG , between
the metacenter and the center of gravity, also called the metacentric height. The closer the
center of gravity comes to the metacenter, the smaller the restoring moment will be, and
the longer the period of rolling oscillations will be. The actual roll period depends also on
the true moment of inertia of the ship around the tilt axis (see Problem 3.11). Whereas the
metacenter is a purely geometric quantity that depends only on the ship’s actual draft, the
center of gravity depends on the way the ship is actually loaded. A good captain should
always know the positions of the center of gravity and the metacenter of his ship before he
sails, or else he may capsize when casting off.
When a series of values for GZ (the ship’s righting
measurements are made.
arm) at successive angles of heel are plotted on a graph,
• KB is the vertical distance from K to the center of the result is a STABILITY CURVE. The stability
buoyancy when the ship is upright. KB is curve, as shown in figure 12-24, is called the CURVE
measured in feet. OF STATIC STABILITY. The word static indicates
• KG is the vertical distance from K to the ship’s that it is not necessary for the ship to be in motion for
center of gravity when the ship is upright. KG is the curve to apply. If the ship is momentarily stopped at
measured in feet. any angle during its roll, the value of GZ given by the
curve will still apply.
• KM is the vertical distance from K to the
metacenter when the ship is upright. KM is NOTE
3. BUOYANCY AND STABILITY
measured in feet.
53
T h e s t a b i l i t y c u r ve i s c a l c u l a t e d
The RIGHTING MOMENT of a ship is W times graphically by design engineers for values
GZ, that is, the displacement times the righting arm. indicated by angles of heel above 7°.
3
0 10 20 30 40 50 60 70 80 90
ANGLE OF HEEL, IN DEGREES
G G Z G Z G Z G
WATERLINE
B B B B
B
o
ANGLE OF HEEL = 0
o
ANGLE OF HEEL = 20 ANGLE OF HEEL = 40
o
ANGLE OF HEEL = 60 o ANGLE OF HEEL = 70
o
The metacenter is only useful for tiny heel angles where all changes are linear in the
angle. For larger angles one uses instead the righting arm which is the horizontal distance
jyG yB j between the center of gravity and the vertical line through the actual center of
buoyancy. The restoring moment is the product of the righting arm and the weight of the ship,
and instability sets in when the righting arm reaches zero for some non-vanishing angle of
heel (see Figure 3.4).
where I1 is the second moment of the open liquid surface. The metacenter position now
becomes
I 1 I1
zM D zB C : (3.34)
V 0 V
The effect of the moving liquid is to lower the metacentric height (or shorten the righting arm)
with possible destabilization as a result. The unavoidable inertial sloshing of the liquid may
further compromise the stability. The destabilizing effect of a liquid cargo is often counter-
acted by dividing the hold into a number of smaller compartments by means of bulkheads
along the ship’s principal roll axis.
Car ferry instability: In heavy weather or due to accidents, a car ferry may inadvertently get
a layer of water on the car deck. Since the car deck of a roll-on/roll-off ferry normally spans the
whole vessel, we have 1 D 0 and I1 I , implying that zM zB < zG , nearly independent of
the thickness h of the layer of water (as long as h is not too small). The inequality zM < zG spells
rapid disaster, as several accidents with car ferries have shown. Waterproof longitudinal bulkheads
on the car deck would stabilize the ro-ro ferry, but are usually avoided because they would hamper
efficient loading and unloading of the cars.
3. BUOYANCY AND STABILITY 55
with the respective area moments I1 D IQ.1 / and I2 D IQ.2 /. The two angles determine
orthogonal principal directions, 1 and 2, in the ship’s waterline area. The principal direction
with the smallest second-order moment around the area centroid has the lowest metacentric
height. If I1 is the smallest moment and the actual roll axis forms an angle with the 1-axis,
we can calculate the moment I for any other axis of roll forming an angle with the x-axis
from
Since I1 < I2 , it follows trivially that if the ship is stable for a tilt around the first principal
axis, it will be stable for a tilt around any axis.
Problems
3.1 A stone weighs F1 D 1;000 N in vacuum and F0 D 600 N when submerged in water of density
0 . (a) Calculate the volume V and (b) average density 1 of the stone.
3.2 A hydrometer is an instrument used to measure the density of a liquid. A certain hydrometer with
mass M D 4 g consists of a roughly spherical bulb and a long thin cylindrical stem of radius a D 2 mm.
The sphere is weighed down so that the apparatus will float stably with the stem pointing vertically
upward and crossing the fluid surface at some point. (a) How much deeper will it float in alcohol with
mass density 1 D 0:78 g cm 3 than in oil with mass density 2 D 0:82 g cm 3 ? You may disregard
the tiny density of air.
3.3 A cylindrical wooden stick with density 1 D 0:65 g cm 3 floats in water (density 0 D
1 g cm 3 ). The stick is loaded down with a lead weight with density 2 D 11 g cm 3 at one end
such that it floats in a vertical orientation with a fraction f D 1=10 of its length out of the water.
(a) What is the ratio M1 =M2 between the masses of the wooden stick and the lead weight? (b) As a
function of the density of the wood, how large a fraction of the stick can be out of the water in hydrostatic
equilibrium (disregarding questions of stability)?
3.4 Prove without assuming constant gravity that the hydrostatic moment of buoyancy equals (minus)
the moment of gravity of the displaced fluid (corollary to Archimedes’ principle).
56 PHYSICS OF CONTINUOUS MATTER
3.5 Assuming constant gravity, show that for a body not in buoyant equilibrium (that is for which
the total force F does not vanish), there is always a well-defined point of attack x0 such that the total
moment of gravitational plus buoyant forces is given by M D x0 F .
3.6 A right rotation cone has half opening angle ˛ and height h. It is made from a homogeneous
material of density 1 and floats in a liquid of density 0 > 1 . (a) Determine the stability condition
on the mass ratio 1 =0 when the cone floats vertically with the peak downward. (b) Determine the
stability condition on the mass ratio when the cone floats vertically with the peak upward. (c) What is
the smallest opening angle that permits simultaneous stability in both directions?
b
* 3.7 A ship of length L has a longitudinally invariant cross-section in the shape of an isosceles triangle
B
B
B with half opening angle ˛ and height h. It is made from homogeneous material of density 1 and floats
B h in a liquid of density 0 > 1 . (a) Determine the stability condition on the mass ratio 1 =0 when the
B ship floats vertically with the peak downward. (b) Determine the stability condition on the mass ratio
B when the ship floats vertically with the peak upward. (c) What is the smallest opening angle that permits
B˛ ˛ d
B simultaneous stability in both directions?
B
B 3.8 Two identical homogenous spheres of mass M and radius a are situated a distance D a apart in
a barotropic fluid. Due to their field of gravity, the fluid will be denser near the spheres. There is no other
gravitational field present, the fluid density is 0 , and the pressure is p0 in the absence of the spheres.
B
B One may assume that the pressure corrections due to the spheres are small everywhere in comparison
B d with p0 . (a) Show that the spheres will repel each other and calculate the magnitude of the force to
˛ ˛ B leading order in a=D. (b) Compare with the gravitational attraction between the spheres. (c) Under
B which conditions will the total force between the spheres vanish?
B
h B
B * 3.9 A barotropic compressible fluid is in hydrostatic equilibrium with pressure p.z/ and density .z/
B in a constant external gravitational field with potential ˆ D g0 z. A finite body having a “small”
B
gravitational field ˆ.x/ is submerged in the fluid. (a) Show that the change in hydrostatic pressure to
lowest order of approximation is p.x/ D .z/ˆ.x/. (b) Show that for a spherically symmetric
Top: Triangular ship of length L
body of radius a and mass M , the extra surface pressure is p D g1 a.z/, where g1 D GM=a2 is the
(into the paper) floating with its
magnitude of surface gravity. (c) Show that the buoyancy force is increased.
peak vertically downward. Bot-
tom: The peak vertically upward.
Note that in this case the draft is 3.10 Let I be a symmetric (2 2) matrix. Show that the extrema of the corresponding quadratic form
h d. n I n D Ixx n2x C 2Ixy nx ny C Iyy ny2 where n2x C ny2 D 1 are determined by the eigenvectors of I
satisfying I n D n.
3.11 Show that in a stable orientation the angular frequency of small oscillations around a principal
tilt axis of a ship is r
Mg0 .zM zG /
!D ;
J
where J is the moment of inertia of the ship around this axis.
* 3.12 A ship has a waterline area that is a regular polygon with n 3 edges. Show that the area
moment tensor (3.36) must have Ixx D Iyy and Ixy D 0.
3.13 (a) Show that the work (or potential energy) necessary to tilt a ship through an angle ˛ to second
order in ˛ is
1 2 1
Z
W D ˛ M0 g0 .zG zB / C 0 g0 z 2 dA;
2 2 A0
where z D ˛y, and (b) show that this leads to the stability condition (3.28).