Flavio Salvati

Fundamentals
of Astronomy
A guide for Olympiads
ii

Copyright c 2020 Flavio Salvati

All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of the author, or as expressly permitted by law, or
under terms agreed with the appropriate reprographics rights organization.
Enquiries concerning reproduction outside the scope of the above should be
sent to the author. You must not circulate this book in any other binding or
cover and you must impose the same conditions on any buyer.

Experience suggests that it is close to impossible for the book to be error-free.

If something looks amiss, please check the webpage www.astrolympiad.com
for a list of typos, updates and additional problems. If you have any questions
or suggestions, don’t hesitate to contact me at [email protected].

ISBN: 979-8688791433

Published on the autumnal equinox, 2020.

To my parents
Contents

Preface ix

I Positional Astronomy 1

1 Celestial Coordinate Systems 2

1.1 Geographical Coordinate System 2
1.2 The Horizontal System 4
1.3 The Equatorial System 6
1.4 The Ecliptic System 8
1.5 The Galactic System 9
1.6 Exercises 9
1.7 Problems 10

2 Transformation of Coordinates 11
2.1 The Three Equations 11
2.2 From the Horizontal to the HA-dec System 14
Upper and Lower Culmination 15
Rising and Setting Times 17
Duration of day and night 17
2.3 From The Ecliptic to the RA-dec System 19
Declination of the Sun throughout the year 20
2.4 Exercises 21
2.5 Problems 22

3 Perturbation of Coordinates 23
3.1 Precession 23
3.2 Nutation 26
3.3 Aberration 26
3.4 Parallax 27
3.5 Refraction 27
3.6 Proper Motion 29
3.7 Other Cycles 30
iv
Contents v

3.8 Exercises 30
3.9 Problems 31

4 Observation and Instruments 32

4.1 Angular Size 32
4.2 The Thin Lens Equation 33
4.3 Optical Telescopes 34
4.4 Angular Resolution 36
4.5 Magnification 38
Minimum Magnification 39
Maximum Magnification 39
4.6 Aperture Ratio 40
4.7 The Role of the Atmosphere 41
4.8 Exercises 41
4.9 Problems 43

5 Time systems 45
5.1 Sidereal and Solar Day 45
Estimating Sidereal Time 47
5.2 Equation of Time 48
5.3 Local Solar Time 51
5.4 Greenwich Mean Time 51
5.5 Time Zones 51
5.6 Daylight Saving Time 52
5.7 Measuring the Length of a Year 52
Sidereal Year 52
Tropical or Solar year 53
5.8 Calendars 53
5.9 Exercises 54
5.10 Problems 55

6 The Moon 56
6.1 Sidereal Month 57
6.2 Synodic Month 57
6.3 Draconic Month 58
6.4 Eclipses 59
Conditions for the Occurrence of Eclipses 59
Total Number of Eclipses in a Year 59
Lunar Eclipse 60
Solar Eclipse 62
Saros cycle 63
6.5 Libration 64
vi Contents

6.6 Exercises 65
6.7 Problems 66

II Radiation Mechanisms 67

7 Electromagnetic Radiation 68
7.1 Wave or particle? 68
7.2 Boltzmann Distribution Law 70
7.3 Maxwell Distribution Law 70
Planetary Atmospheres 71
7.4 Black Body 72
Wien’s Law 72
Stefan-Boltzmann Law 73
7.5 Types of Spectra 74
7.6 Doppler Effect 76
7.7 Harvard Stellar Classification 79
7.8 Yerkes Stellar Classification 80
7.9 Exercises 81
7.10 Problems 82

8 Flux and Magnitude 86

8.1 Flux 86
8.2 Albedo 86
8.3 Stellar magnitudes 87
Absolute Magnitude 89
Magnitude of a Composite System 90
Surface Magnitude 91
Magnitude Systems 91
8.4 Extinction and Optical Depth 92
8.5 Exercises 96
8.6 Problems 98

9 Cosmic Distance Ladder 99

9.1 Parallax 99
Diurnal Parallax 101
Annual Parallax 102
9.2 Spectroscopic Parallax 104
9.3 Standard Candles 105
9.4 Tully-Fisher Relation 106
9.5 Hubble’s Law 108
9.6 Exercises 110
Contents vii

9.7 Problems 112

III Celestial Mechanics 113

10 Gravitation and Kepler’s Laws 114

10.1 Newtonian Mechanics 114
10.2 Universal Law of Gravitation 114
10.3 Kepler’s Laws 118
10.4 Energy of an Orbiting Body 124
Closed Orbits 127
Open Orbits 128
10.5 Escape Velocity 128
Schwarzschild Radius 129
10.6 Virial Theorem 129
10.7 Exercises 130
10.8 Problems 133

11 Motion of the Planets 137

11.1 Apparent Motion of the Planets 137
11.2 Synodic Period 139
11.3 Exercises 141
11.4 Problems 141

12 Orbital Manoeuvres 142

12.1 The Rocket Equation 142
12.2 Oberth Effect 143
12.3 Hohmann Transfer Orbit 143
12.4 Exercises 145
12.5 Problems 146

13 Binary Stars 147

13.1 Distance Between Two Stars 148
13.2 Visual Binaries 148
13.3 Astrometric Binaries 149
13.4 Spectroscopic Binaries 150
13.5 Photometric Binaries 152
13.6 Exercises 154

IV Solutions 155

14 Exercise Solutions 156

viii Contents

15 Problem Solutions 219

Appendix A Mathematics 268

A.1 Vectors 268
A.2 Conic Sections 271
A.3 Plane Trigonometry 276
A.4 Spherical Trigonometry 281
A.5 Special Functions 283
A.6 Derivatives 285
A.7 Integrals 291

Appendix B Kepler’s Laws 297

B.1 Solving the equation of motion 297
B.2 The orbits 299
B.3 Proof of Kepler’s third law 301

Appendix C Virial Theorem 302

Appendix D Tables and constants 304

Suggested resources 306

Index 307
Preface ix

Preface

Chances are you too have gazed in awe and wonder at the night sky, with
endless questions drifting through your mind: how did the universe begin, how
did life spark on Earth, are there any other intelligent beings in the universe
and, if so, why hasn’t anyone visited our planet? These are questions one can
hardly ignore and, while our chances of finding a definitive answer within our
lifetime are slim, we can make progress and enable future generations to do so.
This book will teach you the fundamentals of astronomy, passing on the discov-
eries of countless scientists who lived before us, giving you the tools to continue
along this endless journey. Maybe you will be writing the next chapter.
I have tried to make this book accessible to a large audience, with the only
prerequisites being basic high school mathematics and physics. Slightly more
advanced mathematical tools are covered in the appendix, making the book
self-contained. At the same time, there are plenty of exercises and problems
which will help you become familiar with the concepts explained through-
out the theory. The book is specifically aimed at students preparing for the
regional and national phases of the Astronomy Olympiads, but will still rep-
resent a valuable resource for those chosen to represent their countries at the
international competition (IOAA and IAO). While the book contains a lot of
information, you are encouraged to supplement it with online resources and
other textbooks that might cover some topics in more detail.
The book has been divided into 4 parts. Part I (Positional Astronomy) cov-
ers the celestial coordinate systems, the transformation and perturbation of
coordinates, observation and instruments, the time systems and the Moon.
Part II (Radiation Mechanisms) explains electromagnetic radiation, flux, stel-
lar magnitudes and the cosmological ladder. Part III (Celestial mechanics)
introduces gravitation and Kepler’s laws, the motion of the planets, the rocket
equation, orbital manoeuvres and binary systems. Finally, Part IV (Solutions)
presents detailed solutions to all the exercises and problems.
I decided to distinguish between exercises, which require you to apply the con-
cepts learned throughout the theory (aimed at sharpening your problem solving
skills), and problems, which will guide you through the derivation of entirely
new concepts (targeting your creative thinking). The difficulty of the exercises
and problems is denoted by asterisks, from zero (easy) to four (extremely chal-
lenging). Of course, you may disagree with my judgement of difficulty, but
I think that an arbitrary weighting scheme is better than none at all. Some
problems are extremely hard, so it is normal you will get stuck. In this case,
don’t be discouraged, just set the problem aside for some time and get back
to it later. You will find that thinking about the possible solutions over and
over again will strengthen what you learned and resolve new doubts.
x Preface

If you need some guidance, try reading the solution line-by-line (maybe cover-
ing the rest of the page with a piece of paper), pausing where you think you
can continue by yourself. I have added interesting discussions at the end of
many solutions, so be sure to read them after you solve a question. In the case
of questions with 4 asterisks, my aim was to create an almost insurmountable
challenge, that will always give you something to think about.
I have also included some exercises of various astronomy competitions, from
the national to the international level. The source is stated at the beginning
of each problem (see the next page for a list of references), so that you can
easily find it online.
At the end of the book there are four appendices. Appendix A (Mathematics)
covers the necessary concepts to understand the theory and solve most of the
exercises. My advice is that you start by studying this part, if you are not
familiar with high school mathematics. Appendix B (Kepler laws) contains a
proof of Kepler laws from Newton’s law of gravitation. I suggest you read this
only after having gone through Appendix A and Part III. Appendix C (Virial
theorem) contains a proof of the Virial theorem in its most general form. Fi-
nally, Appendix D (Tables and constants) has been included to speed up the
process of finding known constants when solving exercises and problems.
I want to thank everyone that helped proof-read the book: Alexandra Alexiu
and Amar Shah, who read the whole manuscript and made invaluable sugges-
tions, and Je Qin Chooi and Alisa Hathaway for the additional help in the more
challenging chapters. I also want to thank my parents, Lidwina and Francesco,
and my siblings, Fabiana and Federico, for reading the book and encouraging
me to finish this immense project. Despite scrupulous reviews by myself and
many other people, it is practically impossible for the book to be error-free.
If something looks amiss, visit www.astrolympiad.com for a list of typos and
updates. On the website I will also publish many other problems that, for
the sake of brevity, I could not include in the book. If you discover an error
that has not yet been published, I would be very grateful if you could report
it to me either by filling out the form on the site or by writing to the email
address [email protected]. If you have any questions or suggestions,
don’t hesitate to contact me. Finally, if you liked the book, I would be grateful
if you could leave a review.
— I hope you enjoy the book!
National Olympiads xi

National Olympiads
Part of the exercises were selected from the following national competitions:
– ARAO, the All-Russian Astronomy Olympiad;
– ArmAO, the Armenian Astronomy Olympiad;
– BAAO, the British Astronomy and Astrophysics Olympiad;
– CAO, the Canadian Astronomy Olympiad;
– CzAO, the Czech Astronomy Olympiad;
– INAO, the Indian National Astronomy Olympiad;
– INT and NAZ, the Italian regional and national Astronomy Olympiads;
– MyAO, the Malaysian Astronomy Olympiad;
– SAO, the Singaporean Astronomy Olympiad;
– ONAA, the Romanian Astronomy and Astrophysics Olympiad;
– USAAAO, the U.S.A. Astronomy and Astrophysics Olympiad.
You will also find exercises from the following international competitions:
– IAO, the International Astronomy Olympiads.
– IOAA, the International Olympiads on Astronomy and Astrophysics;
xii Contents
 Part I

Positional Astronomy

1
 1

Celestial Coordinate Systems

In this chapter we aim to study the commonly used astronomical reference

systems. In the past, man believed that the universe was confined within a
spherical shell and that the stars, fixed on this shell, were all equidistant from
the Earth, placed at the centre of the universe. Although we know today
that these assumptions are for the most part false, this simple model is still
useful today, in many aspects, as it was in the past. In fact, since most of
the measurements are taken from Earth, it is useful to place ourselves at the
centre of the universe. Furthermore, since the distance to the stars is very
large, we can neglect the orbital motion of the Earth around the Sun (which
would periodically change the coordinates of the stars) and assume that all the
stars are fixed on a sphere of infinite radius.
Under these hypotheses, for each coordinate system, only a fundamental plane
and the direction perpendicular to it remain to be defined. We begin this
chapter by defining the coordinates of a location on the surface of the Earth,
and then apply a similar reasoning to the celestial vault.

1.1 Geographical Coordinate System

It is important to define the position of an observer on Earth, as the visible

sky depends on the location from which it is observed. Unlike astronomical
coordinates, on Earth we must also define a third coordinate, that is, the height
relative to the surface.
In this coordinate system the fundamental direction is Earth’s rotation axis.
The fundamental plane, perpendicular to Earth’s axis, is called the equator.
The points where the axis of rotation meets the surface of the Earth are called
the north and south poles. Every plane parallel to the equator is called a
parallel ; every great circle passing through the poles is called a meridian.

2
1.1. Geographical Coordinate System 3

Figure 1.1: Geographic coordinates of a location X on Earth.

Let O be the centre of the Earth. Looking at Fig. 1.1, the coordinates of a
location X on Earth are:
• latitude φ, i.e. the angle between the equatorial plane and the straight
line passing through XO;
• longitude λ, i.e. the angle between the Greenwich meridian and the
straight line passing through XO;
• height h, i.e. the vertical distance of X with respect to the average sea
level.
The prime meridian is the Greenwich meridian, which passes through London.
The longitude of a place to the east or west of Greenwich is denoted by x˝ E
or x˝ W respectively, where x˝ is the angle in degrees (normalized to 180˝ ),
between the local meridian and Greenwich. The sky visible to an observer
depends on its geographical position. As we vary the longitude, the culmination
time of the stars (the instant they pass through the meridian) also varies. On
Earth, the sky appears to rotate from east to west (clockwise for an observer in
the northern hemisphere) with an angular velocity of 15˝ per hour. Therefore,
if a star in Rome culminates at 11:00 local time, in London it culminates about
an hour later, as the location is further west by just under 15˝ . As we vary
the latitude, the height of the stars on the horizon also varies. As will be seen
more in detail in the next chapter, some stars are circumpolar (always visible),
others rise and set at different times, while others permanently stay below the
horizon. At the equator, all the stars rise and set over the course of a day;
at the north pole, only half of the celestial vault is visible and all stars are
circumpolar. Finally, varying the altitude modifies the apparent horizon, to be
distinguished from the astronomical horizon, which is the line perpendicular
4 Chapter 1. Celestial Coordinate Systems

Figure 1.2: At a height h it is possible to see an additional patch of sky of angle θ “

pR{pR ` hqq. For this reason, the horizon appears to be farther away when standing on a
very tall building.

to the vertical at each location on Earth (Fig. 1.2). Assuming for the moment
that there is no refraction, the apparent horizon coincides with the geometric
horizon. When standing on a very tall building, we can effectively see an
additional patch of sky of angle θ, and therefore the number of visible stars is
greater. This phenomenon is known as the dip of the horizon. In the limiting
case of an observer at infinite height, the Earth appears as a point, i.e. in
deep space all the stars are visible. Looking back at Fig. 1.2, we see that the
following relationship holds:

pR ` hq cos θ “ R ,

Isolating θ, we find:
´ R ¯
θ “ arccos . (1.1)
R`h

1.2 The Horizontal System

The first and most natural astronomical reference system is the one centred on
the observer. In the horizontal reference system, the fundamental direction is
the vertical line at the point of observation, i.e. the direction of the plumb line
coinciding with the direction of the acceleration of gravity. The intersections
of the vertical with the celestial sphere to the north and south are called zenith
and nadir, respectively. The plane passing through the observer and tangent
to the surface of the Earth is called horizontal plane. The intersection of
the horizontal plane with the celestial sphere marks the astronomical horizon.
1.2. The Horizontal System 5

Figure 1.3: Horizontal coordinates of a star X.

Instead, we call the true horizon the one that is actually visible, which depends
on the dip of the horizon and on the geographic landscape. Evidently, the nadir
and the whole hemisphere below the true horizon are not observable. All the
great circles passing through the zenith and nadir are called verticals. The
great circle passing through the zenith (Z) and the celestial pole (P) is called
the principal vertical. Its intersection with the horizon defines the points north
(N), on the same side of the celestial pole, and south (S), at its antipode. The
vertical circle at 90˝ from the principal vertical is called the prime vertical.
Its intersections with the celestial horizon defines the points east (E) and west
(W). The points north, east, south and west are called cardinal points. We
call upper and lower culmination, respectively, the passage of a star on the
principal vertical, to the south, or on the anti-meridian, to the north.
Let X be a star on the celestial sphere and let X1 denote its projection on the
horizontal plane (Fig. 1.3). Its horizontal coordinates are:
• altitude h, i.e. the angle XX1 , counted in degrees starting from the
horizon, and ranging from 0˝ to 90˝ , if the object is above the horizon,
and from 0˝ to ´90˝ , if it is below the horizon;
• azimuth Az , i.e. the angle SX1 , counted clockwise from 0˝ to 360˝ .
In the horizontal system, each minor circle parallel to the horizon is called an
almucantar. The altitude of the celestial pole with respect to the horizon is
called astronomical latitude, because it is numerically equal to the geographical
latitude. The horizontal system offers the advantage of being intuitive and easy
to set up, but it can be impractical because both the altitude h and the azimuth
Az depend on the position and time of observation. In the next section we will
see how these issues can be overcome.
6 Chapter 1. Celestial Coordinate Systems

1.3 The Equatorial System

There are two equatorial coordinate systems: HA-dec and RA-dec. Both use
the declination δ, which can be regarded as the equivalent to the altitude h for
the horizontal system. In the HA-dec equatorial system, the second coordinate
is the hour angle H, which increases linearly with time; in the RA-dec system
it is the right ascension α, independent of time.

The First Equatorial System, or HA-dec

Figure 1.4: HA-dec coordinates of a star X.

As the name suggests, in this coordinate system the fundamental plane is

the celestial equator, perpendicular to Earth’s rotation axis. The great circles
passing through the celestial north and south poles are called hour circles, as
they all have the same hour angle. Minor circles parallel to the celestial equator
are called parallels of declination. Fig. 1.4 shows what the sky looks like to an
observer at latitude φ who uses the first equatorial system. Let X be a point
on the celestial sphere and let X1 denote its projection on the equator. In the
first equatorial system, its coordinates are:
• declination δ, i.e. the angle XX1 , counted in degrees starting from the
celestial equator, and ranging from 0˝ to 90˝ , if the object is above the
equator (in the direction of the celestial north pole), and from 0 ˝ to
´90 ˝ , if it is below the equator;
• hour angle H, i.e. the angle MX1 , counted clockwise from M and meas-
ured in hours, from 0h to 24h .
The reason why the hour angle is counted in hours, instead of degrees, will
be clarified in the next section. In this coordinate system only the hour angle
H varies with time, increasing steadily because of Earth’s rotation. Instead,
the declination δ is constant to a first approximation, since each point on the
celestial sphere moves on a parallel, and therefore at a constant distance from
1.3. The Equatorial System 7

the celestial equator. This is true only for fixed stars, while the declination of
the Sun, Moon, planets and comets is not constant.

The Second Equatorial System, or RA-dec

We now want to find a coordinate α that is completely independent of time, at
least to a first approximation. In the HA-dec system, the only time-dependent
coordinate was the hour angle. Since the rotation of the Earth is solely respons-
ible for the time variation of the hour angle, it is necessary for the reference
point from which we measure α to move in the sky with the same direction and
velocity as the (apparent) angular velocity of rotation of the celestial sphere.
Let us consider the two points of intersection between the celestial equator
and the ecliptic (the projection in the sky of the plane of Earth’s orbit). For
an observer on Earth, the celestial equator is fixed while the ecliptic rotates
from east to west (in the same direction as the rotation of the celestial sphere);
hence their intersection points rotate from east to west, with the same dir-
ection and velocity as the angular velocity of rotation of the celestial sphere.
Thus, the distances between a fixed star and the intersection points remain
constant. One of these points is called vernal equinox or first point of Aries,
because it used to be in the constellation of Aries (see Sec. 3.1), and is denoted
by the symbol P (the same symbol used for the constellation of Aries). At the
antipode of the vernal equinox, we find the autumnal equinox, denoted by the
symbol V (the same symbol used for the constellation of Libra).
Let X be a point on the celestial sphere and let X1 denote its projection on the
celestial equator. Its second equatorial coordinates are:
• declination δ, defined as for the first equatorial system;
• right ascension α, i.e. the angle P X1 , counted anti-clockwise from P and
measured in hours, from 0h to 24h .
By definition, the declination of the Sun is zero when it passes through the
points P and V. The Sun, however, only transits on the celestial equator
during the equinoxes. We can therefore define P and V as the points through
which the Sun transits during the vernal and autumnal equinoxes, respectively
(hence their names). At the vernal equinox, the declination of the Sun increases
daily (since it is approaching the summer solstice), while it decreases at the
autumnal equinox. For this reason, P and V are also called ascending and
descending nodes, respectively.
Looking at Fig. 1.5, denoting with ST the hour angle of the vernal equinox,
and with H the hour angle of the object X under consideration, the following
relationship holds:
ST “ α ` H . (1.2)
As will be seen in more detail in Sec. 5.1, ST is called the sidereal time.
8 Chapter 1. Celestial Coordinate Systems

Figure 1.5: The sidereal time is equal to the sum of the hour angle and the right ascension
of any object

Because it is a unit of time, the sidereal time is measured in hours; therefore

it is convenient to measure H and α in hours as well. A star transiting on the
meridian has an hour angle of H “ 0h , therefore a clock calibrated in sidereal
time gives us the right ascension of every object on the meridian. To find the
object in question at any other time, we need its hour angle, but this can be
found from the sidereal time and the right ascension, using Eq. 1.2.
The coordinates α and δ are constant to the first order of approximation.
However, in the long run, perturbation phenomena (chapter 3), in addition to
the proper motion of the stars, cause them to vary.

1.4 The Ecliptic System

In the ecliptic coordinate system the fundamental direction is the normal to

the plane of the ecliptic, that is, the normal to Earth’s plane of revolution
around the Sun. Let X be a point on the celestial sphere and let X1 be its
projection on the plane of the ecliptic. Its ecliptic coordinates are:
• ecliptic longitude λ, i.e. the angle P X1 , counted anti-clockwise starting
from P and measured in degrees;
• ecliptic latitude β, i.e. the angle XX1 , measured in degrees starting from X1 .
As will be seen in chapter 5, the ecliptic system is useful in defining the length
of a tropical year.
1.5. The Galactic System 9

1.5 The Galactic System

The galactic coordinate system is of great importance in extragalactic stud-

ies. Unlike other reference systems which are centred on Earth, the funda-
mental plane of the galactic system is the plane of the Milky Way, defined
as the region of maximum emission of the neutral hydrogen 21 cm line. The
galactic north pole is in the direction of the Coma galaxy cluster (αnp “
12.8h , δnp “ 27.4˝ ). The two galactic coordinates are the galactic longitude l
and the galactic latitude b, defined in a similar way to the latitude and lon-
gitude of the other systems. The origin of the galactic system is the galactic
centre (αgc “ 17h 42.4m , δgc “ ´28 ˝ 551 ).

1.6 Exercises

1. (INT 2014, Th.S, q.4) The first accurate measurement of the size of the
Earth was made by Eratosthenes of Cyrene and was obtained by meas-
uring the difference in the altitude of the Sun at the summer equinox in
two locations at a known distance. Assuming that the Earth is spherical,
what is the length of the arc between two locations at latitude `35˝ and
`45˝ (with equal longitude)?
2.* What is the angle formed by the equator with the horizon for an observer
at latitude φ?
3.* What is the condition whereby the altitude of a celestial object remains
unchanged over the course of a day?
4. What is the right ascension and declination of the vernal equinox?
5.* What is the angle formed by the ecliptic and the horizon for an observer
at the north pole?
6.* What is the condition whereby the ecliptic north pole coincides with the
zenith of an observer?
7.* At what latitudes are Mercury and Venus best seen in the evening or
morning?
8.* Where is a star located, if its right ascension is equal to its ecliptic
longitude?
9. Find the hour angle of Capella (α “ 5h 14m 28s ) at ST“ 4h 20m 13s .
10.* The Burj Khalifa, in the United Arab Emirates, is over 829m tall. In the
absence of obstacles and absorption of light, from what distance would
the skyscraper be visible?
10 Chapter 1. Celestial Coordinate Systems

11.* In January 2019 Uranus had an ecliptic longitude of 43˝ . What longit-
ude did the planet have in 1781, the year of its discovery, and which
constellation was it in?

1.7 Problems

1.** The Seasons

The hottest period in the northern hemisphere falls when the Earth is
near aphelion. Explain why the winter in the northern hemisphere is
milder than the winter in the southern hemisphere, while the summer
in the southern hemisphere is slightly warmer. Attempt a quantitative
analysis (Hint: read Ch. 8).
2.** Rotation of the Earth
An object dropped from an appreciable height does not land exactly on
the vertical, but its trajectory appears displaced in the direction of the
Earth’s rotation. This is due to the Coriolis force, or more intuitively to
the fact that the angular velocity at the launch point is the same as that
at the surface of the Earth, and therefore the tangential velocity of the
body is greater than the velocity of the surface. If the object is dropped
from the Leaning Tower of Pisa (h “ 57 m, φ “ 44˝ ), how far does it
land from where it was dropped?
3.* Foucault’s Pendulum
Because of Earth’s rotation, the plane of motion of a pendulum rotates
slowly. Prove that the precession period of Foucault’s pendulum is:
Tsid
T “ , (1.3)
sin φ
where Tsid is the time of rotation of the Earth and φ is the latitude.
4.* Dip of the horizon with refraction
So far we have ignored refraction, and drawn the line of sight to the
apparent horizon as a straight line. However, in the real world, refraction
is non-negligible: this causes the line of sight to the apparent horizon to
be curved, not straight. Assuming we can regard the ray to the horizon
as an arc of a circle, with a curvature k times the Earth’s curvature, show
that Eq. 1.1 still holds if we replace R by Reff , and find an expression
for Reff in terms of R and k.
 2

Transformation of Coordinates

In astronomy it is very useful to change from one coordinate system to an-

other. For example, it may be necessary to convert equatorial to horizontal
coordinates, to know where to find a star. This chapter may be more challen-
ging than others, however it is sufficient to keep in mind only those equations
which will be obtained in Sec. “Upper and Lower Culmination”. These can
also be derived more intuitively, as will be shown in Pr. 2.1.

2.1 The Three Equations

Since all astronomical coordinate systems measure angles on a sphere, instead

of the usual Cartesian system, in astronomy it is particularly convenient to
use spherical polar coordinates. In the Cartesian system, the position of a
point in space is determined solely by the three coordinates x, y and z. In
spherical polar coordinates, however, the position of a point P can be written
as a function of r, ψ and θ (see Fig. 2.1), where:
• r is the distance of P from the origin O;
• ψ is the angle between the x axis and the projection of r on the xy plane;
• θ is the angle between r and the z axis.
Starting from the definition above, it is possible to write x, y, z as a function of
r, ψ, θ. Looking at Fig. 2.1, we see that z “ OHz “ OP cos θ “ r cos θ. OH is
the projection of OP on the xy plane, therefore OH “ Hz P “ OP sin θ “ r sin θ.
Thus, x “ OHx “ OH cos ψ “ r cos ψ sin θ and y “ OHy “ OH sin ψ “
r sin ψ sin θ. To sum up: $
&x “ r cos ψ sin θ
’
y “ r sin ψ sin θ
’
z “ r cos θ .
%

Let xyz and x1 y 1 z 1 be two Cartesian systems, where x1 y 1 z 1 is obtained from

xyz by a rotation of angle χ around its y axis.

11
12 Chapter 2. Transformation of Coordinates

Figure 2.1: Transforming Cartesian coordinates x, y, z to spherical polar coordinates r, ψ, θ.

As can be seen in Fig. 2.2, the two coordinate systems are related by the
following equations:
$ 1
&x “ x cos χ ` z sin χ
’ (2.1a)
1
y “y (2.1b)
’
% 1
z “ z cos χ ´ x sin χ . (2.1c)
Writing xyz and x1 y 1 z 1 in spherical polar coordinates, setting r “ 1 for simplicity:
$ $ 1 1 1
’
& x “ cos ψ sin θ &x “ cos ψ sin θ
’
y “ sin ψ sin θ y 1 “ sin ψ 1 sin θ1
’ ’
% 1
z “ cos θ z “ cos θ1 .
%

Substituting the last two systems of equations in Eqs. 2.1, we find:

$ 1 1
& cos ψ sin θ “ cos ψ sin θ cos χ ` cos θ sin χ
’ (2.2a)
sin ψ 1 sin θ1 “ sin ψ sin θ (2.2b)
’
cos θ1 “ cos θ cos χ ´ cos ψ sin θ sin χ .
%
(2.2c)

These three equations will be sufficient to determine all coordinate transform-

ations. With this in mind, in the next sections we will identify to what angles
χ, ψ, θ, ψ 1 and θ1 correspond in the various celestial coordinate systems, and
then rewrite Eqs. 2.2 accordingly.
2.1. The Three Equations 13

Figure 2.2: Transformation between two Cartesian coordinate systems xyz and x1 y 1 z 1 , where
the latter is obtained from xyz by a rotation of angle χ around its y axis.
14 Chapter 2. Transformation of Coordinates

2.2 From the Horizontal to the HA-dec System

Figure 2.3: Transformation from the horizontal system to HA-dec.

We would now like to find a relationship between the angles ψ, ψ 1 , θ, θ1 , χ that

appear in Eqs. 2.2 , and the angles φ, Az , H, h, δ in Fig. 2.3. In the horizontal
and HA-dec systems, Az and H are measured clockwise, but in the previous
section we defined the angles ψ and ψ 1 anti-clockwise. Therefore ψ “ ´Az and
ψ 1 “ ´H. The angles θ and θ1 are measured starting from the z axis, while
the coordinates h and δ are measured starting from the fundamental planes,
so it is necessary to consider the complementaries: θ “ 90 ˝ ´ h and θ1 “
90˝ ´ δ. Finally, the angle between the two fundamental planes is χ “ 90 ˝ ´ φ.
Summarizing: $
’
’ ψ “ ´Az
’ 1
&ψ “ ´H
’
’
’
θ “ 90˝ ´ h
’
θ1 “ 90˝ ´ δ
’
’
’
’
’
χ “ 90˝ ´ φ .
%

Substituting the above equations in Eqs. 2.2 and using the reduction formulae
(Appendix A.3), we find:
$
& cos H cos δ “ cos Az cos h sin φ ` sin h cos φ
’ (2.3a)
sin H cos δ “ sin Az cos h (2.3b)
’
sin δ “ ´ cos Az cos h cos φ ` sin h sin φ . (2.3c)
%

The above equations allow us to change from the horizontal to the HA-dec
system. To carry out the inverse transformation, we exchange the roles of xyz
2.2. From the Horizontal to the HA-dec System 15

and x1 y 1 z 1 . In doing so, we must remember to change the sign of χ:

$
’
’ ψ “ ´H
’ 1
&ψ “ ´Az
’
’
’
θ “ 90˝ ´ δ
’
θ1 “ 90˝ ´ h
’
’
’
’
’
χ “ ´p90˝ ´ φq .
%

We then obtain:
$
& cos Az cos h “ cos H cos δ sin φ ´ sin δ cos φ
’ (2.4a)
sin Az cos h “ sin H cos δ (2.4b)
’
sin h “ cos H cos δ cos φ ` sin δ sin φ . (2.4c)
%

Since the altitude and declination are in the range [´90˝ ,90˝ ], it suffices to
know the sine of one of these angles to determine the other angle unambigu-
ously. Azimuth and right ascension, however, can take any value from 0˝ to
360˝ (or from 0h to 24h ), so we need both their sines and cosines to determine
them completely.

Upper and Lower Culmination

For an observer in the northern hemisphere, the altitude of a celestial body
is maximum when it passes through the meridian, in the southern direction.
In this case, the body is said to be in upper culmination and its hour angle
is H “ 0h . The altitude is minimum when the object is passing through the
anti-meridian. Now, the body is said to be in lower culmination and its hour
angle is H “ 12h .
In the derivation that follows, we consider only the case of an observer in the
northern hemisphere, so φ ě 0 ˝ . At the end of this section, we will generalize
the result for an observer in the southern hemisphere. When H “ 0h , using
Eq. 2.4c, we find:
sin h “ cos δ cos φ ` sin δ sin φ “ cos pφ ´ δq “ sin p90˝ ´ φ ` δq ,
where we used the cosine addition formula (Eq. A.16). Therefore:
sin h “ sin p90˝ ´ φ ` δq .
The previous equation has two solutions, since sin p180˝ ´ αq “ sin α. There-
fore, the altitude of upper culmination of an object with declination δ, meas-
ured by an observed at a latitude φ ą 0˝ , is:
#
90˝ ´ φ ` δ, if the object culminates south of zenith;
hu “ (2.5)
90˝ ` φ ´ δ, if the object culminates north of zenith.
16 Chapter 2. Transformation of Coordinates

The above formulae give the altitude of the object as measured from the nearest
horizon (north or south). However, unless otherwise stated, the altitude of up-
per culmination is measured from the southern horizon by convention, therefore
it is sufficient to use the first formula. When H “ 12h , Eq. 2.4c implies:

sin h “ ´ cos δ cos φ ` sin δ sin φ “ ´ cos pδ ` φq “ sin pδ ` φ ´ 90˝ q

Therefore:
sin h “ sin pδ ` φ ´ 90˝ q
The altitude of lower culmination, measured by convention from the northern
horizon, is then:
hl “ δ ` φ ´ 90˝ (2.6)
Eqs. 2.5 and 2.6 can be obtained using only elementary geometry, as shown in
Pr. 2.1. Let us examine some special cases:
• when hl ą 0, i.e. δ ą 90˝ ´ φ, the object is circumpolar and it never sets;
• when hu ă 0, i.e. δ ă φ ´ 90˝ , it is never visible;
• when φ ´ 90˝ ă δ ă 90˝ ´ φ, the object rises and sets.
For example, in Helsinki (φ “ 60˝ ), all stars with a declination greater than
30˝ never set and are therefore always visible. On the other hand, stars with
a declination less than ´30˝ can never be observed. Stars with a declination
between those two values rise and set. At the north pole, all stars with δ ą 0˝
are circumpolar, while those with δ ă 0˝ are never visible. At the equator,
no star is circumpolar, but all the stars rise and set. Atmospheric refraction
widens the visible horizon, making stars with declination close to δ “ 90˝ ´ φ
circumpolar and those with a declination close to δ “ φ ´ 90˝ visible.
Imagine we observe a circumpolar star and record its altitudes hu and hl (meas-
ured from the southern and northern horizon) during upper and lower culmin-
ation, respectively. Taking the average between the two altitudes, we obtain
the declination of the star:

1
δ “ phu ` hl q . (2.7)
2

If, instead, we take the difference:

1
φ “ 90˝ ´ phu ´ hl q , (2.8)
2

we obtain the geographical latitude.

2.2. From the Horizontal to the HA-dec System 17

All the previous equations also apply to the southern hemisphere if we sub-
stitute δ Ñ ´δ and φ Ñ ´φ. Indeed, by symmetry, the southern hemisphere
should be equivalent to the northern hemisphere under sign exchange (the
northern hemisphere has a positive sign by convention). Therefore, a star is
circumpolar in the southern hemisphere if δ ă p90˝ ` φq, while it never rises if
δ ą 90˝ ` φ. For example in Sydney (φ “ ´34˝ ), all stars with a declination
less than ´56˝ are circumpolar, while stars with a declination greater than 56˝
are never visible. At the south pole, all stars with δ ă 0˝ are circumpolar,
while stars with δ ą 0˝ are never visible.

Rising and Setting Times

We want to obtain the rising and setting times of an object, that is, the two
moments when its altitude is zero. Let’s assume we know its right ascension
is α. To calculate the sidereal time, with the equation ST “ α ` H, the only
information we are missing is its hour angle. Starting from Eq. 2.4c, we isolate
cos H (for cos δ, cos φ ‰ 0˝ ):

sin h
cos H “ ´ tan δ tan φ ` .
cos δ cos φ
If we neglect atmospheric refraction, we can substitute h “ 0˝ :

cos H “ ´ tan δ tan φ . (2.9)

The previous equation allows us to obtain the hour angle at the moment of
rising and setting, knowing the latitude of the observer and the declination
of the object. Knowing its right ascension, it is then possible to obtain the
sidereal time using Eq. 1.2, and therefore also the local time (see Ch. 5).
Atmospheric refraction increases the visible horizon by about ´341 . In the
case of extended bodies like the Sun and the Moon, we must also consider
their angular radii.

Duration of Day and Night

During the vernal and autumnal equinoxes the declination of the Sun is δ@ “
0˝ , therefore, from Eq. 2.9, we find that the hour angles of rising and setting
are H “ ´90˝ and H “ `90˝ , which are the coordinates of east and west,
respectively. Hence, during the equinoxes, the Sun rises exactly in the east
and sets exactly in the west, and the day lasts 12h everywhere on Earth. From
Eq. 2.9, we see that at the equator, the Sun rises exactly in the east and sets
exactly in the west every day of the year, thus the durations of day and night
are always the same. At the north pole, the altitude of the Sun is always equal
to its declination, hence the Sun is circumpolar and the day lasts 24h when its
declination is positive, whereas the Sun never rises and it is permanently night
18 Chapter 2. Transformation of Coordinates

when its declination is negative.

Using Eq. 2.9, we can obtain the duration of the day for a location at a latitude
φ, when the Sun’s declination is δ@ . The Sun is visible for an hour angle of
2 arccos p´ tan δ tan φq, or converted into hours:
” 1 ı
tday “ 24h ¨ 1 ´ arccos ptan δ @ tan φq , (2.10)
180˝
where we used arccos p´xq “ 180˝ ´arccos x, since cos p180˝ ´ xq “ ´ cos x. In
the case when | tan δ@ tan φ| ą 1, the last identity cannot be used since arccos x
is not defined for arguments greater than unity. In practice, this means that
there is no point of rising and setting, i.e. the Sun is either circumpolar or
never visible. Let us prove this statement.
If tan δ@ tan φ ą 1, then tan δ@ ą 1{ tan φ and, using 1{ tan φ “ tan p90˝ ´ φq,
we find tan δ@ ą tan p90˝ ´ φq. In turn, this implies that δ@ ą 90˝ ´ φ, which
is the condition for the Sun to be circumpolar in the northern hemisphere.
Thus, if tan δ@ tan φ ą 1, the Sun is circumpolar and the day lasts 24h .
Similarly, if tan δ@ tan φ ă ´1, then tan δ@ ă ´1{ tan φ and, using ´1{ tan φ “
tan pφ ´ 90˝ q, we have tan δ@ ă tan pφ ´ 90˝ q. This implies that δ@ ă φ ´
90˝ , hence the Sun never rises. Therefore, if tan δ@ tan φ ă ´1, the Sun
never rises and the duration of the day is 0h . To sum up, in the northern
hemisphere:
1
$ ” ı
h
’
’ 24 ¨ 1 ´ arccos ptan δ@ tan φq if φ ´ 90˝ ď δ@ ď 90˝ ´ φ
& 180˝
tday “ 24h if δ@ ą 90˝ ´ φ
’
’
% h
0 if δ@ ă φ ´ 90˝ .

Similarly, for the southern hemisphere, tan δ@ tan φ ą 1 implies δ ą 90˝ ` φ,

which is the condition for a star to never rise. Instead, if tan δ@ tan φ ă ´1,
then δ@ ă ´p90˝ ` φq, so the Sun is circumpolar. Therefore, in the southern
hemisphere:
1
$ ” ı
h ˝ ˝
&24 ¨ 1 ´ 180˝ arccos ptan δ@ tan φq if ´ p90 ` φq ď δ@ ď 90 ` φ
’
’
tday “ 24h if δ@ ă ´p90˝ ` φq
’
’
% h
0 if δ@ ą 90˝ ` φ .
2.3. From The Ecliptic to the RA-dec System 19

2.3 From The Ecliptic to the RA-dec System

Figure 2.4: Transformation from the ecliptic system to RA-dec.

The equations for changing between the ecliptic and the RA-dec coordinate
systems can be derived in a similar way to the equations converting horizontal
to HA-dec coordinates. From Fig. 2.4, we see that:
$
’
’ ψ “ α ´ 90˝
’ 1 ˝
&ψ “ λ ´ 90
’
’
’
θ “ 90˝ ´ δ
’
θ1 “ 90˝ ´ β
’
’
’
’
’
%
χ “ .
Substituting in Eqs. 2.2, we find:
$
& sin λ cos β “ sin δ sin  ` cos δ cos  sin α
’ (2.11a)
cos λ cos β “ cos δ cos α (2.11b)
’
sin β “ sin δ cos  ´ cos δ sin  sin α . (2.11c)
%

To obtain the inverse transformation, we exchange the coordinates:

$
’
’ ψ “ λ ´ 90˝
’ 1 ˝
&ψ “ α ´ 90
’
’
’
θ “ 90˝ ´ β
’
θ1 “ 90˝ ´ δ
’
’
’
’
’
%
χ “ ´ .
20 Chapter 2. Transformation of Coordinates

Hence, we obtain:
$
& sin α cos δ “ ´ sin β sin  ` cos β cos  sin λ
’ (2.12a)
cos α cos δ “ cos λ cos β (2.12b)
’
sin δ “ sin β cos  ` cos β sin  sin λ . (2.12c)
%

Declination of the Sun throughout the year

Figure 2.5: Declination of the Sun throughout the year starting from the day of vernal
equinox, in the approximation that the Earth’s orbit is circular.

The Sun lies on the ecliptic, hence β@ is constantly equal to zero. Furthermore,
neglecting the eccentricity of the Earth’s orbit, its ecliptic longitude is equal
to λ@ “ p2π{Tt qt, where Tt is the duration of the tropical year (see Ch. 5)
and t is the time elapsed since the last vernal equinox (when λ@ is zero by
definition). Then, using Eq. 2.12c with β@ “ 0˝ , we find:
” ´ 2π ¯ı
sin δ@ “ sin  sin λ@ ñ δ@ “ arcsin sin  sin t . (2.13)
Tt
This equation allows us to compute the declination of the Sun at any given
time of the year, in the approximation that the Earth’s orbit is circular (Fig.
2.5). In particular, we can verify that Eq. 2.13 gives the correct result in the
following cases:
• vernal equinox (t “ 0), sin p2π{Tt ¨ tq “ 0 and δ@ “ 0˝ ;
• summer solstice (t “ Tt {4), sin p2π{Tt ¨ tq “ 1 and δ@ “ ;
• autumn equinox (t “ Tt {2), sin p2π{Tt ¨ tq “ 0 and δ@ “ 0˝ ;
• winter solstice (t “ 3{4Tt ), sin p2π{Tt ¨ tq “ ´1 e δ@ “ ´.
2.4. Exercises 21

2.4 Exercises

1. What is the condition required for a star with declination δ to pass

through the zenith of an observer at latitude φ?
2. At what latitude can you see Sirius (δ “ ´16˝ 421 582 ) graze the horizon
during lower culmination?
3.* What can the maximum and minimum altitudes of the Sun be, in a city
at latitude φ?
4. You are given the upper (hu “ 53 ˝ 541 452 ) and lower (hl “ 34 ˝ 341 232 )
altitudes of culmination of a star. Find the latitude of the observer and
the declination of the star.
5. At what latitudes is the Sun circumpolar at least once a year?
6. (CAO 2018, Th., q.6) On a sunny afternoon on the 22nd of June, an
observer standing upright on a flat surface found that his shadow had a
length equal to his height. At what latitude was the observer located?
7.* (IOAA 2016, Th., q.4) An observer in the northern hemisphere noticed
that the length of the shortest shadow of a 1.000 m vertical stick on a day
was lmin “ 1.732 m. On the same day, the length of the longest shadow
of the same vertical stick was measured to be lmax “ 5.671m. Find the
latitude φ of the observer and the declination of the Sun δ@ on that day.
Assume the Sun to be a point source and ignore atmospheric refraction.
8.* What must the declination of a star be and where should it be observed
so that its azimuth does not change for at least half a day?
9.* What are the ecliptic coordinates of a star with right ascension α “ 6h
and declination δ “ 10 ˝ ?
10. What are the hour coordinates of the four cardinal points and the zenith
for an observer located in the northern hemisphere? What about an
observer in the southern hemisphere?
11. Where do stars which stay above the horizon for more than 12 hours a
day rise?
12. (INT 2011, Th.S, q.1) Marco and Gianna are two young astronomy enthu-
siasts from Rome. Gianna, who is visiting an important foreign capital,
calls Marco to greet him. At one point Gianna remarks “It is a beauti-
ful evening, the Moon is rising right now!”. Marco then replies: “Great!
Here the Moon is passing exactly on the meridian, so I know precisely at
what longitude you are!”. If Marco’s longitude in Italy is 12 ˝ 301 E, what
22 Chapter 2. Transformation of Coordinates

is the longitude of Gianna? Can you tell which city she is in? Neglect
the effect due to the finite distance of the Moon from the Earth (lunar
parallax).
13.* Estimate the declination of the Sun on the 21st of April. At what latit-
udes is the Sun circumpolar?
14.* A star passes through the zenith at 0h 10m of sidereal time, whereas its
altitude on the horizon is 78˝ 121 at 9h 2m of sidereal time. Calculate the
latitude of the observer.
15.* You want to spend a romantic evening with your partner, so you take
them to the beach to see the sunset with a mechanical lift. The elevator
can be raised to a height of 6 m, at a speed of v. What is the minimum
v so that the sunset can be observed twice, once at sea level, and the
second time at 6 m?

2.5 Problems

1.* Upper and Lower Culmination

Obtain the formulae for the upper and lower culmination of a star in two
different ways using only elementary geometry.
In the first case, consider an inertial frame in two instants separated by
12h . The direction of the star and of the Earth’s axis are fixed, but the
observation point has rotated by 180˝ .
In the second case, consider the frame of the observer. Now the ob-
server is fixed, while the star appears to be moving in a circle around the
direction of the Earth’s axis.
 3

Perturbation of Coordinates

Even though a star can appear to be stationary with respect to the Solar Sys-
tem, its coordinates change over time due to various perturbation effects. Of
course, the altitude and the azimuth are constantly changing, but the declin-
ation and the right ascension are also slowly varying.

3.1 Precession

The Sun and Moon, as well as the other planets in the Solar System, orbit very
close to the ecliptic plane and tend to pull Earth’s equatorial bulge towards
this plane. The Earth reacts to this effect by rotating like a top around the axis
perpendicular to the ecliptic. As a consequence, the Earth’s axis traces out a
pair of cones, with a period of 25765 years. This motion is called precession of
the equinoxes since the vernal equinox, moving clockwise along the ecliptic by
50” per year, anticipates the date of the next equinox. As a result, the ecliptic
longitude of all bodies on the celestial sphere increases by 50” each year.
In our century, Earth’s rotation axis points very close to Polaris (α Ursae
Minoris), about a degree away. In 3000 BC, Earth’s axis pointed towards the
fainter Thuban, in the constellation of Draco. In about 12, 000 years, it will
point in the direction of Vega, which will then become the new pole star. In-
stead, the celestial south pole currently sits in a region of sky particularly clear
of bright stars.
The equatorial coordinates are also affected by precession. To adjust for
this variation, stellar catalogues must be updated every century. Currently,
most maps and catalogues give the coordinates of stars at the time (or epoch)
J2000.0, which corresponds to January 1, 2000.
We now derive the expressions that relate the change in declination and right
ascension to the annual variation of ecliptic longitude of a celestial object,
which is stationary with respect to the Solar System. We assume that only
λ, δ and α vary over time (β is fixed because it is perpendicular to λ, the direc-
tion along which the precession proceeds). This is a good approximation, since
 changes much more slowly than the other factors involved (see Sec. 3.7).
23
24 Chapter 3. Perturbation of Coordinates

Consider Eq. 2.12c:

sin δ “ sin β cos  ` cos β sin  sin λ .

Differentiating both sides:

cos δ dδ “ cos β sin  cos λ dλ .

Substituting cos β cos λ “ cos α cos δ (Eq. 2.12b), we find:

dδ “ dλ sin  cos α . (3.1)

This equation relates the variation dδ in declination to the variation dλ in

ecliptic longitude. If, instead, we differentiate Eq. 2.12b:

´ sin α cos δ dα ´ cos α sin δ dδ “ ´ cos β sin λ dλ .

Substituting dδ “ dλ sin  cos α (Eq. 3.1), it follows that:

sin α cos δ dα “ dλ pcos β sin λ ´ sin  cos2 α sin δq ,

which, according to Eq. 2.11a, can also be written as:

sin α cos δ dα “ dλ psin δ sin  ` cos δ cos  sin α ´ sin  cos2 α sin δq .

Substituting cos2 α “ 1 ´ sin2 α (Eq. A.12, Appendix A.3) and simplifying:

dα “ dλ psin α sin  tan δ ` cos q . (3.2)

This equation relates the variation dα in right ascension to the variation dλ
in ecliptic longitude. Thus, given that dλ{dt “ 502 {year, the annual changes
dδ{dt and dα{dt for an object with declination δ and right ascension α are
given by the equations:
$
’ dδ “ dλ sin  cos α
’
& (3.3a)
dt dt
’ dα dλ
’
% “ psin α sin  tan δ ` cos q . (3.3b)
dt dt

The International Astronomical Union (IAU) divides the sky into 88 official
constellations with precise boundaries, so that each point on the celestial sphere
belongs to one and only one constellation. The constellations visible from
northern latitudes mainly draw their origins from the Greek culture, and their
names recall various mythological figures. During the vernal and autumnal
equinoxes, the Sun passes through the points P and V, respectively. At the
3.1. Precession 25

time of the ancient Greeks this happened when the Sun was in the constella-
tion of Aries and Libra, respectively (hence the use of these symbols). Today
however, due to the precession of the equinoxes, the Sun passes through the
zodiac constellations with a delay of about a month (corresponding to one
constellation), compared to the convention adopted at the time of the Greeks.
Hence, the points P and V have now moved to the adjacent constellations of
Pisces and Virgo, respectively. To be precise, we should be using the symbols
f and ` for the vernal and autumnal equinoxes, but convention stuck with
the tradition of the ancient Greeks. It might amuse you that astrology makes
an error of about a month in associating your zodiac sign with the actual pas-
sage of the Sun through your constellation. In Tab. 3.1, the current dates of
the Sun’s passage through the zodiacal constellations are listed.

Capricornus d
20 January Ñ 16 February
Aquarius e
16 February Ñ 11 March
Pisces f
11 March Ñ 18 April
Aries P
18 April Ñ 13 May
Taurus ]
13 May Ñ 21 June
Gemini ^
21 June Ñ 20 July
Cancer _
20 July Ñ 10 August
Leo 
10 August Ñ 16 September
Virgo `
16 September Ñ 30 October
Libra V
30 October Ñ 23 November
Scorpius b
23 November Ñ 29 November
Ophiuchus
29 November Ñ 17 December
Sagittarius c
17 December Ñ 20 January
Table 3.1
26 Chapter 3. Perturbation of Coordinates

3.2 Nutation

Nutation is the oscillatory motion of Earth’s rotation axis in the direction

perpendicular to precession. Nutation is due to the fact that the precession
angular momentum is added to that of the rotation and, consequently, the
resulting angular momentum is not exactly directed along the axis of symmetry
of the rotating object. The main component of nutation is associated to the
regression of the Moon’s nodal line and has the same period of 18.6 years. The
amplitude of this motion is very small and in the order of 9.2 ”.

3.3 Aberration

Figure 3.1: a) The angle of aberration is equal to the ratio of the tangential component of
the velocity of the observer (v sin θ) to the velocity of light (c). b) Due to aberration, stars
appear to be moving on ellipses whose eccentricity increases with decreasing ecliptic latitude.

Due to the finite speed of light, all stars appear slightly displaced in the direc-
tion of the observer’s motion. Therefore, as a consequence of Earth’s orbital
motion, we see a periodic change in the apparent position of the stars, called
aberration. In the framework of special relativity, it can be shown that, for
small velocities (v ! c), the aberration angle is approximately:
v
A“ sin θ , (3.4)
c
where A is measured in radians and θ is the angle between the true position
of the body and the direction of the observer’s velocity. On Earth, the largest
value of aberration is v{c “ 202 , called the constant of aberration. Aberration
associated with the rotation of the Earth is called diurnal aberration, but is
usually negligible (only 0.32 in amplitude) compared to annual aberration.
3.4. Parallax 27

Looking at Fig. 3.1 a), one can intuitively justify Eq. 3.4. Imagine a light
ray propagating to a telescope. The projection of the velocity of the telescope
along the direction perpendicular to the ray is v sin θ, whereas the velocity
of the ray is c. Hence, stars appear displaced by an angle equal to the ratio
of these velocities. Fig. 3.1 b) shows that, due to aberration, stars move on
ellipses whose eccentricity increases with decreasing ecliptic latitude. Stars on
the ecliptic poles describe a circle; those on the ecliptic, a segment.

3.4 Parallax

When looking at an object from different angles, we see it in different direc-

tions. This phenomenon, known as parallax, leads to a periodic variation of
the coordinates of celestial objects during the year. The angle of parallax is
related to the object’s distance from Earth, therefore by measuring the angle
of parallax we can obtain its distance. Parallax will be discussed in more detail
in Ch. 9.

3.5 Refraction

Light is refracted by the atmosphere, therefore the apparent position of celestial

objects differs from the true one. Even though refraction varies significantly
with temperature and pressure, we can assume that conditions are roughly
constant at the same height, i.e. that the refractive index of air depends only on
the height above ground. Near the zenith, Earth’s curvature is negligible and
we can therefore divide the atmosphere into infinitesimal layers of air parallel
to each other. Let α be the angle at which the ray enters the atmosphere and
β the angle at which we observe it (measured from the zenith). The refractive
index of a medium is defined as the ratio of the speed of light in vacuum to
the speed of light in the medium. The refractive index of space (effectively
a vacuum) is ns “ 1, whereas the refractive index of Earth’s atmosphere at
sea level is about n0 “ 1.0003. The refractive index is always greater than
unity. A ray of light that travels through a medium of refractive index ni and
is incident on a surface of refractive index nr , with angle αi measured relative
to the normal, is refracted by an angle αr , such that:

sin αr ni
“ . (3.5)
sin αi nr

This equation is called Snell’s law. Looking at Fig. 3.2, we apply Eq. 3.5 to a
light ray entering the atmosphere:
28 Chapter 3. Perturbation of Coordinates

Figure 3.2: Refraction of a light ray passing through the atmosphere.

ns sin α “ nk sin βk ,
nk sin βk “ nk´1 sin βk´1 ,
... ,
n1 sin β1 “ n0 sin β .

Or, equivalently:
sin α “ n0 sin β ,
where we have used ns “ 1. Let R “ α ´ β be the angle of deviation. Close
to the zenith, R ! 1 rad, therefore sin R « R and cos R « 1, giving:

α“R`β
ñ sin α “ sin pR ` βq “ sin R cos β ` cos R sin β
« R cos β ` sin β .

Substituting sin α “ n0 sin β in the last equation, we find:

R “ pn0 ´ 1q tan β . (3.6)

The above equation is only valid close to the zenith, where Earth’s curvature
can be neglected. At the horizon R « 341 , slightly larger than the angular
diameter of the Sun. When the lower part of the Sun grazes the apparent
horizon, the Sun has already set on the astronomical horizon.
3.6. Proper Motion 29

3.6 Proper Motion

Figure 3.3: The proper motion changes the shape of the constellations. a) Big dipper today,
b) in 30000 years and c) in 60000 years.

Many stars move in directions that seem to be constant in time. This effect
is due to the relative motion between the Sun and the star, and is therefore
called proper motion. Due to this motion, the shape of the constellations
changes slowly with time, as shown in Fig. 3.3.
The speed of a star relative to the Sun can be divided into two components,
one radial and one tangential. The tangential component is responsible for
the proper motion, which can be measured by taking pictures of the star at
different moments in time. The tangential motion can be further divided into
the component that changes the right ascension (uα cos δ), and the component
that changes the declination (uδ ). The factor cos δ appears because a star with
declination δ moves on a parallel of declination, which has a radius of cos δ.
Therefore, if the rate of change of the right ascension measured on the celestial
equator is uα , the velocity on the parallel of declination is actually uα cos δ.
To sum up, the total angular velocity of a star is:
b
u “ u2α cos2 δ ` u2δ . (3.7)

The tangential velocity is equal to the angular velocity multiplied by the dis-
tance of the star:
vt “ u ¨ r . (3.8)
Barnard’s star, with an angular velocity of about 10.32 per year, is the star
with the largest proper motion observed so far. While it is usually necessary to
observe a star for a few years to measure its proper motion, the radial motion
can be easily determined with a single observation, taking advantage of the
Doppler effect (Sec. 7.6). The radial velocity can be determined from the
displacement of the spectral lines z “ ∆λ{λ, according to Eq. 7.20:

p1 ` zq2 ´ 1
vr “ c, (3.9)
p1 ` zq2 ` 1
or, in the non-relativistic limit:

vr “ c z . (3.10)
30 Chapter 3. Perturbation of Coordinates

Hence, the total linear velocity of the star is:

b
v “ vr2 ` vt2 . (3.11)

3.7 Other Cycles

In addition to the relatively fast motions described above, there are other slower
ones. Among these, the most important are the variation of the inclination of
Earth’s axis and the cycle of eccentricity.
The axis of the Earth is inclined by 23˝ 271 , on average, with respect to the
ecliptic, but this varies from a minimum of 22.1˝ , to a maximum of 24.5˝ , with a
periodicity of 41,000 years. This effect therefore adds to that of nutation, which
instead has a much shorter period of 18.6 years. Currently, the inclination
of the Earth’s axis relative to the ecliptic is decreasing at a rate of 472 per
century. Finally, the eccentricity of Earth’s orbit varies from essentially zero
(0.000055) to mildly eccentric (0.0679). The components of this cycle combine
into a 100,000-year period. The present eccentricity is 0.0167 and is decreasing.
The eccentricity varies primarily due to the gravitational attraction exerted by
Jupiter and Saturn; however, the semi-major axis remains unchanged.

3.8 Exercises

1. How long does it take the Barnard to travel a distance in the sky equal
to the angular diameter of the Moon?
2. Aldebaran’s proper motion is u “ 202 {yr and its parallax is πp “ 0.0482 .
The spectral line of iron λ “ 440.5 nm is shifted by 0.079 nm towards
red. What are the radial and tangential velocities of the star? (Hint: use
Eq. 9.3).
3.** The coordinates of Sirius at the epoch J1900.0 were α “ 6h 40m 45s ,
δ “ ´16˝ 351 , and its proper motion had components uα “ ´0.037 s/a,
uδ “ ´1.122 /a. Find its coordinates at the epoch J2000.0. Take into
account the precession of Earth’s axis.
4.* The parallax of Sirius is πp “ 0.3752 and its radial velocity is ´8 km/s
(i.e. directed towards the Earth). By using the data from the previous
exercise, find its tangential velocity. How long will it take Sirius to reach
the closest point to the Sun? What will the values of its tangential and
radial velocities be at that time? What about its parallax?
5.* (ArmAO 2019, Th., q.1) The atmospheric refractive index of a planet
3.9. Problems 31

varies with the height from the surface according to:

n0
nphq “ ,
1 ` h
where n0 and  are constants. A ray emitted in the horizontal direction
indefinitely circles the planet at any height (if we neglect absorption by
the atmosphere). What is the radius of the planet?
6.** (IAO 2014, Th.α, β, q.4) Hydroplanet consists of a rocky “core” of radius
R and a thick layer of water surrounding it from all sides. Local humans
live at the bottom of this world’s ocean (i.e. on the surface of the “core”),
and the hydrosphere is an analogue of our atmosphere for them. Local
scientists observe astronomical objects from the bottom of the ocean.
Like on Earth, the duration of the day-night period on Hydroplanet is
24h .
• Find the minimum depth h of the ocean, for which celestial bodies
will be visible at the horizon.
• What will the duration of the day be for inhabitants of the planet’s
equator? The disk of the central star can be considered as a point
source of light.
• Calculate the value of “atmospheric” refraction at the horizon on
such a planet. The outer surface of the ocean is smooth, with no
waves or ripples.

3.9 Problems
1.**** Precession of the Equinoxes
Assuming the Earth to be an ellipsoid of revolution with moment of
inertia along its axis Ip and equator Ie , such that Ie {pIp ´ Ie q « 300,
estimate the period of precession of the equinoxes. Also prove that the
influence of the Moon is approximately double that of the Sun.
 4

Observation and Instruments

Until the end of the Middle Ages, the eye was the most important instrument
for the observation of the sky. With the aid of different devices it was pos-
sible to record the position of celestial bodies. The telescope was invented in
the Netherlands at the beginning of the 17th century, but only in 1609 was
it first used, by Galileo Galilei, to conduct astronomical observations. Since
then, great strides have been made in both the quantity and quality of obser-
vations. While, initially, observation was limited to within the Solar System,
the telescope now allowed observations of celestial objects hundreds of parsecs
away. Since Galileo, we have enjoyed 400 years of exponential improvement
and growing sophistication of the tools at our disposal. In 1800, William Her-
schel discovered that visible light is only a small part of the electromagnetic
spectrum. Soon enough, a variety of new tools were developed that would
extend the study of the night sky to previously inaccessible regions of the
electromagnetic spectrum.

4.1 Angular Size

Since all celestial objects appear infinitely far away for an observer on Earth,
their first distinctive feature is the angular size. Let h0 be the diameter of the
object and d its distance from the observer. Its angular size is then:
h0
tan α0 “ . (4.1)
d
For small α0 (which is always the case when observing celestial objects), we
can use the approximation tan α0 « α0,rad , so that:
h0
α0,rad “ , (4.2)
d
where α0,rad is measured in radians. To convert this angle to degrees, we simply
multiply by 360˝ {2π (Eq. A.11, Appendix A.3). Hence:
180˝ h0
α0,deg “ . (4.3)
π d
32
4.2. The Thin Lens Equation 33

4.2 The Thin Lens Equation

Figure 4.1: An object is placed at a distance p from a lens with refractive index n, and
surfaces with radii of curvature of R1 and R2 . The image is formed at a distance q (the
thickness of the lens is exaggerated). In this case, the curvature has opposite direction on
either side of the lens, hence R1 is positive and R2 is negative.

Let us place an object at a distance p from a lens with refractive index n, and
surfaces with radii of curvature of R1 and R2 (Fig. 4.1). We would like to know
at what distance q its image is formed. By applying Snell’s law (Eq. 3.5) to a
ray leaving the object and refracting through the first and second surfaces of
the lens; it is possible to prove that (pr. 4.2), in the limit where the thickness
of the lens tends to zero:
1 1 ´ 1 1 ¯
` “ pn ´ 1q ´ . (4.4)
p q R1 R2
If we define the focal length f as:
1 ´ 1 1 ¯
“ pn ´ 1q ´ , (4.5)
f R1 R2
we obtain the thin lens equation:

1 1 1
` “ . (4.6)
p q f

Eq. 4.5 is called the lens maker equation, because it can be used to determine
the values of R1 and R2 needed for a given refractive index and a desired focal
length f . Taking p Ñ 8, i.e. 1{p Ñ 0, Eq. 4.6 reduces to:
1 1
“ ñ q“f. (4.7)
f q
Therefore, the image of a distant object is formed in the focal plane of the lens.
34 Chapter 4. Observation and Instruments

Ray Diagrams for Thin Lenses

Figure 4.2: Rays are drawn to understand the type of image that is formed. In a) the image
is real and inverted, in b) it is virtual and upright.

Ray diagrams are very convenient for locating images formed by thin lenses or
systems of lenses. The general procedure is to draw three rays, whose paths
are particularly simple to sketch (Fig. 4.2):
• ray 1 is drawn parallel to the principal axis. After being refracted by the
lens, this ray passes through the focal point on the back side of the lens;
• ray 2 is drawn through the focal point on the front side of the lens and
emerges parallel to the principal axis;
• ray 3 is drawn through the centre of the lens and continues in a straight
line.
Properties or rays 1 and 2 follow from Eq. 4.7, since an object at infinity
(whose light rays are incident on the lens parallel to the principal axis) has its
image in the focal point. Property of ray 3 follows from symmetry, since both
sides of the lens are identical. An example of application of this principle is
shown in Fig. 4.2. In a), the image is real and inverted because it forms on
the back side of the lens and it is upside-down. In b) instead, the rays would
diverge. It is therefore necessary to consider their continuation (dashed lines)
on the same side of the object. In this case the image is called virtual, since it
cannot be captured on a screen.

4.3 Optical Telescopes

The term telescope usually refers to the optical telescope, operating in the
visible region of the electromagnetic spectrum. However, there are many
other telescopes that detect different wavelengths, from gamma rays to radio
waves.
4.3. Optical Telescopes 35

Figure 4.3: Galileo’s refracting telescope.

Figure 4.4: Newton’s reflecting telescope.

In general, every telescope has two important functions in astrophysics:

• it collects a greater amount of light by using a larger surface than the
naked eye, making it possible to observe fainter objects;
• it increases the angular size of the object and the angular resolution.
Two fundamentally different types of optical telescopes exist: the refracting or
Galilean telescope and the reflector or Newtonian telescope. In order to focus
the image of a celestial object, the former uses a set of lenses (Fig. 4.3), taking
advantage of the phenomenon of refraction. The latter uses a set of mirrors,
according to the phenomenon of total reflection (Fig. 4.4).
The operation of a Galilean telescope is shown in Fig. 4.5 (overleaf). Light
from a distant object is incident on the first lens, called the objective, and is
focused at a distance fo , equal to the focal length of the objective (the ob-
ject is essentially at infinity). The objective forms a real and inverted image
close to the second lens, called the eyepiece. To provide the largest possible
magnification, the rays leaving the eyepiece must be parallel, hence the im-
age created by the objective must be located at the focal point of the eye-
piece, at a distance fe from the eyepiece. Thus, the distance between the
lenses must be exactly d “ fo ` fe . This is the length of the telescope tube.
The main parameters of a telescope are:
• D, diameter of the objective;
• fo , focal length of the objective;
• fe , focal length of the eyepiece.
36 Chapter 4. Observation and Instruments

Figure 4.5: The object has angular size α0 . When viewed from the eyepiece of a telescope,
its image has an angular size of α1 .

Let α0 be the angular size of a celestial object. The linear dimension h1 of its
image formed on the focal plane of the objective is:
h1
tan α0 “ ñ h1 “ fo ¨ tan α0 .
fo
In the approximation α0 ! 1, then:

h1 “ fo ¨ α0, rad . (4.8)

For example, the angular size of the Sun is approximately α0 “ 301 « 0.0087 rad.
Its linear size on the focal plane of a telescope with focal length fo “ 15 m is
h1 “ 0.0087 ¨ 15 m“ 13 cm. If we are given the distance d and dimension h0
of an object, we know its angular size α0,rad « h0 {d. Substituting in Eq. 4.8:

h1 h0
“ . (4.9)
fo d

4.4 Angular Resolution

An important feature of a telescope is its angular resolution. To understand

the physical mechanism that constrains the resolution of a telescope, let us
consider in detail the image of a point source formed by a telescope.
When two light rays of equal amplitude interfere in phase with each other, the
resulting ray has twice the amplitude of the original rays. However, when two
rays interfere out of phase, the resulting amplitude is zero. Clearly, intermedi-
ate cases are also possible, so that the resultant amplitude can be any number
between zero and two times the amplitude of the original rays. Because of
the periodic nature of a wave, we expect two light rays to interfere construct-
ively if the difference in their path lengths is any integer number of wavelengths
(which corresponds to being completely in phase); destructively if the difference
4.4. Angular Resolution 37

in their path lengths is any half integer (completely out of phase). Because of
its wavelike property, light is diffracted by the circular aperture of a telescope,
giving rise to a visible diffraction pattern (Fig. 4.6). At the centre of this
pattern, all light rays have travelled the same distance from the aperture, and
are therefore in phase with each other. They interfere constructively, giving
rise to a bright spot. Moving to the side in any direction, light rays need to
travel slightly different paths, and therefore reach the new point with slightly
different phases. Hence, light intensity is reduced at the sides.

Figure 4.6: a) Image of a star as viewed with a telescope (often only the inner circle is
visible). b) Intensity of the diffraction pattern.

At some particular distance from the centre, all light rays are out of phase,
hence the intensity is zero. Moving further away from the centre, the phase
difference will again be an integer wavelength at some other distance, leading
to constructive interference and to a bright circular annulus. For this reason,
point sources (like stars) produce, on the focal plane of a telescope, a diffrac-
tion pattern of alternating dark and bright zones, called Airy disks. In fact,
this is the same phenomenon observed by Thomas Young in his double slit ex-
periment. The general physical theory that predicts the shape of the pattern
is called Fraunhofer diffraction. The effect of this diffraction depends on the
ratio between the wavelength of light and the diameter of the telescope.
In order for two objects to be resolved, the central bright spots of their images
must not overlap. In more general terms, according to Rayleigh’s criterion,
the distance between the two central maxima must be greater than the dis-
tance between the first maximum and first minimum of either image. The
cross-section of the diffraction pattern of two point sources is shown in Fig.
4.7 (overleaf). In particular, in a) the two sources cannot be resolved (they
appear as a single source), since the distance d between their first maxima is
less than the distance dmin . On the contrary, in b) the sources can be just
resolved, since d is equal to dmin . It can be shown (pr. 4.3) that the angular
distance between the central bright spot and the first minimum is 1.22λ{D,
where D is the diameter of the objective and λ the wavelength. It then follows
38 Chapter 4. Observation and Instruments

Figure 4.7: According to Rayleigh’s criterion, two sources are resolved when the distance
between the two central maxima is greater than the distance between the first maximum and
first minimum of either image. Hence, a) cannot be resolved, whilst b) can be just resolved.

that the angular resolution is given by:

λ
R « 1.22 , (4.10)
D

where R is measured in radians. This formula is valid for all light across the
electromagnetic spectrum. Often, the resolution of Earth-based telescopes is
limited by the conditions of the atmosphere, rather than by the instrument
itself. This effect is called seeing and generally does not allow for a better res-
olution than a few arcseconds. If the telescope collects images on photographic
plates, the resolution is further reduced by the finite size of the emulsion grains.
These typically range between 10 to 30 micrometers. For a focal length of 1
m, the scale is 1 mm “ 2062 , so 0.01 mm corresponds to about 2” The size of
a pixel in CCDs (charge-coupled devices) can be as low as 3 ´ 5 micrometers,
which corresponds to around one arcsecond. Therefore, the CCD is not usually
the limiting factor in the resolution of an Earth-based telescope.

4.5 Magnification

The angular magnification (ω) is defined as the ratio between the angular size
of the image formed by the eyepiece and the angular size of the object. Looking
at Fig. 4.5:

h1
tan α0 “ « α0,rad ,
fo
h1
tan α1 “ « α1,rad .
fe
4.5. Magnification 39

Hence the magnification is:

α1,rad fo
ω“ « . (4.11)
α0,rad fe

For example, if the objective has focal length fo “ 100 cm and we use an
eyepiece with a focal length of fe “ 2 cm, the magnification is ω “ 100{2 “ 50x.
The magnification is not a fundamental property of a telescope, since it depends
on the eyepiece, which can easily be changed.

Minimum Magnification

Figure 4.8: If we were to use a magnification lower than the minimum, the diameter of the
outgoing light would be larger than the diameter of our pupil, which means losing photons.

The minimum magnification is obtained from the condition that the diameter
of the outgoing light from the eyepiece is less than or equal to the diameter of
our pupil (Fig.4.8). Thus:
fe D
děD¨ “ .
fo ω
Hence, the minimum magnification is:
D
ωmin “ , (4.12)
d
where D and d are the diameters of the objective and of our pupil, respectively.
At night, the diameter of the pupil is typically 8 mm, therefore the minimum
magnification of a telescope with diameter 100 mm is approximately 13x.

Maximum Magnification
The maximum magnification ωmax is the largest magnification that is worth
using with a telescope of diameter D. Its value is equal to the ratio of the
angular resolution of the human eye (e « 21 “ 5.8 ¨ 10´4 rad) and the angular
resolution (R) of the telescope:
e eD D
ωmax “ « « .
R λ 1 mm
40 Chapter 4. Observation and Instruments

Thus:
D
ωmax “ . (4.13)
1 mm
For example, the maximum magnification of a telescope with aperture 100 mm
is 100x. There is no point in using a magnification greater than the maximum,
since the resolution of the image would then be worse than the resolution of
our eyes.

4.6 Aperture Ratio

Consider an object with angular size α0 , viewed with a telescope of diameter

D and focal length fo . As we saw in Sec. 4.3, the linear size of its image on
the focal plane of the telescope is:

h1 “ fo tan α0 .

Assuming the image is circular, the area it covers is:

A “ πh21 “ πfo2 tan2 α0 .

Let F be the incoming light flux from the object. The power collected by the
telescope is:
D2
P “F ¨ .
4
Therefore, the intensity of the image is:

P F D2
I“ “ ¨ .
A 4 tan2 α0 fo2

For a given object, the intensity depends only on D and fo :

D2
I9 .
fo2

We define the aperture ratio Rf as:

fo
Rf “ . (4.14)
D
Hence, the perceived brightness of an object is directly proportional to the
reciprocal of the aperture ratio squared:

1
I9 . (4.15)
Rf 2
4.7. The Role of the Atmosphere 41

Telescopes with a long focal length and a small diameter are therefore slower
than telescopes with a short focal length and a large diameter. The lower the
speed of the telescope, the longer the exposure time required to obtain an image
of equal brightness on a CCD. For this reason, telescopes with a lower aperture
ratio are preferred in photography. The aperture ratio is usually denoted by
f {Rf . Therefore, a telescope with an aperture ratio of Rf “ 8 is written as
f {8. For fast telescopes, the focal ratio can be as low as f {1 ´ f {3, but is
usually smaller and around f {8 ´ f {15.

4.7 The Role of the Atmosphere

Astronomical observation from the surface of the Earth has to deal with the
atmosphere. Not only is light subject to dispersion, absorption and re-emission,
but the effect of the atmosphere is strongly dependent on the wavelength, as
well as the time and local conditions. The atmosphere reduces the brightness
of an object (Sec. 8.4) and the maximum possible resolution. In the best
conditions, the maximum possible resolution is a few arcseconds. Without any
modifications, a 10 m telescope would have the same angular resolution as
a 20 cm amateur telescope. Actually, the resolution of the former would be
worse, because of the distortion caused by the mirror’s own weight. Active
optics are used to correct the distortion of the mirror; adaptive optics are used
to correct atmospheric turbulence. In both cases, a thin deformable rubber
layer with tens or hundreds of thousands of piezoelectric crystals is positioned
under the mirror. They behave like small actuators, expanding or contracting
when a potential difference is applied to their ends. To correct the effect of
the atmosphere, these crystals make adjustments on the order of micrometers,
a few hundred times per second. To determine the necessary corrections, the
telescope automatically observes a reference star near the object to be observed.
In the event that a sufficiently bright star does not exist near the object of
interest, a carefully calibrated laser can be used to excite sodium atoms at an
altitude of 90 km, thus allowing the required calibration.

4.8 Exercises

1. What is the smallest angle that can be resolved at a wavelength of 500 nm

by the Keck telescope, which has a mirror with a diameter of 10 m? What
should be the size of a radio telescope, observing at the wavelength of
20 cm, if its resolution is the same?
2. A spiral galaxy at a distance of D “ 55¨106 ly, is observed perpendicular
to its galactic plane and has an angular diameter of α “ 4002 . What is
the diameter of the galaxy in light years? How large is its image on the
focal plane of a telescope with a focal length of f “ 1 m?
42 Chapter 4. Observation and Instruments

3.* Calculate the apparent maximum and minimum angular sizes of the Sun
from Mars. Suppose Mars has a satellite with the same orbital character-
istics as the Moon. What should the minimum diameter of this satellite
be in order for solar eclipses to be visible from Mars?
4. Compute the ratio of the angular diameters of Mars when seen in oppos-
ition and quadrature.
5. A telescope is pointed at a coin of diameter 16.25 mm, which is at a
distance of 15.1 ˘ 0.1 m. The diameter of the image of the coin formed
on the focal plane of the telescope is 1.35 mm. Find the focal length fo .
6.* (NAZ 2016, Th. S, q.3) On the evening of the 20th of April 2016, in Milan,
it is possible to observe three stars of equal apparent magnitudes, with
coordinates: Star 1 (α2016 “ 6h 30m , δ2016 “ `35˝ 201 ), Star 2 (α2016 “
6h 30m , δ2016 “ 34˝ 401 ) and Star 3 (α2016 “ 6h 24m , δ2016 “ `35˝ 201 ).
The stars are observed with a telescope of diameter D “ 200 mm and
an aperture ratio of f/10. A camera is placed on the focal plane of the
telescope with a CCD of dimension 4096x4096 pixels. Each pixel is a
square with sides lpix “ 6.4µm. Is it possible to obtain an image in
which Star 1 and 2 appear together? What about Star 1 and 3 ?
7.* (CAO 2018, Th., q.4) Amateur astronomers use following method to
estimate the field of view (FOV) of their telescopes: they locate a star
with known declination, adjust the telescope so that the star crosses the
field of view along its diameter, measure the time taken for the star
to cross the FOV and, hence, estimate the FOV. For a telescope (d“
400 mm, f“ 4000 mm) the crossing time of αAur (δ “ 46˝ 01 14.42 ) is 2.5
minutes. Can the Moon be seen in full through this telescope?
8.* (USAAAO 2020, Th.3, q.1) An astronomer used his f/5 telescope with a
diameter of 130 mm to observe a binary system. He is using an eyepiece
with a field of view of 45˝ and a focal length of 25 mm. In this system,
star A and B have masses of 18.9 and 16.2 solar masses, and apparent
magnitudes of 9.14 and 9.60, respectively. The period of the system is
108 days, and its distance from the Solar System is 2.29 kpc. The binary
system has an edge-on orbit relative to the Solar System. Find the FOV
of the telescope, its limiting magnitude, its angular resolution and the
angular separation between the stars. Will the astronomer be able to
observe both stars as distinct points in the telescope? (Hint: read Chs.
8 and 10).
9.* (SAO 2019, Th.1, q.3) An achromatic lens is used to ensure that light
of different wavelengths have the same focal lengths, i.e. to correct chro-
matic aberration. An achromatic lens was made by combining plano-
4.9. Problems 43

convex (one side flat and one side convex) and plano-concave (one side
flat and one side concave) lenses, made from two different types of glasses,
A and B. These two glasses have the following refractive indices: nA,red “
1.48, nA,blue “ 1.50, nB,red “ 1.66, nB,blue “ 1.70. Find the radii of
curvature RA , RB necessary to produce a combination equivalent to a
converging lens with focal length of 600 mm (Hint: use Eq. 4.5, taking
into account that A and B have different refraction indices).

4.9 Problems

1.* Combination of thin lenses

Show that if two lenses with focal lengths f1 and f2 are combined, they
behave like a single lens with an effective focal length of:
1 1 1
“ ` . (4.16)
feff f1 f2

2.** Lens maker’s equation

Consider two transparent media having indices of refraction n1 and n2 ,
where the boundary between the two media is a spherical surface of radius
R1 and index of refraction n2 . When an object is placed at a distance p
from the boundary (inside n1 ), an image is formed at a distance q (inside
n2 ). By applying Snell’s law, show that:
n1 n2 n2 ´ n1
` “ . (4.17)
p q R1
Now, imagine that the other side of the lens has a curvature of R2 . The
image formed by the first boundary acts as the object for the second
boundary. In the approximation that the thickness of the lens is much
smaller than the radii of curvature, prove Eq. 4.4.
3.**** Airy Disk
Show that the amplitude dψp of a wavelet reaching the diffraction screen
at P“ px0 , y0 q, which is a distance r away from the aperture element
(x, y), is proportional to:
eikr
dψp 9 hpx, yq dxdy ,
r
where hpx, yq is the aperture function, which is equal to unity if px, yq
belongs to the aperture and zero otherwise; and k “ 2π{λ. In the ap-
proximation that L " x0 , y0 , x, y (Fig. 4.9 on the next page), show that
the total amplitude at P is approximately;
ż
x0 x`y0 y
ψp 9 hpx, yqe´ik R dxdy , (4.18)
44 Chapter 4. Observation and Instruments

Figure 4.9: Fraunhofer diffraction

where R2 “ L2 ` x20 ` y02 . Now consider a circular aperture of diameter

D. Using the identities:
ż 2π
e´iqρ cos φ dφ “ 2πJ0 pqρq , (4.19)
0

and ż qD{2
qD ´ qD ¯
qρJ0 pqρq dpqρq “ J1 (4.20)
0 2 2
where J1 and J0 are the first and zero-order Bessel functions of the first
kind respectively, show that:
a
J1 pπD{λ sin2 θ ` sin2 ψq
ψp 9 a (4.21)
πD{λ sin2 θ ` sin2 ψ

where sin θ “ y0 {L and sin ξ “ x0 {L. Knowing that the first zero of
J1 pxq is at x “ 3.8317, prove Eq. 4.10. What is the diffraction pattern
of an elliptical aperture, with major axes 2a and 2b?
 5

Time systems

There are two essentially different ways to measure time. The first is based
on the Sun, and is therefore called solar time. The second is based on the
rotation of the Earth, and is called sidereal time. They each have their own
advantages. Solar time is best adopted for day-to-day time-keeping, as our
lives are heavily based on the day-night cycle. In the next section we will show
that the sidereal day is approximately 4 minutes shorter than the solar day. If
calendars were to adopt sidereal time, after about half a year, clocks would lag
behind the Sun by 12 hours: we would be working at night and sleeping during
the day (as if this doesn’t already happen). On the other hand, sidereal time is
particularly useful in astronomy. As explained in Sec. 1.3, sidereal time allows
us to find the position of any celestial object whose right ascension is known.
What this also means is that the celestial vault appears to rotate clockwise (in
the same direction as the apparent motion of the Sun) by about 4 minutes (or
1˝ ) every day. This is why most stars are only visible at certain times during
the year, and every season is characterized by different stars. For instance,
for an observer in the northern hemisphere, the most distinctive feature of the
summer sky is the Summer Triangle, whose defining vertices are Altair, Deneb,
and Vega (in the constellations of Aquila, Cygnus, and Lyra, respectively).
Instead, in winter, you should immediately recognize the constellation of Orion,
or alternatively, the winter triangle, formed by Sirius, Procyon and Betelgeuse
(in Canis Major, Canis Minor and Orion, respectively).

5.1 Sidereal and Solar Day

The solar day is the time it takes the Sun to return to the same meridian. The
length of the day measured in this way is called true solar day. The duration
of the true solar day is equal to 86400 s on average, varying from a minimum
of 86378 s (mid-September), to a maximum of 86430 s (end-December).
The sidereal day is the time it takes a fixed star to return to the same meridian,
which is also equal to the period of rotation of the Earth about its axis. In
Sec. 1.3, the sidereal time was defined as the hour angle of the vernal equinox.
45
46 Chapter 5. Time systems

Figure 5.1: The solar day is longer than the sidereal day: after completing a revolution about
its axis, the Earth needs to rotate by an additional degree (i.e. by the same amount it covered
in its orbit around the Sun), in order to see the Sun at the same meridian. This additional
rotation takes 4 minutes, which is the difference between the solar and sidereal day.

Since P is fixed with respect to the stars, the sidereal day can also be defined
as the time between two consecutive passages of the vernal equinox on the
same meridian.
The lengths of the solar and sidereal days are different. After a sidereal day, the
Earth has covered about one degree of its orbit around the Sun (or, more pre-
cisely, 360 ˝ {365.25 “ 0.986˝ ). Therefore, after completing a rotation about its
axis, the Sun does not appear exactly at the meridian, but the Earth needs to
rotate an additional degree (Fig. 5.1). This takes around 4 minutes (24h {360),
therefore the solar day is about 4 minutes longer than the sidereal day. How-
ever, the difference between a solar and a sidereal day is not constant. It
depends on the eccentricity of the Earth’s orbit and on the declination of the
Sun. In any case, the solar day is always longer than the sidereal day, since
the orbital and rotational motions proceed in the same direction.
Now, let us carry out a more rigorous analysis. Let ω be the angular velocity of
the Earth about its own axis and Ω the angular velocity of revolution around
the Sun. Since the orbital and rotational motions proceed in the same direction,
an observer on Earth sees the Sun moving clockwise with angular velocity ωr “
ω´Ω. Let tsol be the duration of the mean solar day, tsid the length of the sider-
eal day and trev the period of Earth’s orbit around the Sun (equal to a year).
Since ωr “ 2π{tsol , ω “ 2π{tsid and Ω “ 2π{trev , we have:
1 1 1
“ ´ . (5.1)
tsol tsid trev
5.1. Sidereal and Solar Day 47

Isolating tsid , substituting trev “ 365.25 tsol and tsol “ 86400 s, we find:

tsol ¨ trev
tsid “ « 86164.1 s “ 23h 56m 4.1s of mean solar time .
tsol ` trev

The mean solar day is therefore 3m 56s longer than the sidereal day. The
sidereal day can be taken as a constant to a very good approximation, but
in the long term it also varies, mainly due to the tidal force exerted by the
Moon. As a consequence, the Moon gains energy, moving away from the Earth
(at a current rate of 3.8 cm each year), whilst the Earth’s rotation is slowed
down. If the Earth–Moon system continued its motion undisturbed, after a
sufficiently long time, the Earth would always show the same side to the Moon,
i.e. the rotation and the orbital motions would be synchronised. There are a
few examples of celestial bodies in the Solar System in which perfect synchrony
has already been established. One of the most widely known examples is the
dwarf planet Pluto and its largest moon, Charon.
Earlier we showed that one sidereal day, i.e. 24h “ 86400s of sidereal time, is
equal to 23h 56m 4.1s “ 86164.1s of mean solar time. We can then write two
equations to convert from sidereal to solar time and vice-versa:

86164.1
tsol “ tsid ñ tsol « 0.99727 tsid , (5.2)
86400
and:
86400
tsid “ tsol ñ tsid « 1.002738 tsol . (5.3)
86164.1
Indeed, using Eq. 5.2, we see that a sidereal day (tsid “ 24h “ 86400 s of
sidereal time) is equal to 86164.1 s of solar time. Conversely, using Eq. 5.3 we
find that a solar day (tsol “ 24h “ 86400 s of solar time) is equal to 86636.5 s
of sidereal time.

Estimating Sidereal Time

Since the sidereal time is defined as the hour angle of the vernal equinox, it
follows that ST“ 0h when P passes through the meridian. During the vernal
equinox, the Sun coincides with P, therefore they culminate at the same time.
Since the Sun culminates at 12 : 00 of local time, it follows that:

ST “ Tt ` 12h ,

where Tt is the true local solar time. This equation has the accuracy of a few
minutes. Since the sidereal day is shorter than the solar day by 3m 56s , after
n solar days from the vernal equinox, the sidereal time will be:

ST “ Tt ` 12h ` p3m 56s q ¨ n .

48 Chapter 5. Time systems

Often we only know the value of Tm , i.e. the mean local solar time. Measuring
Tt would require a sundial or a similar instrument, which I bet you don’t carry
in your pocket. In this case we can use the equation of time Tt “ Tm ` E.T.
(Sec. 5.2), knowing that, during the vernal equinox, E.T. “ ´7m :

ST “ Tm ` 11h 53m ` p3m 56s q ¨ n . (5.4)

The vernal equinox usually falls on the 21st of March, but it can vary in either
direction by about two days. For this reason, unless we are given the day of
vernal equinox, it is not possible to compute the sidereal time with an accuracy
greater than 5 minutes. When talking about the time at a given place, we
usually refer to the (mean solar) local time, therefore it is more likely you will
use Eq. 5.4.

5.2 Equation of Time

The length of the solar day varies throughout the year. Due to the eccentricity
of its orbit, Earth’s velocity varies from a maximum at perihelion to a minimum
at aphelion. Since the daily angle covered by the Earth around the Sun is
proportional to the Earth’s velocity, the true solar time is longer at perihelion
and shorter at aphelion. Another factor that periodically changes the length of
the true solar day is the obliquity of the ecliptic. Since Earth’s rotation axis is
perpendicular to the celestial equator, we are only interested in the projection
of the Sun’s motion along this plane. Let dλ@ {dt be the daily angle covered by
the Sun, when its declination is δ@ , as viewed by an observer on Earth. First,
let us work out the projection of the Sun’s motion on the parallel of declination
δ@ . This is just pcos { cos δ@ q dλ@ {dt: indeed, it is dλ@ {dt during a solstice
(locally the Sun moves along the parallel), and cos  dλ@ {dt during an equinox
(when the Sun’s path forms an angle  with respect to the parallel, which is
the equator in this case). We now need to project this motion on the celestial
equator. This adds a factor of 1{ cos δ@ , i.e. the ratio of the circumference
of the celestial equator to that of the parallel (see Sec. 3.6). Therefore, the
projection of the Sun’s motion on the equator is proportional to cos { cos2 δ@ .
The eccentricity gives a maximum deviation of around 5 s per day, while the
obliquity of the ecliptic is the dominant factor and gives rise to a maximum
deviation of around 20 s per day.
Our clocks are based on the mean solar time, that is, on a fictitious Sun that
moves on the celestial equator at a constant angular speed. By definition, the
average solar time Tm is equal to the hour angle Hm of the fictitious Sun plus
12 hours (so that every day starts at midnight):

Tm “ Hm ` 12h . (5.5)
5.2. Equation of Time 49

Figure 5.2: The equation of time (E.T.) gives the difference between the mean solar time
and the true solar time. The value of E.T. varies throughout the year, due to the eccentricity
of the Earth’s orbit and the obliquity of the ecliptic.
50 Chapter 5. Time systems

Figure 5.3: The analemma on the local meridian, as seen by an observer at an intermediate
latitude in the northern hemisphere.

We can also write the same equation for Tv and Hv . The equation of time,
denoted by E.T., is the difference between the mean solar time Tm and the
true solar time Tt :
E.T. “ Tt ´ Tm . (5.6)
During a year, the small differences between each true and mean solar day
sum up to significant values of E.T.: from a maximum of 16 minutes on the
2nd of November, to a minimum of -14 minutes on the 11th of February. The
equation of time is zero four times per year: on the 15th of April,14th of June,
1st of September and 24th of December.
If we take a picture of the Sun every day throughout the year, at the same
mean solar time, and superpose these images, we see that the Sun follows a
long and slender figure-of-eight path, with one lobe much larger than the other,
called the analemma (Fig. 5.3). The vertical component of this motion is easy
to explain: since the declination of the Sun changes during the year, its height
on the horizon varies accordingly, being lower in winter and higher in summer.
What about the horizontal component? At the same mean solar time, the hour
angle of the true Sun changes during the year in accordance with the equation
of time (Eq. 5.6), therefore its horizontal position varies accordingly. For an
observer in the northern hemisphere, the Sun is located in the highest part
of the analemma during the summer solstice, in the lowest part during the
winter solstice and in the part halfway in between (which does not correspond
to the point where the lobes meet) during the vernal and autumnal equinoxes.
The easternmost and westernmost points are occupied by the Sun when the
equation of time is minimum and maximum, respectively. The Sun passes
through the four points halfway in between the horizontal extrema when the
equation of time is zero.
5.3. Local Solar Time 51

5.3 Local Solar Time

Both the average and true solar times are local times, i.e. they depend on the
hour angle of the Sun, either real or artificial. The time on our watch is likely
to be different from the true solar time (measured, for example, by a sundial).
For centuries, the length of the day was directly related to the Sun’s passage on
the meridian. This meant that each place had its own local solar time, which
could vary significantly within the same country. Clearly, our clocks must be
based on a different time, one that does not vary throughout the year and that
is possibly the same within the same region or country.

5.4 Greenwich Mean Time

As railways progressed across each country, the time difference between cities
became problematic. The United Kingdom, the most industrialized European
nation at that time, decided to standardize time within its country according
to the time marked by the clock in Greenwich, central London. This time
standard became known as the Greenwich Mean Time (GMT).
The Greenwich Mean Time was formally replaced by the Universal Time (UT)
in 1928 (although the title has not yet entered common usage), but was es-
sentially the same as GMT until 1967, when the definition of the second was
changed. At the beginning, the universal time was obtained directly from
observations (UT0). After correcting for the perturbation motions of the
Earth’s rotation axis (precession, nutation, motion of the poles), we obtain
UT1. This system also suffers some irregularities, caused by small variations
in the Earth’s rotation speed. These disturbances have at least two periods,
one year and half a year. By removing these effects we obtain UT2, with
a precision of about 10´7 s. UT2 is also not exact, due to the decrease in
the Earth’s velocity of rotation, caused by the tidal force with the Moon.
For this reason, astronomers have abandoned the rotation of the Earth as
a method of measuring time. In 1967, the one-second duration was defined
as 9192631770 times the period of oscillation of Caesium 133 that transits
from the hyperfine level F “ 4 to F “ 3. Later, this definition was modi-
fied to include small relativistic effects due to the Earth’s gravitational field.
The precision of atomic time is approximately 10´12 s.

5.5 Time Zones

In 1884, the time zone system was proposed, according to which time is de-
termined only for the 24 main meridians, separated in longitude by 15˝ . A
time zone is defined as the sector on the Earth’s surface delimited by the two
great circles at a distance of 7.5˝ from a main meridian. However, the bound-
52 Chapter 5. Time systems

aries of each time zone coincide with the theoretical ones only in the open sea,
while, within continents, they are modelled by borders. As a consequence of
the rotation of the Earth from west to east, every time zone to the west of
Greenwich is an hour of GMT earlier; to the east, an hour later. Imagine we
leave Greenwich at 12:00 GMT on the 1st of January, and travel west at an
arbitrarily large speed. Since we travel through 24 time zones before finally
landing back in Greenwich, it looks like we need to turn our watch back by 24h .
In this case, our watch would then read 12:00 GMT, 31st of December, which
is clearly a paradox, since no time has actually elapsed. To avoid this, the time
zone system requires a date change at a specific meridian. Furthermore, when
crossing this meridian from east to west, we need to decrease the date by one;
if crossing from west to east (as in the previous example), increase it by one.
By convention, this line roughly follows the Greenwich anti-meridian (the main
meridian 180˝ from it), and is called the International Date Line (IDL). If you
will, the IDL is a branch cut that makes time continuous. In many places, the
Date Line follows exactly the Greenwich anti-meridian, in others it deviates
to avoid crossing inhabited areas. The islands of the republic of Kiribati, in
the archipelago of the equatorial Sporades, is the place with the farthest time
zone from Greenwich (GMT + 14:00).

5.6 Daylight Saving Time

The Daylight Saving Time was introduced for the purpose of reducing elec-
tricity consumption, by taking advantage of daylight during summer. It is
generally adopted from 2:00 a.m. on the first Sunday of spring and involves
setting the clocks forward by one hour, compared to standard time. It returns
to standard time at 2:00 a.m. on the first Sunday of autumn.

5.7 Measuring the Length of a Year

Sidereal Year
The sidereal year is the time it takes the Sun to return to the same position with
respect to the fixed stars. It is equal to Earth’s orbital period, given by Kepler’s
third law (Sec. 10.3). The duration of the sidereal year is 365.2564 average
solar days (365 days, 6 hours, 9 minutes and 10 seconds). The sidereal year is
not based on the relative position of the Earth and Sun; therefore, if we were to
use the sidereal year in our calendars, after a sufficiently long time the seasons
would no longer fall in the same months of the year, but would shift by about
twenty minutes per year due to the precession of the equinoxes. The period of
revolution of a celestial body is usually expressed in sidereal years.
5.8. Calendars 53

Tropical or Solar year

The tropical year or solar year is the time it takes the Sun to return to the
same position relative to the Earth. In other words, the tropical year is the
time interval between two consecutive passages of the Sun through the vernal
equinox. Because it is based on the motion of the Sun, the tropical year is
adopted by calendars. The tropical year is equal to 365.2422 average solar
days (365 days, 5 hours, 48 minutes and 46 seconds), but is subject to small
perturbations, the most significant of which is due to the Moon (pr. 6.5).
The length of the tropical year can be obtained from the length of the sidereal
year (which we simply call year) and the precession period of the equinoxes
(Tpr “ 25765 years). Since the orbital and precessional motions proceed in
opposite directions, the angular velocity of the Sun with respect to the vernal
equinox is given by ωr “ ωyr ` ωpr . Hence:
1 1 1
“ ` ,
Ttr Tyr Tpr
Tpr,yr
Ttr “ « 0.9999612 Tsid “ 365.2422 mean solar days .
1 ` Tpr,yr

5.8 Calendars

Our calendar is the result of a long evolution. In general, a calendar must have
two properties:
• the number of days in a year must be an integer, since it would be
inconvenient for some years to start at night, others during the day;
• the average duration of a year must be close to that of the tropical year,
so that the seasons always fall in the same period of the year.
The main problem calendars have to deal with is the incompatibility between
the units of the day, month and year, since the number of days and months
in a year are not integers. Our calendar draws its origins from the Roman
calendar which, in its original form, was based on the phases of the moon.
Around 700 BC, the calendar was based on the apparent motion of the Sun
around the Earth and had a duration of twelve months. However, each month
was equal to a lunar cycle, so the duration of the year was 354 days and the
calendar was out of phase with the seasons. Then, under Julius Caesar, the
duration of the year was extended to 365 days. To better synchronize the
year with the seasons, it was also decided to add one day to every fourth year
(leap year ). Therefore, in the Julian calendar, the average duration of the year
was 365 days and 6 hours. Even this calendar was not perfect: the tropical
year is shorter by about 11 minutes and 14 seconds, so that, after 128 years,
54 Chapter 5. Time systems

the Julian calendar was already lagging behind by one day. The difference
already amounted to 10 days in 1582, when Pope Gregory XIII proposed a
new calendar. In the Gregorian calendar, every fourth year is a leap year.
However, years divisible by 100 are never leap years except when they are
also divisible by 400. For example, 1900 was not a leap year, but 2000 was.
The Gregorian calendar was adopted slowly, at different times by different
countries, and the transition period continued until the twentieth century. As
you might have guessed by now, even the Gregorian calendar is not perfect:
the difference with the tropical year accumulates every year and, in about 3300
years, it will amount to one day.

5.9 Exercises

1. Is the difference in longitude between two places equal to the difference

in sidereal or solar time?
2. Find the local time in Rome (λ “ 12˝ 281 54.482 E) at 12:00 GMT. What
is the difference with the time shown by a clock?
3.* When in Greenwich the local time is 10h 17m 14s , in Naples it is 11h 14m 15s .
What longitude is Naples located at?
4.** Compute the hour angle of the true Sun Ht for an observer at a place
with longitude 128˝ 151 W on the 10th of February 2003, at 08h 46m 22.0s
GMT, knowing that the equation of time that day was E.T. “ ´14m 13s .
5. On what day of the year are the solar and sidereal times the same?
6. This evening, a star rises at 22:00. At what time will it rise in a month?
7.* The sidereal time in Greenwich on the 18th of February 2003, at 0h GMT,
was 9h 50m 12s . At what time of GMT did a star with right ascension
18h transit the meridian that day?
8.* What is the best month of the year to observe (around midnight) a star
with right ascension α “ 10h ?
9.* What is the sidereal time in Rome (λ “ 12˝ 281 54.482 E) at 12:00, on the
13th of April 2000? That year, the vernal equinox occurred at 7 : 35
GMT on the 20th of March.
10.* The coordinates of Arcturus are α “ 14h 15.7m and δ “ 19˝ 1.11 . Find
the sidereal time when the star sets in Boston (φ “ 42˝ 191 , λ “ 71˝ ). At
what (meridian) time does Arcturus set on the 10th of January?
11. What would the relation between solar and sidereal time be if the Earth
rotated about its axis in the opposite direction?
5.10. Problems 55

12.* What is the accumulated error in the Gregorian calender after 100, 500
and 1000 years?
13.** (IAO 2012, Th.α, β, q.1) Recently, on the 6th of June 2012, an infrequent
astronomical phenomenon, the transit of Venus across the solar disc, took
place. The next transit of Venus will take place only in 2117. Calculate
the date of that transit (Hint: read Ch. 11 first).

5.10 Problems

1.* Effect of ecliptic obliquity on the equation of time

Using the equations obtained in Ch. 2, prove that:
dα cos  dλ@
“ . (5.7)
dt cos2 δ@ dt

2.*** Equation of time

Knowing that the equation of an ellipse in plane polar coordinates with
centre in the rightmost focus is (see Pr. 10.12):

ap1 ´ e2 q
rpθq “ , (5.8)
1 ` e cos θ
applying conservation of energy and angular momentum, find the tan-
gential velocity of the Earth at each point in its orbit. Hence, prove that
the daily angle covered by the Earth around the Sun is:
c
dθ GM p1 ` e cos θq2
“ . (5.9)
dt a3 p1 ´ e2 q3{2

Therefore, compute the daily change in the Sun’s right ascension dα@ {dt,
in terms of θ, γ, , e, where γ is the angle of the vernal equinox measured
from the direction of perihelion,  is the obliquity of the ecliptic and e is
the eccentricity of Earth’s orbit. Finally, using Mathematica/Matlab or
similar, compute the equation of time. Sketch the analemma, observed
at a mean solar time Tm , from a place at latitude φ (Hint: read Ch. 10).
 6

The Moon

The Moon is Earth’s only natural satellite. It orbits the Earth anti-clockwise
(as seen from the north pole) in a slightly elliptical orbit with eccentricity
e “ 0.055 and semi-major axis a “ 384400 km. Its orbital plane is inclined by
5˝ 91 with respect to the ecliptic, which it intersects along the node line. The
period of revolution of the Moon around the Earth is the same as the period
of rotation about its axis, hence the Moon always shows the Earth its same
side. However, due to its orbital inclination and eccentricity, it is possible to
see 59% of its surface at one time or another. This effect is called libration,
and will be discussed in detail in Sec. 6.5.
The Moon reflects only 8% of the light incident upon it (comparable to a lump
of coal), making it one of the least reflective objects in the Solar System. Even
with the naked eye, one can clearly see two distinct regions on the visible side of
the Moon (the near side, as opposed to the far side): light areas, called terrae
(or highlands), and darker areas, called maria. The maria were once thought
to be seas and oceans, and were given beautiful names such as Oceanus Pro-
cellarum, Mare Tranquillitatis and Sinus Iridum (the Ocean of Storms, the Sea
of Tranquillity and the Bay of Rainbows). When the far side of the Moon was
first photographed by the Soviet probe Luna 3 in 1959, a surprising feature was
its almost complete lack of maria. Today we know that the maria are made
up of basaltic rocks, formed by volcanic activity, which seems to be prevalent
on the near side. The surface of the Moon is covered in lunar regolith. It has
a thickness of about 3–5 m in the maria regions and 10–20 m in the terrae.
The Moon has been studied more than any other body in the Solar System; it
has been imaged from above by lunar orbiters, and its surface studied by a num-
ber of landers – the first being Lunar 9 in 1965, followed by the Russian lunar
rovers and the NASA Surveyor craft. Lunar exploration culminated in the
NASA Apollo programme, when six spacecraft landed men on the Moon.

56
6.1. Sidereal Month 57

6.1 Sidereal Month

The sidereal month is the time interval it takes the Moon to return to the
same position with respect to the fixed stars. The duration of this period is
27.32166d , i.e. 27d 7h 43m 12s . The most evident lunar period is, however, the
synodic month, related to the so-called lunation.

6.2 Synodic Month

The synodic month is the time interval it takes the Moon to return to the same
position relative to the Sun. After a synodic month the Earth has travelled
about 27˝ in its orbit around the Sun, therefore the Moon must travel that
same (additional) angle around the Earth, before being aligned again with the
Earth and the Sun. Since the Moon covers about 13˝ every day, the synodic
period is about two days longer than the sidereal one.
Following the same strategy adopted in Sec. 5.1, to compute the relation
between sidereal and solar days, let us consider the relative angular velo-
city between the Moon and the Sun. Denoting with ω$ and ω@ the angu-
lar velocities of the Moon and the Sun, respectively, their relative angular
velocity is ωr “ ω$ ´ ω@ . Substituting ω$ “ 2π{Tsid , ω@ “ 2π{Tyr and
ωr “ 2π{Tsyn :
1 1 1
“ ´ .
Tsyn Tsid Tyr
Solving for Tsyn :

Tsid Tyr 365.2564 ¨ 27.32166

Tsyn “ “ « 29.53059d “ 29d 12h 44m 3s .
Tyr ´ Tsid 365.2564 ´ 27.32166

This is the mean duration of the synodic month. Due to the eccentricity
of Earth’s orbit, the synodic month varies from a minimum of 29.27d , near
summer solstice, to a maximum of 29.84d , near winter solstice. A lunar year
comprises 12 mean synodic months, and has a duration of 354d 8h 48m . In the
lunar calendar, for convenience, the year is assigned an integer number of days.
To better approximate the real period, one day is added every third year, so
that 2 years with 354 days are followed by one with 355 days. The error is
about 48 minutes every year.

Lunar Phases
The lunar phases result from the different positions occupied by the Moon,
relative to the Sun, in its rotation around the Earth. Every lunation is char-
acterized by four important moments: new moon, first quarter, full moon and
last quarter.
58 Chapter 6. The Moon

New moon occurs when the Moon is in between the Earth and the Sun. During
new moon, the Moon appears completely dark and it rises and sets together
with the Sun. Approximately seven days after new moon, the Moon is in quad-
rature with the Earth and the Sun; this moment is called first quarter. Only
the western part of the Moon is illuminated — the part to the east being com-
pletely dark. The Moon rises around noon and sets around midnight, lagging
behind the Sun by six hours. Another seven days later, the Moon is aligned
with the Sun and the Earth, on the same side of the Earth, and appears com-
pletely illuminated. During a full moon, the Moon rises at 18:00 (sunset), and
sets at 6:00 (sunrise), therefore it precedes the Sun by 12 hours. Finally, in the
last quarter, the Moon is again in quadrature, appearing bright to the east and
dark to the west. The Moon now rises at midnight and sets at noon, making
it 6 hours ahead of the Sun. The entire cycle lasts one synodic month.

6.3 Draconic Month

The draconic month is the time interval it takes the Moon to complete a full
revolution with respect to the node line. This period is slightly shorter than
the synodic month due to the precessional motion of the node line, which
is in the opposite direction to the orbital motion of the Moon. The value
of the draconic month (Tdra ) can be calculated from the precessional period
(Tpr “ 18.61 years) and from the sidereal period (Tsid ):

1 1 1
“ ` ,
Tdra Tsid Tpr
Tpr Tsid
ñ Tdra “ « 27.2122d “ 27d 5h 5m 34s .
Tpr ` Tsid

Instead, the draconic year is useful for calculating eclipses. It is defined as the
time between two consecutive passages of the Sun through the same lunar node.
This period can be determined from the relative angular velocity between the
Sun and the node line. Let Ttr be the duration of the tropical year, and Tpr
the period of the precession of the lunar nodes. Then:
1 1 1
“ ` ,
Tyr, dra Ttr Tpr
Tpr Ttr
ñ Tyr, dra “ « 346.6201d “ 346d 14h 53m .
Tpr ` Ttr
6.4. Eclipses 59

6.4 Eclipses

Conditions for the Occurrence of Eclipses

If the orbital plane of the Moon coincided with the plane of the ecliptic, there
would be one solar and one lunar eclipse every synodic month. However, the
plane is tilted by 5˝ 91 ; therefore, for an eclipse to occur, the Moon must be
close to the node line during full or new moon. In the next sections we will
show that, for a (partial or total) lunar eclipse to occur, the Sun-node distance
must be smaller than 10˝ and, only in exceptional cases, can it reach 12˝ . For
a total lunar eclipse this distance must be less than 4.5˝ . In order for a (partial
or total) solar eclipse to occur, the distance of the Sun from the node must
be less than 15˝ , therefore solar eclipses are more frequent than lunar eclipses.
Since it takes the Sun six months to travel from one node to the other, the
eclipses are distributed in two groups, the so-called eclipse seasons.

Total Number of Eclipses in a Year

During a synodic period, the Sun moves by about 30˝ along the ecliptic. Ima-
gine there is a new moon at the beginning of a synodic month. Half a month
later, the Moon has covered half of its orbit, so there is a full moon, while
the Sun has moved by 15˝ . At the end of the month, there will be a new
moon again, and the Sun will have moved by 30˝ relative to the starting pos-
ition. Taking into account the position of the node line, it is then possible to
determine the types of eclipses that can occur in a synodic month:
1. 2 solar and 1 lunar. This happens when the Sun is initially at ´15˝ from
the node line. At the beginning of the month, a solar eclipse occurs. Half
a month later there is a full moon, and the distance of the Sun from the
node is around 0˝ , therefore a lunar eclipse occurs. At the end of the
month, the Sun will be at 15˝ from the node line, and a second solar
eclipse will occur.
2. 1 solar and 1 lunar. In this case, the initial distance of the Sun from the
node must be between ´15˝ and ´5˝ . After half a month, the Sun-node
distance is just right (0˝ ´ 10˝ ) for a lunar eclipse to occur. At the end
of the month, this distance will be greater than 15˝ , therefore no other
solar eclipse occurs.
3. 1 solar. The initial distance of the Sun from the node line must be less
than 5˝ . The first solar eclipse occurs at the beginning of the month.
However, at the next full moon, the Sun’s distance from the node line is
greater than 10˝ , so no lunar eclipse occurs. Likewise, at the end of the
month, no other solar eclipse will occur.
In any case, we see that the number of solar eclipses must be greater than or
60 Chapter 6. The Moon

equal to the number of lunar eclipses. The maximum number of eclipses that
can occur in a year is 7: 5 solar and 2 lunar, or 4 solar and 3 lunar. In the
first case, at the beginning of the year there are two solar eclipses, the first
one towards the beginning of January and the second during the next new
moon, with a lunar eclipse between them (case 1). After half a year, in the
second eclipse season, there are again two solar eclipses with a lunar eclipse in
between. At the end of the year, after 354 days (or 12 synodic months) from
the first eclipse, the last solar eclipse occurs.
In the second case, a lunar eclipse occurs shortly after the beginning of January,
and is followed by a solar eclipse (case 2). In the second eclipse season, there
are two solar eclipses with a lunar eclipse between them (case 1). Then, at
the end of the year, there will be the fourth solar eclipse and the third lunar
eclipse (case 2), 354 days after the first lunar eclipse. However, observing seven
eclipses in a year is very rare. The most frequent event is that of 2 solar and
2 lunar eclipses in a year. The minimum number of eclipses in a year is 2: one
solar eclipse (case 3), in each of the two eclipse seasons.

Lunar Eclipse

Figure 6.1: Lunar Eclipse.

Assume that the Earth, Moon and Sun are spherical bodies. Due to illumin-
ation by the Sun, the Earth casts a large cone of shadow (umbra) into space.
Let V be the vertex of this cone and l “ VE, its length (Fig. 6.1). The
quantities SE “ dC , STs “ R@ and ETe “ RC are the Earth-Sun distance,
the radius of the Sun and the radius of the Earth, respectively. Consider the
similar triangles Ts SV and Te EV, then:
l ` dC l RC dC
“ ñ l“ .
R@ RC R@ ´ RC
Since R@ “ 695510 km, RC “ 6378 km and dC “ 149.6 ¨ 106 km, we get
l “ 1.383 ¨ 106 km. Hence, the length of the umbra is 3.5 times the Earth-
Moon distance.
6.4. Eclipses 61

In addition to the umbra, the Earth projects a larger penumbra into space.
When the Moon crosses the umbra, a total lunar eclipse is observed. If, instead,
the Moon only crosses the penumbra, a partial eclipse is observed, and the
brightness of the Moon is only slightly attenuated.
Let us now calculate the radius of the umbra at the distance of the Moon
(EM “ d$ , in Fig. 6.1). Assuming that the angle in V is small, the following
relation holds:
RC l
“ .
r l ´ d$
Isolating r and substituting l found earlier:
d
r “ RC ´ $ pR@ ´ RC q . (6.1)
dC
Evaluating the above equation, we find r “ 4560 km, i.e. around 2.65 times
the radius of the Moon. In fact, due to the eccentric orbits of the Moon and
the Earth, this value can range from 2.6 to 2.8. The Moon covers a distance
equal to its diameter every hour (see Ex. 6.6), therefore the totality of a lunar
eclipse can last up to two hours.
A total lunar eclipse takes place as follows: when the Moon first enters the
penumbra cone, its brightness decreases slightly. At the moment of first con-
tact, the Moon enters Earth’s umbra. As the eclipse continues, Earth’s shadow
progressively covers a greater portion of the Moon until, at the moment of to-
tality, it is completely obscured. Subsequently, the Moon emerges again, its
brightness slowly increasing. The moment when the Moon leaves the umbra is
called last contact, and marks the end of the eclipse. Partial eclipses, in which
the Moon only crosses the penumbra of the Earth, are very difficult to observe,
because of the slight, almost imperceptible, decrease in brightness.
Let us now calculate the maximum distance between the Sun and the node
line, for which a partial lunar eclipse can take place. In the limiting case, the
Moon barely grazes Earth’s umbra, therefore it is at a maximum distance of
r ` R$ from the node line. The angle from the node line is then:

r ` R$
tan α « αrad “ ñ α “ 0.94˝ ,
d$

where r is given by Eq. 6.1. Now, the maximum distance of the Sun from the
node line is exactly equal to this angle, since only the relative positions of the
Sun, Earth and Moon matter. However, the Moon’s orbit is inclined by 5.145˝
relative to the plane of the ecliptic, so the maximum distance of the Sun, in
ecliptic longitude, is:
α
∆λ “ “ 10.5˝
sin 5.145˝
62 Chapter 6. The Moon

which is approximately the angle quoted at the beginning of this section.

The limit for a total eclipse is obtained by taking r ´ R$ as the maximum
distance of the Moon from the node line. In this case, we get α “ 0.42˝ and
∆λ “ 4.7˝ . Since the orbits of the Earth and the Moon are elliptical, these
limits vary depending on the positions of the celestial bodies (Ex. 6.10).

Solar Eclipse

Figure 6.2: Solar eclipse: unlike lunar eclipses, visible from the entire hemisphere on Earth
facing the Moon, total solar eclipses can be observed only in a region of 200 km. The
penumbra is much larger, its radius being around twice that of the Moon.

When the Moon covers the disk of the Sun, a solar eclipse takes place. For a
solar eclipse to be observed, there must be a new moon, and both the Moon
and the Sun must be close to the node line. Since both the orbits of the Earth
and the Moon are elliptical, the Moon can appear larger or smaller than the
Sun. If the angular diameter of the Moon is greater than that of the Sun,
a total solar eclipse occurs. If, instead, the angular diameter of the Moon is
smaller, a very thin portion of the Sun’s disk is left uncovered: in this case,
an annular solar eclipse is observed. On average, the angular diameter of the
Sun is slightly larger, hence, for a total solar eclipse to occur, either the Moon
must be close to perigee, or the Earth close to aphelion.
Although solar eclipses occur more often than lunar eclipses, the former are
much more difficult to observe. In fact, during a solar eclipse, the length of
the umbra on the surface of the Earth is just 200 km (Ex. 6.12), which makes
it visible only in very few places. On the contrary, a total lunar eclipse can be
seen from the entire hemisphere on Earth facing the Moon.
What is the maximum distance of the Sun from the node line, for a (partial
or total) solar eclipse to be observed? In this case, the Moon must be at point
M in Fig. 6.3. The distance EV « Te V is just l, computed in the previous
section. Also, ETe “ RC and ME “ d$ are the radius of the Earth and
the distance Moon-Earth, respectively. Since α is small, we can approximate
6.4. Eclipses 63

Figure 6.3: For a (partial or total) solar eclipse to occur, the distance of the Moon from the
node line must be less than α.

MH « d$ αrad and HE « d$ . The triangles EVTe and HVM are similar,

therefore:
d$ αrad RC RC ´ d ¯
“ ñ αrad “ 1` $ , (6.2)
l ` d$ l d$ l
giving α « 1.21˝ . Adding to this the angular radius of the Moon, the total angle
is then 1.47˝ . The maximum distance of the Sun from the node line is:
1.47˝
∆λ “ « 16˝
sin 5.145˝
which is an overestimation of the limit stated at the beginning of this section.

Saros cycle
At the time of the Chaldeans it was already known that the eclipses repeated
periodically every 18 years and 11.3 days, the so-called Saros cycle. Indeed,
eclipses repeat in the same order, when the same phase of the Moon repeats
itself at a distance from the node line equal to the distance at which the
first eclipse occurred. It follows that the Saros cycle is equal to the time
interval necessary for the synodic month, draconic month and draconic year
(29.53, 27.21 and 346.62 days, respectively) to return in phase with each other.
In particular, 242 draconic months correspond approximately to 223 synodic
months and 19 draconic years. More exactly:
• 242 draconic months = 6585.36 mean solar days
• 223 synodic month = 6585.32 mean solar days
• 19 draconic years = 6585.78 mean solar days
However, 223 synodic months are 0.04 days shorter than 242 draconic months,
after 6585 days the Moon will return to a slightly different position (the cycle
is not exact). Since the Saros cycle is an integer number of days plus approx-
imately 1/3, the eclipses will be visible from locations shifted by around 120˝
64 Chapter 6. The Moon

in longitude, compared to the previous cycle. For this reason, the Greeks had
defined another period, known as the exeligmos, equal to three Saros cycles.
Each Saros cycle comprises 70 eclipses: 41 solar and 29 lunar.

6.5 Libration

In addition to the motions of revolution around the Earth and precession of

the node line, the Moon is subject to a rotational motion around its axis which
is synchronous with the orbital motion around the Earth. As a consequence,
the Moon always shows the Earth its same side (approximately). In fact, the
Moon is also subject to an apparent motion, called libration. As a result of this
motion, it is possible to see 59 % of the lunar surface at one time or another,
instead of the 50 % that we would have expected.

Longitudinal Libration
While the rotation of the Moon around its axis is uniform (to a good approxim-
ation), its orbital velocity is not. For example, if the Moon is close to perigee,
its speed is higher than average, and it is possible to see a thin area towards
west, that wouldn’t have been visible otherwise. This longitudinal motion has
a period equal to the anomalistic month (the time it takes the Moon to re-
turn to the same position with respect to the line passing through perigee and
apogee, called the apse line), with a length of 27d 13h 18.6m (see Pr. 6.2). The
maximum amplitude of longitudinal libration is 7˝ 541 (Pr. 6.6).

Latitudinal Libration
The rotation axis of the Moon is inclined by 6˝ 411 with respect to the per-
pendicular to its orbital plane. Since the direction of its axis stays unchanged
(because of the conservation of angular momentum), every draconic month the
Moon periodically shows the Earth its north and south poles. The amplitude
of this motion is equal to the inclination of the Moon’s axis.

Diurnal Libration
Diurnal libration is a secondary effect compared to the previous ones and is
caused by parallax. Observers at different locations on Earth see the Moon
at slightly different angles. As a result, they see different portions of it. The
amplitude of diurnal libration is about 1˝ .
6.6. Exercises 65

6.6 Exercises

1. What planets can be eclipsed by the full moon?

2. Is the full Moon higher in summer or in winter?
3. For what range of latitudes is it possible to see the Moon at the zenith?
4. At what latitudes is the Moon circumpolar at least once a year?
5. When is the Moon closest to you: at the zenith or at the horizon?
6. How long does it take for the Moon to cover a distance equal to its
diameter? (Hint: use Eq. 10.18).
7. (INT 2012, Th.S, q.2) In the morning, you look out the window of your
home (in Italy) and see the moon half illuminated. Is it a first or last
quarter moon? At what time does it set?
8.** (NAZ 2015, Th.S, q.5) A small outer planet travels around the Sun in
a circular orbit that lies exactly on the plane of the ecliptic. It has a
peculiarity: the Moon can never eclipse it. Find the radius of its orbit
(Hint: use Kepler’s law, Eq. 10.13).
9. On the 1st of January there was a new moon. What will be the position
of the Moon, the same day of the following year?
10.* Compute again the maximum distance of the Sun from the node line
for a lunar eclipse to occur. Take into account the eccentricities of the
orbits.
11.* The same question as in the previous exercise, but for a solar eclipse.
12.* Compute the maximum lengths of the umbra and penumbra on Earth,
during a solar eclipse. Take into account the eccentricities of the orbits.
66 Chapter 6. The Moon

6.7 Problems

1.* Meton cycle

In the fifth century BC, the Athenian astronomer Meton calculated how
many years it would take the Moon to return to the same phase on the
same day of the year. Find the length of this cycle, knowing the duration
of the synodic month (29.53d ) and of the tropical year (365.24d ).
2. Anomalistic month
The apse line of the Moon completes a full rotation every 8.8504 years.
Find the duration of the anomalistic month, defined as the time it takes
the Moon to return to the same position with respect to the apse line.
3.* Measuring the distance of the Moon
When the Moon passes through the zenith, its angular diameter is
αz “ 311 3611 . When it is on the horizon, its angular diameter is 311 42 .
Calculate the distance of the Moon from Earth. (The effects of atmo-
spheric refraction have already been accounted for).
4.** Distance of the Moon, another method
The Moon is observed from a place of latitude φ “ 41˝ 551 . When the
Moon transits on the meridian and has maximum declination, its distance
from the zenith is zu “ 13˝ 331 ; when it has minimum declination, its
distance is zl “ 71˝ 251 . Find the distance of the Moon from Earth.
Knowing that the inclination of the ecliptic is 23˝ 271 , find the inclination
of the Moon’s orbit. (You will need to solve an equation numerically.)
5.* Lunar influence
Among the numerous factors that affect the duration of the tropical
year, we want to estimate the contribution of the Moon. Assume that
the centre of mass of the Earth-Moon system moves in a circular orbit
around the Sun and that the Moon moves in a circular orbit around the
Earth, on the plane of the ecliptic, completing (approximately) 12 and
a half revolutions every year. Show that the year is longer than average
if there is a first quarter moon during the spring equinox. Evaluate the
difference in minutes between that year and the following year.
6.** Longitudinal libration
Estimate the amplitude of longitudinal libration. (Hint: read Ch. 10
first, and use Eq. 5.9).
 Part II

Radiation Mechanisms

67
 7

Electromagnetic Radiation

What we generally call light is only a small part of the electromagnetic spec-
trum, in particular the visible region between 400 and 700 nanometers in
wavelength. By studying the radiation emitted by a body, it is possible to
obtain a large amount of information, including its temperature, composition,
surface gravity, electric and magnetic fields. In this chapter we illustrate how
this can be achieved. We start by discussing the properties of electromagnetic
radiation, moving on to introduce some concepts of statistical physics, which
will come in handy in the study of the black body spectrum. Only a brief
summary of the essential results, without delving too much into derivations,
can be given here. However, some proofs will be presented in the problems. In
case you wish to gain a deeper insight in this topic, you might want to con-
sult some statistical physics and quantum mechanics textbooks (see “Suggested
Resources”).

7.1 Wave or particle?

Electromagnetic waves can be defined by three physical parameters:

• the period T , i.e. the time interval it takes a crest to reach the position
previously occupied by the preceding crest;
• the frequency f , i.e. the number of crests that pass through a given
point every second. The frequency is the reciprocal of the period, that
is: f “ 1{T ;
• the wavelength λ, i.e. the distance between two adjacent crests or troughs,
equal to the distance covered by the wave in a period.
In vacuum, the velocity of all electromagnetic radiation is equal to the speed
of light, c. Since velocity is equal to space (λ) divided by time (T ), it follows
that:
λ
c “ “ λf . (7.1)
T
68
7.1. Wave or particle? 69

When taking into consideration phenomena involving energy exchange between

electromagnetic radiation and matter, the wave model becomes inadequate and
unable to justify some experimental observations. In this case, a corpuscular
model is used instead. According to this model, radiation is made up of energy
packets, called photons. The energy carried by each photon is directly propor-
tional to its frequency, and the constant of proportionality is called Planck’s
constant, equal to h “ 6.626 ¨ 10´32 J s. The energy of a photon is:

E “ hf . (7.2)

Since f “ c{λ, we also have E “ hc{λ. What distinguishes different elec-

tromagnetic waves is their energy, i.e. their frequency. High-frequency (short-
wavelength) radiation is composed of high-energy photons, whereas low-
frequency (long-wavelength) radiation is composed of low-energy photons. The
set of all electromagnetic waves is called the electromagnetic spectrum.
According to de Broglie’s law, each photon carries a momentum of:

h
p“ . (7.3)
λ

Since λ “ c{f , we can also write p “ hf {c “ E{c. A body illuminated

with light is subject to radiation pressure, due to the exchange of momentum
between the object and the electromagnetic field. Since force equals rate of
change in momentum, p “ E{c implies that the force is equal to the power
incident on the surface, divided by the speed of light. But pressure is force
divided by area, therefore, defining the flux (F ) as the power per unit area
(and therefore energy per unit area, per unit time), it follows that:

F
Prad “ . (7.4)
c

This expression makes the implicit assumption that the body absorbs all of the
incident (collimated) light. If, instead, the body is completely reflecting, the
radiation pressure is double the value given by Eq. 7.4. Indeed, the photons
bounce back with the same initial momentum, hence the change in momentum
of the body is twice the initial momentum of the photons. The radiation force
is generally too small to be noticed under everyday circumstances. However,
in outer space, it is usually the main force acting on an object, besides gravity.
The discovery that light carries momentum has led to the development of new
propulsion methods, such as spaceships capable of deploying huge solar sails
or small probes propelled by reflecting the collimated light of powerful laser
beams (see Ex. 7.5).
70 Chapter 7. Electromagnetic Radiation

7.2 Boltzmann Distribution Law

Boltzmann’s law describes the most likely energy distribution of a system. In

this and the next section, we will consider the probability distribution function
gpEq, defined so that gpEq dE equals the probability that the system’s energy
is in the range [E, E ` dE]. From statistical arguments, it can be shown that
the most probable distribution is:

gpEq 9 e´E{pkB T q , (7.5)

where E is the energy of the state, T is the temperature and kB is the

Boltzmann constant, equal to kB “ R{Na “ 1.381 ¨ 10´23 J/K.

7.3 Maxwell Distribution Law

The Maxwell distribution law describes the most likely velocity distribution
of particles within a system. We define f pvq, called the Maxwell distribution
function, such that f pvq dv is equal to the probability of finding a particle
with velocity in the range [v, v ` dv]. This function can be derived from the
Boltzmann distribution (Pr. 7.2):
´ m ¯3{2 2 ´mv2 {2kB T
f pvq “ 4π v e , (7.6)
2πkB T
where m is the mass of a particle. Note the occurrence of the Boltzmann factor
e´E{pkB T q , where E is the kinetic energy mv 2 {2.
The most probable velocity corresponds to the maximum of f pvq. Equating
the derivative of f pvq to zero, it can be shown (Ex. 7.8) that:
c
2kB T
vmp “ . (7.7)
m
The root mean square (rms) velocity,aas the name suggests, is the square root
of the average velocity squared, i.e. xv 2 y. It is given by (Ex. 7.12):
c
3kB T
vrms “ . (7.8)
m
It is useful for calculating the kinetic energy of a system:
´1 ¯ 1 3
xKy “ m xv 2 y ¨ N “ mN vrms 2 “ nRT .
2 2 2
where we used kB “ R{Na and n “ N {Na , with n denoting the number of
moles. Therefore:
3
xKy “ nRT , (7.9)
2
7.3. Maxwell Distribution Law 71

This formula is valid for a gas composed of monoatomic particles. The equipar-
tition theorem states that each (quadratic) degree of freedom contributes equally
to the total kinetic energy. A monoatomic gas has three degrees of freedom, one
for each of the three spatial dimensions. Therefore, each degree must contribute
an energy of p1{2 qnRT . A diatomic gas has two additional degrees of freedom.
These correspond to the rotation around the two axes perpendicular to the line
joining the centres of the atoms (the energy associated with the rotation about
the axis joining the atoms is negligible, since the atoms are very small). Hence,
the kinetic energy of a diatomic gas is p5{2q nRT . Therefore, at the same tem-
perature, a diatomic gas stores more energy than a monoatomic gas.
Planetary Atmospheres
The atmosphere of a planet is made up of several molecules: some lighter,
like helium and hydrogen, others heavier, like carbon dioxide, ammonia and
methane. Since the average kinetic energy of molecules in a gas is proportional
to the temperature (Eq. 7.9), it follows that molecules move faster in a warmer
atmosphere. At a given temperature, the average velocity of a molecule is
inversely proportional to the square root of its mass. Therefore, hydrogen
molecules (m “ 2 u) move four times faster than oxygen molecules (m “ 32 u).
If a molecule located in the upper atmosphere moves at a sufficiently high
speed, it can exceed the escape velocity of the planet. Therefore, small and
hot planets may have easily lost all of their lighter molecules, while massive
and cold planets could still retain their primordial atmosphere.
However, the situation is not that simple: Maxwell’s curve (Eq. 7.6) shows that
there are molecules with much higher velocities than the average. In fact, about
one molecule in a million has a velocity three times higher than the rms velocity,
while one molecule in 1016 has a velocity five times higher. It can be shown
that, if the rms velocity of a molecule is less than 1/6 of the escape velocity
from a planet, this molecule does not escape from the planet’s atmosphere in
significant quantities during the lifetime of the Solar System. For example, the
mean square velocities of molecular nitrogen (506 km/s) and oxygen (473 km/s)
in the Earth’s atmosphere are well below 1/6 of the escape velocity from Earth
(11.2 km/s). Assuming the Moon had an atmosphere with approximately the
same temperature as the Earth’s, the rms velocities of oxygen and nitrogen
would have been only 1/5 of the escape velocity from the Moon (2.4 km/s).
Therefore, no trace of such an atmosphere would remain today. The same
arguments apply to the hydrogen in the Earth’s atmosphere, which has a mean
square velocity of 2 km/s, just under 1{6 of Earth’s escape velocity. Hydrogen
has had sufficient to escape, and indeed today only constitutes 0.000055 % of
the Earth’s atmosphere. On the other hand, the average speed of hydrogen is
1{60 of the escape velocity from Jupiter. This explains why Jupiter has been
able to maintain hydrogen, which is its major constituent.
72 Chapter 7. Electromagnetic Radiation

7.4 Black Body

Figure 7.1: The power emitted by a black body as a function of the wavelength. With
increasing temperature, the wavelength of maximum emission shifts to the left and, at the
same time, the maximum power increases.

The term black body, first introduced by Gustav Kirchhoff in 1860, describes a
theoretical body that absorbs all incident electromagnetic radiation. In equi-
librium, because of energy conservation, a black body must re-emit all of the
absorbed energy into space, in the form of black body radiation. The curve
obtained by recording the amount of energy emitted by the black body as a
function of the wavelength is called black body curve or continuous spectrum
curve (Fig. 7.1).
Continuous spectra are produced by all bodies above absolute zero (0 K) and
do not depend on the composition of the emitting body, but are a function of
its temperature only. In Fig. 7.1, we see that, for long wavelengths, the power
emitted by a black body increases initially, reaches a maximum at a certain
wavelength λmax , then decreases again, going to zero at infinity. As the tem-
perature is decreased, λmax moves towards longer wavelengths and, at the same
time, the maximum power decreases. Instead, as the temperature is increased,
λmax moves towards shorter wavelengths and the maximum power increases.
Indeed, the area enclosed by the curve is equal to the total power emitted by the
black body, which decreases with decreasing temperature. These observations
are summarized in Wien’s and Stefan-Boltzmann laws, respectively.

Wien’s Law
The relationship between the colour of a hot body and its temperature was
first noticed in 1792, by porcelain manufacturer Thomas Wedgewood. All his
ovens, regardless of size, shape or construction, turned red when heated above
7.4. Black Body 73

a certain temperature. It looked like their colour was exclusively related to

their temperature. Using purely thermodynamic arguments, in 1893, Wilhelm
Wien showed that the wavelength (λmax ) at which the emission of energy is
maximum is related to the temperature (T ) according to:

λmax T “ k , (7.10)

where k is Wien’s constant:

k “ 2.897 ¨ 10´3 m K . (7.11)

As the temperature is raised, a body turns first red (620 ´ 750 nm), then
orange (590 ´ 620 nm), then yellow (570 ´ 590 nm) and finally blue-white
(450 ´ 475 nm). In the same way, colder stars appear red and orange, whereas
warmer ones appear yellow and white-blue. This does not mean that a star
emits only at the wavelength corresponding to its perceived colour, but that
the other wavelengths are surpassed by that of maximum emission.
For example, the surface temperature of the Sun is T@ “ 5778 K. The max-
imum emission wavelength is then:

k 2.898 ¨ 10´3 m K
λmax “ “ “ 501 ¨ 10´9 m .
T@ 5778 K
Surprisingly, 501 nm corresponds to green, although it is very close to yellow
560 nm. The wavelength of maximum emission does not always correspond to
the colour perceived by our eyes. This is due to the fact that the atmosphere
disperses smaller wavelengths more effectively (see Sec. 8.4), and our eyes are
the most sensitive to yellow.

Stefan-Boltzmann Law
The second black body law is the Stefan-Boltzmann law, experimentally ob-
tained by Josef Stefan in 1879 and theoretically derived by Ludwig Boltzmann
in 1884, using only thermodynamic arguments. The law states that the power
emitted by a black body, per unit area, is directly proportional to the fourth
power of its absolute temperature. The proportionality constant σ is called
the Stefan-Boltzmann constant:

σ “ 5.67 ¨ 10´8 W/m2 K4 .

Therefore, the power emitted by a spherical body with radius R on the entire
electromagnetic spectrum is:

L “ 4πR2 σT 4 . (7.12)
74 Chapter 7. Electromagnetic Radiation

In general, a body can deviate significantly from the black body model. To
improve the theoretical prediction, the emissivity e is introduced. Quantitat-
ively, the emissivity is equal to the ratio of the radiation emitted by the body
to the radiation emitted by a black body at the same temperature; therefore,
the emissivity can also be interpreted as the ability of a body to absorb in-
cident light. An ideal black body absorbs all incident light, so e “ 1. On the
contrary, a mirror reflects most of the incident radiation (at least the visible
radiation), therefore e is close to 0. In any case, e can range from 0 to 1. In a
more general form then, the Stefan-Boltzmann law can be written as:

L “ 4πR2 e σT 4 . (7.13)

If an object is at temperature T and the space surrounding it is at tem-

perature T0 , the body absorbs energy from the surroundings at a rate of
Pin “ 4πR2 e σT0 4 and emits energy at a rate of Pout “ 4πR2 e σT 4 . Therefore,
the total rate of exchanged energy is:

Ptot “ 4πR2 e σpT 4 ´ T0 4 q .

If an object is in equilibrium (i.e. constant temperature) with the surroundings,

it radiates and absorbs the same power, so that T “ T0 . If an object is warmer
than the environment, it radiates more energy than it absorbs and, in the
absence of other mechanisms, its temperature decreases. The temperature
of empty space, associated with the cosmic background radiation, is about
T0 “ 2.725 K.

7.5 Types of Spectra

There are three different kinds of spectra: continuous, absorption and emission.
The spectrum of a star can be regarded as the superposition of a continuous
spectrum, also called black body spectrum (as it depends only on the temper-
ature of the star), and an absorption spectrum (or, more rarely, an emission
spectrum) due to the absorption (or emission) of radiation by the gas in the
photosphere. The continuous component of the spectrum gives information
about the temperature of the photosphere, while the absorption (or emission)
lines make it possible to determine the chemical composition of the star, its
surface gravity and the electric and magnetic fields.
The continuous spectrum originates from a process called bremmstrahlung,
from the German bremsen “to brake, decelerate” and strahlung “radiation”,
which literally translates to braking or deceleration radiation. Because of the
very high temperature, matter is in the plasma state at the centre of stars,
and electrons move chaotically amongst ionized atoms. An electron approach-
ing a charged atom is scattered and accelerates away from it, consequently
7.5. Types of Spectra 75

emitting radiation in the form of electromagnetic waves (Pr. 7.1). Since

the electron-atom distance and the electron velocity can vary continuously
within a star, the resulting electromagnetic radiation is distributed continu-
ously over all wavelengths, therefore the spectrum it produces is also continu-
ous. Bremmstrahlung is thus characterized by a continuous radiation distribu-
tion that becomes more intense (and moves towards higher frequencies) with
increasing velocity (thermal energy) of the electrons, i.e. with increasing tem-
perature. Bremmstrahlung is sometimes referred to as free-free radiation: it is
created by charged particles that are not bound to any ion, atom or molecule,
and are therefore free, both before and after the deflection.
Another phenomenon that contributes to the continuous spectrum is the re-
combination of an electron with an ion. As before, the initial energy of an
electron is free to take any value, therefore, even though the final energy must
be quantized (i.e. bound to have the same energy as a given orbital in the
atom), the energy emitted in the form of electromagnetic radiation can take
any value, giving rise to a continuous spectrum. The recombination of an elec-
tron with an ion is referred to as free-bound radiation. As long as there is a
free component, the total energy is not quantized and the resulting spectrum
is continuous.
The emission spectrum is produced by the emission of energy from excited
atoms. Normally, the electrons inside an atom occupy the orbitals which cor-
respond to the lowest overall energy (ground state) but, when excited, they
move to higher-energy orbitals. Returning to the ground state, the atom emits
a photon whose energy is equal to the energy difference between the excited
and the ground state. Since the energy of an electron in an atom is quant-
ised, and the electron does not leave the atom after the transition, it follows
that the energy (and therefore the wavelength) of the emitted photon must be
quantized. Because the energy levels vary from atom to atom (they depend on
nuclear charge and degree of ionization), each element leads to different emis-
sion lines. Therefore, the emission (and absorption) spectrum of each element
is unique. In this sense, it can be regarded as the fingerprint of an element:
by examining the emission spectrum of an unknown object and comparing it
with other previously recorded spectra, it is possible to determine the chemical
composition of the source.
The absorption spectrum is the opposite of the emission spectrum. Electrons
within an atom can absorb the incident radiation and move to higher-energy or-
bitals, then returning to the ground state by emitting exactly the same energy.
Therefore, each atom absorbs the same wavelengths that it emits. At these
wavelengths, the continuous spectrum appears darker, since light is absorbed
and then re-emitted in a random direction (which rarely coincides with the
direction of the observer). The analysis of absorption lines, called Fraunhofer
lines, provides information on the chemical composition of stars.
76 Chapter 7. Electromagnetic Radiation

7.6 Doppler Effect

Figure 7.2: A spectrum is “blueshifted” when all the absorption (or emission) lines are shifted
towards shorter wavelengths (compared to the spectrum at rest); “redshifted” when they are
shifted towards longer wavelengths.

As seen in the previous section, the emission and absorption lines of atoms
and molecules can be recorded in the laboratory. By comparing these with the
spectra of celestial bodies, it is possible to determine their chemical composi-
tion. We often notice a shift of all these emission (or absorption) lines by a few
nanometers towards longer (red) or shorter (blue) wavelengths, as in Fig. 7.2.
This is due to the relative motion between the source and the observer. The
law that relates the relative velocity to the shift of the spectral lines is called
Doppler effect. From the discovery of exoplanets and spectroscopic binaries,
to the expansion of the universe, this law has permanently changed the way
we think of our place in the universe.
If you already studied the Doppler effect, it was probably applied to a wave
that propagates in a medium, such as sound in air. You should have then learnt
two different formulae, depending on whether the observer or source are in mo-
tion. However, light shows a different behaviour. Electromagnetic radiation
propagates in vacuum, therefore there is no absolute reference system (such as
air, for sound) against which to measure velocity. The only information we can
possibly have is the relative velocity between two bodies. Therefore, we expect
the Doppler effect for electromagnetic radiation to depend only on the relative
velocity between source and observer; hence there must be only one formula
that describes this effect. The hypotheses just mentioned are actually the pos-
tulates of special relativity, therefore the equation we are looking for can only
be obtained within this theory. We first derive an approximate formula, using
classical mechanics, and then the exact one, using special relativity.
In the classical approximation, let us only consider the case of a moving source
and a stationary observer, the proof for the other case being very similar. The
7.6. Doppler Effect 77

Figure 7.3: Doppler effect for a source in motion with velocity v, which emits light with
wavelength λ. After a period, the source will have moved vT away from the observer,
therefore the observer perceives that light is emitted with a wavelength λ1 “ λ ` vT .

source moves with velocity v away from the observer. Light has wavelength
λ, period T and velocity c, thus λ “ cT . After a period T , the source has
moved away by a distance vT , hence the next crest is emitted at a distance vT
farther away from the observer. Therefore, as shown in Fig. 7.3, the perceived
wavelength is slightly longer: λ1 “ λ ` vT “ pc ` vq{f . Since light propagates
with velocity c in both cases, we find c “ λ1 f 1 “ pc ` vqf 1 {f , so that:
1 ´ v¯
f1 “ f « 1´ f, (7.14)
1 ` v{c c

where we have assumed v{c to be small and used the approximation p1 ` xqα «
1 ` αx for x ! 1 (Appendix A.6), since classical physics is only valid for v ! c.
In the derivation that follows, we will make use of the fact that, given two
inertial reference systems K and K 1 , moving with relative velocity v, the time
∆t measured by an observer in K is related to the time ∆t1 measured by an
observer in K 1 by the following equation:
1
∆t1 “ γ∆t where γ “ a . (7.15)
1 ´ pv{cq2

This relation is called time dilation. Indeed, since 1 ´ pv{cq2 is always less than
unity, it follows that γ ě 1. Therefore, the time measured in K 1 is longer than
the time in K.
Let ∆t “ 1{f be the emission time in the reference system of the source. The
time perceived by the observer is ∆t1 “ γ∆t, because of time dilation. In the
observer’s reference system, after a period ∆t1 , the photons have travelled a
distance equal to c∆t1 “ c γ∆t, while the source has travelled away by v∆t1 “
78 Chapter 7. Electromagnetic Radiation

v γ∆t. Therefore, the perceived wavelength is slightly longer: λ1 “ pc ` vqγ∆t.

Then, from f 1 “ c{λ1 , we find:
a
1 p1 ´ v 2 {c2 q 1
f “c
c`v ∆t
a
p1 ´ v{cqp1 ` v{cq
“ f
1 ` v{c
d
1 ´ v{c
“ f.
1 ` v{c

Setting β “ v{c:
d
1´β
f1 “ f . (7.16)
1`β

If β ă 0, i.e. if the source is moving closer to the observer, Eq. 7.16 tells us that
f 1 ą f . In this case, we say that light is blueshifted, since blue is the colour with
the highest frequency. If, instead, β ą 0, i.e. the source is moving away from
the observer, it follows that f 1 ă f . In this case, we say that light is redshifted,
since red is the colour with the lowest frequency. The energy of a photon
is proportional to its frequency, therefore redshifted light carries less energy.
Conversely, blueshifed light carries more energy. Let Le be the luminosity
emitted by a star in its reference system. The observed luminosity is:

f1
Lo “ Le ¨ . (7.17)
f

In the classical limit (β ! 1), Eq. 7.16 reduces to:

1
´ v ¯´ v¯ v
f « 1´ 1´ f «1´ ,
2c 2c c
where we have used the approximation p1 ` xqα « 1 ` αx for x “ v{c ! 1.
This agrees with the classical result found earlier (Eq. 7.14). Eq. 7.16 can also
be written as a function of the wavelength:
d
1`β
λ1 “ λ.
1´β

From this, we can then compute the wavelength shift ∆λ “ λ1 ´ λ:

ˆd ˙
1`β
∆λ “ ´ 1 λ. (7.18)
1´β
7.7. Harvard Stellar Classification 79

In the non-relativistic limit, this reduces to:

∆λ “ βλ . (7.19)

The term ∆λ{λ is usually denoted by z, called the redshift parameter. If z is

known, β can be found from Eq. 7.18:

p1 ` zq2 ´ 1
β“ , (7.20)
p1 ` zq2 ` 1

which, in the non-relativistic limit (z ! 1), becomes:

β“z . (7.21)

7.7 Harvard Stellar Classification

A star can be classified based on its surface temperature, which can be estim-
ated from the wavelength of maximum emission, using Wien’s law. Another
way to estimate the temperature is by studying the ionized elements in the
photosphere, since the type and degree of ionization depends on the temper-
ature. According to the Harvard stellar classification system, the stars can be
divided into 7 classes: O, B, A, F, G, K and M (there are several mnemonics
used to remember this sequence, the most famous one being: “Oh Be A Fine
Girl/Guy, Kiss Me”).

Class Temperature / K Colour Luminosity / L@ Mass / M@

O ě 33000 blue ě 30000 ě 16
B 10000 – 33000 light blue 25 – 30000 2.1 – 16
A 7500 – 10000 white 5 – 25 1.4 – 2.1
F 6000 – 7500 white-yellow 1.5 – 5 1.04 – 1.4
G 5200 – 6000 yellow 0.6 – 1.5 0.8 – 1.04
K 3700 – 5200 orange 0.08 – 0.6 0.45 – 0.8
M ď 3700 red ď 0.08 0.08 – 0.45
Table 7.1

However, stellar classification is not that simple. Every class is further divided
into 10 subclasses, numbered from 0 to 9. In the Harvard classification system,
the Sun is a G2 star. More recently, the classes L and T have been added.
These correspond to brown dwarfs. Tab. 7.1 gives the range of temperatures,
luminosities and masses for the 7 classes O – M.
80 Chapter 7. Electromagnetic Radiation

7.8 Yerkes Stellar Classification

The Harvard scheme specifies only the surface temperature and some spectral
features of the star. A more precise classification would also include the lumin-
osity of the star, since two stars with similar temperatures, but different radii,
can have very different luminosities. It is possible to obtain the luminosity by
examining the spectrum of the star. The mass of giant and dwarf stars is of
the same order of magnitude, but their radii are very different. Therefore, the
gravitational acceleration on the surface of a giant star is much lower than the
acceleration on a dwarf star, since g “ GM {R2 . Given the lower gravity, gas
pressures and densities are much lower in giant stars than in dwarf stars. It can
be shown that the width of spectral lines is proportional to the pressure (Pr.
7.4), therefore measuring this width ultimately allows us to find the radius of
the star. By estimating the temperature of a star using Wien’s law, it is then
possible to find its luminosity using Stefan-Boltzmann’s equation. The Yerkes
scheme divides the stars into six luminosity classes:
• Ia : Most luminous supergiants
• Ib : Less luminous supergiants
• II : Luminous giants
• III: Normal giants
• IV: Subgiants
• V : Main sequence stars (dwarfs)
For example, the Sun belongs to the luminosity class “V”. Thus, the Sun would
be more fully specified as a G2V-type star.
7.9. Exercises 81

7.9 Exercises
1. At what wavelength does a star with surface temperature T “ 4000 K
emit the most energy?
2. Compute the total brightness of a star with surface temperature
T “ 6800 K and radius R “ 2.5R@ . Write your answer in terms of
the solar luminosity L@ .
3.* Assuming that dust particles (ρ « 103 kg/m3 ) behave like black bodies,
determine the diameter of a spherical particle that is in equilibrium (at
rest) at a distance of one astronomical unit from the Sun.
4. Estimate the number of neutrinos, produced by the Sun, that arrive on
Earth. Each nuclear reaction produces 26.8 MeV of energy and 2 electron
neutrinos. On Earth, early experiments detected only 1{3 of the expected
number of neutrinos. This was known as the solar neutrino problem, and
was finally solved by realizing that an electron neutrino can oscillate into
the other two types of neutrinos (muon and tau).
5.* (ARAO 2019, Th.X/XI, q.1) A future project aims to propel small space-
ships with a powerful laser beam, sending them over long distances.
What speed can a spaceship be propelled at, if its perfectly reflecting
base has a diameter of 1 mm, its total mass is 1 mg, and the optical laser
has a power of 1 MW and divergence of 52 ? Assume that the base of the
spaceship is oriented perpendicularly to the laser beam, that the beam
itself is very thin when exiting the laser, and neglect the gravitational
action of all the bodies surrounding the spaceship.
6.* At sea level, the Earth’s atmosphere has an average temperature of
T “ 14˝ C. Compute:
• vrms for molecular hydrogen, oxygen and nitrogen;
• the percentage of each of these molecules with velocities greater
than the escape velocity from the Earth at sea level (you will need
to solve an integral numerically).
7.** Prove that the probability of finding a molecule with 3 and 5 times
the root mean square velocity of the gas is 5.88 ¨ 10´6 and 3.62 ¨ 10´16 ,
respectively (you will need to solve an integral numerically).
8.* Prove a
that the most probable speed of molecules in a gas is
vmp “ 2kB T {m.
9.** Assume that the atmosphere has a uniform composition with an effective
molar mass of u “ 28.9 g/mol and a constant temperature of T “ 14˝ C.
Find the ratio of atmospheric pressure at height h and at sea level.
82 Chapter 7. Electromagnetic Radiation

10.** A space station is in the shape of a cylinder with radius R0 and is filled
with air. The cylinder rotates around its axis of symmetry, providing an
acceleration equal to g at the outer edge. If the temperature inside the
space station is constant and equal to T , what is the ratio of the pressure
P0 at the outer edge of the station, to Pc , the pressure at the centre?
11.** Consider a spaceship with internal pressure P0 “ 1 atm, constant tem-
perature T0 “ 285 K and volume V0 “ 100 m3 . Let µ “ 28.9 g/mol be
the effective molar mass of air. Suddenly, the spaceship is struck by an
object that makes a hole of area A “ 1 cm2 . Find the pressure of air
inside the spaceship as a function of time. How long does it take for the
pressure to halve?
12.** Prove the formula for the root mean square speed of a molecule in a gas.
You may want to use the fact that the average of v 2 , for the distribution
f pvq, is given by the integral:
ż8
2
xv y “ v 2 f pvq dv (7.22)
0

7.10 Problems

1.* Radiation emitted by accelerated charge

The energy emitted by an accelerated charge depends exclusively on
the vacuum permittivity 0 , the speed of light c, the electric charge q
and the acceleration a. Using dimensional analysis, derive a formula for
Ep0 , c, q, aq. It can be shown that the numerical factor is 1{6π.
2.** Maxwell distribution
Let gpvx q be the velocity distribution in the x direction, so that gpvx q dvx
is equal to the fraction of particles with velocities in the interval [vx ,vx `
dvx ]. Using the fact that gpvx q must be proportional to the Boltzmann
factor, where E “ mvx2 {2, and normalizing gpvx q, show that:
c
m 2
gpvx q “ e´mvx {2kB T (7.23)
2πkB T

Applying a similar reasoning to vy and vz , find an expression for the frac-

tion of particles with velocities in the interval [pvx , vy , vz q, pvx ` dvx , vy `
dvy , b
vz ` dvz q]. Thus, determine the velocity distribution f pvq, where
v“ vx2 ` vy2 ` vz2 , proving the Maxwell distribution.

3.** Natural width of emission lines

Excited atoms spontaneously emit photons when relaxing back to the
7.10. Problems 83

ground state. Consider a hydrogen atom that emits visible light. By

equating the difference in energy between two energy levels with the
energy radiated by an accelerated electron (Pr. 7.1), estimate the time ∆t
necessary for the transition. Using the Heisenberg uncertainty principle
applied to energy and time (∆E∆t ě h̄{2), find the uncertainty in the
energy ∆E. This is the natural thickness of the emission lines for a
hydrogen atom that emits visible light.
4.* Doppler broadening

Figure 7.4: Emission lines from stellar spectra are widened because of the (thermal) motion
of atoms on the surface.

In addition to the natural width of emission lines, the random motion of

atoms on the surface of a star causes, due to the Doppler effect, a further
increase in width (Fig. 7.4). Let x be the axis of observation. Starting
from the Boltzmann distribution law, show that the fraction of atoms
with mass m, emitting between λ and λ ` dλ, is:

mc2 pλ ´ λ0 q
" *
Ipλq9 exp ´ , (7.24)
2kB T λ20

where λ0 is the wavelength of emission at rest. Therefore, show that

the width for which the intensity is at least half the maximum value
(Ipλq ą Ipλqmax {2) is:
c
∆λ kB T
“ 2 2 ln 2 . (7.25)
λ mc2

5.* Wien and Stefan-Boltzmann laws

The blackbody distribution law can be obtained from the Boltzmann
distribution, where the energy is that of a quantum harmonic oscillator
84 Chapter 7. Electromagnetic Radiation

E “ p 21 ` nqh̄ω, since electromagnetic waves inside the body form stand-

ing waves. It is then possible to obtain the following expression for the
power emitted in the frequency interval [f , f ` df ]:
2πh f3
Pf df “ df . (7.26)
c2 hf
e kB T
´1
Compute the expression for Pλ dλ, defined as the energy emitted in the
interval [λ, λ`dλ]. Prove that the maximum of Pλ is at λT “ 0.0029 K m
(Wien’s law). Also prove that the maximum of Pf is at f T “ 0.0051 K/s.
By integrating over all frequencies, prove that P “ σT 4 (Stefan-Boltzmann
law), where σ is given by:
2π 5 kB
4
σ“ . (7.27)
15c2 h3
You can use the following integral:
x3 π4
ż8
x
dx “ .
0 e ´1 15

6. Ultraviolet catastrophe
Obtain an approximate expression for Pλ , in the limiting case of long
wavelengths. Show that, in this approximation:
2πckB T
Pλ “ . (7.28)
λ4
This is the Rayleigh-Jeans law. It is possible to derive this equation
using only classical physics (in fact, Planck’s constant does not appear
in Eq. 7.28). Before quantum mechanics, this was believed to be the
distribution law governing black body radiation. Explain why people
talked about the “ultraviolet catastrophe”.
7.** Why do stars emit energy?
Understanding the physical process that allows stars to emit energy is
essential for estimating their lifetime. Over time, numerous models have
been proposed. Explain, with reasoning supported by calculations, why
the first two proposals cannot be correct, while the third is believable.
• 1841, Mayer’s hypothesis. The Sun takes energy from the impact
of asteroids on its surface. (Hint: first solve Ex. 10.13)
• 1854, Kelvin-Helmholtz hypothesis. According to the so-called con-
traction theory, the Sun emits energy at the expense of its gravit-
ational potential energy. Over time, the star contracts, until all of
its potential energy is exhausted (Hint: first solve Pr. 10.4).
7.10. Problems 85

• In the nucleus of the star, mass is converted directly into energy

through a process known as nuclear fusion, according to E “ mc2 .
Furthermore, only 10 % of the initial mass of hydrogen in the star
can be transformed into helium. In the hydrogen-helium conversion,
0.72 % of the hydrogen mass is converted into energy.
8.* Inside a star
It is possible to describe the interior of a star with four differential equa-
tions that govern the distribution of pressure, mass, energy production
and temperature gradient. Assume throughout that the star is spheric-
ally symmetric.
• The equation of hydrostatic equilibrium governs the pressure inside
a star. Assuming that the star is in equilibrium with its internal
pressure P prq and gravitational force, prove that:

dP Gmprqρprq
“´ , (7.29)
dr r2
where ρprq is the density at a distance r from the centre of the star
and mprq is the mass contained within that distance (Hint: first
solve Pr. 10.2).
• The second equation governs the mass distribution. Prove that:

dmprq
“ 4πr2 ρ . (7.30)
dr

• The third equation governs the production of energy. Let  be the

coefficient of energy production per unit mass and unit time, and
let Lr the luminosity produced at a distance r from the centre of
the star. Prove that:
dLr
“ 4πr2 ρ  . (7.31)
dr

• The fourth equation governs the temperature gradient inside the

star. Let kprq be the coefficient of energy absorption, such that
k ρ dr is equal to the energy absorbed by a layer dr of the star,
with density ρ. Given that the radiation pressure of a photon gas
is Prad “ 1{3 aT 4 , where a is the radiation constant, show that:

dT 3 kρ Lr
“ . (7.32)
dr 4ac T 3 4πr2
It can be proven that a “ 4σ{c.
 8

Flux and Magnitude

In this chapter we introduce the basic concepts that characterize electromag-

netic radiation.

8.1 Flux

The flux is the amount of energy incident on a surface per unit time and
unit area. Assuming that a star radiates isotropically with power L (intrinsic
brightness), at a distance d from the star its energy is distributed on a sphere
of radius d, with surface area 4πd2 . Therefore, the flux at this distance is:

L
F “ . (8.1)
4πd2
The flux is directly proportional to the intrinsic brightness and inversely pro-
portional to the square of the distance. Consider, for example, the Sun ob-
served from Earth. Its brightness is L@ “ 3.828 ¨ 1026 W, while the Sun-
Earth distance is dC “ 149.6 ¨ 109 m. Hence, the solar flux on Earth is
k “ L@ {p4πd2C q “ 1367 W. This is known as the solar constant.
If we know the radius and the temperature of a star, we can replace L in
Eq. 8.1 with the expression given by the Stefan-Boltzmann law. The previous
equation then becomes:
σeR2 T 4
F “ . (8.2)
d2

8.2 Albedo

The albedo (A) of a body is the ratio of the reflected (Lr ) to the incident (Li )
power:
Lr
A“ .
Li
Only a part of the radiation incident on a body is again reflected into space,
while the rest is absorbed. As a result, the body warms up, emitting the
86
8.3. Stellar magnitudes 87

absorbed energy as thermal radiation. The difference between the incident

and the reflected power represents the absorbed power: Li ´ Lr “ Li p1 ´ Aq.
A body that reflects most of the incident radiation has a high albedo, while
a darker body is characterized by a low albedo. Snow has an albedo of 0.9,
the oceans — 0.06. On land, the albedo varies from 0.1, in forests, to 0.4,
in deserts. Because the Earth is mainly covered by water, the average albedo
of its surface is about As “ 0.08. The albedo of the atmosphere is instead
Aatm “ 0.24, due to the high reflectivity of the clouds. The average albedo of
the Earth is about AC “ 0.31 (a little less than the sum of As and Aatm ).

8.3 Stellar magnitudes

Since ancient times astronomers have recorded the positions of stars on the
celestial sphere and measured their brightnesses. The first stellar catalogue
was compiled in 160 BC by the Greek astronomer Hipparchus. He divided
the stars into six classes of brightness, from the first magnitude, formed by the
brightest stars, to the sixth, formed by the stars barely visible to the naked eye.
However, when more sensitive instruments became available, it was clear that
two stars with a magnitude difference of one had a flux ratio of approximately
2.5. In 1854, Norman Pogson replaced the Hipparchus scale with a more precise
one, while remaining as faithful as possible to the old classification method.
According to this new scale, it was established that the ratio between the fluxes
of a first and sixth magnitude star should be 100. Therefore,
? two stars with a
magnitude difference of one would have a flux ratio of 5 100 “ 2.512. Adopting
the convention used by Hipparchus, we can write:
Fx ?
5
“ p 100q´∆m , (8.3)
F0
where ∆m represents the difference in magnitude between an object with flux
Fx and another with flux F0 . The minus sign in front of ∆m signifies that
higher magnitudes correspond to lower fluxes.
In the following derivation, we will use the properties of the logarithm given in
Appendix A.5. In Eq. 8.3, we take the logarithm of both sides, isolating ∆m:
Fx ∆m
log “ log 100´ 5
F0
2∆m
“ log 10´ 5
2
“ ´ ∆m ,
5
Fx
ñ ∆m “ ´2.5 log .
F0
88 Chapter 8. Flux and Magnitude

In general, for two stars with magnitudes m1 and m2 :

F1
m1 ´ m2 “ ´2.5 log . (8.4)
F2

While the inverse formula is:

F1
“ 10´0.4pm1 ´m2 q . (8.5)
F2

Sirius, the brightest star in the night sky, has an apparent magnitude of -1.5.
The full Moon has a magnitude of -12.74; the Sun of -26.74. If we know the
radii, temperatures and distances of two stars from Earth, we can substitute
Eq. 8.2 for the fluxes:
R1 2 T1 4 d2 2
m1 ´ m2 “ ´2.5 log ,
R2 2 T2 4 d1 2
which simplifies to:
R1 T1 2 d2
m1 ´ m2 “ ´5 log . (8.6)
R2 T2 2 d1
If the stars share some parameters, a simpler equation can be obtained. Note
that Eq. 8.6 compares the bolometric magnitudes (mbol ), that is, the fluxes
integrated over the whole spectrum. Since our eyes record light in a small win-
dow of the electromagnetic spectrum, the bolometric magnitude is, in general,
different from the visual magnitude mv (see “Magnitude Systems”).
The flux, recorded by light-sensitive devices, is proportional to the brightness
of the body. Our eyes perceive brightness in a slightly different way, and it
is for this reason that we use the magnitude scale. The magnitude is directly
proportional to the logarithm of the brightness (i.e. to the logarithm of the
flux).
The fact that the human eyes perceive light in a logarithmic way can be cla-
rified with the following argument. Let us imagine we have a large number
of light bulbs, all of the same brightness. At first we only turn on one, then
another, and compare the overall brightnesses in the two cases. You will agree
that the perceived brightness doubles. Now, let us turn on 100 light bulbs, and
see how the brightness changes by adding another one. For a photosensitive
device, the increase in brightness is the same as it was in going from one to
two bulbs. However, for the human eye, the brightness does not noticeably
change. We would perhaps need to turn on another 100 bulbs to notice the
same increase in brightness. In other words, what our eyes perceive is not the
absolute change in flux ∆F , but rather the relative
ş change ∆F {F . Hence the
logarithmic relation, which is just the integral of dF {F .
8.3. Stellar magnitudes 89

Absolute Magnitude
The flux of a star depends on both its intrinsic brightness and on its dis-
tance from Earth. Therefore, an intrinsically brighter star, placed at a greater
distance than an intrinsically fainter star, may appear less luminous. Let us
imagine moving both stars to the same distance from Earth. If we now find
that the first star is brighter than the second, we conclude that the former is
also intrinsically brighter. For this reason, it is convenient to introduce the
quantity called absolute magnitude, equal to the magnitude that a star would
have if it were placed at the conventional distance of 10 parsec. Therefore, by
comparing the absolute magnitudes, we are effectively comparing the intrinsic
luminosities of the stars, rather than their fluxes.
Starting from this definition, let us derive the formula that relates the absolute
magnitude of a star to its apparent magnitude and distance from Earth. If a
star was placed at a distance of 10 pc, its flux would be:
L
FM “ .
4πp10 pcq2
However, the actual distance of the star (in parsec) is d, therefore the flux is:
L
Fm “ .
4πd2
Denoting with M the absolute magnitude of the star, and applying Eq. 8.4:
L{p4π ¨ 10 pcq2
M ´ m “ ´2.5 log
L{p4πd2 q
´ d ¯2
“ ´2.5 log
10
“ ´5 log d ` 5 .

Therefore, the absolute magnitude is:

M “ m ´ 5 log d ` 5 , (8.7)

where d is measured in parsec. If, instead of the distance, we know the parallax
of the star, it is possible to use the formula d “ 1{πp 2 (Eq. 9.5):
1
M “ m ´ 5 log ` 5.
πp 2
Using the logarithm property log 1{x “ ´ log x, the last equation becomes:

M “ m ` 5 log πp 2 ` 5 . (8.8)
90 Chapter 8. Flux and Magnitude

Magnitude of a Composite System

We want to find the overall magnitude of multiple light sources that appear to
the observer as a single, brighter object. Since the magnitude scale is logar-
ithmic, to obtain the magnitude of a composite system we cannot simply add
up the magnitudes of each source. We can, however, add up their fluxes, which
are proportional to their luminosities. Using Eq. 8.5 for each source:

F1 “ 10´0.4pm1 ´m0 q ¨ F0 ,
F2 “ 10´0.4pm2 ´m0 q ¨ F0 ,
... ,
´0.4pmn ´m0 q
Fn “ 10 ¨ F0 ;

where m0 and F0 are the magnitude and the flux of a reference object (intro-
duced for convenience, but will simplify at the end). The flux of the system is
obtained by summing the fluxes of the components. Therefore, applying Eq.
8.4 to the composite system and to the reference object:
F1 ` F2 ` ... ` Fn
msys ´ m0 “ ´2.5 log .
F0
Substituting the fluxes as a function of the magnitudes:
” 10´0.4pm1 ´m0 q ` 10´0.4pm2 ´m0 q ` ... ` 10´0.4pmn ´m0 q ı
msys ´m0 “ ´2.5 log F0 .
F0
Simplifying F0 :

msys ´ m0 “ ´2.5 log r10´0.4pm1 ´m0 q ` 10´0.4pm2 ´m0 q ` ... ` 10´0.4pmn ´m0 q s .

Dividing by the factor 100.4 m0 :

msys ´ m0 “ ´2.5 log r100.4 m0 ¨ p10´0.4 m1 ` 10´0.4 m2 ` ... ` 10´0.4 mn qs .

Finally, using the logarithm property log px ¨ yq “ log x ` log y:

msys ´ m0 “ ´2.5 log 100.4 m0 ´ 2.5 logr10´0.4 m1 ` 10´0.4 m2 ` ... ` 10´0.4 mn s .

But ´2.5 log 100.4 m0 “ ´m0 , therefore the last equation simplifies to:

msys “ ´2.5 logr10´0.4 m1 ` 10´0.4 m2 ` ... ` 10´0.4 mn s . (8.9)

8.3. Stellar magnitudes 91

Integrated Magnitude
The relationships given above refer to point objects, such as stars and planets.
The total brightness of an extended source, such as a galaxy or a star cluster,
can be expressed in terms of the integrated magnitude, equal to the magnitude
that the object would have if all of its light were concentrated into a point
source. The integrated magnitude can be obtained from the surface brightness
and the area of the object under consideration. The surface brightness, with
units of mag/arcsec2 , is a measure of the brightness of a portion of the surface
with an area equal to one arcsecond squared.
Let Fs be the flux from one arcsecond squared of the object and A its total area
(expressed in arcsecond squared). Let mint and ms be the integrated magnitude
and surface brightness respectively. From Eq. 8.4, it follows that:
Fint
mint ´ ms “ ´2.5 log .
Fs
Since Fint “ A ¨ Fs , the last equation can be written as:

mint “ msup ´ 2.5 log A . (8.10)

The apparent magnitude of an astronomical object is generally quoted as an

integrated value. Therefore, if a galaxy has a magnitude of 12.5, it means we
receive the same total amount of light from the galaxy as we would from a star
with magnitude 12.5. However, the star is so small that it can be effectively
considered as a point source (the largest angular diameter, that of R Doradus, is
0.057 ˘ 0.005 arcsec), whereas the galaxy may extend over several arcseconds
or arcminutes. Therefore, the galaxy will be more difficult to see than the
star. Surface brightness can be useful to understand the visibility of a certain
object. It can be compared to the surface brightness of the night sky, which is
21.9 mag/arcsec2 at the zenith, in the V band.

Magnitude Systems
Usually, when talking about brightness, we refer to the bolometric brightness,
i.e. the brightness of the object summed over all wavelengths. However, prac-
tically speaking, the bolometric brightness cannot be measured directly, since
each measurement only records light in a narrow region of the electromagnetic
spectrum. For this reason, it is customary to specify the brightness of an ob-
ject in a particular region of the electromagnetic spectrum. The sensitivity of
the human eye varies with the wavelength. In daylight, it has a maximum at
the wavelength of 550 nm (yellow), while it decreases towards red and violet.
The magnitude that corresponds to the part of the electromagnetic spectrum
visible to the naked eye is called visual magnitude, and is denoted by mv .
Photographic plates are usually more sensitive to blue and violet and can also
92 Chapter 8. Flux and Magnitude

record radiation not visible to the naked eye. The photographic magnitude
is usually denoted by mpg , and is generally slightly different from the visual
magnitude. However, the sensitivity of the eye can be simulated using a yellow
filter and plates sensitive to yellow and green light. If, ideally, we were able to
measure the magnitude of the star over its entire spectrum, we would obtain
the bolometric magnitude, denoted by mbol . In practice, this is difficult since
a large part of the electromagnetic spectrum is absorbed by the atmosphere,
and different wavelengths require different sensors. Bolometric magnitude and
visual magnitude are related by:
mbol “ mv ´ BC , (8.11)
where BC is called the bolometric correction. By definition, the bolometric
correction is zero for Sun-type stars (G2V spectral class). In all other cases,
for both warmer and colder stars, the bolometric magnitude is greater than
the visual magnitude, therefore BC is always negative.
One of the most common systems used to classify the magnitude of stars, in
the region of the spectrum close to visible light, is the UBV system, developed
by Harold L. Johnson and William W. Morgan in 1950. In this system, mag-
nitudes are measured using three different filters: U“Ultraviolet, B“Blue and
V“Visual, hence the name UBV. In the UBV system it is usual to give the
visual magnitude of the star, together with the quantities U´B and B´V,
called the colour indices. The colour indices were chosen to be zero for an A0-
type star (like Vega). Therefore, Vega has V“ 0.03, B´V“U´B“0. Instead,
the Sun has V“ ´26.8, B´V“ 0.62 and U´B“ 0.10. Later, this system was
expanded to UBVRI, in order to include the red R and infrared I.

8.4 Extinction and Optical Depth

All the above equations are valid if the space between the source and the ob-
server is completely empty, i.e. in the absence of a medium (such as interstel-
lar dust or the atmosphere), which absorbs and scatters light. The interstellar
dust is composed of grains of various sizes, the smallest of which are planar mo-
lecules, such as aromatic hydrocarbons (PAHs), formed by a few dozen atoms.
Then, there are small three-dimensional grains (SG, between 100 ´ 2000 nm)
and big grains (BG, over 2000 nm) mostly composed of silicates and graphite.
If the wavelength is less than or comparable to the grain size, light is mostly ab-
sorbed. If, instead, the wavelength is significantly greater, it can pass through
the interstellar medium without appreciable absorption. Therefore, PAHs ab-
sorb ultraviolet radiation, while BG can also absorb light emitted in the red
band by relatively old stars. Since the interstellar medium alters the colour of
the stars, it is usually possible to trace back the value of the extinction (ab-
sorption and scattering), by comparing the observed colour with the expected
8.4. Extinction and Optical Depth 93

one (from the spectral class of the star). Since blue light is scattered more
than red light, the B´V difference increases in the presence of an interstellar
medium.
Additionally, light is scattered and absorbed by the atmosphere. The most
important wavelength range for which the atmosphere is transparent is the
so-called optical window, as it corresponds to the sensitivity range of the hu-
man eyes (400 ´ 700 nm). In the visible region, light is mainly absorbed and
scattered by fine dust particles. At wavelengths below 300 nm, ozone is the
main absorber, and it prevents ultraviolet light from reaching Earth’s surface.
Ozone is concentrated in a small band at a height of about 20´30 km, protect-
ing the Earth from harmful, high-energy radiation. At shorter wavelengths,
the molecules mostly responsible for absorption are molecular oxygen and ni-
trogen. Almost all radiation below 300 nm is absorbed by the atmosphere. At
wavelengths greater than visible light, however, the atmosphere is quite trans-
parent up to 1.3 µm. At this wavelength, electromagnetic radiation reaches
the Earth in only a few thin windows. Wavelengths between 20 µm-1 mm are
completely absorbed by the atmosphere. The radio window corresponds to the
range 1 mm-20 m. Beyond 20 m (actually the upper limit depends on the daily
conditions of the atmosphere), the ionosphere absorbs all radiation.
Light scattering by air molecules is inversely proportional to the fourth power
of the wavelength (Rayleigh scattering). Therefore, the sky appears blue during
the day because this is the colour with the shortest wavelength (and, hence, the
most scattered by the atmosphere). At sunset, light from the Sun has to travel
a greater distance through the atmosphere before reaching our eyes. In this
case, both blue and red light are highly scattered, but blue light is scattered
too early to be observed this far, therefore the intensity of red light reaching
our eyes is greater.
Let us now derive a relationship between the extinction of light and the dis-
tance it travels through an opaque medium. It is reasonable to assume that,
in a small interval rr, r ` drs, the extinction is proportional to the incident
brightness L and to the distance dr. Therefore:

dL “ ´κL dr ,

where κ is called the opacity, and is a measure of the amount of radiation

absorbed by the medium per unit length. The opacity of vacuum is zero, while
it is greater when the medium is more opaque or turbid. If the opacity is
constant along the line of sight, it is possible to integrate both sides of the
previous equation, while bringing κ outside the integral:
ż Lprq żr
dL
“ ´κ dr ,
L0 L 0
94 Chapter 8. Flux and Magnitude

Lprq
ln “ ´κr .
L0
We then obtain:
L “ L0 e´κr . (8.12)
The above equation gives the brightness of a star at a distance r, in the pres-
ence of a medium with opacity κ. The ratio L{L0 is called the transmittance,
and is often denoted by the letter e, followed by the wavelength at which it
is measured. For instance, the transmittance of the Earth’s atmosphere for
yellow light is approximately 0.8, and it is smaller for ultraviolet and infrared.
The plot of the transmittance as a function of wavelength is just the absorption
spectrum of the gas constituting the medium. The quantity κr is called the
optical depth, and is often denoted by the Greek letter τ .
In order to determine the effective magnitude of a star in the presence of an
opaque medium, it is necessary to add a correction factor to the equations in
the previous sections. Let F0 be the light flux on the surface of the star (with
radius R) and Fd the flux at a distance d. Isolating the luminosity in Eq. 8.1:

Ld “ 4πd2 Fd , L0 “ 4πR2 F0 .

Using Eq. 8.12 to write Ld {L0 , taking the ratio of the above equations:

R2 ´κd
Fd “ F0 e .
d2
To compute the absolute magnitude, we need the flux (measured without ex-
tinction) at a distance of 10 pc:

R2
F10 pc “ F0 .
p10 pcq2
Therefore, measuring d in parsecs:
F10 pc
M ´ m “ ´2.5 log
Fd
F0 R2 {p10 pcq2
“ ´2.5 log
F0 R2 e´κd {d2
d
“ ´5 log ´ 2.5 log eκd
10
“ ´5 log d ` 5 ´ 2.5 log eκd .

Taking extinction into account, the equation that relates the apparent and
absolute magnitudes is:

M “ m ´ 5 log d ` 5 ´ 2.5 log eκd . (8.13)

8.4. Extinction and Optical Depth 95

Compared to the formula without extinction, we note the additional factor

Apdq “ ´2.5 log eκd . Let us examine some limiting cases. For κ equal to zero,
the term eκd is equal to unity, therefore the logarithm and the correction term
are zero. This situation corresponds to the absence of an interstellar medium,
in which case Eq. 8.13 reduces to Eq. 8.7. For κ that tends to infinity, eκd
tends to infinity, therefore the correction factor is minus infinity. Now, consider
two possibilities. If we fix the absolute magnitude of the star, it follows that
the apparent magnitude is infinite, i.e. the apparent brightness is zero. On the
other hand, if we fix the apparent magnitude, we find an absolute magnitude
of minus infinity. Indeed, the intrinsic brightness of the object must be infinite
to give a finite apparent magnitude in the presence of a medium that lets no
light through.
Let us now compute the variation of the luminosity of a star as a function of its
altitude. Let RC be the radius of the Earth, h the height of the atmosphere and
90˝ ´ z the altitude of the star (where z is the zenith distance). Since most of
the air is concentrated in the first 10 km of the atmosphere, h « 10 km ! RC .
From Fig. 8.1, it can be seen that, for stars close to the horizon, the distance
which light travels through the atmosphere is h{ sin z. Therefore:

mspace “ msurface ´ 2.5 log eκatm h{ sin z .

Taking into account extinction by the interstellar medium and the atmosphere,
we obtain the most general formula:

M “ msurface ´ 5 log d ` 5 ´ 2.5 log eκint d ´ 2.5 log eκatm h{ sin z . (8.14)

Figure 8.1: For small zenith distances (z), the distance which light travels through the
atmosphere is h{ sin θ, where h « 10 km is the height of the atmosphere with significant
density.
96 Chapter 8. Flux and Magnitude

8.5 Exercises

1. (NAZ 2015, Th.S, q.4) Approximating the Sun as a blackbody, with a

temperature of Te “ 5778 K, compute its luminosity L@ , the light flux
on Earth and the fraction of energy that is incident on the Earth.
2.* (ONAA 2019, Th.S2, q.10) The decay of radioactive elements inside the
Earth gives rise to an average heat flux of 5 ¨ 10´2 W/m2 on the surface
of the Earth. Express this flux as a fraction of the thermal radiation
absorbed from the Sun, corresponding to a temperature of 300 K. If
these radioactive decays were the only source of energy, what would the
temperature of the Earth be?
3. From what distance do we need to observe a 100 W lamp so that it
appears as bright as the Sun?
4. With a magnitude of ma “ ´1.43, Sirius is the brightest star in the night
sky. Actually, Sirius has a small companion star (a white dwarf), with
an apparent magnitude of mb “ 8.68. How many times is Sirius brighter
that its companion?
5. The magnitude of the faintest star that can be observed with the naked
eye is mb “ 6, that of the brightest, mt “ ´1.5. What is the ratio of the
fluxes of the two stars?
6. The apparent magnitude of the Sun, as seen from Earth, is ´26.8. Com-
pute the apparent magnitudes of the Sun as seen from Mercury, Venus,
Mars, Jupiter and Saturn.
7.* (NAZ 2015, Th.S, q.3) After a geomagnetic storm, the Sun’s brightness
temporarily increases by 10 %. At this moment, Jupiter is in opposition.
Find the apparent magnitudes of the Sun and Jupiter. What are the
time intervals between the occurrence of the storm and the increase in
brightness of the Sun, and between the observed increase in brightness
of the Sun and the increase in brightness of Jupiter? Assume that the
orbits of Earth and Jupiter are circular.
8. The difference in apparent magnitudes of two identical stars is ∆m “ 3.5.
Compute the ratio of their distances from Earth.
9. A star of magnitude m “ ´1.45 is d “ 8.6 ly away from Earth. What is
its absolute magnitude? Express its luminosity in terms of L@ .
10. Arcturus (α Bootes) has apparent and absolute magnitudes of m “ ´0.05
and M “ ´0.3, respectively. Compute its distance from Earth. What is
its parallax?
8.5. Exercises 97

11. The absolute magnitudes of the Sun and Sirius are M@ “ 4.83 and
Ms “ 1.4, respectively. At what distance from the Earth should Sirius
be, to appear as bright as the Sun?
12.* Estimate the number of photons that reach your eyes every second, hav-
ing been emitted by a star with apparent magnitude m “ 6. Assume
that the star only emits energy at the wavelength of λ “ 550 nm.
13.* Find the albedo of the Moon. You are given the apparent magnitudes of
the full moon m$ “ ´12.74 and the Sun m@ “ ´26.74, and the angular
diameter of the Moon α “ 321 . Assume the Moon only reflects light
coming from the Sun.
14.* (IAO 2019, Th.β, q.3) Supergiant UY Scuti is the largest and one of the
fastest burning stars currently known, with a volume of around 5 bilion
times that of the Sun. The mass lost per unit time, due to radiation, is
only 0.04% of the total mass lost over the same time, and only 0.5% of
the light passes through the upper layers of the star (that is, reaches the
observer). Estimate the absolute magnitude, the temperature, and the
remaining life time of UY Scuti.
15.* (IAO 2013, Th.β, q.2) Gliese 581g is the most Earth-like planet found
outside the Solar System, and the exoplanet with the greatest recognized
potential for harbouring life. The parallax of Gliese 581 is πp “ 0.132 ,
its magnitude m “ 8.0 and its mass M “ 0.31M@ . Estimate the orbital
period of the planet.
16.** Altair is a main sequence star with mass M “ 1.7M@ , magnitude m “
0.77 and parallax πp “ 0.1952 . Estimate its density.
17. A system is composed of three stars with apparent magnitudes of 3.67,
4.65 and 5.12. What is the magnitude of the composite system?
18. A spectroscopic binary is comprised of two stars which periodically ec-
lipse each other: one is large and cold, the other is small and hot. If the
maximum and minimum magnitudes of the system are mmax “ 3.70 and
mmin “ 4.85, respectively, what are the magnitudes of the two stars?
What is the ratio of the maximum and minimum fluxes on Earth?
19.* (IAO 2015, Th.β, q.4) According to an ancient legend of the Middle
Volga there was a constellation called White Leopard (Pardus Album)
in the sky in the very distant past, in which the number of stars was
exactly equal to the number of letters in the Greek alphabet, and the
stars had magnitudes mαPaA “ 0.10, mβPaA “ 0.20, mγPaA “ 0.30,
mδPaA “ 0.40 and so on, adding 0.1 every time until ωPaA. Calculate
the total magnitude of the stars of this constellation.
98 Chapter 8. Flux and Magnitude

20. Two stars of magnitudes 3 and 8 are observed with a telescope. A picture
is taken of the first star, with an exposure time of 10 seconds. What
should the exposure time for the second star be, if we want it to appear
as bright as the first star?
21. On an especially foggy day, the Sun appears as bright as the full moon
(observed on a clear day). Compute the optical thickness τ .
22.** (INAO 2017, Th., q.3) Human life can survive on Earth only if the
equilibrium temperature remains below 333 K. Assume that the Milky
Way contains 1011 stars uniformly distributed across the galaxy, with
and average number density of 0.14 pc´3 , and that there is a supernova
every 30 years, with luminosity 1011 L@ . What is the probability that a
supernova wipes out life on Earth, in the total life span of the Sun?

8.6 Problems

1.* Olbert’s paradox

Assuming that the stars are uniformly distributed in the universe, show
that the amount of energy which arrives on Earth from a spherical shell
(of universe) of radius r and thickness dr is the same as that from a
spherical shell of radius R and thickness dr, with r ‰ R. What is the
total brightness of the sky, as observed from the Earth? How can you
explain this result?
2.* Earth’s temperature
Compute the mean temperature on Earth, as a function of the solar
radius (R@ ) and temperature (T@ ), the Earth-Sun distance (dC ) and the
albedo of the Earth (AC ). Compute the numerical value for AC “ 0.31,
neglecting the greenhouse effect of the atmosphere.
3.*** Earth’s temperature, another model
As can be seen from the previous problem, the greenhouse effect is non-
negligible. Let ev and ei be the transmittances of the atmosphere for
the solar (visible) and terrestrial (infra-red) radiations, and let Aatm and
As be the albedos of the atmosphere and of the surface of the Earth,
respectively. Assume that the atmosphere is a thin envelope surrounding
the Earth, with radius RC and uniform temperature Tatm . Compute, as
a function of R@ , T@ , dC , As , Aatm , ev , and ei , the new value for the
Earth’s temperature. Estimate the temperature for As “ 0.08, Aatm “
0.24, ev “ 0.8, ei “ 0.20.
 9

Cosmic Distance Ladder

In this chapter we will examine the techniques that astronomers use to measure
the distances of celestial bodies. Many different methods exist, however each
of them is only applicable to a specific length scale. In fact, a direct distance
measurement is only possible for up to a few hundred parsecs, using the method
of diurnal parallax, for objects within our Solar System, and annual parallax,
for the closest stars. At greater distances, several methods rely on standard
candles (i.e. objects of known intrinsic brightness), such as Cepheid variables
or Type Ia supernovae. Even binary systems, where the orbital characteristics
can be used to calculate the masses of the component stars (and hence their
luminosities), can be used as standard candles. At even greater distances, it
becomes impossible to resolve each star, therefore it is necessary to estimate
the intrinsic brightness of whole galaxies. For example, the Tully-Fisher rela-
tion links the luminosity of a spiral galaxy with its rotational velocity, while
the Faber-Jackson method relates the luminosity of an elliptical galaxy to the
dispersion of the velocities of stars around its centre. At extremely large scales,
we can measure the redshift of a galaxy to obtain its recession velocity, and
therefore its distance, from Hubble’s law. The cosmic distance ladder is the
succession of these, and many other methods. The ladder analogy arises be-
cause no single technique can measure distances at all ranges encountered in
astronomy. Each rung of the ladder provides information that can be used to
determine the distances at the next, higher, rung.

9.1 Parallax

Parallax is a very common phenomenon that manifests itself as the change in

the angular position of an object as seen by an observer, due to the relative
motion between them. Try placing one of your fingers in front of your eyes:
first, look at it with your right eye only, then with your left eye only. Your
finger will appear in different directions, with respect to the background. In
this example, the distance between your eyes is the parallax baseline. The closer
your finger is to your eyes, or the greater the parallax baseline (for instance, you
99
100 Chapter 9. Cosmic Distance Ladder

Figure 9.1: Triangulation method: the distance of an object (such as a ship) from the
baseline, can be obtained by recording the angles φ1 and φ2 it forms with the baseline at
points P1 and P2 , which are a distance b apart.

could move your head sideways before observing the second time), the farther
apart the two images appear. It is clear that the parallax angle must depend
on the ratio between the parallax baseline and the distance of the object.
This simple concept proves immediately useful. The parallax baseline and the
parallax angle are quantities easily measurable from our position, therefore the
distance of the object can be easily obtained. In addition to being an essential
tool in astronomy, the parallax method is also used in terrestrial triangulation.
Let P1 H “ x and HP2 “ b ´ x in Fig. 9.1. Then:
$
d
“ tan φ2 (9.1a)
’
’
&
b´x
’ d
% “ tan φ1
’ (9.1b)
x
Dividing the first equation by the second:
x tan φ2
“
b´x tan φ1
´ tan φ2 ¯ tan φ2
x 1` “b
tan φ1 tan φ1
tan φ2
ñx“b .
tan φ1 ` tan φ2
Substituting x in Eq. 9.1b, we get d:

tan φ1 ¨ tan φ2
d“b . (9.2)
tan φ1 ` tan φ2
9.1. Parallax 101

This equation can be used to find the distance of an object from the baseline,
given the angles φ1 and φ2 that it forms with the parallax baseline at points P1
and P2 , which are a distance b apart. In astronomy, usually P1 H “ HP2 “ b{2,
therefore φ “ φ1 “ φ2 . Eq. 9.2 then simplifies to:
b
d“ tan φ .
2
For celestial objects d " b, therefore φ « 90˝ . Since the position of a star is
measured with respect to the background of fixed stars, it is useful to define
the parallax angle πp as the complementary of φ: πp “ 90˝ ´ φ, where πp ! 1
(hence, the parallax angle is just the angle subtended by the baseline). Since
tan p90˝ ´ πp q “ 1{ tan πp « 1{πp,rad , we find:

b{2
d“ . (9.3)
πp,rad

In the next sections, we will look at two different choices for the parallax
baseline: Earth’s diameter, and the diameter of the Earth’s orbit around the
Sun. These give the diurnal and annual parallaxes, respectively.

Diurnal Parallax

Figure 9.2: In order to measure the distance to a celestial object using the method of diurnal
parallax, we record the angles φ1 and φ2 between the object under consideration and the
reference object, for two observation sites P1 and P2 on the equator, at a distance α in
longitude. From the parallax angle πp “ pφ1 ´ φ2 q{2, it is then possible to obtain the
distance, using Eq. 9.4.

The diurnal parallax of an object (within the Solar System) can be measured
by simply performing two observations 12h apart, or by observing the same
102 Chapter 9. Cosmic Distance Ladder

object from two different locations on Earth at the same time. In the first
case, after twelve hours, the Earth has completed half a rotation, and the
observation site (preferably at the equator) has moved by a distance equal to
the Earth’s diameter. Either way, if the observation site is at the equator,
the parallax baseline is Earth’s diameter. Half of the angle between the two
images of the object is called the diurnal parallax. This method allows us to
determine the distances of most bodies in the Solar System, but the diurnal
parallax is too small to measure the distances of stars, which appear fixed.
To find the diurnal parallax, we take a fixed star, and measure the angles φ1
and φ2 it forms with the object under consideration, at two observation sites
on the equator separated in longitude by α (Fig. 9.2). The angle of diurnal
parallax πp is then πp “ pφ1 ´ φ2 q{2. Using Eq. 9.3, where b{2 “ RC sin pα{2q:

RC sin pα{2q
d“ . (9.4)
πp,rad

Since the Moon is the celestial body closest to us, lunar parallax measurements
began rather early compared to those for the Sun and the planets. One of the
most famous measurements was the one conducted by Hipparchus. This was
carried out during a total solar eclipse in Syene and during a partial solar
eclipse in Alexandria. At the same time, an observer in Syene saw the solar
disk completely eclipsed by the Moon, while one in Alexandria was able to
observe 1{5 of the solar disk, that is, 1{5 of 30 arcminutes. Assuming the Sun
to be at an infinite distance, Hipparchus estimated the Earth-Moon distance
to be in the range of 59 to 67 Earth radii: an impressive result, considering
that the accepted value today is 60.2. The reason for this uncertainty lies in
the fact that Hipparchus had not been able to determine the parallax of the
Sun. In fact, the solar parallax was more difficult to measure, since the Sun is
so bright that it is impossible to record its position relative to the stars. One
of the first methods used to calculate the distance of the Sun was the transit of
Venus on the solar disk. An alternative way is estimating the parallax of Venus
at the moment of maximum elongation (Pr. 9.1), or estimating the parallax
of Mars in opposition.

Annual Parallax
Calculating the distances of celestial objects outside the Solar System requires
a larger baseline than the Earth’s diameter. The widest basis we can use is
the diameter of the Earth’s orbit, which can be calculated using the diurnal
parallax, as described in the previous section (in this sense, the diurnal parallax
allows us to climb one rung up the ladder). In this case, we can simply carry
out two observations of an object 6 months apart (Fig. 9.3). The annual
parallax is then half of the angle between the two images of the object against
9.1. Parallax 103

the background of fixed stars. This method allows us to record parallax angles
for stars at most a few hundred parsecs away. Distant stars appear fixed and
constitute the background on which we observe the succession of all celestial
phenomena.

Figure 9.3: With the method of annual parallax, the distance to a celestial object can be
obtained by carrying out two observations six months apart. The parallax baseline is the
diameter of Earth’s orbit.

Looking at Fig. 9.3, we see that b{2 “ dC . From Eq. 9.3, it follows that:

dC
“ πp,rad .
d
Since πp is a small angle, it is best to measure it in arcseconds. In a radian
there are p180{πq¨60¨60 “ 206264.8062 « 206265 arcseconds. Therefore:

dC πp2
“
d 206265
1
ñ “ πp2 .
d{p206265 ¨ dC q

What prevents us from defining a new unit? Since we are dealing with the
par allax of an arcsecond, why not call it parsec! To get rid of that annoying
numerical factor, we then decide that 1 pc“ 206265 au (where au is the astro-
nomical unit, equal to the Earth-Sun distance, dC ).
Therefore, d{p206265 ¨ dC q is equal to the distance d expressed in parsec:
104 Chapter 9. Cosmic Distance Ladder

1
πp 2 “ . (9.5)
dpc
The parsec is therefore defined as the distance of a star with an annual parallax
of one arcsecond. We also note that 1 pc“ 3.262 ly. Proxima Centauri, the
closest star to the Solar System, has a parallax of 0.768 arcseconds. Applying
Eq. 9.5, we see that its distance from Earth is about 1.3 pc, or 4.23 ly.
The annual parallax is a tiny effect and was only measured in 1838, by Friedrich
Wilhelm Bessel. With the Hipparcos satellite (1989 – 1993), the European
Space Agency (ESA), was able to measure the parallax of 118,000 stars with an
accuracy of 0.001”, significantly better than the resolution allowed by telescopes
on Earth. Despite the impressive precision of the Hipparcos satellite, the
distances it can measure are still small compared to the distance of the galactic
centre, about 8kpc away from the Solar System. A more recent satellite, Gaia,
measured the parallax of 1 billion stars with a precision of 10 µarcsec.

9.2 Spectroscopic Parallax

The spectroscopic parallax does not have much in common with the trigono-
metric parallax, but the use of the term “parallax” refers to the fact that this
method is also used to measure the distances to celestial bodies.
Consider Eq. 8.13:
m ´ M “ 5 log d ´ 5 ` Apdq ,
where d is the distance measured in parsec and Apdq is the correction term for
extinction. If we know m ´ M , we can obtain d, in principle.
While m can be directly calculated from observation, finding M is a much
harder task. The absolute magnitude can be estimated from the spectral class
of the star, but the relation that links spectral class and intrinsic brightness is
only a rough approximation. Indeed, the intrinsic brightness depends on sev-
eral factors, such as the percentage of heavy elements, which can vary slightly
within the same spectral class. At the same time, this relation is not perfect
because the main sequence is not a thin line, but it extends by one mag-
nitude in either direction. Then, there is the matter of estimating the term
Apdq. Usually, it can be found by comparing the theoretical colour of the star
(based on its spectral class) with the one observed. While the colours would be
the same with no extinction, the interstellar medium primary absorbs shorter
wavelengths, therefore making the star appear redder (see Sec. 8.4). An estim-
ate of the extinction can therefore be obtained from the colour indices (B´V,
U´B) of the star.
9.3. Standard Candles 105

9.3 Standard Candles

A standard candle is an astronomical object whose intrinsic brightness can be

accurately calculated from measurements conducted on Earth. Examples of
standard candles are Cepheid variables and supernovae.
Cepheid variables are a particular type of variable stars, whose brightness var-
ies in a predictable and regular way due to the instability of their surface. In
fact, even a small increase in pressure tends to expand the surface of the star
and lower its temperature, thus making it appear fainter (Pr. 9.2). In turn,
the initial increase in volume causes a decrease in pressure, which then leads
the star to compress, increasing its temperature and, therefore, its brightness.
By modelling Cepheid variables as thermodynamic engines, Eddington was the
first to explain how this periodic cycle sustains itself. Imagine a combustion
engine, where a spark triggers the combustion when the gas is maximally com-
pressed by the piston. In the same way, the upper layers of the star become
more opaque during the expansion phase, therefore absorbing more energy
and being pushed outwards. Then, during the compression phase, they be-
come more transparent, absorbing less energy and being pushed inwards by
gravity. This cycle is called κ mechanism, since “κ” is the Greek letter which
conventionally denotes the opacity of a medium.
The name of this class of stars comes from δ Cephei, discovered in 1784 by
John Goodricke. The relationship between period and luminosity was first
discovered in 1912 by Henrietta Swan Leavitt. After studying thousands of
variable stars in the Magellanic Clouds, she understood that the period of
variation of brightness was only a function of the intrinsic brightness of the
Cepheid, given that all of the stars she was observing were approximately at
the same distance from Earth. For Cepheid variables, the logarithm of the
period of variation of brightness is proportional to the logarithm of the aver-
age brightness (Fig. 9.4, overleaf). Owing to this relationship, and to the great
precision with which the pulsation period can be measured, Cepheid variables
can be used as standard candles in determining the distance of globular clusters
and galaxies in which they are contained. The period-luminosity relationship
for classic Cepheid variables is:

xLy
log “ 1.15 log Pd ` 2.47 , (9.6)
L@

where xLy denotes the average luminosity and Pd is the period of pulsation,
measured in days.
Edwin Hubble was the first to identify a Cepheid variable in the Andromeda
galaxy. He estimated its distance at 9 ¨ 105 light years, thus proving it had to
be located outside our galaxy. This discovery showed that the Milky Way is
just one of countless galaxies in the universe.
106 Chapter 9. Cosmic Distance Ladder

Today Cepheid variables are divided into two types: Type I Cepheids (or
Classical Cepheids) and Type II Cepheids. While the former belong to the
(younger) population I stars, the latter belong to (older) population II stars,
usually poor in heavy metals, and hence fainter. Because Type I Cepheids
are brighter, they can be used to measure greater distances compared to Type
II Cepheids. There is also another class of variable stars, called RR Lyrae.
However, these are fainter than both types of Cepheids.

Figure 9.4: Period-luminosity relationship for three types of variables.

9.4 Tully-Fisher Relation

All spiral galaxies rotate around their centre. This rotation gives rise to a
blueshifted spectrum in the part of the galaxy that is rotating towards us,
and a redshifted spectrum in the part that is rotating away from us. Let vc
be the radial velocity of the centre of the galaxy, vrot the maximum velocity
of rotation of the stars around the centre and i the angle between the line
of sight and the perpendicular to the plane of the galaxy. To an observer on
Earth, the extremes appear to be moving at radial velocities v1 “ vc ´ vrot sin i
and v2 “ vc ` vrot sin i (Fig. 9.5). The displacements in the spectral lines
∆λ1 “ λ1 v1 {c and ∆λ2 “ λ2 v2 {c are:
$
’ ∆λ1 vc ´ vrot sin i
’
& “ (9.7)
λ c
’ ∆λ2 vc ` vrot sin i
’
% “ (9.8)
λ c
Subtracting the two equations:
c
vrot “ p∆λ2 ´ ∆λ1 q . (9.9)
λ sin i
9.4. Tully-Fisher Relation 107

Figure 9.5: Consider a galaxy whose plane forms an angle 90˝ ´ i with the line of sight,
and in which the maximum velocity of stars is vrot . To an observer on Earth, the extremes
appear to be moving at radial velocities vc ´ vrot sin i and vc ` vrot sin i, where vc is the
velocity of the centre of the galaxy.

The above equation allows us to find the rotational velocity of the galaxy from
quantities measurable on Earth.
The faster the galaxy rotates, the greater ∆λ1 and ∆λ2 . A galaxy rotates
because of its own gravity, therefore, the greater its mass, the greater its rota-
tional velocity. The rotational velocity is related to the mass (Eq. 10.18) by:

GM v2 R
2
vrot “ ñ M “ rot .
R G
The greater the mass, the larger the number of stars, and therefore the brighter
the galaxy. Hence, galaxies rotating faster are brighter. Assuming spiral galax-
ies to be comprised (on average) of Sun-like stars, the luminosity-to-mass ratio
must be a constant for all galaxies, hence CLM “ L{M . Then:
2 R
vrot
L “ CLM .
G
Assuming that all galaxies have approximately the same surface luminosity,
we substitute CLR “ L{R2 for R:
2
CLM 4
vrot 4
L“ “ Cvrot ,
CLR G2
2 {C
where C “ CLM LR . Hence, we obtain the Tully-Fisher relation:

4
L “ Cvrot . (9.10)
108 Chapter 9. Cosmic Distance Ladder

An alternative form relates the magnitude to the rotational velocity. Taking

as a reference a galaxy with rotational velocity vrot,0 and absolute magnitude
M0 , we use Eq. 8.4:
L
M “ M0 ´ 2.5 log
L0
4
vrot vrot
“ M0 ´ 2.5 log 4 “ M0 ´ 10 log .
vrot,0 vrot,0

Hence, the Tully-Fisher relation can also be written as:

vrot,1
M1 ´ M2 “ ´10 log . (9.11)
vrot2

It is therefore possible to compute the absolute magnitude of a galaxy from

the absolute magnitude of another galaxy and the ratio of their rotational
velocities, which can in turn be determined from Eq. 9.9. However, Eq. 9.11
is not exact, and the constant in front of the logarithm (10, in this case)
actually varies according to the type of spiral galaxy under consideration. We
can divide spiral galaxies into at least three classes (Sa, Sb, Sc), depending
on their surface luminosity CLR and luminosity-to-mass ratio CLM . We then
obtain three equations:
$
&M “ ´9.95 ¨ log vrot ` 3.15 for Sa galaxies
’ (9.12a)
M “ ´10.2 ¨ log vrot ` 2.71 for Sb galaxies (9.12b)
’
M “ ´11.0 ¨ log vrot ` 3.31 for Sc galaxies . (9.12c)
%

Although not exact, the theoretical model described in this section gives a
reasonable estimate for the absolute magnitude of a spiral galaxy.
The spectral line that is often used to measure the Doppler shift is the hydrogen
line at 21 cm. Being very narrow, this line allows the measurement of the
smallest displacements. At the same time, because of its long wavelength, it
is not significantly absorbed by the interstellar medium. This method allows
us to measure distances up to 200 Mpc, beyond which the displacement in the
spectral lines becomes too tiny to observe.

9.5 Hubble’s Law

In 1929, cosmology achieved one of its most important discoveries when Ed-
win Hubble found a relationship that linked the velocities of galaxies to their
distances from Earth, thus showing that the universe is expanding. It was the
first experimental verification that confirmed what had already been predicted
by Georges Lemaître, in 1927. The linear relationship obtained at that time is
9.5. Hubble’s Law 109

Figure 9.6: An observer on Earth sees A and B receding with velocities v~a “ H0 r~a and
v~b “ H0 r~b , respectively. Therefore, an observer in A sees B moving away with a velocity of
v~b ´ v~a “ H0 pr~b ´ r~a q (directed radially outwards from A). Hence, also A can be regarded
as the centre of expansion of the universe.

actually valid only at relatively small distances from Earth, where the acceler-
ation of the expansion of the universe is negligible. Hubble’s law states that:

v “ H0 d , (9.13)
where H0 is Hubble’s constant and d is the distance. According to the latest
measurements, H0 “ 68 ˘ 5.5 km s´1 Mpc´1 . In the case of small velocities,
the redshift z “ ∆λ{λ is given by:
v H0 d
z“
“ .
c c
In the more general case, when special relativity is taken into account, we can
use Eq. 7.20 to obtain: b
H0 d
1` c
z“b ´ 1.
H0 d
1´ c
We now want to show that every point in the universe can be regarded as
the centre of this expansion, described by the same Hubble’s law. Let us place
the Earth at the centre of a Cartesian system and consider two galaxies A
and B, with distances r~a and r~b from Earth (Fig. 9.6). According to Hubble’s
law, the magnitudes of the recession velocities of A and B are va “ H0 ra and
vb “ H0 rb , respectively. Since the recession velocities are directed along the
radial direction (Earth-galaxy, in both cases), they can be written in vector
form as:
v~a “ H0 r~a
v~b “ H0 r~b .
110 Chapter 9. Cosmic Distance Ladder

The recession velocity of galaxy B, observed from A, is then:

v~b ´ v~a “ H0 r~b ´ H0 r~a “ H0 p~

rb ´ r~a q .

But r~b ´ r~a is the distance of B from A, and is directed radially outwards from
A. Hence, also A can be regarded as the centre of expansion of the universe,
with the same Hubble’s law.

9.6 Exercises

1. (NAZ 2015, Th.J, q.4) Compute the distance to Procyon (α CMi) in

parsecs and in light years, given its parallax πp “ 0.2862 . Procyon has
an apparent magnitude of m “ 0.34. If its distance were 10 times greater,
would it still be visible to the naked eye?
2. A spiral galaxy in the Virgo Cluster rotates around its centre with a max-
imum velocity of vrot,1 “ 175 km/s, and its apparent magnitude is mv,1 “
11.87. Another spiral galaxy, in the Coma Cluster, has vrot,2 “ 245 km/s
and mv,2 “ 14.63. Find the ratio of the distances to the galaxies.
3. If the uncertainty in the measurement of the absolute magnitude of a
Cepheid variable is ∆M “ 0.5, what is the uncertainty in its distance?
4.* The observed rotational velocity of a spiral galaxy around its centre is
vobs “ 300 km/s. The ratio between its semi-major and semi-minor axes
is a{b “ 1.74, and the apparent magnitude is m “ 18 (H band). The
redshift is z “ 0.20. Obtain the value of Hubble’s constant. You are
given the Tully-Fisher relation in the H-band:

MH “ ´9.5 log vrot ` 2.08

5. (CAO 2018, Th., q.5) A given elliptical galaxy has an apparent mag-
nitude of m “ 18 and redshift of z “ 0.1. Estimate the mass of the
galaxy. Interstellar absorption is neglected.
6. (MyAO 2018, Th., q.10) Estimate the age of the Universe from Hubble’s
constant.
7.* (CzAO 2018, Th.AB, q.C) Suppose that the universe contains only mat-
ter that interacts only by gravity. Modelling the universe as a sphere
of radius R that expands radially due to Hubble’s law, assuming it has
a constant density of ρ, determine the total energy of a spherical shell
with radius r and thickness dr. What is the critical density for which
the energy of the universe is zero? (Hint: use Eq. 10.22).
9.6. Exercises 111

8.* (USAAAO 2020, Th.2, q.3) The Lyman-break galaxy selection technique
makes use of the fact that any light from galaxies with wavelength shorter
than the Lyman limit (the shortest wavelength in the Lyman series) is
essentially totally absorbed by neutral gas surrounding the galaxies. The
ionization energy of hydrogen is 13.6 eV. Suppose that we are observing
galaxies in the V band, whose effective midpoint is 551 nm and bandwidth
is 99 nm.
• At what range of redshifts would we begin to see galaxies “disap-
pear” (break) from images in the V band?
• What range of recessional velocities (km/s) and distances (Mpc)
does this correspond to? Assume only Hubble expansion contributes
to the radial velocity and redshift.
9.** (IAO 2012, Th.α, β, q.6) Astronomers have discovered a distant galaxy
that appears to have the same colour as  Eridani, but which is 1000
times fainter. It appears, however, that this galaxy is only composed of
stars similar to the Sun in physical characteristics. Find the number of
stars in the galaxy. You know the following about  Eridani: temperature
T “ 4900 K, magnitude m “ 3.74, parallax πp “ 0.3112 and mass M “
0.82M@ .
10.* (IOAA 1016, Th., q.6) The star β-Doradus is a Cepheid variable star
with a pulsation period of 9.84 days. We make a simplifying assumption
that the star is brightest when it is most contracted (radius being R1 )
and it is faintest when it is most expanded (radius being R2 ). For sim-
plicity, assume that the star maintains its spherical shape and behaves
as a perfect black body at every instant during the entire cycle. The
bolometric magnitude of the star varies from 3.46 to 4.08. From Doppler
measurements, we know that during pulsation the stellar surface expands
or contracts at an average radial speed of 12.8km/s. Over the period of
pulsation, the peak of (intrinsic) thermal radiation of the star varies from
531.0 nm to 649.1 nm. Find:
• the ratio of radii of the star in its most contracted and most expan-
ded states;
• the radii of the star (in metres) in its most contracted and most
expanded states;
• the flux of the star when it is in its most expanded state.
112 Chapter 9. Cosmic Distance Ladder

9.7 Problems

1.* The Astronomical Unit

Determining the astronomical unit (au) was one of the biggest problems
in celestial mechanics. Although the ratios between the distances of the
planets from the Sun had already been obtained, a solution had not yet
been found to measure one of the distances involved. By observing the
apparent motion of Venus with a telescope, it was possible to measure a
maximum elongation of  “ 46.3˝ . Limiting ourselves, for simplicity, to
the plane of the orbits, consider two stations P1 and P2 on the equator,
separated by a longitude difference of 120 ˝ . From these stations, we
observe Venus (at maximum elongation) and a star at a much greater
distance. From station P1 , the angle between the star and Venus is
φ1 “ 0 ˝ 371 152 ˘ 32 ; from station P2 it is φ2 “ 0 ˝ 361 522 ˘ 32 . Determine
the value of the astronomical unit, estimating the relative error, given
that the Earth’s radius is RC “ 6380 km. (Hint: read Ch. 11 first).
2.* Luminosity-radius relationship
By differentiating the Stefan-Boltzmann law, derive a relationship between
the infinitesimal variations of brightness, radius and temperature.
Assuming that the star compresses adiabatically (T V γ´1 is constant),
find a relationship between the infinitesimal variations of brightness and
radius only. Hence, show that its brightness increases when the star
compresses (Hint: assume γ “ 5{3).
3.** Period-density relationship
The radial oscillations within a star are the result of sound waves that
propagate within it. An estimate of the period of oscillation can be
obtained by considering the time that a wave takes to cross the star.
Knowing that the speed of the waves is:
d
γP
v“
ρ

where γ is the adiabatic expansion coefficient, P the pressure and ρ the

density; using for P the expression of hydrostatic equilibrium (Eq. 7.29),
compute the period of oscillation as a function of density.
 Part III

Celestial Mechanics

113
 10

Gravitation and Kepler’s Laws

The primary task of celestial mechanics is to explain and predict the motions of
the planets and their satellites. In the past, several empirical models, such as
the epicycles and Kepler’s laws, were employed to describe these motions, but
none were based on a firm theoretical ground. In 1687, Newton came up with
a simple explanation: the law of universal gravitation. We begin the chapter
with a summary of Newtonian mechanics and the universal law of gravitation,
from which Kepler’s laws can be derived. Although the exact proof of Kepler’s
laws is given in Appendix B, in this chapter we show, when possible, some
simple derivations.

10.1 Newtonian Mechanics

Newtonian mechanics accurately describes nature for most of the phenomena

we are familiar with. It is based on the following four principles, of which the
first is attributed to Galileo (which we will call law zero), and the other three
to Newton:
0. The laws of physics are the same in all inertial reference systems, i.e.
systems moving at a constant velocity with respect to one another.
1. In an inertial reference system, every object in a state of uniform motion
remains in that state of motion unless acted upon by an external force.
2. The resultant force acting on a body is equal to its mass times its accel-
eration: F~ “ m~a.
3. If A exerts a force F~ on B, then B exerts a force ´F~ on A (same mag-
nitude, opposite direction).

10.2 Universal Law of Gravitation

In 1687, Newton published the “Philosophiae Naturalis Principia Mathematica”,

in which, for the first time, he described the law of universal gravitation. Ac-
cording to this law, each point mass m1 attracts every other point mass m2
114
10.2. Universal Law of Gravitation 115

Figure 10.1: Left: gravitational field created by a spherically symmetric body. Right: the
same field is produced by a point mass, of the same mass as the body, placed at its centre.

with a force F which is directly proportional to the product of the masses and
inversely proportional to the square of their distance r:

Gm1 m2
F “ , (10.1)
r2

where F is directed along the radius vector connecting the two bodies and is
always attractive, and G is the universal gravitational constant, first calculated
by Cavendish in 1798, with a value of:

G “ 6.674 ¨ 10´11 Nm2 kg´2 .

Two important points, which you can prove in Pr. 10.1, are:
1. the gravitational field created outside a spherically symmetric body is
the same as that created by a point mass, of the same mass, placed at
its centre (Fig. 10.1);
2. if the system has particular symmetries, we can consider a surface (usu-
ally a sphere or a cylinder) such that only the mass inside it contributes
to the gravitational field, while the mass outside gives a net zero contri-
bution (Fig. 10.2, overleaf).
From the first point, it follows that the gravitational force created by a spher-
ically symmetric planet is equal to that created by a point mass, of the same
mass, placed at its centre, which can be calculated using Eq. 10.1. Therefore,
the gravitational force acting on a body of mass m0 on the surface of a planet
of mass M and radius R can be obtained by substituting m1 “ M , m2 “ m0
and r “ R in Eq. 10.1. According to Newton’s second law, the gravitational
force is also equal to F “ m0 g, where g is the acceleration of the body. Hence,
116 Chapter 10. Gravitation and Kepler’s Laws

Figure 10.2: At a distance r from the centre of a spherical symmetric body, only the mass
mr inside a sphere of radius r contributes to the gravitational field. Outside, part of the
mass attracts the body to the left and part of it to the right. As it turns out, these two
contributions cancel each other.

the gravitational acceleration on the surface of the planet is:

GM
g“ . (10.2)
R2
Note that the acceleration does not depend on the mass of the body m0 .
Therefore, it is possible to define a vector field, called the gravitational field,
equal at each point in space to the acceleration that a hypothetical body would
have if it were placed at that point. On Earth, MC “ 5.972 ¨ 1024 kg and
RC “ 6378 km, hence:
g “ 9.81 m/s2 ,
which is the mean value of the gravitational acceleration on Earth.
The second point can greatly simplify a problem if particular symmetries are
present. Let us find, as an example, the gravitational field at a distance r from
the centre of a spherically symmetric planet, of mass M and radius R. For r ě
R, the answer is given by Eq. 10.2, with R replaced by r. However, inside the
body, at distance r from the centre, the only contribution to the gravitational
field is given by the mass mr inside a sphere of radius r (Fig. 10.2). Assuming
constant density, mass is proportional to volume, which in turn is proportional
to the radius cubed, i.e. mr “ pr{Rq3 M . Using Eq. 10.2 (with M replaced by
mr and R by r), we finally obtain:
GM
g“ r for r ă R .
R3
Let us now discuss an extension of the second point. It is possible to divide
~
any closed surface into many infinitesimal elements having surface area ∆A,
with magnitude equal to the area of that element, direction perpendicular to
 10.2. Universal Law of Gravitation 117

Figure 10.3: It is possible to divide a spherical surface into many infinitesimal elements with
~ so that ~g ¨ ∆A
surface area ∆A, ~ is constant for every element.

the surface (which, at that point, is effectively planar) and pointing outwards.
We then define the flux of a vector field ~g on the surface S, as the sum of the
~ for ∆A that tends to zero on the whole surface. In this
scalar product ~g ¨ ∆A,
limit, the sum becomes an integral:
ÿ ż
Φ “ lim ~ “ ~g ¨ dA
~g ¨ ∆A ~, (10.3)
~
∆AÑ0 S
ş
where S denotes the integral over the entire (not necessarily closed) surface S.
Now, from the second point, it follows that the flux of the gravitational field on
a closed surface S depends only on the mass inside the surface. The form of this
dependence can be found by considering the simple case of a point mass placed
at the centre of a spherical surface (Fig. 10.3). In this case, the flux is:
¿ ¿ ¿
~ GMint ~ GMint GMint `
4πR2 “ ´4πGM ,
˘
Φ “ ~g ¨ dA “ ´ 2
r̂ ¨ dA “ ´ 2
dA “ ´ 2
R R R

where we have taken the constant GMűint {R2 outside the integral, and equated
the integral over the closed surface ( dA) to the area of the sphere (4πR2 ).
The negative sign arises because r̂ and ~g have opposite directions, i.e. because
gravity is attractive. Therefore, in general:
¿
Φ“ ~ “ ´4πGMint .
~g ¨ dA (10.4)

To use the above equation, it is necessary to find a surface with the right
symmetry (otherwise the integral cannot be easily solved). If working with a
spherical distribution, take a sphere; for a cylindrical distribution (for instance,
a cylinder or a plane), take a cylinder (Ex. 10.18).
118 Chapter 10. Gravitation and Kepler’s Laws

10.3 Kepler’s Laws

Based on the observations made by the Danish astronomer Tycho Brahe, after
16 years of work, Kepler was able to calculate the parameters of the orbit of
Mars and, subsequently, those of the Earth. The transition from observations
to the derivation of an empirical law required a great deal of work, as the orbit
of the planet is observed from the Earth, which, in turn, orbits the Sun. In
what Kepler himself called a flash of genius, he realized that every 687 days
(the orbital period of Mars), Mars returned to exactly the same position in the
Solar System. Kepler’s analysis of the planetary motion can be summarized in
three laws, which bear his name:
1. All planets move in elliptical orbits, with the Sun at one focus.
2. The radius vector joining any planet to the Sun sweeps out equal areas
in equal times.
3. The square of the orbital period of any planet is directly proportional to
the cube of the semi-major axis of its orbit.

First Law
Kepler’s first law states that each planet moves in an elliptical orbit around
the Sun. The Sun is in one focus of the ellipse, with nothing in the other
focus. In Appendix B we prove this statement, starting from Newton’s law
of gravitation. Here, we briefly discuss the properties of the ellipse which
are useful in astronomy, while you should refer to Appendix A.2 for a more
mathematical treatment.
In astronomy, the points of the ellipse closest and farthest from the focus in
which the Sun is located are called perihelion and aphelion, respectively. For
bodies orbiting the Earth, these points are instead called perigee and apogee.
By convention, a is the semi-major axis, b the semi-minor axis, c the semi-focal
distance and e the eccentricity, defined as e “ c{a (Fig. 10.4). If the Sun is
at F1 , the distances of the planet to the Sun in perihelion and aphelion are,
respectively:
dp “ AF1 “ AO ´ F1 O “ a ´ c ,
da “ F1 B “ OB ` F1 O “ a ` c .

In astronomy we rarely know b, while it is more common to be given some

combination of a, e, da and dp . If we know a and e, then we can find dp and
da using the formulae:
dp “ a ´ c “ ap1 ´ eq ,
da “ a ` c “ ap1 ` eq .
10.3. Kepler’s Laws 119

Figure 10.4: a is the semi-major axis, b the semi-minor axis and c the semi-focal distance.
The Sun can be in either F1 or F2 .

If, instead, we know da and dp , we can then find a and e by taking the sum
and difference of the above equations:
da ` dp
a“ , (10.5)
2

da ´ dp
c“ . (10.6)
2
Hence, the eccentricity e can be written as:
c da ´ dp
e“ “ . (10.7)
a da ` dp
In the case e Ñ 1, the orbit becomes a line. Therefore, Kepler’s laws can also
be used to study the motion of a body that falls vertically in the gravitational
potential (Ex. 10.10).

The Second Law

With regard to Kepler’s second law, it is important to define the angular
momentum of a body, i.e. the vector product of its position vector and the
linear momentum:
~ “ m~r ˆ ~v .
L (10.8)
Let θ be the angle between the position and velocity vectors. Then, the mag-
nitude of the angular momentum is L “ m r v sin θ (see Appendix A.1). In
a circular orbit, the position vector is always perpendicular to the velocity,
therefore θ “ π{2, hence L “ m r v. In an elliptical orbit, the position vector
is perpendicular to the velocity only at aphelion and perihelion.
We define the torque, with respect to a fixed origin, as the vector product of
the position vector (lever arm) and the force:

~τ “ ~r ˆ F~ . (10.9)
120 Chapter 10. Gravitation and Kepler’s Laws

Figure 10.5: It can be proven that the rate at which the radius vector ~r sweeps out area is
equal to L{2m, where L is the angular momentum. Since L is constant, it follows that the
radius vector sweeps out equal areas in equal times, which is Kepler’s second law.

As before, the magnitude of torque is τ “ F r sin θ. Note that, if a body is

only subject to a central force (i.e. directed radially inwards, such as gravity),
θ “ 0, hence τ “ 0.
The principal equation of rotational mechanics relates L ~ and ~τ :

~
dL
“ ~τ , (10.10)
dt
which follows from the definition of the force as the rate of change of mo-
mentum. But why consider angular momentum? In an isolated system in
which the only force is gravity, the torque acting on a body is zero, as pre-
viously shown. Then, from Eq. 10.10, it follows that the rate of change of
angular momentum is zero: the angular momentum is constant. In physics,
we are especially interested in all the quantities that are constant in time, and
angular momentum, like energy, is one of those.
Could it be that the constant rate at which the radius vector sweeps out area
(Kepler’s second law) is a consequence of the angular momentum being con-
stant? After an infinitesimal time dt, the planet covers a distance of ds “ v dt
along its orbit (Fig. 10.5). Since ds is small compared to r, it follows that
=ABS « θ, where θ is the angle between the radius vector and the velocity.
The area swept by the radius vector joining the planet to the Sun is equal to
the area of triangle ABS, which in turn is equal to half its base ds times its
height r sinp=ABSq:
1 1 L
dA “ ds r sinp=ABSq “ r v sin θ dt “ dt .
2 2 2m
Dividing both sides by dt, we find:
dA L
“ .
dt 2m
10.3. Kepler’s Laws 121

Since L is constant, it follows that dA{dt is constant, thus proving Kepler’s

second law.
Kepler’s second law can then be used to compute the time t that it takes a
planet to cover an area of As . Let T be the period of revolution and A “ πa b
the total area of the orbit. We then have the relation:
t As As
“ ñ t“T¨ . (10.11)
T A A
Because of the conservation of angular momentum, we expect the velocity of a
planet to be maximum at perihelion and minimum at aphelion. Indeed, since
the velocity is perpendicular to the radius vector at those two points:

da va “ dp vp , (10.12)

where da , dp and va , vp are the distances and velocities at aphelion and peri-
helion, respectively. In Ex. 10.9 you will use Eq. 10.12, together with the
conservation of energy, to compute va and vp , given only a and e.

The Third Law

Kepler’s third law states that, for every body in the Solar System, the ratio of
the square of the orbital period (T ) and the cube of the semi-major axis (a) is
constant:
T2
“ k. (10.13)
a3
Writing the previous equation for a generic planet and for the Earth:
2
T2 TC
“ ,
a3 a3C

or, equivalently:
´ T ¯2 ´ a ¯3
“ .
TC aC
But T {TC is just the period T measured in years, whilst a{aC is the semi-major
axis of the planet measured in astronomical units. Hence:

Tyr 2 “ aau 3 . (10.14)

In the following section, we derive Kepler’s third law from Newton’s law of
gravitation. We will only consider the case of circular orbits, while the general
case is treated in Appendix B. We will also assume that the mass of the planet
is much smaller than the mass of the Sun. This is a good approximation for
the Solar System, but it is not strictly valid for binary systems, in which the
orbiting bodies can have similar masses.
122 Chapter 10. Gravitation and Kepler’s Laws

Proof of Kepler’s Third Law

The derivation of Kepler’s third law in the case of circular orbits is a good
opportunity to look at an application of Newton’s law of gravitation, and to
define some recurring quantities.
It is useful to define the instantaneous angular velocity, equal to the angle
travelled per unit of time:
∆θ dθ
ω “ lim “ . (10.15)
∆tÑ0 ∆t dt
In uniform circular motion, ω is constant and equal to ω “ 2π{T , since the
body covers an angle of 360˝ or 2π rad every period T .
It is possible to write ω as a function of the instantaneous tangential velocity
and the radius vector. After the body covers an infinitesimal angle dθ, the
distance dst covered in the tangential direction is:

dst “ R dθ .

Dividing both sides by dt, and recognizing vt “ dst {dt, ω “ dθ{dt:

vt “ ωR . (10.16)

In general, vt “ v sin θ, where θ is the angle between the velocity and the radius
vector. In the special case of circular motion, v “ vt , since the path (and
therefore the velocity) is always tangent to the radius vector. The velocity
is then equal to the ratio of the circumference and the period of rotation,
i.e. v “ 2πR{T , which gives ω “ 2π{T , as before. Let us now find an
expression for the force necessary to keep a body moving in a curved orbit.
After a time dt, the body covers an angle of
dθ, and its velocity changes by dv “ vt dθ
(see Fig. 10.6). Therefore, its acceleration
is:
dv vt dθ dθ
ac “ “ “ vt “ vt ω .
dt dt dt
Using Eq. 10.16, the previous formula can
be written as:

vt2
ac “ “ ω2R . (10.17)
Figure 10.6: After the body covers an
R
angle of dθ, its velocity changes by dv “
vt dθ. Eq. 10.17 gives the required acceleration as
a function of the instantaneous tangential
10.3. Kepler’s Laws 123

velocity (or angular velocity) and the radius of curvature. This is known as
the centripetal acceleration, since it is directed radially towards the centre
of rotation. The centripetal force is then given by Newton’s second law, i.e.
Fc “ mac . If a planet is found to rotate in a circular orbit, the necessary
centripetal force is provided by the force of gravity:
GmM
“ mac .
R2
Using Eq. 10.17 to substitute ac in terms of the velocity:
c
GM
v“ . (10.18)
R

Since the circumference is 2πR, the time (period of revolution) taken to cover
it is T “ 2πR{v. Substituting for v, given by Eq. 10.18:

4π 2 3
T2 “ R . (10.19)
GM
This is the canonical form of Kepler’s third law. Eq. 10.19 implies:
T2 4π 2
“ “ k,
R3 GM
which is what we wanted to show. Eq. 10.19 is also valid for elliptical orbits,
but R must be replaced with the semi-major axis a.
Let us now derive Kepler’s third law without neglecting the gravitational
field created by the planet itself. Let m1 ,m2 be the masses of the two bod-
ies and r1 , r2 the radii of their circular orbits. The bodies rotate around
their common centre of mass with the same angular velocity ω, being dia-
metrically opposite and at a constant distance r1 ` r2 from each other at
all times. In the Solar System, even if all planets aligned, the centre of
mass would still be below the surface of the Sun. For the Earth-Moon
system however, this phenomenon is of such importance that it gives rise
to two tides every day. In accordance with Newton’s law of gravitation,
the force that the bodies exert on each other is:
Gm1 m2
Fg,1 “ Fg,2 “ .
pr1 ` r2 q2
For the bodies to remain in a circular orbit, the centripetal force must have
the constant magnitude of:

Fc,1 “ m1 ω 2 ¨ r1 ,
Fc,2 “ m2 ω 2 ¨ r2 .
124 Chapter 10. Gravitation and Kepler’s Laws

But the centripetal force is the gravitational force, hence:

Gm2
“ ω 2 ¨ r1 ,
pr1 ` r2 q2
Gm1 2
2 “ ω ¨ r2 .
pr1 ` r2 q
Adding the above equations:
Gpm1 ` m2 q 2
2 “ ω ¨ pr1 ` r2 q
pr1 ` r2 q
ñ Gpm1 ` m2 q “ ω 2 ¨ pr1 ` r2 q3 .
Let r1 ` r2 “ a be the distance between the bodies. Since ω “ 2π{T , we find:

4π 2
T2 “ a3 . (10.20)
Gpm1 ` m2 q
In the case of the Solar System, the mass of the Sun is much greater than the
mass of any planet, therefore the sum of the two masses at the denominator
reduces to the mass of the Sun only, and we recover Kepler’s third law.
Eq. 10.20 is also valid for elliptical orbits, but in this case a must be interpreted
as the sum of the semi-major axes a1 and a2 . We require the centre of mass to
be fixed, hence a1 m1 “ a2 m2 , and substituting either a1 or a2 in a “ a1 ` a2 ,
we get a “ p1 ` m1 {m2 qa1 “ p1 ` m2 {m1 qa2 .
Kepler’s third law brings with it a great application: by measuring the orbital
period of the planets, it is possible to obtain the ratios of their semi-major
axes to that of the Earth. However, unless we measure one of the distances
involved, it is as if we had a map with no scale. Try solving Pr. 9.1, for an
example of how one of these distances can be measured.

10.4 Energy of an Orbiting Body

The total energy of an orbiting body is equal to the sum of its potential and
kinetic energy. Imagine two bodies, at a distance r from each other. The
gravitational potential energy of this configuration is equal to the work required
to bring one of them to a distance r from the other (which remains fixed),
starting from infinity. During this process, we must apply a force equal and
opposite to that of gravity, so that the speed remains constant, and the change
in kinetic energy is zero (we want to isolate the contribution of the potential
energy). Therefore, the potential energy is the opposite of the work done by
the force of gravity to bring one of the two bodies at a distance r from the
other, starting from infinity:
∆U “ ´∆W . (10.21)
10.4. Energy of an Orbiting Body 125

Eq. 10.21 is valid for any conservative force (i.e. a force for which the work
done is independent of the path taken, such as gravity). Work is defined as the
scalar product of force and displacement, hence, considering an infinitesimal
displacement dx towards the fixed body, the work done by the force of gravity
is dWg “ Fg dx. The total work ∆Wg is given by the sum (integral) of all
these infinitesimal contributions dWg . Since gravity is a conservative force, to
evaluate ∆Wg , we can chose a path that is a straight line:
ż żr ż8 ż8
Gm1 m2 dx
∆Wg “ dWg “ ´ Fg dx “ 2
dx “ Gm1 m2 2
8 r x r x
” 1 ı8 ”1 ı Gm m
1 2
“ Gm1 m2 ´ “ Gm1 m2 ´ 0 “ .
x r r r
As expected, ∆Wg is positive, since the force of gravity and the displacement
both point towards the body. Therefore, the potential energy is:

Gm1 m2
U “´ . (10.22)
r

If the body only moves in a small region of space, we can regard the gravita-
tional field ~g as a constant. Therefore:

∆U “ ´∆W “ ´F~ ¨ ~h “ mgh

ñ U phq “ mgh ` const. ,

where h is the projection of the distance covered by the body in the direction
of the gravitational field (positive if moving away from it), and U phq is defined
within an arbitrary constant, which depends on the choice of zero-height. In
Eq. 10.22, the arbitrary constant is zero, since, by convention, an object at
infinity has zero potential.
The kinetic energy is defined as the work done by a force to increase the velocity
of a body from zero to v, while at constant potential:

∆K “ ∆W , (10.23)

where:
żv żv
dv 1 1
ż ż ”1 ıv 1
∆W “ F dx “ ma dx “ m v dt “ mv 1 dv 1 “ mv 12 “ mv 2 .
0 dt 0 2 0 2

In the above derivation, we used v 1 as a variable of integration, to distinguish

it from the upper limit v. Hence:

1
K “ mv 2 .
2
126 Chapter 10. Gravitation and Kepler’s Laws

Figure 10.7: The dashed lines represent the gravitational and centrifugal potentials; the solid
line their sum, known as the effective potential. The horizontal lines E1 , E2 , E3 , E4 are
the energies of an hyperbolic, parabolic, elliptic and circular orbit, respectively, and their
intersections with the effective potential give the turning points of the orbits.

Therefore, the total energy of a body of mass m, orbiting at a distance r from

a body of mass M (with m ! M ), is:
GmM 1
E “U `K “´ ` mv 2 . (10.24)
r 2
The total energy is conserved if the system is isolated, i.e. if no dissipative
forces are present. Indeed, a conservative force changes U and K by ∆U “
´∆W (Eq. 10.21) and ∆K “ ∆W (Eq. 10.23). Hence ∆U ` ∆K “ 0, i.e.
the energy is conserved. Decomposing the velocity into radial and tangential
components (v 2 “ vr 2 ` vt 2 ), and substituting vt as a function of the angular
momentum (vt “ L{mr), we finally obtain:
´ L2 GmM ¯ 1
E“ ´ ` mvr 2 . (10.25)
2mr2 r 2
In the above equation, the terms in brackets depend only on the variable r,
therefore they can be regarded as an effective potential:

L2 GmM
Ueff “ 2
´ .
2mr r
The effective potential is plotted in Fig. 10.7. The dashed lines represent
the gravitational (´GmM {r) and centrifugal (L2 {2mr2 ) potential terms; the
solid line their sum, the effective potential. The horizontal lines show the
different energies an orbiting body may have. Since the kinetic energy must
always be positive, the body can never be found in regions where the effective
10.4. Energy of an Orbiting Body 127

potential is greater than the total energy. The intersection points between
the horizontal lines and the effective potential are therefore known as turning
points, since the body must invert its (radial) motion at those points. Looking
at Fig. 10.7, when E ă 0 (E3 ), the horizontal line always intersects the
effective potential at two points, which (from left to right) are the minimum
and maximum distances. When E has the smallest possible value (E4 ), the
two intersection points coincide: this is the case of a circular orbit. When
E “ 0 (E2 ) or E ą 0 (E1 ), there is only one turning point and the orbit is
unbound: it reaches a minimum distance from the centre, but moves infinitely
far away afterwards. Hence, we essentially have four types of orbits:
$
’
’
’ E ą 0, hyperbola ;
’
&E “ 0 , parabola ;
’
’
’ E ă 0, ellipse ;
’
%E min , circle .

Closed Orbits
As the name suggests, a body in a closed orbit periodically rotates around the
centre of attraction. Closed orbits can be either circular or elliptical.
In the case of circular orbits, the potential energy is constant and equal to:
GM m
Uc “ ´ .
R
According to Eq. 10.18, the velocity is:
c
GM
vc “ .
R
Hence the kinetic energy is:
1 1 GM GmM
Kc “ mvc2 “ m “ .
2 2 R 2R
The total energy is the sum of the potential and kinetic energies:
GmM GmM
Ec “ Uc ` Kc “ ´ `
R 2R
GmM
ñ Ec “ ´ .
2R
The equation above is also valid for an elliptical orbit, if R is replaced by the
semi-major axis a:
GmM
E“´ . (10.26)
2a
128 Chapter 10. Gravitation and Kepler’s Laws

Hence, it follows that the total energy of circular and elliptical orbits is always
negative. As a tends to infinity, the energy goes to zero; for small a, it grows
increasingly negative. Eq. 10.26, used in conjunction with the conservation of
angular momentum, should allow you to solve (almost all) problems in Celestial
Mechanics.

Open Orbits
In the Solar System, the planets are gravitationally bound to the Sun, travelling
in elliptical, almost circular orbits. If a body does not rotate periodically
around the centre of attraction, we say that its orbit is unbound, and it can
be either parabolic or hyperbolic. In this case, the body first approaches the
centre, reaches a minimum distance, and then moves infinitely far away. Since
the body can reach arbitrarily large distances from the centre of attraction, for
some point infinitely far away its potential energy must be zero. If the body
is at rest when it is at an infinite distance, then its total energy is E “ 0,
and it can be shown that it travels in a parabolic orbit. Of course, the body
is never exactly at infinity, and the velocity is never exactly zero (therefore
it is actually able to cover the orbit). The situation is reversible, so that the
minimum amount of energy that a body needs in order to escape from a system
is E “ 0. If, on the other hand, the velocity of the body at an infinite distance
is non-zero, then the total energy is E ą 0, and it can be shown that the body
follows a hyperbolic orbit (Appendix B). An object in a parabolic or hyperbolic
orbit must come from outside the Solar System.

10.5 Escape Velocity

The escape velocity is the minimum velocity required to move indefinitely far
away from the centre of attraction. If the initial velocity of a body is exactly
equal to the escape velocity, as it moves away from the centre, the velocity
decreases until, asymptotically, it tends to zero. Therefore, the final kinetic
and potential energies are zero, hence the total energy is zero as well. Because
energy is conserved:
GmM 1 2
´ ` mvesc “ 0,
R 2
where m ! M and R is the initial distance between the body and the centre
of attraction. Solving for vesc , we find:
c
2GM
vesc “ . (10.27)
R
If a body has a velocity lower than the escape velocity, it follows a circular or
an elliptical orbit, being gravitationally bound to the centre of attraction. If
10.6. Virial Theorem 129

the velocity is exactly equal to the escape velocity, the orbit is parabolic, if
it is greater, the orbit is hyperbolic. In general, the escape velocity is greater
for planets with a larger mass. This aspect is important in the discussion of
planetary atmospheres, covered in Sec. 7.3.

Schwarzschild Radius
The Schwarzschild radius is the size of the event horizon of a black hole, i.e.
the distance from the singularity within which the escape velocity from the
black hole is greater than the speed of light. At a distance from the singu-
larity equal to the Schwarzschild radius (Rs ), the escape velocity is exactly
equal to c: c
2GM
vesc “ c “
Rs

2GM
ñ Rs “ . (10.28)
c2
Although this equation was obtained in the framework of classical physics, it
turns out that this is the same result predicted by the theory of general relativ-
ity. The above derivation is given to facilitate the memorization of Eq. 10.28.

10.6 Virial Theorem

For an inverse square force law, such as gravity, the Virial theorem relates the
average potential and kinetic energies:

U “ ´2K . (10.29)
The total energy, as usual, is the sum of the potential and kinetic energies:

U U
E “U `K “U ´ “ . (10.30)
2 2
For example, for a body in circular orbit, the potential and kinetic energies
are constant. Since U “ Uc “ ´GmM {a, it follows that the energy is also
constant and equal to:
1 GmM
Ec “ Uc ñ Ec “ ´ ,
2 2a
which is the result given in Eq. 10.26. In Appendix C, we prove the Virial
theorem for a general potential of the from U “ ´αr´n , in which case:
n
K“´ U . (10.31)
2
For the gravitational potential, n “ 1, hence we recover Eq. 10.29.
130 Chapter 10. Gravitation and Kepler’s Laws

10.7 Exercises

1. Compute the ratio between the gravitational acceleration on the surface

of the Sun and the Earth.
2. (INT 2015, Th.S, q.3) The commander of the interstellar mission Kout-Iz-
Paut wants to attempt a descent on the planet Sweets-3 and it is essential
for him to know the gravitational acceleration gSW3 on the surface of the
planet. The only data he is given is the density (ρSW3 “ 4ρE ) of the
planet Sweets-3, and its radius (RSW3 “ RE {4). Compare the value of
gSW3 with the gravitational acceleration on Jupiter gE .
3.* (IOAA 2012, Th., q.5) Compute the ratio between the mean densities of
the Earth and Sun, using only the following data:
• mean angular diameter of the Sun (θ@ “ 321 ), as seen from Earth;
• gravitational acceleration on the surface of the Earth (g “ 9.80 m/s2 );
• length of the sidereal year (Tsid “ 365.2564d );
• the fact that one degree in longitude corresponds to s “ 111 km on
the surface of the Earth.
4. The Hubble Space Telescope rotates around the Earth, in a circular orbit,
at an average distance of H “ 560 km from the surface. Calculate the
average velocity vc and the orbital period T of HST.
5. Determine the distance d of a geostationary satellite from the surface of
the Earth, and its velocity v.
6. (NAZ 2015, Th.S, q.2) Consider a satellite that orbits the Earth along
the equatorial plane, in the same direction as Earth’s rotation. At what
minimum height (hmin ) from the Earth’s surface should the satellite be,
in order for an observer to see it move in the sky from east to west?
7. CoRoT-7b is the closest exoplanet to its star that has ever been dis-
covered. The semi-major axis of its orbit is just 0.0172 au, while its
period of revolution is only 0.8536 days. Estimate the mass of the star
CoRoT-7, assuming that the mass of the planet is negligible compared
to the mass of the star.
8. In its orbit around the Sun, a comet reaches a maximum distance of
31.5 au, and a minimum of 0.5 au. What is its orbital period? At what
rate does the radius vector joining the comet to the Sun sweep out area,
in square kilometres per year?
9.* A planet is in an elliptical orbit around the Sun, with semi-major axis a
and eccentricity e. Compute its velocities at perihelion and aphelion.
10.7. Exercises 131

10.* Two asteroids of masses M1 and M2 (negligible compared to that of the

Earth) are on a collision course with Earth, at a distance of 1.2 ¨ 106 km
and 106 km, respectively. Assuming their initial velocities are zero, find:
• the time it takes the first asteroid to impact the Earth;
• the time between the two impacts;
• if the impacts happen on the equator, what is the distance (meas-
ured along the Earth’s surface) between the two craters?
11.* (IOAA 2011, Th., q.8) A spaceship orbits very close to the asteroid
Seneca, transmitting data on Earth. Due to the relative motion between
the asteroid and the Earth, the time taken by a signal transmitted by the
spaceship to arrive on Earth varies between 2 and 39 minutes. Assuming
that the Earth moves around the Sun in a circular orbit that does not
intersect the orbit of Seneca, calculate:
• the semi-major axis as and the eccentricity es of Seneca’s orbit;
• the period Ts of Seneca’s orbit.
12.* (IAO 2015, Th.α, q.5) A satellite moves around the Earth in a slightly
elliptical, equatorial, orbit and passes through the perigee point at a
height of dp “ 428.0 km above sea level, with a speed 0.6 % greater than
that of a circular orbit at the same height. Estimate the time taken for
the satellite to reach a height of d “ 498.0 km.
13.* (BAAO, 2019, Th., q.1, adapted) Calculate the time the Parker Solar
Probe (PSP) spends at a distance from the Sun less than the length of
the latus rectum of its orbit (i.e. the point where the angle between the
radius vector and the semi-major axis, known as the true anomaly, is
90˝ ). First, solve this exercise using Kepler’s second law and Eq. A.7
(Appendix A.2); then using the equations given in Pr. 10.12.
14.** (IOAA 2014, Th., q.2) In a gravitational catastrophe, the mass of the
Sun suddenly halves. Find the new orbital period if, at that time, the
Earth was at aphelion or perihelion.
15. What should the radius of a planet be, if it has the same density as the
Earth and if a human can escape from it with a simple jump?
16. Estimate the number of stars in a globular cluster of diameter 40 pc,
if the escape velocity from its outer edge is 6 km/s. Assume that, on
average, the cluster is composed of Sun-like stars.
17. If the Sun collapsed to form a non-rotating black hole, what would the
radius of its event horizon be? Calculate the average density.
132 Chapter 10. Gravitation and Kepler’s Laws

18.* A planet is shaped as a long cylinder. Its mean density, radius and
rotational period are the same as on Earth.
• Compute the first cosmic velocity vc,1 (i.e. the orbital velocity on
the surface of the planet).
• What is the height of a geostationary satellite, measured from the
surface?
• What can you say regarding the second cosmic velocity (i.e. the
escape velocity)?
19.* A homogeneous sphere of radius a contains a cavity of radius a{4, whose
centre is 3a{8 from the surface. The diameter passing through the centre
of the sphere and the cavity meets the surface at points A and B. Cal-
culate the ratio of the gravitational acceleration at these two points.
20.** A projectile is launched from the surface of the Earth with the first
cosmic velocity, at an angle θ “ 30˝ with respect to the local horizon.
Neglecting friction with air, find:
• the semi-major axis a of the orbit;
• the maximum height of the projectile with respect to the Earth’s
surface, in units of RC ;
• the range of the projectile, i.e. the distance, along the Earth’s
surface, between the launch and landing points;
• the time of flight.
21.** A probe is launched from the north pole of the Earth with the first
cosmic velocity, in such a way that it lands at the equator. Neglecting
friction with air, find:
• the semi-major axis a of the orbit;
• the maximum height of the probe, with respect to the Earth’s sur-
face, in units of RC ;
• the time of flight.
22.** In the absence of atmosphere on Earth, what are the maximum and
minimum velocities of Sun-orbiting asteroids that impact our planet?
23.*** Find the maximum time a comet, following a parabolic trajectory
around the Sun, can spend within the orbit of the Earth. Assume that
the Earth’s orbit is circular and in the same plane as that of the comet.
24.*** What should the shape of a planet be (of fixed volume and density),
such that it produces the maximum possible gravitational acceleration
10.8. Problems 133

at a certain point in space? How much greater is the gravitational ac-

celeration in this case, compared to that of a spherical planet with the
same mass and density?
25.*** Two comets with identical masses and speeds are found to be ap-
proaching the Sun along parabolic trajectories that lie in the same plane.
The comets collide at their common perihelion P, and break into many
pieces that then travel in all directions, but with identical initial speeds.
What shape is the envelope of the subsequent trajectories of the pieces
of debris?

10.8 Problems

1.** Planets as point masses

For a spherical shell, show that:
• at its exterior, the gravitational field is the same as that produced
by a point mass, of the same mass, placed at its centre;
• at its interior, the gravitational field is zero.
Hence, prove the two points stated in Sec. 10.2.
2.* Using Gauss’ theorem
Compute the gravitational acceleration at a distance r from the centre
of the Earth, assuming its density is constant.
3.* Deep tunnels
On a future lunar base, one wonders what the fastest way is to transport
a load from a place on the surface of the Moon to its antipode. Two
solutions have been proposed:
• introducing the load into a circular orbit;
• dropping the load into a straight tunnel connecting the two places.
Which method enables a faster transport? Assume the density of the
Moon is constant.
4.** Potential energy of a spherical body
Compute the gravitational potential energy of a star with mass M and ra-
dius R. You can imagine assembling the star in spherical shells, brought
from infinity.
5.** Escape velocity
Obtain an expression for the escape velocity in the case that the masses
of the two bodies are comparable.
134 Chapter 10. Gravitation and Kepler’s Laws

6.** Roche limit

Obtain an expression for the Roche limit, defined as the minimum dis-
tance from the centre of a celestial body such that a secondary body
(held together only by the force of gravity) does not disintegrate due to
tidal forces. Your answer must be a function of the density of the two
bodies and the radius of the central body. Give a numerical value for the
Roche limit in the case of the Sun-Earth and Earth-Moon systems.
7.** Jeans limit
Obtain an equation for the Jeans limit, i.e. the critical mass above which
a nebula collapses to form a star. Write the result as a function of the
temperature T , the molar mass µ of the gas that constitutes the nebula
and its density ρ.
8.** Hill sphere and Lagrangian points L1 , L2
Calculate the distance to the Sun of the Lagrangian points L1 and L2 , i.e.
those points on the line joining the Earth and the Sun, placed in between
the Sun and the Earth and beyond the Earth, respectively, having the
same orbital period as the Earth. Hence, find the radius of the Hill
sphere, defined as the distance within which a secondary body can retain
satellites while orbiting around a central body. Give a numerical value
for the radius of Earth’s Hill sphere.
9.*** Position of Lagrangian point L3
Find the position of the Lagrangian point L3 , i.e. that point on the line
joining the Earth and the Sun, on the opposite side of the Earth relative
to the Sun, having the same orbital period as the Earth.
10.*** The three-body problem and Lagrangian points L4 , L5
A system consists of three bodies with masses m1 , m2 and m3 placed at
distances r1,2 , r1,3 and r2,3 . Find a relationship between their distances so
that the triangle formed by the three masses always maintains a constant
shape. Hence, calculate the orbital period of the system. Explain how
this problem is related to the calculation of the Lagrangian points L4
and L5 , positioned (on Earth’s orbit) 60 ˝ before and after the Earth,
and having the same orbital period.
11.** The fate of abandoned satellites
At the end of their useful lives, satellites lose energy in the upper layers
of the atmosphere, before finally burning up when they reach the denser,
lower, layers. It can be shown that satellites originally moving along cir-
cular trajectories will continue to travel in approximately circular orbits,
with their orbital radii slowly decreasing. Consider a satellite with mass
m “ 500 kg, in an approximately circular orbit, when it is abandoned.
10.8. Problems 135

The friction force on the satellite is approximately Fa “ c ρ v 2 , where

c “ 0.23 m2 , ρ is the density of air at the height of the satellite and v is
its instantaneous velocity.
• Does the satellite decelerate or accelerate as a result of the air drag?
Explain your answer from the point of view of dynamics.
• Find a (simple) equation that relates the drag force to the tangential
acceleration of the satellite.
• What is the density of air at an altitude of 200 km, if the orbital
radius of the satellite decreases by 100 m during a single revolution?
12.*** Kepler’s equation

Figure 10.8: True (θ), eccentric (E) and mean (M ) anomalies.

Let O be the centre of an elliptical orbit, F the focus around which the
body orbits and P the position of the body (see Fig. 10.8).
The true anomaly (θ) is defined as the angle between the direction of
perihelion and the position of the body, as seen from F . Prove that the
distance r of P from F is related to the true anomaly by:

ap1 ´ e2 q
rpθq “ . (10.32)
1 ` e cos θ

Let E be the intersection between the perpendicular to the major axis

passing through P and the circle. The eccentric anomaly (E) is defined
as the angle between the line OE and the direction of perihelion. Prove
that the distance r is related to the eccentric anomaly by:

r “ ap1 ´ e cos Eq . (10.33)

The mean anomaly (M ) is defined as the angle corresponding to the

fraction of the period of an elliptical orbit that has elapsed since the body
136 Chapter 10. Gravitation and Kepler’s Laws

a
passed through perihelion. Hence, M “ ωpt ´ t0 q, where ω “ GM {a3
is the mean angular velocity of revolution and t0 is the time at perihelion.
Prove that the mean anomaly is related to the eccentric anomaly by:

M “ E ´ e sin E , (10.34)

this is known as Kepler’s equation (Hint: use Eqs. 10.32, 10.33 and 5.9).
13.**** Relaxation time
Consider a system of stars, such as a galaxy or a cluster. In these sys-
tems, collisions between stars are very rare, since the average distance
between them is much greater than their size. A physical model can
approximate the system with a continuous distribution of matter, thus
neglecting its granularity. However, the real force acting on each star is
slightly different from the force obtained using the physical model. We
want to estimate the time scale for which the deviation from the con-
tinuous distribution model becomes significant. Systems for which this
deviation is small are called “collision-free”, and the time scale, called
relaxation time, is usually larger than the age of the universe.
You can use the following model: a star of mass m and velocity v ap-
proaches a stationary star (of the same mass m) with impact parameter
b. If the distance between the two stars is large, their interaction is small
and the star moves approximately on a straight line at a distance b, un-
dergoing a small deviation in the direction perpendicular to its velocity.
Find the variation of the square of the velocity in the perpendicular dir-
ection as a function of the impact parameter b. Then, integrating over
all the values of b, estimate the relaxation time, i.e. the time necessary
for the variation in the mean square speed to be equal to the square of
2
the velocity of the star: dvKtot « v2.
 11

Motion of the Planets

The Sun and the Moon always rise in the east and set in the west. However,
over the course of a year, they describe great circles on the celestial sphere in
the opposite direction to the apparent daily rotation of the sky, with a variable
but almost constant velocity. The trajectory described by the planets on the
celestial sphere is, instead, more complex: it contains knots and loops, that is,
points in which the motion reverses, from direct (same direction as the Sun,
from west to east) to retrograde, or vice-versa.

11.1 Apparent Motion of the Planets

According to their apparent motion, we distinguish inferior planets (such as

Mercury and Venus) from superior planets (all others, excluding Earth).
Inferior planets always appear to be moving close to the Sun, their angular
separation never exceeding the maximum elongation (max ). Considering the
triangle SEw T in Fig. 11.1, we see that:

sin max “ d{dC , (11.1)

where d and dC are the Sun-planet and Sun-Earth distances, respectively. Be-
cause of the eccentricity of the orbits, max varies according to the position of
the Earth and the planet, taking the values max “ 18˝ ´ 28˝ for Mercury and
max “ 45˝ ´ 48˝ for Venus.
At the point of maximum eastern elongation Ee , the planet is visible in the
evening, shortly after sunset. In fact, since the sky rotates from east to west,
the Sun, further west, sets earlier than the planet, which is therefore visible for
a brief period of time during twilight. Subsequently, the planet moves closer to
the Sun until, at the moment of inferior conjunction Ci , it is aligned with the
Sun and the Earth. If the alignment is perfect, the planet appears completely
dark, eclipsing the portion of the Sun it passes through. If, instead, the planet
does not transit on the solar disk, it is still possible to see a small illuminated
portion of it. The planet then moves away from the Sun, with a retrograde
motion towards the west, and is now visible just before dawn. Its apparent
137
138 Chapter 11. Motion of the Planets

Figure 11.1: Positions occupied by an inferior planet. The angular distance between the
Sun and the planet never exceed the maximum elongation max . The planet appears to
be stationary, as viewed from Earth (E), at the points of maximum eastern and western
elongation (Ee and Ew , respectively), while it appears to have the greatest velocity in superior
and inferior conjunction (Cs and Ci , respectively).

velocity decreases, until it appears stationary at the point of maximum (west-

ern) elongation Ew . Here, the planet reverses its motion, now approaching the
Sun at an increasing velocity. The moment when the planet is on the opposite
side of the Earth, relative to the Sun, is called upper conjunction Cs . If the
alignment is perfect, the planet disappears behind the solar disk. The planet
continues its motion in the direct direction, moving away from the Sun. Then,
at the point of maximum (eastern) elongation, it reverses its motion again,
and the cycle starts over.
The motion of the superior planets on the celestial sphere is somewhat dif-
ferent, since their elongation can take any value. Now, the moment when the
planet is on the opposite side of the Earth relative to the Sun is called conjunc-
tion, denoted by C in Fig. 11.2. When in conjunction, the planet is difficult to
observe, since it rises and sets very close to the Sun. It then proceeds in the
direct direction, moving with a decreasing velocity away from the Sun, until it
becomes stationary at the point of eastern quadrature Qe . Now the motion is
retrograde, and the planet moves away at an increasing velocity. The velocity
of the planet is maximum when it is aligned with the Sun and the Earth, on
the same side of the Earth. This is the point of opposition O. Thereafter, the
planet moves at a decreasing velocity until it stops, at the point of western
quadrature Qw . It then moves with direct motion again, until reaching the
point of conjunction, where the cycle repeats itself.
Explaining the change in direction of the motion of a planet is relatively simple
if we assume that the Earth orbits around the Sun. In the past, however, it
was believed that the Earth sat motionless at the centre of the universe (geo-
11.2. Synodic Period 139

Figure 11.2: Positions occupied by a superior planet. Unlike inferior planets, whose angular
distance from the Sun cannot exceed the maximum elongation, superior planets can have
any elongation. The planet appears stationary, as viewed from Earth (E), at the points
of eastern and western quadrature (Qe and Qw , respectively), while its apparent velocity
reaches a local maximum in conjunction and opposition (C and O, respectively).

centric model). The Ptolemaic system (the best known geocentric model) tried
to explain direct and retrograde motions by assuming that planets rotate on
circles called epicycles, which in turn move on greater circles, called defer-
entials. Moreover, to explain the difference in the length of the seasons, the
centre of the Sun’s deferential was assumed to be at some distance from the
Earth, on a point called the eccentric. Often, the simplest explanation (i.e.
the heliocentric model) is the right one.

11.2 Synodic Period

The synodic period is the interval of time between two successive conjunctions
or oppositions of a planet. After a sidereal period, the planet has completed an
orbit around the Sun, returning to the same position within the Solar System.
However, during that time, the Earth has moved around the Sun, therefore
the position of the planet as seen from Earth will, in general, be different.
All planets orbit the Sun anti-clockwise when viewed from the north pole of the
Earth. Hence, to calculate the relative angular velocity, we take the difference
between the angular velocities of the Earth (ωe ) and the planet (ωp ). For
inferior planets ωp ą ωe , therefore:

ωr “ ωp ´ ωe ,

while, for superior planets ωp ă ωe , thus:

ωr “ ωe ´ ωp .
140 Chapter 11. Motion of the Planets

Since ω “ 2π{Tsyn , ωp “ 2π{Tp and ωe “ 2π{Tyr , we find, for inferior and

superior planets, respectively:
1 1 1
“ ´ ,
Tsyn Tp Tyr
1 1 1
“ ´ .
Tsyn Tyr Tp

For inferior planets:

Tp Tyr
Tsyn “ . (11.2)
Tyr ´ Tp
For superior planets:
Tp Tyr
Tsyn “ . (11.3)
Tp ´ Tyr
These formulas can be applied to any pair of bodies in orbit around a common
centre, if we substitute for Tp and Tyr the periods of their orbits. If they travel
in opposite directions, the relative angular velocity is the sum of the angular
velocities, hence the minus sign at the denominator in Eqs. 11.2 and 11.3
becomes a plus sign, and we are left with one equation only.
We can express Tsyn directly in years by dividing both sides of the previous
equations by Tyr . For inferior planets then:

Tsyn Tp {Tyr Tp,yr

“ ñ Tsyn, yr “ .
Tyr 1 ´ Tp {Tyr 1 ´ Tp,yr

While, for superior planets:

Tsyn Tp {Tyr Tp,yr

“ ñ Tsyn, yr “ .
Tyr Tp {Tyr ´ 1 Tp,yr ´ 1
11.3. Exercises 141

11.3 Exercises

1. (INT 2014, Th.S, q.1) Two consecutive oppositions of an external planet

have been observed. The time interval between the two events is 398.85
days. Which planet is it?
2. A planet has a synodic period of 584 days. What is its distance from the
Sun and what planet is it?
3. (INT 2013, Th. S, q.3) Two space stations are in circular polar orbit
around the Earth, travelling in the same direction. Stations A and B are
located at heights h “ 200 km and H “ 400 km. Determine the time
between two consecutive alignments of the two stations.
4. The elongation of a planet is 150˝ . Is it an inferior or superior planet?
5. In the evening, which planets rise in the east?
6. The synodic period of a planet is equal to its sidereal period. What is
its sidereal period?
7. Calculate the time it takes Mars to move from opposition to quadrature.
You may assume the orbits to be circular.
8.* The largest and smallest maximum elongations of Mercury are emax,1 “
28˝ and emax,2 “ 18˝ , respectively. Find the eccentricity of its orbit.
9.* (CAO 2019, Th., q.3) An observer on earth measured the apparent mag-
nitude of an asteroid at every opposition. The asteroid’s period is 3.9
years. Estimate the eccentricity of its orbit, if the difference between
the highest and lowest apparent magnitudes is ∆m “ 2.5. Assume the
Earth’s orbit is circular.
10.* (INAO 2020, Th., q.1) On the evening of the autumnal equinox, Siddhant
noticed that Mars was exactly along the north-south meridian at the
moment when the sun was setting. If the orbital radius of Mars is 1.52 au,
at what time will Mars rise on the next autumnal equinox?

11.4 Problems

1.* Great opposition

The orbit of Mars is more eccentric than Earth’s orbit. Of the various
oppositions, some occur when Mars is at perihelion. These (great) op-
positions are very favourable for observing the planet. How often do they
occur?
 12

Orbital Manoeuvres

In this chapter we present a simplified approach for placing into orbit and chan-
ging the orbits of satellites. Most of the time the limiting factor is the available
fuel, which is related, through the rocket equation, to a quantity called ∆v.

12.1 The Rocket Equation

Figure 12.1: After the exhaustion of mass ´dm at velocity ve ´ v, the mass of the rocket
will be m ` dm and its velocity v ` dv.

Let us assume that, in the reference system of the rocket, the exhaustion speed
of the fuel is constant and equal to ve . This velocity depends on the amount of
energy generated by chemical reactions in the fuel. For an inertial observer, the
instantaneous velocity of the rocket is v (Fig. 12.1). The rocket then expels a
mass of fuel ´dm with velocity v ´ ve , and consequently its velocity and mass
increase by dv and dm, respectively (note that we will be integrating from
the initial to the final mass, hence the increment dm is negative). Requiring
momentum to be conserved:
mv “ pm ` dmqpv ` dvq ´ p´dmqpve ´ vq .
Simplifying:
dm
dv “ ´ve .
m
142
12.2. Oberth Effect 143

Integrating both sides: ż vf ż mf

dm
dv “ ´ve .
vi mi m
Hence:
mi
∆v “ vf ´ vi “ ve ln . (12.1)
mf
An important result, which we will use to justify the Oberth effect, is that ∆v
does not depend on the initial speed of the rocket, but only on the exhaustion
speed and the amount of fuel burned.

12.2 Oberth Effect

The Oberth manoeuvre consists in firing the rocket when the velocity is max-
imum, for example at perihelion. In some cases, it may be convenient to bring
the rocket closer to a celestial body to take advantage of this effect.
As shown in Sec. 12.1, ∆v depends only on the amount of fuel burned. For
sufficiently small ∆v, the change in the energy of the rocket is:
∆K “ Kf ´ Ki « mv∆v .
Therefore, for a fixed ∆v, the energy gained is directly proportional to the ve-
locity of the spaceship. Note that m remains fixed, since we are only interested
in the kinetic energy of the rocket (and not in that carried by the fuel).

12.3 Hohmann Transfer Orbit

Hohmann’s manoeuvre describes the most efficient way of transferring a satel-
lite between two circular orbits, around the same centre of force and coplanar
to each other. It consists of two stages (Fig. 12.2):
• at point P the probe is given an impulse to bring it into an intermediate
elliptical orbit, tangent to the final circular orbit;
• at point A the probe is given a second impulse, to introduce it into the
final circular orbit.
Let R1 and R2 be the radii of the initial and final circular orbits, respectively.
We want to find the two ∆v-s that the probe must be given in P and A.
After the first impulse, the probe travels on an ellipse with semi-major axis
2a “ R1 ` R2 . The energy of this orbit is given by Eq. 10.26:
GmM
E“´ .
R1 ` R2
The kinetic energy of the probe in P is then:
GmM GmM GmM ´ 1 ¯
K “E´U “´ ` “ 1´ ,
R1 ` R2 R1 R1 1 ` R2 {R1
144 Chapter 12. Orbital Manoeuvres

which corresponds to a velocity of:

c d d
2 2GM ´ 1 ¯ 2R2 {R1
vP “ K“ 1´ “ v0 ,
m R1 1 ` R2 {R1 1 ` R2 {R1

where v0 is the velocity of the circular orbit with radius R1 . The ∆v in P is:

ˆd ˙
2R2 {R1
∆vP “ v0 ´1 . (12.2)
1 ` R2 {R1

The velocity of the probe in A can be obtained applying conservation of angular

momentum at perihelion and aphelion (Eq. 10.12):

R1
mvP R1 “ mvA R2 ñ vA “ vP .
R2
a
The final velocity is that of a circular orbit with radius R2 , i.e. vf “ GM {R2 .
Hence, the ∆v that must be applied in A is:
ˆ d ˙
2
∆vA “ vf ´ vA “ vf 1 ´ .
1 ` R2 {R1
a
Since vf “ R1 {R2 v0 , the last equation can be rewritten in terms of v0 :
c ˆ d ˙
R1 2
∆vA “ v0 1´ . (12.3)
R2 1 ` R2 {R1

Assuming that both impulses are instantaneous and using Kepler’s third law,
we find the time required to enter the new circular orbit:

π2
T2 “ pR1 ` R2 q3 . (12.4)
8GMC

The Hohmann transfer orbit requires the starting and arrival points to be in
specific positions, therefore the probe can be launched only in certain windows
of time. For example, the required alignment for space missions between Earth
and Mars occurs every 26 months. An even more favourable time is during
great oppositions (Pr. 11.1), when the Earth-Mars distance is minimum. Of-
ten, Hohmann’s manoeuvre is not the fastest way to achieve a transfer, but it
does allow the rocket to carry the largest payload. Due to its reversible nature,
this manoeuvre can also be used to bring a satellite from a higher to a lower or-
bit. In this case, the impulses must be directed opposite to the velocity.
12.4. Exercises 145

Figure 12.2: Hohmann transfer between two circular orbits of radii R1 and R2

12.4 Exercises

1. Calculate the velocity increments ∆vP and ∆vA that a probe must be
given, to be transferred from a circular orbit of radius R1 “ 1 au around
the Sun, to another circular orbit of radius R2 “ 5.2 au (Jupiter’s orbit)
and coplanar with the first. Determine the transfer time T .
2.* The sum of the positive speed increments necessary to transfer a probe
around the Sun from a circular orbit of radius R1 to another circular
orbit of radius R2 and coplanar to the first is 84.73 km/s. Determine the
radii of the two orbits if the transfer time is 8h 3m 10s . You will need to
solve an equation numerically.
3.* A probe is intended to leave Earth’s gravitational field but the fuel in the
main engine is a little less than necessary. The probe is also equipped
with an auxiliary engine, capable of operating for a short period of time.
When is it convenient to operate the auxiliary engine: at take-off, or
when the probe is almost stationary, or is it not important?
4.** A construction project for a space lift requires it to remain fixed relative
to some point on Earth and suspended in equilibrium, without falling
towards the Earth or moving away into space. What should the length
of the lift be? What is the maximum tension in the rope?
5.* (ARAO 2018, Th.X, q.2) A probe was launched to another planet in the
Solar System along an energetically optimal trajectory. After flying close
to the planet, the probe immediately set off on its way back to Earth.
Throughout the mission, the probe (with no engines) made one revolution
around the Sun and returned to our planet, at exactly the same starting
146 Chapter 12. Orbital Manoeuvres

position. For which planet closest to the Sun is this possible?

6.* (BAAO 2018, Th., q.4, adapted) The most energetic chemical reaction
we can use in a rocket is hydrogen-oxygen, which gives ve = 4.46 km/s,
and engineering limits us to a rocket design with a maximum of 96 % of
launch mass being fuel. Assuming that the density of rocky planets is:
Rp
ρ “ 2.32 ` 3.18 g/cm3 ,
RC
derive the maximum Rp above which any alien civilization would be
unable to escape their planet’s gravity using simple propulsion systems.

12.5 Problems

1.* Circular orbits

What is the minimum velocity required to place a satellite into a circular
orbit around the Earth? How much faster should the satellite be, if it is
to be placed in a polar orbit? Where should the satellites be launched
from, in the two cases? What is the minimum speed required to escape
the Earth’s gravitational field?
2.* Changing the plane of the orbit
Find the minimum ∆v required to change the inclination of the circular
orbit of a satellite, moving with velocity v0 , by an angle i. Estimate the
additional mass of fuel required, at launch, to change the inclination of
the circular orbit of a 1000 kg satellite by 1˝ . Assume the satellite orbits
the Earth 200 km above the surface, and that ve “ 4.5 km/s.
3.** Third cosmic velocity
What is the minimum ∆v required for a probe, launched from Earth,
to escape the Solar System? The gravitational field of the Earth is not
negligible compared to that of the Sun. What requires more energy:
leaving the Solar System or reaching the Sun?
4.** Rocket in a gravitational field
Find the rocket equation in a constant gravitational field of magnitude
g, if the (constant) rate at which fuel is expelled is r.
5.** Reaching the centre
For what n does the potential U “ αr´n allow a body to reach the centre
of attraction (r Ñ 0), if the angular momentum is non-zero?
6.*** Gravitational slingshot
What is the best planet to use as a gravitational slingshot? Taking
advantage of this planet, what is the minimum ∆v required to escape
the Solar System, if launching from Earth?
 13

Binary Stars

Often, stars appear very close to each other, and one might wonder whether
their proximity is only apparent, or if they really do form a binary system.
If the linear distance between the stars is large (the limit depends on the mass
of the components, but is rarely greater than a few light years), they do not
orbit each other and we call them optical binaries. In fact, just under half of
all stars are in isolated systems, thus you are more likely to observe a system
of two or more stars, rather than a single star.
Based on the method of their discovery, binary systems fall into different cat-
egories. Two stars form a visual binary if their angular distance is greater than
the minimum resolution of a telescope on Earth. Even if the system cannot be
directly resolved, other methods can still be used to identify the binary.
In a binary system, the stars orbit around their common centre of mass. This
motion gives rise, as a result of the Doppler effect, to an observable shift in
the emission or absorption lines in their spectra. Binary systems that are
discovered on the basis of the Doppler shift are known as spectroscopic binar-
ies. By analysing the spectra of both stars, we can obtain the minimum mass
of each component, as well as their mutual distance. However, if one of the
stars is much brighter than its companion, the spectrum of the system can be
completely dominated by that of the brightest star. In this case, the system
is called an astrometric binary. Unless the mass of the brightest star can be
estimated from its spectral type, it is not possible to compute the mass of its
companion. If the orbital plane of the system coincides with the plane of ob-
servation, the stars periodically eclipse each other and are called photometric
binaries. From the variation in brightness of the system, it is possible to infer
the period of rotation and the relative sizes of the stars. We can then analyse
their spectrum to obtain more information.
Binary systems can also be classified according to the distance between the
components. If the two stars are very far apart, we call them distant binar-
ies, with a typical period of revolution of around tens or thousands of years.
If, instead, the stars are at a distance on the order of one astronomical unit,
they are called close binaries. In this case, tidal forces may be strong enough
147
148 Chapter 13. Binary Stars

to deform the stars significantly, giving rise to an anisotropic emission of en-

ergy from their surface, which can be observed as a periodic variation in their
brightnesses. Finally, if the stars of a close binary system come into contact,
we call them contact binaries. In the following sections, we will study in more
detail each of the cases listed above.

13.1 Distance Between Two Stars

Let S1 and S2 be two stars, at distances d1 , d2 from Earth, with declinations

δ1 , δ2 and right ascensions α1 , α2 , respectively. Considering a spherical polar
coordinate system centred on Earth, their positions can be written as:

S1 “ pd1 cos δ1 cos α1 , d1 cos δ1 sin α1 , d1 sin δ1 q ,

S2 “ pd2 cos δ2 cos α2 , d2 cos δ2 sin α2 , d2 sin δ2 q .

Then, the distance between S1 and S2 is:

d2 “ p∆xq2 ` p∆yq2 ` p∆zq2
“ d21 ` d22 ´ 2d1 d2 cos δ1 cos δ2 cos α1 cos α2 ´ 2d1 d2 cos δ1 cos δ2 sin α1 sin α2
´ 2d1 d2 sin δ1 sin δ2
Applying the cosine addition formula (Eq. A.16), we find:

d2 “ d21 ` d22 ´ 2d1 d2 rcos δ1 cos δ2 cos pα2 ´ α1 q ` sin δ1 sin δ2 s . (13.1)

When analysing possible candidates for binary systems, we usually take into
consideration stars with very similar right ascensions and declinations. Let
δ1 “ δ, δ2 “ δ`∆δ and α2 ´α1 “ ∆α, with ∆α, ∆δ ! 1. In this approximation,
Eq. 13.1 becomes (see Ex. 13.1):
” ∆δ 2 ∆α2 ı
d2 “ d21 ` d22 ´ 2d1 d2 1 ´ ´ cos2 δ . (13.2)
2 2
If ∆δ “ ∆α “ 0, the stars are on the same line of sight, hence Eq. 13.1 reduces
to d “ |d1 ´ d2 |, as expected. If d is less than a few light years, there is a good
chance that the stars do form a binary system. It is then possible to observe
the relative motion of the visual binary to obtain the shape of the orbits.

13.2 Visual Binaries

The motion of visual binaries can be studied using the generalized version of
Kepler’s law (Eq. 10.20):
4π 2
T2 “ a3 ,
Gpm1 ` m2 q
13.3. Astrometric Binaries 149

where m1 and m2 are the masses of the two stars and a “ a1 ` a2 is the sum of
the semi-major axes of their orbits. The stars orbit around the common centre
of mass, placed in one of the two foci of their elliptical orbits (Fig. 13.1). The
semi-major axes obey the relationship:
a1 m2
m1 a1 “ m2 a2 ñ “ .
a2 m1
By observing the system, it is possible to obtain the period T , while a1 and a2
can be estimated only if we know the distance of the system from Earth. In
this case, it is possible to determine the mass of both components:

4π 2
m1 “ pa1 ` a2 q2 a2 ,
GT 2
4π 2
m2 “ pa1 ` a2 q2 a1 .
GT 2

13.3 Astrometric Binaries

The stars that form this type of binary system have a significant difference in
brightness, therefore the spectrum of the faintest star is completely dominated
by that of the brightest. If the mass of the brightest star can be estimated
from its spectral class, then the mass of the faintest can also be determined.
If not, there is no way to determine the mass of the system. Sirius was the
first star to be classified as an astrometric binary, following the observation of
its periodic motion, which indicated the presence of a smaller companion —
the first example of a white dwarf. This was later directly observed thanks to
more powerful telescopes.

Figure 13.1: Orbits of two stars in a visual binary.

150 Chapter 13. Binary Stars

13.4 Spectroscopic Binaries

Spectroscopic variables appear as a single star even to the most powerful tele-
scopes, but their spectra reveal a periodic displacement of the absorption or
emission lines. Unlike astrometric binaries, in this case it is possible to observe
the spectra of both stars.
As seen in Sec. 7.6, for small velocities, the displacement of the spectral
lines is directly proportional to the radial velocity, according to the equation
∆λ “ λpv{cq. Therefore, the displacement is maximum when the stars are
directly approaching or moving away from the observer. Let us assume that
the perpendicular to the orbital plane of the binary system forms an angle i
with the direction of observation. Let v be the velocity obtained from the shift
∆λ of the spectral lines. The velocity measured in this way is actually the
projection along the line of sight of the velocities of the two stars (see Sec. 9.4,
for a similar reasoning). Therefore, the observed velocity (v) is related to the
orbital velocity (v0 ) by:
v “ v0 sin i . (13.3)
From the periodic shift of the spectral lines, we can obtain the orbital period T
of the system. Let us assume that the orbits of the stars are circular, with radii
a1 and a2 , respectively. The maximum orbital velocity of the first star is:
2πa1
v0,1 “ .
T
The observed velocity is given by Eq. 13.3:
2πa1 sin i
v1 “ . (13.4)
T
In the case of spectroscopic binaries, we cannot directly observe the two com-
ponents, therefore the empirical determination of a1 and a2 is precluded. Let
a “ a1 ` a2 be the distance between the two bodies. Considering the centre of
mass frame, we find:
m1
m1 a1 “ m2 a2 ñ a2 “ a1 . (13.5)
m2
We want to express a1 as a function of m1 , m2 and a only:
´ m1 ¯ m2
a “ a1 ` a2 “ a1 1 ` ñ a1 “ a .
m2 m1 ` m2

From Kepler’s third law (Eq. 10.20):

4π 2
T2 “ a3
Gpm1 ` m2 q
” GT 2 pm ` m q ı1{3
1 2
ña“ .
4π 2
13.4. Spectroscopic Binaries 151

Hence:
m2 ” GT 2 pm1 ` m2 q ı1{3 ” GT 2 ı1{3
a1 “ “ m2 .
m1 ` m2 4π 2 4π 2 pm1 ` m2 q2
Substituting a1 in Eq. 13.4
” 2πG ı1{3
v1 “ m2 sin i . (13.6)
pm1 ` m2 q2 T
Hence, we find:
m2 3 sin3 i v1 3 T
“ . (13.7)
pm1 ` m2 q2 2πG
If the spectrum of the system is dominated by that of the brightest star (astro-
metric binaries), only T and v1 can be empirically determined, therefore it is
not possible to compute the total mass or the mass of the individual compon-
ents. If, however, it is possible to record the spectral lines of both stars, then
v1 , v2 and T can be obtained, and the mass of both stars can be determined,
as long as we know i. Indeed, since the stars are always diametrically opposite,
their angular velocities are the same:
v1 a1
ω “ v1 {a1 “ v2 {a2 ñ “ . (13.8)
v2 a2
Using Eq. 13.5, we can write a1 {a2 “ m2 {m1 , hence:
v2
m1 “ m2 (13.9)
v1
Substituting m1 in Eq. 13.7, we find:
m2 sin3 i v1 T
“ . (13.10)
pv1 ` v2 q2 2πG
If we know the inclination i, it is straightforward to find m2 and, hence, m1 . If
the two stars eclipse each other (eclipsing binaries), it is possible to determine
the inclination, which is close to i « 90˝ . In all other cases, we can only state
a lower limit for the mass of the two stars. Isolating m1 and m2 :
T
m1 “ v2 pv1 ` v2 q2 ,
2πG sin3 i
T
m2 “ v1 pv1 ` v2 q2 .
2πG sin3 i
The minimum values for m1 and m2 are obtained by setting i “ 90˝ :
T
m1,min “ v2 pv1 ` v2 q2 ,
2πG
T
m2,min “ v1 pv1 ` v2 q2 .
2πG
152 Chapter 13. Binary Stars

This reasoning is only valid for circular orbits, whereas for increasing values
of eccentricity the behaviour of the system deviates from the circular case. In
fact, most of the spectroscopic binaries have nearly circular orbits, since tidal
forces tend to reduce the eccentricity on a time scale which is short compared
to the lifetime of the system.

13.5 Photometric Binaries

In photometric binaries, the presence of the secondary star can be inferred by

analysing the light curve of the system. The brightness of the binary system
can change over time for several reasons. If the stars periodically eclipse each
other, we call them eclipsing binaries. It is possible to observe a variation in
brightness even if the two stars do not eclipse each other. In these systems,
called ellipsoidal binaries, at least one component has been significantly de-
formed by the gravity of the other, so that, at different moments in time, the
projection of the surface area of the star in the direction of the observer is dif-
ferent. This effect, together with the fact that the temperature, and therefore
the energy radiated per unit area, is lower for parts of the surface farther from
the centre, produces a periodic variation in the amount of energy that reaches
the observer. Ellipsoidal binaries are often close or contact binaries.
Based on the characteristics of its light curve, it is possible to determine the
mechanism underlying the change in brightness of a system. Eclipsing binaries
are also called Algol stars (β Persei), named after the first observed system
of this type. Ellipsoidal binaries are instead divided into β Lyrae, which are
mainly close binaries, and WUMa, mainly contact binaries.
Eclipsing binaries usually consist of stars with very different brightnesses (for
example, a giant and a dwarf star). For the stars to eclipse each other, the
angle i between the perpendicular to the plane of the system and the line of
sight must be close to 90˝ . The light curve is roughly constant during most of
the period, when the stars are far from each other and both are visible from
Earth. The light curve (Fig. 13.2) is characterized by two minima of very
different depths: the deeper corresponds to the large and cold star totally ec-
lipsing the small and hot star; the shallower corresponds to the small and hot
star partially eclipsing the large and cold one. In fact, according to the Stefan-
Boltzmann law, the hotter star emits more energy per unit area compared to
the colder star, whilst the area of the eclipsed region, equal to the cross-section
of the smaller star, is the same in both cases. The shape of the minimum is
different depending on whether the eclipses are partial or total. In a partial
eclipse, the light curve varies gently; in a total eclipse, the minimum is flat and
constant throughout the time when one of the two stars is completely eclipsed.
Consider an eclipsing binary system, at a distance d from Earth, consisting of
stars having temperatures T1 , T2 and radii R1 , R2 , with R1 ą R2 . When both
13.5. Photometric Binaries 153

Figure 13.2: Periodic variation in the brightness of an eclipsing binary.

stars are visible, the total flux reaching the Earth is:
σ
F0 “ F1 ` F2 “ pR1 2 T1 4 ` R2 2 T2 4 q .
4d2
The first minimum corresponds to the first star covering the second completely.
If T2 ą T1 , this is the deeper minimum, otherwise, if T2 ă T1 , this is the
shallower minimum. The flux is:
σ
Fmin,1 “ pR1 2 T1 4 q .
4d2
The second minimum corresponds to the second star partially eclipsing the
first. The area of the eclipsed portion is πR2 2 , hence light from the first star
only reaches the observer from a portion of area πpR1 2 ´ R2 2 q:
σ “
pR1 2 ´ R2 2 qT1 4 ` R2 2 T2 4 .
‰
Fmin,2 “ 2
4d
The ratios of the minima to the average luminosities are thus:

Fmin,1 R1 2 T1 4
“ , (13.11)
F0 R1 T1 4 ` R2 2 T2 4
2

Fmin,2 pR1 2 ´ R2 2 qT1 4 ` R2 2 T2 4

“ . (13.12)
F0 R1 2 T1 4 ` R2 2 T2 4
154 Chapter 13. Binary Stars

13.6 Exercises

1. Prove Eq. 13.2 from Eq. 13.1.

2. Mizar and Alcor have parallaxes of 0.039362 , 0.039912 , and their coordin-
ates are αm “ 13h 23m 56s , δm “ `54˝ 551 312 and αa “ 13h 25m 14s , δa “
54˝ 591 172 . What is the distance between them in parsecs?
3. An eclipsing binary consists of two stars of radii R1 and R2 , with
R1 “ 4R2 and T1 “ 3T2 . Compute the ratio between the maximum
and minimum fluxes. Hence, find the maximum magnitude difference.
4. By what percentage does the luminosity of a Sun-like star decrease during
the transit of a gas giant (i.e. a planet similar to Jupiter) and an Earth-
like planet? What about a red dwarf?
5.* In an eclipsing binary, the difference between the two lowest magnitudes
of the system is ∆1 “ 1.5, whilst the difference between the maximum
and minimum magnitudes is ∆2 “ 2.3. The system is made up of a
large, cold star, and a small, warm star. Find the ratios of their radii
and temperatures.
6.* Two stars with masses m and 2m orbit around their common centre
of mass. If the orbits are circular, they do not intersect. What is the
smallest value of the eccentricity for which they intersect?
7.* The presence of a planet orbiting a star can be revealed by a small
periodic shift in its absorption or emission lines. What is the order of
magnitude of the frequency shift that an optimally positioned observer,
outside the Solar System, should be able to measure, if the presence
of Jupiter is to be inferred in this way? Assume that the frequency of
observation is f « 6 ¨ 1014 Hz.
8.** (ARAO 2018, Th.XI, q.6) Suppose that, in our galaxy, there is a special
class of binaries, whose components are identical to the Sun, and rotate in
circular orbits, separated by 1 au. The concentration of such systems in
space is constant (in particular, does not depend on the distance from the
disk of the Galaxy) and is equal to 0.001 pc´3 . There are observatories at
your disposal in the northern and southern hemispheres on Earth. Each
of them has a photometer that can observe stars up to a magnitude of 15 ,
with an accuracy of 0.001 magnitudes, a spectrograph with a resolution
λ{∆λ “ 105 and limiting magnitude 12 , and an astrograph with an
angular resolution of 0.12 and limiting magnitude 20. How many such
pairs will be discovered as optical, eclipsing and spectroscopic binaries,
respectively? Neglect the interstellar absorption of light.
 Part IV

Solutions

155
 14

Exercise Solutions

1. Celestial Coordinate Systems

1.1
Assuming the Earth to be perfectly spherical, the length (l) of the arc that subtends an
angle α “ φ2 ´ φ1 “ 10˝ is:
l 2πRC π
“ ñ l“ RC « 1, 113 km .
10˝ 360˝ 18

1.2

Figure 14.1: Height of the celestial equator on the (astronomical) horizon.

As you can see in Fig. 14.1, the height of the equator on the (astronomical) horizon is:

he “ 90 ˝ ´ φ .

1.3
Stars move on minor circles parallel to the celestial equator (i.e. their declination is constant),
hence the altitude of a star is constant only if the horizon and the equator coincide. This
only happens at the poles, therefore the condition is φ “ ˘90˝ .

156
 157

1.4
The vernal equinox (P) is one of the two points of intersection between the celestial equator
and the ecliptic. The declination is defined as the angle to the celestial equator, hence the
declination of P is zero. By definition, the right ascension is the angle to the vernal equinox,
therefore the right ascension of P is also zero. To sum up, δγ “ αγ “ 0.

1.5
At the north pole, the celestial equator coincides with the horizon. Since the ecliptic is
inclined by an angle of  “ 23˝ 271 (ecliptic obliquity) to the celestial equator, this is also the
angle between the ecliptic and the horizon, at the north pole.

1.6
The north ecliptic pole is inclined by an angle of  “ 23 ˝ 271 with respect to the axis of the
Earth. Hence, the observer must be located at a distance  from the north pole, i.e. at a
latitude of φ “ 90 ˝ ´  “ 66 ˝ 331 (Fig. 14.2).

Figure 14.2: The north ecliptic pole appears at the zenith for an observer at a latitude of
φ “ 90 ˝ ´  “ 66 ˝ 331 .

1.7
The best viewing conditions are obtained for objects at the zenith, since the distance which
light travels through the atmosphere is the smallest. The planets orbit the Sun very close to
the ecliptic, hence we require the ecliptic to appear at the zenith. This happens at latitudes
in the interval ´ ď φ ď , where  “ 23 ˝ 271 , i.e. for places in between the tropics.

1.8
Both the ecliptic longitude and right ascension are taken anti-clockwise from the vernal
equinox, however, the former is measured along the ecliptic, the latter along the celes-
tial equator. Therefore, when λ “ α “ ˘90 ˝ , the two coordinates have the same value
(Fig. 14.3, overleaf). This is the great circle that passes through the celestial and ecliptic
poles (P, EP) and the summer and winter solstices (Ss , Sw ), and is called the solstitial
colure. On the other hand, the great circle passing through the poles and the equinoxes is
the equinoctial colure.
158 Chapter 14. Exercise Solutions

Figure 14.3: The solstitial colure is the great circle passing through the celestial and ecliptic
poles (P, EP), and the summer and winter solstices (Ss , Sw ).

1.9
Using Eq. 1.2, to compute H from α and ST:
H “ TS ´ α “ 4h 20m 13s ´ 5h 14m 28s
“ ´p4h 74m 28s ´ 4h 20m 13s q “ ´p54m 15s q

1.10
By symmetry, the problem is equivalent to finding the distance of the horizon for an observer
on top of the skyscraper. Looking at Fig. 1.2, we are interested in finding the length of the
segment PT:
a
d “ pR ` hq2 ´ R2
?
“ 2Rh ` h2
?
« 2Rh « 103 km ,
which is approximately the distance from Dubai to the Iranian sea coast.

1.11
The orbital period of Uranus is 84 years. Its orbit is inclined by only 0.7˝ relative to the
ecliptic, so it is reasonable to assume that its orbital plane coincides with that of the Earth.
In 238 years, Uranus completes 238{84 “2.833 orbits, hence its longitude must have been
43˝ ´ 2.833 ¨ 360˝ “ ´977˝ . Normalizing to 360˝ , we get λ “ 103˝ . Therefore, Uranus was
in the constellation of Gemini.

2. Transformation of Coordinates

2.1
For the star to appear at the zenith, its altitude during upper culmination must be hu “ 90˝ .
Then, from Eq. 2.5:
hu “ 90˝ ` δ ´ φ ñ δ “ φ .
Hence, the latitude must be equal to the declination of the star.
 159

2.2
Setting hl “ 0 ˝ in Eq. 2.6:

hl “ δ ` φ ´ 90˝
ñ φ “ 90 ´ δ “ 90 ´ p16 421 582 q “ 89˝ 591 602 ´ p16˝ 421 582 q “ 73˝ 171 22 .
˝ ˝ ˝

2.3
For an observer in the northern hemisphere, the altitude of the Sun is maximum when its
declination is maximum. This happens during the summer solstice, when the declination is
equal to the ecliptic obliquity (δmax “ 23˝ 271 ). Conversely, the altitude of the Sun is min-
imum during winter solstice, when δmin “ ´p23˝ 271 q. Using Eq. 2.5, we find the maximum
and minimum altitudes of the Sun, measured from the southern horizon:

hmax “ 90˝ ` 23 ˝ 271 ´ φ “ 113˝ 271 ´ φ ,

hmin “ 90˝ ´ p23 ˝ 271 q ´ φ “ 66˝ 331 ´ φ .

Whenever hmax ą 90˝ , i.e. for φ ă 23˝ 271 , the maximum altitude of the Sun is 90˝ . In this
case, the Sun passes through the zenith twice a year, before and after the summer solstice.
For an observer in the southern hemisphere, we replace φ, δ with ´φ, ´δ, obtaining:

hmax “ 90˝ ` 23 ˝ 271 ` φ “ 113˝ 271 ` φ ,

hmin “ 90˝ ´ p23 ˝ 271 q ` φ “ 66˝ 331 ` φ .

The altitude is maximum when the declination of the Sun is minimum, i.e. during winter sol-
stice. If hmax ą 90˝ , i.e. for φ ą ´p23˝ 271 q, the maximum altitude of the Sun is 90˝ . In this
case, the Sun passes through the zenith twice a year, before and after winter solstice.

2.4
From Eq. 2.7, we find the declination of the star:
hl ` hu 34˝ 341 232 ` 53˝ 541 452
δ“ “ “ 43.5˝ 441 342 .
2 2
The latitude can be obtained from the altitude of upper culmination, using Eq. 2.5:

φ “ 90˝ ` δ ´ hu “ 90˝ ` 43.5˝ 441 342 ´ p53˝ 541 452 q “ 80˝ 191 492 .

Alternatively, we can use Eq. 2.8:

hu ´ hl 53˝ 541 452 ´ p34˝ 341 232 q
φ “ 90˝ ´ “ 90˝ ´
2 2
˝ 19˝ 201 222
“ 90 ´ “ 90 ´ p9 40 11 q “ 80˝ 191 492 .
˝ ˝ 1 2
2

2.5
In the northern hemisphere, the Sun is circumpolar when δ ą 90˝ ´ φ, hence φ ą 90˝ ´ δ.
Since the maximum declination of the Sun is equal to the ecliptic obliquity  “ 27˝ 271 , the
Sun is circumpolar at least once a year for φ ą 90˝ ´ 23˝ 271 “ 66˝ 331 . The solution is similar
for the southern hemisphere, thus in general |φ| ą 66˝ 331 .
160 Chapter 14. Exercise Solutions

2.6
Since the height of the observer is equal to the length of their shadow, it follows that the
altitude of the Sun is h “ 45˝ . If the Sun is in upper culmination at this moment, its altitude
can be related to φ and δ through:

hs “ 90˝ ´ φ ` δ ,
hn “ 90˝ ` φ ´ δ ,

if the Sun culminates to the south or to the north of the observer, respectively. Using
δ “  “ 23˝ 271 since it is summer solstice, we finally find:

φ “  ˘ p90˝ ´ hq .

Hence, the latitude can be either 68.4˝ or ´21.6˝ .

2.7
If the Sun were to set, its minimum altitude would be zero, and the length of the shade
projected by the vertical stick would be infinite. Since lmax is finite, we deduce that the
Sun is circumpolar at the place of observation. Therefore, it is indeed possible to find its
maximum and minimum altitudes:

hmax “ 90˝ ` δ@ ´ φ ,
hmin “ δ@ ` φ ´ 90˝ .

The length of the shade is related to the length of the stick and the altitude of the Sun:
l ´ l ¯
“ tanphmax q ñ 90˝ ` δ@ ´ φ “ arctan ,
lmin lmin
l ´ l ¯
“ tanphmin q ñ δ@ ` φ ´ 90˝ “ arctan .
lmax lmax
Summing and subtracting the two equations, we obtain, respectively:
1” ´ l ¯ ´ l ¯ı
δ@ “ arctan ` arctan “ 20˝ ,
2 lmin lmax
1” ´ l ¯ ´ l ¯ı
φ “ 90˝ ´ arctan ´ arctan “ 80˝ .
2 lmin lmax
As we can verify, the Sun is indeed circumpolar (δ@ ą 90˝ ´ φ).

2.8
For an observer at latitude φ, all stars rotate around the celestial north pole, which is
inclined by an angle φ to the northern horizon. The stars describe circles on the celestial
sphere with radii |90˝ ´ δ| and inclinations 90˝ ´ φ relative to the horizon. Hence, a star
with declination δ “ ˘90˝ appears stationary wherever it is observed from, and its azimuth
is always constant. At the equator (φ “ 0˝ ), all stars travel in circles perpendicular to the
horizon. Then, a star with Az “ 6h has constant azimuth until its passage on the meridian,
when its azimuth changes to Az “ ´6h . To sum up, the azimuth is constant for at least half
a day if: δ “ ˘90˝ and the observer is located anywhere on Earth, or Az “ ˘6h , and the
observer is at the equator.
It is also possible to solve the problem using Eq. 2.3b to isolate cos h and substituting in Eq.
2.3c. If we then require all terms containing the hour angle H to be zero, we find a system
of four equations, whose solutions give the desired result.
 161

2.9
This case is simple enough that we do not need to use the equations to convert from ecliptic
to equatorial coordinates. As we saw in Ex. 1.8, the point α “ 6h is at λ “ 90˝ . At this
point, the ecliptic is above the celestial equator by an angle of  “ 23˝ 271 , while the star is
δ “ 10˝ above the celestial equator. Hence, the ecliptic latitude of the star is β “ δ ´  “
10˝ ´ p23˝ 271 q “ ´p13˝ 271 q.

2.10
In the northern hemisphere: In the southern hemisphere:
• S: H “ 0h , δ “ ´p90˝ ´ φq ; • S: H “ 12h , δ “ ´p90˝ ` φq ;
• W: H “ 6h , δ “ 0˝ ; • W: H “ 6h , δ “ 0˝ ;
• N: H “ 12h , δ “ 90˝ ´ φ ; • N: H “ 0h , δ “ 90˝ ` φ ;
• E: H “ 18h , δ “ 0˝ ; • E: H “ 18h , δ “ 0˝ ;
h
• Z: H “ 0 , δ “ φ . • Z: H “ 0h , δ “ φ .

2.11
A star that rises exactly in the east and sets exactly in the west stays above the horizon for
12h . As the points of rising and setting shift north, the star is visible for longer, because
more of its path will be above the horizon. Therefore, stars that are visible for more than
12h rise in the north-east direction.

2.12
Neglecting refraction and parallax, assuming the Moon has zero declination, Gianna will
observe the Moon at the meridian once the Earth has rotated by 90 ˝ , that is, after 1/4 of
a sidereal day. Therefore, Gianna’s longitude is λ “ ´p12˝ 301 q ` 90˝ “ 77 ˝ 301 E. The only
capital city with this longitude is Washington.

2.13
On the 21st of April, on average 30 days have passed since the vernal equinox. Applying Eq. 2.13:
” ´ 2π ¯ı
δ@ “ arcsin sin  sin 30 “ 11˝ 191 « 11˝ .
365.25
The Sun is circumpolar at:

φ ě 90 ˝ ´ δ@ “ 90 ˝ ´ 11˝ “ 79˝ .

2.14
A star can pass through the zenith only if its declination is equal to the latitude, hence
δ “ φ. In this case, the passage happens during upper culmination, when Hc “ 0. When
the sidereal time is 9h 2m , the hour angle of the star is H “ 0h ` p9h 2m ´ 10m q “ 8h 52m .
Setting δ “ φ in Eq. 2.4c:

sin h “ cos H cos2 φ ´ sin2 φ ñ 1 ´ sin h “ cos Hp1 ´ cos φq .

Isolating φ and substituting h “ 78˝ 121 :

c
1 ´ sin h
φ“ “ 83˝ 34m .
1 ´ cos H
162 Chapter 14. Exercise Solutions

2.15
During the equinoxes, the path of the Sun is perpendicular to the horizon, hence its velocity
relative to the horizon is maximum. By examining this case, we obtain the minimum velocity
that guarantees a successful observation at any other time of the year. The apparent angular
velocity of the Sun is:
360˝
ω“ “ 4.178 ¨ 10´3 ˝ {s .
86164.1s
At the height of h “ 6 m, the dip of the horizon is:
´ R ¯
C
θ “ arccos “ 0.0786˝ .
RC ` h
The Sun covers this angle in a time ∆t “ θ{ω “ 18.81 s. Hence, the velocity must be greater
than vc “ 6 m{∆t “ 0.32 m/s. If the velocity is exactly vc , during the equinoxes, the sunset
is prolonged by approximately 20 s.

3. Perturbation of Coordinates

3.1
The Barnard star has an angular velocity of ub “ 10.32 /yr. The angular diameter of the
Moon is approximately θ$ “ 301 , i.e. 30 ¨ 602 “ 18002 . Hence, the time taken is:
θ 1800
∆t “ $ “ « 174.8 years .
ub 10.3

3.2
Aldebaran’s distance from Earth is d “ 1{πp2 “ 20.83 pc, hence its tangential velocity is:
vt “ u ¨ d “ 4.167 pc ¨ arcsec/yr .
By definition, pc ¨ arcsec “ 1 au, therefore:
vt “ 4.167 au/yr « 20 km/s .
The redshift parameter is z “ ∆λ{λ “ 1.79 ¨ 10´4 . Since z ! 1, we use Eq. 7.21:
vr “ c ¨ ∆λ{λ « 54 km/s .
The total velocity is then: b
vt “ vr2 ` vt2 « 58 km/s .

3.3
Let upr “ dλ{dt “ 502 /yr. From Eqs. 3.3:
dδpr
“ upr sin  cos α ,
dt
dαpr
“ upr psin α sin  tan δ ` cos q .
dt
The total rates of change of the coordinates, dαtot {dt and dδtot {dt, are equal to the sum of
the perturbations due to the proper motion of the star and to the precession of Earth’s axis:
$
dαtot
& dt “ uα ` upr psin α sin  tan δ ` cos q
’
’
(14.1)
% dδtot “ uδ ` upr sin  cos α .
’
’
dt
 163

In principle, it is possible to solve this system of coupled differential equations. First, using
both equations, solve for α and dαtot {dt as a function of δ and dδtot {dt only. Then, take
the derivative of the second equation with respect to time, and substitute your previous
expressions for α and dαtot {dt. You will be left with a second order differential equation,
with the only variable being δ and its time derivatives. Although not pretty, it can be solved.
In this case however, the period under consideration is short enough so that we may regard
dδtot {dt and dαtot {dt as being constant. The approximate solution is thus:
dαtot
αJ2000.0 “ αJ1900.0 ` ∆t « 6h 46m 26s ,
dt
dδtot
δJ2000.0 “ δJ1900.0 ` ∆t « 16˝ 381 422 .
dt

3.4

Figure 14.4: I and F are the initial and final positions of Sirius, E is the observer on Earth.

Sirius’ distance is d “ 1{πp2 “ 2.667 pc. The tangential velocity is given by Eqs. 3.8 and 3.7:
b
vt “ u ¨ d “ d u2α cos2 δ ` u2δ “ 1.603 ¨ 10´6 pc/yr .

The radial velocity is vr “ 8.182 ¨ 10´6 pc/yr. The total velocity is then:
b
v “ vr2 ` vt2 “ 8.337 ¨ 10´6 pc/yr .

Looking at Fig. 14.4, the distance Sirius has to travel to be the closest to the Sun is:
vr
∆d “ d cos θ “ d ¨ .
v
The time taken is therefore:
∆d vr 8.182
∆t “ “ d ¨ 2 “ 2.667 ¨ ¨ 106 years “ 3.14 ¨ 106 years .
v v 8.3372
At the point of closest approach, the radial velocity is zero. Indeed, if it were negative, the
star would be getting closer to the Sun, but this is absurd since we already chose the point
of closest approach. If it were positive, because the velocity was initially negative, some
time earlier the velocity must have been zero, but then the body was closer at that time.
Hence, the velocity must be entirely tangential, i.e. vr,f “ 0, vt,f “ v. The final distance is
2
df “ d sin θ “ dvt {v, thus πp,f “ 1{df “ v{pvt ¨ dq “ 8.337{p1.603 ¨ 2.667q “ 1.952 .
164 Chapter 14. Exercise Solutions

3.5
As seen in Sec. 3.5, we can write Snell’s law between the layer at h “ 0 and that at h:
nphq sin αi
“ .
nph “ 0q sin αr
Since the ray circles the Earth at a height h, the refraction angle is αr “ 90˝ . Instead, the
incident angle is sin αi “ R{pR ` hq. Substituting nphq in the previous equation:
n0 1 R
¨ “ sin αi “ .
1 ` h n0 R`h
Taking the reciprocal of both sides:
h
1 ` h “ 1 ` ñ R “ 1{ .
R

3.6

Figure 14.5: Refraction of light on Hydroplanet.

As can be seen in Fig. 14.5, for an object seen on the horizon, Snell’s law gives:
sin αr nw 4
“ “ ,
sin αi ns 3
where nw “ 1.33 « 4{3 and ns “ 1 are the refractive indices of water and space, respectively.
If the ray joining P and A is totally reflected, then it is not possible to see any object on the
horizon. This happens when αr ą 90˝ , hence the stars are visible on the horizon only if:
sin αi 3 3
ď ñ sin αi ď .
sin 90˝ 4 4
Considering triangle OPA in Fig. 14.5, we see that sin αi “ R{pR ` hq, hence the condition
on the height is:
h 4 R
1` ě ñ hě .
R 3 3

From now on we will consider the case h “ R{3, for which αr “ 90˝ . Now, the outgoing
 165

ray (PA) coincides with the apparent horizon, which forms an angle of αr ´ αi with the
astronomical horizon. As long as the star is above the apparent horizon, which covers an
angle of 180 ` 2pαr ´ αi q “ 360 ´ 2αi , it is day on Hydroplanet. Denoting with T the period
of rotation, the length of the day is:
360 ´ 2 arcsin p3{4q
t“ ¨ T “ 0.73 ¨ T “ 17h 31m 12s .
360
The value of atmospheric refraction at the horizon is:
´3¯
R “ αr ´ αi “ 90˝ ´ arcsin “ 41.4˝ ,
4
much larger than 341 , on Earth.

4. Observation and Instruments

4.1
From Eq. 4.10, we obtain the angular resolution:
λv 500 ¨ 10´9
R « 1.22 “ 1.22 “ 6.1 ¨ 10´8 rad .
Dv 10
If a radio telescope observes at λr “ 20 cm, with the same resolution, its diameter is:
λr λr 0.2
Dr « 1.22 “ Dv “ ¨ 10 m “ 4 ¨ 106 m .
R λv 500 ¨ 10´9
This shows that a single radio telescope cannot achieve the same resolution of an optical
telescope, but other strategies, such as interferometry, need to be used.

4.2
Using Eq. 4.2:
h0 400
αrad “ ñ h0 “ d ¨ αrad “ 55 ¨ 106 ly ¨ “ 1.06 ¨ 105 ly .
d 206264.8
From Eq. 4.8, we then have:
400
h1 “ αrad ¨ fo “ m “ 1.94 mm .
206264.8

4.3
The maximum and minimum distances of Mars from the Sun are, respectively:
da “ ap1 ` eq “ 1.5237 ¨ p1 ` 0.0934q au “ 2.492 ¨ 1011 m ,
dp “ ap1 ´ eq “ 1.5237 ¨ p1 ´ 0.0934q au “ 2.067 ¨ 1011 m .
The radius of the Sun is R@ “ 695475 km, hence its maximum and minimum angular
diameters are, respectively:
αmax “ 2R@ {dp “ 6.74 ¨ 10´3 rad ,
αmin “ 2R@ {da “ 5.59 ¨ 10´3 rad .
The distance to the Moon is d$ “ 384.4 ¨ 106 m. If a solar eclipse is just visible, αmin is equal
to the angular size of the satellite, as seen from Mars. Therefore, the diameter is:
d “ αmin ¨ d$ « 2150 km .
166 Chapter 14. Exercise Solutions

The angular sizes of the Sun observed from Mars and from Earth are in the same ratio as
the Mars-Sun and Earth-Sun distances. When Mars is in aphelion, this ratio is equal to
1.5237p1 ` 0.0934q “ 1.666, hence the satellite can be 1.666 times smaller than the Moon, i.e
it can have a minimum diameter of 2R$ {1.666 « 2085 km; which is approximately the same
result we obtained earlier. This result is actually more accurate, since it makes no reference
to the Earth-Moon distance and the radius of the Sun. The reason for the discrepancy is
that total solar eclipses on Earth only occur when the Moon is close to aphelion, or the
Earth is close to perihelion (hence d$ is actually a bit smaller than 2150 km).

4.4
The ratio of the angular diameters is equal to the ratio of the distances
a from which Mars is
observed. During quadrature, the Earth-Mars distance is dQ “ d2D ´ d2C , during opposition
dO “ dD ´ dC . Hence, the ratio of the angular diameters is:
d ´d
a D2 C2 “ 0.46 .
dD ´ dC

4.5
Using 4.9, we have:
h1 fo h1
“ ñ fo “ d “ p1254 ˘ 10q mm .
h0 d h0

4.6
As explained in Sec. 3.6, the angle ∆α on the celestial equator corresponds to the angle
∆α cos δ on the parallel of declination δ, since the radius (hence, the arc length) of the
parallel is smaller by a factor of cos α. In this case the computation is easy, because both
pairs of stars have one common coordinate. Otherwise, we would need to use the spherical
law of cosines (Appendix A.4). The angular distances are:

∆1,2 “ δ1 ´ δ2 “ 401 “ 24002 ,

360˝
∆1,3 “ pα1 ´ α2 q cos δ1 “ 1.224˝ “ 44052 .
24h
The focal length of the telescope is fo “ 10 ¨ 200 mm?“ 200 cm. The CCD is a square, with
length L “ 4096 ¨ lpix “ 2.62 m and diagonal D “ 2L “ 3.71 cm. To capture the widest
possible angle, we use its diagonal:
D D
“ tan ∆max ñ ∆max “ arctan “ 1.06˝ « 38162 .
fo fo
Hence, it is possible to capture Star 1 and 2 together, but not Star 1 and 3.

4.7
The field of view is:
360
θFOV “ p2.5 ¨ 60q ¨ cos p46˝ 01 14.42 q “ 261 ,
86164.1
where we have used the fact that the angular velocity of a star with declination δ is ωsid cos δ
(see Sec. 3.6). The angular diameter of the Moon is approximately 301 , hence we cannot see
the full Moon using this set up.
 167

4.8
To find an approximation to the FOV of the telescope, we divide the FOV of the eyepiece
by the magnification:
FOVeyepiece FOVeyepiece 45˝
FOVtelescope “ “ “ “ 1.7˝ .
ω ftelescope {feyepiece 130 ¨ 5{25
The faintest star that is observable with the naked eye has an apparent magnitude of meye “
6, and the diameter of the pupil at night is approximately d “ 8 mm. On the other hand,
the telescope has a diameter of D “ 130 mm, hence a star viewed with the telescope can be
D2 {d2 fainter, and barely visible. According to Eq. 8.4:
D2 D
meye ´ mtelescope “ ´2.5 log “ ´5 log
d2 d
D 130
ñ mtelescope “ meye ` 5 log “ 6 ` 5 log « 12 .
d 8
To calculate the angular resolution, we use Eq. 4.10:
1.22λ
θ“ « 5 ¨ 10´6 rad ,
D
where λ “ 550 nm has been taken as the average wavelength for visible light. Using Eq. 10.14,
taking into account that the total mass of the system is Mtot “ p18.9 ` 16.2qM@ “ 35.1M@ ,
we compute the separation between the stars in astronomical units:
2
Tyr 1 2
3
“ ñ a “ pTyr ¨ Mtot, in M@ q1{3 « 1.45 au .
aau Mtot, in M@
Therefore, the angular separation is:
a 1.45
α“ “ “ 3.1 ¨ 10´9 rad . (14.2)
d 2.29 ¨ 103 pc ¨ 206265au/pc
The angular separation between the stars is smaller than the angular resolution, therefore
the astronomer will observe both stars as a single point in his telescope.

4.9
The curvature on either side of the system of lenses is in the same direction, hence RA and RB
are both positive. Taking into account the different indices of refraction in Eq. 4.5:
1 pnA ´ 1q pnB ´ 1q
“ ´ .
f RA RB
Considering red and blue light, respectively:
1 nA,r ´ 1 nB,r ´ 1
“ ´ ,
fr RA RB
1 nA,b ´ 1 nB,b ´ 1
“ ´ .
fb RA RB
Since fr “ fb “ f , equating both sides:
nA,b ´ nA,r nB,b ´ nB,r nB,b ´ nB,r
“ ñ RB “ RA .
RA RB nA,b ´ nA,r
Inserting the numerical values, we get RB “ 2RA . Using this condition in the equation for fr :
1 1 ” ı 2nA,r ´ nB,r ´ 1
“ 2pnA,r ´ 1q ´ pnB,r ´ 1q “ .
f 2RA 2RA
Hence:
f {2
RA “ “ 1m.
2nA,r ´ nB,r ´ 1
Therefore, RB “ 2RA “ 2 m.
168 Chapter 14. Exercise Solutions

5. Time Systems

5.1
The difference in longitude is equal to the difference in sidereal time. Indeed, the (fixed)
stars return to the same position with respect to the meridian every complete rotation of
the Earth, that is, every sidereal period. Hence, the time difference between two locations
with a difference in longitude of ∆λ is:
∆λ
∆Tsid “ ¨ 86400s .
360˝
Using Eq. 5.2, this can be converted in solar time:
86164.1 ∆λ
∆Tsol “ ¨ ∆Tsid “ ¨ 86164.1s .
86400 360˝

5.2
The difference in longitude between Rome and Greenwich is ∆λ “ 12 ˝ 281 54.482 , which
corresponds to a difference in sidereal time of 49m 55.63s , or p86164.1{86400q ¨ p49m 55.63s q “
49m 47.45s of solar time. Since Rome is to the east of Greenwich, the time in Rome is greater
by 49m 47.45s . When the time in Greenwich is 12 : 00 GMT, in Rome the local time is
12h 49m 47.45s . Since Rome belongs to the time zone GMT+1, when in Greenwich the time
is 12 : 00 GMT, in Rome the meridian time is 13 : 00 GMT+1. The difference between
GMT+1 and the local time in Rome is therefore 10m 12.55s .

5.3
The difference in local time is ∆t “ 11h 14m 15s ´10h 17m 14s “ 57m 1s . Local time is measured
in solar time, but the difference in longitude is equal to the difference in sidereal time. By
converting in sidereal time, using Eq. 5.3, we find ∆λ “ ∆t ¨ p86400{86164.1q « 3431s “
57m 11s . The longitude of Naples is therefore ∆λ “ 14˝ 171 392 .

5.4
The 24 principal meridians corresponding to the time zones are separated in longitude by 15˝ ,
starting from Greenwich at λ “ 0˝ . The meridian closest to the place under consideration is
the one at 135 ˝ W. The difference in longitude between the place and the meridian is thus
∆λ “ 6˝ 451 “ 27m . This corresponds to a difference in solar time of p86164.1{86400q ¨ 27m “
26m 55.5s . Hence, the local time is Tm “ 8h 46m 22.0s ` 26m 55.5s “ 9h 13m 17.5s . From the
equation of time, we can then determine the true solar time:

E.T. “ Tt ´ Tm ñ Tt “ E.T. ` Tm “ ´p14m 13s q ` 9h 13m 17.5s “ 8h 59m 4.5s .

Hence, the hour angle of the Sun is:

Ht “ Tt ´ 12h “ ´p3h 0m 55.5s q .

5.5
Solar and sidereal time are equal during the autumnal equinox, when P culminates at midnight.

5.6
Each solar day the stars move west by 3m 56s . Hence, they rise 3m 56s earlier every day, or
1h 58m earlier after one month. Therefore, a month later, the time of rising will be 20 : 58.
 169

5.7
A star transits the meridian when its hour angle is zero. At this moment, the sidereal
time is equal to the right ascension of the star (Eq. 1.2), i.e. ST“ α “ 18h . There-
fore, we need to wait 18h ´ p9h 50m 12s q “ 8h 9m 48s of sidereal time, which corresponds to
p86164.1{86400q ¨ p8h 9m 48s q “ 8h 8m 27.8s of solar time. Hence, the star transits the meridian
at 0h ` 8h 8m 27.8s “ 8h 8m 27.8s GMT.

5.8
We want the star to be in upper culmination, so that the path which light travels through
the atmosphere is the shortest possible. At the moment of upper culmination, the star is on
the meridian, hence the sidereal time is equal to the right ascension: ST “ α “ 10h . Since
we want to observe the star at midnight, we require the sidereal time at 12:00 on that day
to be 10h ´ p86400{86164.1q ¨ 12h “ ´2.03h “ 21.97h (normalized to 24h ). Starting from the
21st of March, when ST“ 0, the sidereal time increases by about 2h each month. Therefore,
ST “ 21.97h corresponds to the end of February.

5.9
When the time in Rome is 12:00 GMT+1, in Greenwich the time is 11:00 GMT. The time
elapsed between the vernal equinox and 11:00 GMT on the 13th of April is 24d 3h 25m “
24.1424d . Using Eq. 5.4, we find the sidereal time in Greenwich:
ST “ Tm ` 11h 53m ` 24.1424 ¨ p3m 56s q “ Tm ` 1h 34m 58s .
In Rome, the local time is 11h ` p12˝ 281 54.482 q{15˝ “ 11h 49m 56s . Hence, the sidereal time
in Rome, at 12 : 00 GMT+1 on the 13th of April 2000, was:
ST “ 11h 49m 56s ` 1h 34m 58s “ 13h 24m 54s .

5.10
Neglecting atmospheric refraction, the hour angle of rising and setting can be computed
from Eq. 2.9:
cos H “ ´ tan δ tan φ ñ H “ ˘108.47˝ “ ˘7h 14m .
Stars rise to the east of the meridian, hence the hour angle of rising must be negative. There-
fore, the hour angles of rising and setting are Hr “ ´7h 14m and Hs “ 7h 14m , respectively.
The sidereal times of the two events are then:
STr “ Hr ` α “ ´p7h 14m q ` 14h 16m “ 7h 2m ,
STs “ Hs ` α “ 14h 16m ` 7h 14m “ 21h 30m .
Note that the answer is independent of the date, indeed a star rises and sets at the same
sidereal time every day. Between the 10th of January and the vernal equinox, there are
around 70 days, therefore the sidereal time on the 10th of January is (Eq. 5.4):
ST “ Tm ` 11h 53m ´ 70 ¨ p3m 56s q “ Tm ` 7h 18m ,
Isolating the mean solar time, we find:
Tm,r “ STr ´ 7h 18m “ 7h 2m ´ p7h 18m q “ 23h 44m ,
Tm,s “ STs ´ 7h 18m “ 21h 30m ´ p7h 18m q “ 14h 12m .
The closest main meridian to Boston is the one at 75˝ W, hence the meridian time is smaller
by 4˝ {15˝ “ 16m than the local time in Boston (actually 86164{86400 ¨ 16m “ 15m 57s ).
Therefore, the times of rising and setting are 23 : 30 and 14 : 00 GMT+5 respectively, with
a precision of 10m .
170 Chapter 14. Exercise Solutions

5.11
If the Earth rotated in the opposite direction, the velocity of the Sun relative to the celestial
sphere would be ωr “ ω ` Ω, where ω and Ω are the angular velocities of rotation and
revolution, respectively. Therefore:
1 1 1 tsol triv 365.2564
“ ` ñ tsid “ “ “ 1.002745d “ 1d 3m 57.2s .
tsol tsid triv triv ´ tsol 365.2564 ´ 1
Hence, the sidereal day would be 3m 57s longer than a solar day.

5.12
The length of a year is 365 days. Every fourth year is a leap year, the years divisible by 100
are not leap years, except the ones that are also divisible by 400. Considering an interval
of time of 400 years, starting from 0 until 399, we see that year 0 is a leap year, but years
100, 200, 300 are not. Hence, compared to the Julian calendar, there are three less leap years
every 400 years. Therefore, the average length of a year in the Gregorian calendar is:
1 3
T “ 365 ` ´ “ 365.2425d .
4 400
The length of the tropical year is 365.2422d , so the difference between the true and mean
year is 3 ¨ 10´4 days, every year. Hence, after 100, 500 and 1000 years, the error in the
Gregorian calendar will be, respectively:

∆t100 “ 100 ¨ 3 ¨ 10´4 “ 0.03 days ,

∆t500 “ 500 ¨ 3 ¨ 10´4 “ 0.15 days ,
∆t1000 “ 1000 ¨ 3 ¨ 10´4 “ 0.3 days .

5.13
The orbital period of Venus is tsid “ 0.61519 years. From Eq. 11.2, the synodic period is:
tsid
Tsyn “ « 1.59869 years
1 ´ tsid
Between 2012 and 2117 there are 105 years. Of these, tp2117 ´ 2012q{4u ´ 1 “ 25 are
leap years, since 2100 is not a leap year. Let x be the number of days between the 1st of
January 2117 and the day when Venus transits the Sun that same year. Since the last transit
happened on the 6th of June 2012, i.e. 157 days after the 1st of January, the difference in
days between the two eclipses is 365 ¨ 105 ` 25 ` px ´ 157q “ 38193 ` x. The period of a planet
is measured in sidereal years, therefore the synodic period is 1.59869 ¨ 365.2564 “ 583.9317d .
The transit happens again when the Earth and the Sun return to the same relative position,
hence the number of synodic years between the two transits must be an integer. Then,
p38193 ` xq{583.9317 “ n, with n integer. Since 0 ď x ď 365, the only possible value is
n “ 66, and we obtain x “ 66 ¨ 583.9317 ´ 38193 “ 346.49. Hence, the next transit will be
in 2117, 346 days after the 1st of January, i.e. on the 12th of December. Furthermore, the
eclipse will occur approximately 12h later than the previous one.

6. The Moon

6.1
The angular distance of the full moon from the Sun (also called elongation) is 180˝ .
Bodies that orbit the Sun at a distance smaller than the Earth-Sun distance, cannot have an
 171

elongation greater than 90˝ . For example, the maximum elongations of Mercury and Venus
are 28˝ and 48˝ , respectively (see Ch. 11). Therefore, Mercury and Venus can never be
eclipsed by the full moon. On the other hand, superior planets (all the others, except the
Earth) can have any elongation, and can therefore be eclipsed by the full moon.

6.2
During summer, the declination of the Sun is close to the maximum value of  “ 23˝ 271 ;
during winter, it is close to the minimum of ´. The full moon is on the opposite side of
the Earth relative to the Sun. Hence, if the ecliptic is δ@ above the celestial equator in the
direction of the Sun, it is δ@ below the equator in the direction of the full moon. Since the
orbit of the Moon is inclined by an angle i “ 5˝ 91 to the ecliptic, it follows that the declination
of the Moon during summer and winter is close to ´ ˘ i and  ˘ i, respectively. Hence, for
an observer in the northern hemisphere, the full moon is higher in winter than in summer.
Similarly, the new moon, as well as the Sun, are higher in summer than winter.

6.3
Since the orbit of the Moon is inclined by i “ 5˝ 91 to the ecliptic, and the ecliptic is inclined
by 23˝ 271 to the celestial equator, it follows that the declination of the Moon can take any
value between δmax “ 28˝ 351 and δmin “ ´28˝ 351 . The condition for a body with declination
δ to appear at the zenith is hu “ 90˝ “ 90˝ ´ φ ` δ, i.e. φ “ δ. Hence, the Moon can be
seen at the zenith from places with latitude:

´p28˝ 351 q ă φ ă 28˝ 351 ñ |φ| ă 28˝ 351 .

6.4
The Moon is circumpolar if φ ą 90˝ ´ δ. Since the maximum declination of the Moon is
δmax “ 28˝ 351 (see Ex. 6.3), we find φ ą 61˝ 251 . By symmetry, in the southern hemisphere,
φ ă ´p61˝ 251 q. Hence, in general |φ| ą 61˝ 251 .

6.5

Figure 14.6: Distance of the Moon at the zenith and horizon.

Looking at Fig. 14.6, when the Moon is at the zenith, its distance from Earth is:

dz “ d$ ´ RC “ 377629 km .
172 Chapter 14. Exercise Solutions

When it is on the horizon:

b
dh “ d2$ ´ RC
2
“ 383947 km .

Hence, the Moon is closer when at the zenith.

6.6
Using Eq. 10.18, we find the orbital velocity of the Moon:
d
GM$
v“ “ 1017 m/s .
d$

Since 2R$ “ 3474.2 km, the time taken to cover a distance equal to its diameter is:
2R$
∆t “ “ 3416 s « 1h .
v

6.7
As explained in Sec. “Lunar Phases”, the Moon is visible in the morning either if it rises at
6 : 00 and sets at 18 : 00 (new moon), or if it rises at 0 : 00 and sets at 12 : 00 (last quarter).
We are told that only half of the Moon is visible, hence there must be a last quarter moon.
In this case, the side to the east is illuminated.

6.8
For a planet to be eclipsed by the full moon, it must orbit the Sun close to the plane of the
ecliptic, at a distance to the Sun greater than the Sun-Earth distance (see Ex. 6.1). Then,
the eclipse occurs when both the full moon and the planet are close to the node line. Since
the node line precesses with a retrograde motion of period 18.61 years, the sidereal period
of the planet must be exactly 18.61 years, if it is not to be eclipsed by the full moon. Then,
from Kepler’s third law (Eq. 10.14):

Tpr, yr 2 “ a3au ñ a “ 18.612{3 au “ 7.022 au .

6.9
The length of the tropical year is 365d 5h 48m 46s , that of the lunar year is 354d 8h 48m . Hence,
the lunar year is shorter than the tropical year by 10d 21h 46s . Since the synodic period of the
Moon is 29d 12h 44m 33s “ 29.5309d , after 10d 21h 46s “ 10.8755d , the Moon forms an angle of
10.8755{29.5309 ¨ 360˝ “ 132.6˝ with the Earth and the Sun. Therefore, the Moon will be in
between the first quarter (90˝ ) and full moon (180˝ ).

6.10
From Eq. 6.1, we see that the radius of the umbra projected by the Earth at the position of
the Moon is:
d
r “ RC ´ $ pR@ ´ RC q .
dC
Hence, the maximum distance of the Sun from the node line can be:
r ` R$ 1 1 ´ RC ` R
$ ´ R@ ´ RC .
¯
∆λ “ ˝
ñ ∆λ “ ˝
d$ sin 5.145 sin 5.145 d$ dC
Now, ∆λmax is obtained when d$ is minimum and dC is maximum, whilst ∆λmin is obtained
when d$ is maximum and dC is minimum. Let eC , e$ be the eccentricities of the orbits of
 173

the Earth and Moon, respectively. Then, dC,max “ p1 ` eC qdC , dC,min “ p1 ´ eC qdC , while
d$,max “ p1 ` e$ qdC , d$,max “ p1 ´ e$ qd$ . Hence:
1 ” RC ` R
$ ´ R@ ´ RC « 11.38˝ ,
ı
∆λmax “ ˝
sin 5.145 p1 ´ e$ qd$ p1 ` eC qdC
1 ” RC ` R
$ ´ R@ ´ RC « 9.79˝ .
ı
∆λmin “
sin 5.145˝ p1 ` e$ qd$ p1 ´ eC qdC

6.11
We can solve the exercise both from scratch, using a similar approach to that described in
Sec. “Solar Eclipse”, or by using Eq. 6.1 with the substitution d$ Ñ ´d$ , since, during
a solar eclipse, the Moon is on the opposite side of the Earth compared to a lunar eclipse.
Adopting the second strategy:
d 1 ´ RC ` R
$ ` R@ ´ RC .
r “ RC ` $ pR@ ´ RC q ñ ∆λ “
¯
dC sin 5.145˝ d$ dC
Following a similar reasoning to Ex. 6.10:
1 ” RC ` R
$ ` R@ ´ RC « 17.3˝ ,
ı
∆λmax “ ˝
sin 5.145 p1 ´ e$ qd$ p1 ´ eC qdC
1 ” RC ` R
$ ` R@ ´ RC « 15.7˝ .
ı
∆λmin “
sin 5.145˝ p1 ` e$ qd$ p1 ` eC qdC

6.12
The distance from the Moon to the surface of the Earth is d$ ´ RC , whereas the Sun-Moon
distance during a solar eclipse is dC ´ d$ . To compute the radius of the umbra on the
surface of the Earth, it is possible to use Eq. 6.1, substituting RC Ñ R$ , d$ Ñ d$ ´ RC
and dC Ñ dC ´ d$ . The radius of the umbra is maximum when the Earth is at aphelion
and the Moon is at perigee:
d$ p1 ´ e$ q ´ RC
ru “ R$ ´ pR@ ´ R$ q « 116 km .
dC p1 ` eC q ´ d$ p1 ´ e$ q
Therefore, the maximum length of the umbra is about 230 km.
If we take the distances dC , d$ equal to their average values (i.e. the semi-major axes), we
obtain a negative value of r. This means that, for a total solar eclipse to occur, the Earth
must be close to aphelion or the Moon close to perigee (otherwise the eclipse is anular).
Let us now derive the radius of the penumbra on Earth. Looking at Fig. 6.4, it suffices to
replace R@ in the previous formula with ´R@ (or, equivalently, R$ Ñ ´R$ , but then we
also need to take the opposite sign overall). We then get, on average:
d ´ RC
rp “ R $ ` $ pR@ ´ R$ q .
dC ´ d$
Neglecting RC and R$ in the two factors at numerator, and d$ at denominator, we obtain:
d R@ {dC ¯
rp « R$ ` $ R@ “ R$ 1 `
´
,
dC R$ {d$
but R@ {dC and R$ {d$ are the angular diameters of the Sun and the Moon, therefore the
fraction is approximately unity (actually 1.03, on average). Hence, it follows that the radius
of the penumbra on Earth is approximately twice the radius of the Moon: rp « 2R$ .
Depending on the relative positions of the Earth and the Moon, the radius of the penumbra
can vary, deviating from the average by at most a few hundred kilometres.
174 Chapter 14. Exercise Solutions

7. Radiation Mechanisms

7.1
Applying Wien’s law (Eq. 7.10):
k 2.897 ¨ 10´3 m K
λmax “ “ « 724 nm .
T 4000 K
which corresponds to the colour red.

7.2
Applying Stefan-Boltzmann law (Eq.7.13) to the star under consideration and the Sun:
L “ 4πσR2 T 4 ,
L@ “ 4πσR@ 2 T@ 4 .
Dividing the two equations and using T@ “ 5778 K:
L ´ R ¯2 ´ T ¯4 ´ 6800 ¯4
“ “ 2.52 ¨ « 12 .
L@ R@ T@ 5778

7.3
Let r be the radius of a dust particle and a its distance to the Sun. Light is incident on an
area of πr2 , hence the radiation force is given by Eq. 7.4:
F L@ L@ r2
Fr “ ¨ πr2 “ 2
πr2 “ .
c 4πa c 4a2 c
The gravitational force, written as a function of the density and radius of the dust particle is:
GM@ 4 GM@
Fg “ m “ πr3 ρ ¨ .
a2 3 a2
Equating the forces:
L@ r2 4π 3 GM@
“ ρr .
4a2 c 3 a2
Isolating r, we find:
3L@
r“ « 0.6 µm .
16πρGM@ c
Both the gravitational force and the radiation pressure grow as a´2 , hence the radius of
equilibrium does not depend on the distance a. The gravitational force is proportional to
the mass of the particle, which grows as r3 , while the radiation force is proportional to the
cross section area, which grows as r2 . For this reason, there is only one value of r for which
these two forces are equal.

7.4
The electronvolt eV is equal to the kinetic energy gained by an electron accelerating from
rest through an electric potential difference of one volt. Therefore:
1eV “ e ¨ p1 J/Cq ñ 1eV “ 1.6 ¨ 10´19 J ,
since the electron charge is e “ ´1.6 ¨ 10´19 C. Every reaction releases an energy of:
Er “ 26.8 ¨ 106 ¨ 1.6 ¨ 10´19 J “ 4.3 ¨ 10´12 J ,
and produces two neutrinos, hence Nn “ 2Nr . The solar constant is approximately
k “ 1367 W/m2 , therefore:
2k
Nn “ « 6.4 ¨ 1014 neutrinos/(s ¨ m2 q .
Er
 175

7.5
The action of the laser beam is different in the “near zone”, where the entire beam is incident
on the base of the spaceship, and in the “far zone”, where the ship will only intercept part of
the beam. The distance at which the size of the beam is equal to the base diameter is:
D0
d0 “ ,
Θ
where D0 is the base diameter and Θ is the divergence of the laser beam, expressed in
radians. As long as the distance of the spaceship is less than d0 , the laser beam gives rise to
a constant force of:
2P
F0 “ ,
c
where P is the power emitted by the laser, and the factor of 2 appears because the base of
the spaceship is perfectly reflecting. After travelling a distance d0 , the spaceship will have
gained a kinetic energy of:
2P D0
K0 “ F0 d0 “ .
cΘ
When the spaceship moves at distances greater than d0 , only part of the laser beam will be
incident on its base. Hence, the total power reflected by the spaceship will be:
´ d ¯2
0
P pdq “ P .
d
As a result, the force decreases with distance according to:
2P ´ d0 ¯2
F pdq “ .
c d
We can find the total kinetic energy gained by the spaceship in the “far zone” either by
integrating the force from d0 to infinity, or by recognizing that this force is similar to the
gravitational force, as it is proportional to the inverse of the square of the distance, but has
opposite direction. Hence, we can associate a “radiation potential” of:

2P d20
U“ .
cd
Since, at infinity, U Ñ 0, while initially U0 “ 2P d0 {c, from the conservation of energy it
follows that the kinetic energy must increase by U0 (equal to K0 ). Hence, at an infinite
distance from the laser, the total kinetic energy of the spaceship will be:
4P D0
Kf “ 2K0 “ ,
cΘ
and its velocity: c c
2Kf 4P D0
v“ “ « 1 km/s .
m mc Θ

7.6
In Eq. 7.8, we can substitute the mass m (measured in kg) and kB , with µ (measured in
kg/mol) and R, respectively. Indeed m “ µ{Na , where Na is Avogadro constant, while the
Boltzmann constant is numerically equal to kB “ R{Na . Therefore, the root mean square
velocity can be written as: c
3RT
vrms “ (14.3)
µ
176 Chapter 14. Exercise Solutions

Substituting µH2 “ 2.016 g/mol, µO2 “ 32.000 g/mol e µN2 “ 28.013 g/mol:
d
3RT
vrms,H2 “ “ 1884.4 m/s ,
µH 2
d
3RT
vrms,O2 “ “ 473.0 m/s ,
µO2
d
3RT
vrms,N2 “ “ 505.5 m/s .
µN2

a
At sea level, the escape velocity from the Earth is vf “ 2GM {R “ 11182 m/s. Hence, we
need to numerically evaluate the integral:
´ µ ¯3{2 ż 8 2
f “ 4π v 2 e´µv {2RT dv .
2πRT vf

Which gives:

fH2 “ 1.72 ¨ 10´24 ,

fO2 “ 7.28 ¨ 10´356 ,
fN2 “ 1.86 ¨ 10´311 .

The fractions of O2 and N2 with velocities greater than the escape velocity are extremely
small. In practice, the probability that a molecule close to the surface will escape is insigni-
ficant. This is due to the escape velocity being greater at the surface, and to the fact that
the mean free path of a molecule is very small when the gas density is high. Thus, molecules
are more likely to leave from the uppermost layers of the atmosphere. The critical layer is
defined as the height at which a molecule, moving upward, has a probability 1/e of hitting
another molecule, before escaping. The region of the atmosphere above the critical layer
is called the exosphere. Earth’s exosphere starts at a height of 500 km, where the kinetic
temperature of the gas is 1500 – 2000 K.

7.7
Similarly to the previous exercise, we need to evaluate the integral:
´ µ ¯3{2 ż 8 2
f “ 4π v 2 e´µv {2RT dv .
2πRT vf
a
Using the fact that vf “ 3vrms “ 3 3RT {µ:
´ 3?3 ¯3 ż 8 ´ 272 v 2 ´ 33 ¯ 18
v2 e dv “ 1 ´ erf ? ` ? e´27{2 “ 5.8874 ¨ 10´6 .
2v
f “ 4π f
2πvf vf 2 6π
a
In the case vf “ 5 3RT {µ, we have:
´ 5 ?3 ¯ 3 ż 8 ´ 752 v 2 ´ 53 ¯ 30
v2 e dv “ 1 ´ erf ? ` ? e´75{2 “ 3.6233 ¨ 10´16 .
2v
f “ 4π f
2πvf vf 2 6π
 177

7.8
Taking the derivative of Eq. 7.6 with respect to v, we find:
27 2
´ µ ¯3{2 ” µ ı ´ 2vf2 v
4π 2 ´ v2 ve “ 0.
2πRT RT
It follows that: c
µ 2RT
2 ´ v2 “0 ñ v“ .
RT µ

7.9
Let us consider an element of atmosphere in the shape of a parallelepiped, with infinitesimal
height dh, base surface A and mass dm “ ρphqA dh. The pressure on the upper face is P `dP ,
and that on the lower face is P . The pressure difference gives rise to a force (directed upwards)
that keeps the parallelepiped afloat, against the gravitational force exerted by the Earth:
Fg “ Fp ñ gρA dh “ A dP ,
where g can be taken as a constant. Indeed, we are assuming the temperature is constant as
well, hence our model will be restricted to relatively small heights. From the ideal gas law:
Pµ
P V “ nRT ñ ρ “ .
RT
Substituting the density in the first equation:
Pµ
dP “ ´g dh .
RT
Separating variables and integrating:
ż Ph żh
dP µ
“ ´g dh
P0 P RT 0
Ph µ
ln “ ´g h.
P0 RT
It follows that: ! µ )
Ph “ P0 exp ´g h .
RT
Near the surface of the Earth, the pressure and density of air decrease exponentially. This
model is only valid for heights up to a few kilometres, since the temperature actually decreases
by about 10K/km. A better approximation is the model of the adiabatic atmosphere.
Consider an element of air that, due to a small turbulence, is displaced upwards slightly.
Air pressure decreases with height, so the cube expands. The air is a bad heat conductor,
therefore the exchange of energy between the element and the surrounding air is small and
the process can be regarded as adiabatic. If the temperature of the air around the element
is now lower that the temperature of the element itself, the element will be denser than the
surrounding air and, consequently, will be pushed down by gravity, returning to its original
position. In the adiabatic model, therefore, the atmosphere is in a continuous dynamic
balance. In the limiting case, we can assume that the temperature of the layers above the
element is the same as the temperature the element would acquire if it rose adiabatically. In
this case, the element will just find a new equilibrium position slightly upwards. In reality,
this model is valid only in the absence of convective motion and if the air is perfectly dry.
It can be shown that convective motion takes place when the temperature gradient satisfies
´dT {dz ą g{cp µ « 0.2857 K/km, i.e. if the temperature decreases by more than 1 degree
every 3.5 km, approximately. For air, the temperature gradient is more than 30 times higher.
If the air is humid, the water inside the cube condenses as it rises. Condensation releases
heat, which heats up the cube, making it less dense than the surrounding air and therefore
able to rise even higher. This is the phenomenon that gives rise to clouds.
178 Chapter 14. Exercise Solutions

7.10
Consider a circular annulus of radius r, thickness dr and mass dm “ ρprq ¨ 2πr dr. Let P
be the pressure on the inner rim and P ` dP that on the outer rim. The force caused by
the pressure difference must equal the centripetal force due to the rotation of the spacecraft.
The density of air can be found from the ideal gas law:
Pµ
P V “ nRT ñ ρ “
RT
where µ is the molar mass of the gas. Hence:
Pµ
dm “ 2π r dr .
RT
The pressure differential is due to the centrifugal force:
Ω2 r ´ Pµ ¯ ´ Ω2 r ¯ P µΩ2
dP “ dm “ 2π r dr ¨ ñ Pc “ r dr .
A RT 2πr RT
Integrating:
ż Pr żr
dP µΩ2
“ r1 dr1
Pc P RT 0
2
Pr µΩ 2
ln
“ r
Pc 2RT
µΩ2 2
" *
ñ Pr “ Pc exp r .
2RT
Since the station provides an acceleration equal to g at the outer edge, we have Ω2 R0 “ g.
The pressure P0 on the outer edge is:
! µg )
P0 “ Pc exp R0 .
2RT

7.11
Space debris can sometimes move at high relative velocities with respect to satellites, there-
fore constituting a serious danger to orbiting bodies. Let P be the pressure inside the
spaceship and assume the pressure outside to be zero. We will also assume that the air near
the hole obeys Bernoulli’s law, whereby its velocity (v) is related to its density (ρ) and the
pressure difference (P ) by: c
1 2P
P “ ρv 2 ñ v “ .
2 ρ
The rate at which air leaves the spaceship through the hole of area A is:
c
2P
dV “ A ¨ v “ A dt .
ρ
The air density can be written as a function of the pressure using the ideal gas law:
mtot nµ Pµ
ρ“ “ “ ,
V V RT
where µ is the molar mass of the gas. Differentiating the ideal gas law, we obtain the change
in pressure following the expulsion of an infinitesimal volume of air, assuming nRT to be
instantaneously constant:
V
P V “ const ñ P dV ` V dP “ 0 ñ dV “ ´ dP .
P
 179

Substituting ρ and dV in the initial equation, we obtain:

c
dP A 2RT
“´ dt .
P V µ
Integrating: c
ż P ptq
A 2RT t
ż
dP
“´ dt .
P0 P V µ 0
Hence:
c " c *
P ptq A 2RT A 2RT
ln “´ t ñ P ptq “ P0 exp ´ t « P0 e´0.000405¨t .
P0 V µ V µ
Hence, the pressure halves after a time:
ln 2
t“ « 1710 s “ 28m 30s .
0.000405

7.12
Using the formula suggested:
ż8
xv 2 y “ v 2 f pvq dv
0
ż8
´ m ¯3{2 2
“ 4π v 4 e´mv {2kB T
dv .
2πkB T 0
The gaussian integral is: ż8 c
2 1 π
I0 “ e´αx dx “ .
0 2 α
If, instead, we have:
ż8
2
” 1 ´αx2 ı8 1
I1 “ xe´αx dx “ ´ e “ .
0 2α 0 2α
Adding another x in the integrand:
ż8
1 8 ´αx2
ż
2
” 1 2 8
ı 1 ? ´3{2
I2 “ x2 e´αx dx “ ´ xe´αx ` e dx “ πα ,
0 2α 0 2α 0 4
2
were we have used integration by parts on v 1 “ xe´αx dx and u “ x. It is easy to see that:
3 ? ´5{2
I4 “ πα .
8
Hence:
´ m ¯3{2 3 ? ´ 2k T ¯5{2
B
xv 2 y “ 4π ¨ π
2πkB T 8 m
3 ? 2kB T
“ 4 π ππ ´3{2
8 m
3kB T
“ .
m
Therefore, the root mean square speed is:
c
a 3kB T
xv 2 y “ ,
m
as we wanted to show. Another interestingaresult is that the root mean square speed for
molecules hitting the walls of a container is 4kB T {m, which is larger than the result found
above. Indeed, faster molecules hit the wallsş more often, hence the mean speed is larger. In
8
this case, the integral we need to calculate is 0 v 3 f pvq dv, where the additional v comes from
the number of collisions with the wall being proportional to the particle’s velocity. Try it!
180 Chapter 14. Exercise Solutions

8. Flux and Magnitude

8.1
According to the Stefan-Boltzmann law (Eq. 7.13):

L@ “ 4πσR@ 2 Te 4 “ 3.84 ¨ 1023 W .

The flux arriving on Earth is:

L@
F “ “ 1367 W .
πd2C
This is the solar constant k “ 1367 W, equal to the power incident on Earth per square meter.
2
The Earth only intercepts photons on an effective cross-sectional area of πRC . Hence, the
fraction of the Sun’s luminosity that arrives on Earth is:
2
πRC
f“ « 4.5 ¨ 10´10 .
4πd2C
A hypothetical sphere that completely encompasses a star is called a Dyson sphere. Some
have hypothesized that, once the energy requirements of an intelligent civilization exceed
what can be provided by their home planet’s resources alone, a Dyson sphere would be built,
in order to gather most of the energy generated by fission inside their star. To get a feeling
of the amount of energy produced by the Sun, imagine we were able to gather all of its
energy for just one second: this would be enough to meet our energy demands for the next
half-million years (assuming energy consumption stays the same).

8.2
According to the Stefan-Boltzmann law (Eq. 7.13), the power emitted by the Earth is:
4
Pout “ 4πRC σTC ,

where TC “ 300 K. Since the Earth is in thermal equilibrium, the Sun provides a thermal
power of Pout . Hence, the power incident on the surface of the Earth, per unit area, is:

Pout 4
F “ 2
“ σTC « 460 W/m2 . (14.4)
4πRC
The ratio between the flux from the Sun and the flux due to radioactive decay is therefore
460{p5 ¨ 10´2 q « 104 . If the only energy
? source were radioactive decay, the temperature of
4
the Earth would be approximately 104 “ 10 times smaller, i.e. only 30 K.

8.3
The solar constant is equal to k “ 1367 W/m2 . Approximately 40 % of the solar radiation is
reflected back into space or absorbed by clouds, whilst 15 % is absorbed by the atmosphere.
Hence, the energy arriving on the surface is 45 % of the solar constant, which corresponds to
a flux of keff “ 0.45 ¨ k “ 615 W/m2 . The distance at which a light bulb of power L “ 100 W
would have to be placed to appear equally as bright as the Sun is:
c
L L
keff “ ñ d “ “ 0.114 m .
4πd2 4πkeff

8.4
Using Eq. 8.5:
Fa
“ 10´0.4pma ´mb q “ 11070 .
Fb
 181

8.5
The ratio of the fluxes is found using Eq. 8.5:
Fb
“ 10´0.4pmb ´mt q “ 10´3 .
Ft

8.6
We know the apparent magnitude of the Sun, as seen from Earth, and the ratio of the planet-
Sun to Earth-Sun distances (i.e. the distance of the planet written in au). The apparent
magnitude of the Sun, as viewed from each planet, is found using Eq. 8.4:

de 2
mp “ me ´ 2.5 log ñ mp “ me ` 5 log dp,au .
dp 2
Substituting the numerical values, we find:

mmer “ ´26.8 ` 5 log 0.3871 “ ´28.9 ,

mven “ ´26.8 ` 5 log 0.723 “ ´27.5 ,
mmar “ ´26.8 ` 5 log 1.524 “ ´25.9 ,
mjup “ ´26.8 ` 5 log 5.203 “ ´23.2 ,
msat “ ´26.8 ` 5 log 9.539 “ ´21.9 .

8.7
Let F0 be the initial flux and F1 “ 1.1 F0 the flux after the increment. Using Eq. 8.4:
F1
∆m “ m1 ´ m0 “ ´2.5 log “ ´0.103 ,
F0
for both the Sun and Jupiter. On Earth, the increase in brightness of the Sun is observed
after light has travelled a distance equal to the Earth-Sun distance. Hence:
dC
∆t@ “ “ 499 s “ 8m 19s .
c
After reaching the Earth, light must travel to Jupiter and back. The time this takes is:
2pdE ´ dC q
∆tE “ “ 4195s “ 1h 9m 55s .
c

8.8
The intrinsic luminosity of the stars is the same, hence the ratio of their fluxes depends only
on their distances from Earth:
F2 L{p4πd2 2 q ´ d ¯2
1
“ 2 “ .
F1 L{p4πd1 q d2
Using Eq. 8.4, we find the difference in magnitudes from the flux ratio:
´ d ¯2 d1
1
∆m “ m2 ´ m1 “ ´2.5 log “ ´5 log .
d2 d2
Isolating d1 {d2 :
d1
“ 10´0.2∆m « 0.2 .
d2
182 Chapter 14. Exercise Solutions

8.9
From Eq. 8.7, writing the distance d in parsecs:
´ 8.6 ¯
M “ m ´ 5 log d ` 5 “ ´1.45 ´ 5 log ` 5 “ 1.44 .
3.262
Knowing that M@ “ 4.83, we use Eq. 8.5 to find the ratio of the luminosities:
L
“ 10´0.4pM ´M@ q “ 22.7 .
L@

8.10
Using Eq. 8.7, isolating the distance d:
m´M
`1
M ´ m “ ´5 log d ` 5 ñ d “ 10 5 “ 11.22 pc .

The parallax is given by Eq. 9.3:

1 1
πp2 “ “ “ 0.08912 .
d 11.22

8.11
Writing down Eq. 8.7 for Sirius and the Sun, respectively:

Ms ´ ms “ ´5 log ds ` 5 ,
M@ ´ m@ “ ´5 log d@ ` 5 .

The stars must appear equally as bright, therefore their apparent magnitudes must be the
same, i.e. ms “ m@ . Subtracting the two above equations and using the logarithm property
log ds ´ log d@ “ log ds {d@ , we find:
ds ds
Ms ´ M@ “ ´5 log ñ “ 10´0.2pMs ´M@ q “ 4.85 .
d@ d@
Hence, Sirius would be equally as bright as the Sun if its distance were 4.85 au.

8.12
The maximum diameter of the pupil at night is around 2r “ 8 mm. Since the apparent
magnitude of the Sun is ´26.74, and its flux on Earth is k “ 1367 W/m2 , the flux of a star
with magnitude m “ 6 would be:

F “ k ¨ 10´0.4pm´m@ q “ 1.1 ¨ 10´10 W/m2 .

The energy of a photon with wavelength λ is E “ hf “ hc{λ. Assuming that the star only
emits photons with wavelength λ, the rate of incident photons per square meter (effectively,
the flux of photons) is:
F Fλ
Ff “ “ .
E hc
Hence, the rate of incident photons on our pupils, with total area A “ 2πr2 , is:
Fλ
Ff ¨ A “ ¨ 2πr2 « 3 ¨ 104 photons/s .
hc
 183

8.13
The distance of the Moon to the Sun is approximately equal to the Sun-Earth distance
dC , therefore the flux incident on the Moon is equal to the solar constant k “ 1367 W/m2 .
2
Hence, the total incident energy on the Moon is Lin “ kπR$ , and the reflected energy is
2
Lrif “ ALin “ πkAR$ . Assuming that the Moon reflects light isotropically (like a perfect
sphere), the flux arriving on Earth is:
Lrif kA ´ R$ ¯2 1 2
F$ “ 2
“ “ kAαrad .
4πd$ 4 d$ 4

where αrad is the angular radius of the Moon as seen from Earth. The ratio of the fluxes of
the Moon and Sun are:
F$ ´0.4pm ´m@ q
“ 10 $ .
k
Hence, substituting in the first equation:
1 2 ´0.4pm ´m@ q
$ 4 ´0.4pm ´m@ q
$
Aαrad “ 10 ñ A “ 2 10 « 0.53 .
4 αrad
In fact, the Moon does not behave like a perfect sphere, but reflects more strongly in the
same direction as the incident light. Therefore, the coefficient 4 in the above formula should
be replaced by 1, approximately, in which case A “ 0.13. We have therefore found an upper
bound on the albedo of the Moon, whose measured value is A « 0.12.

8.14
The absolute magnitude of UY Scuti, if we only considering only the light reaching us, can
be obtained from Eq. 8.8:

Mapp “ m ` 5 ` 5 log πp2 “ ´3.25 .

Since only 0.5% of the light passes through the envelope surrounding the star, the true value
of the absolute magnitude is:

M “ Mapp ` 2.5 log 0.005 “ ´9 .

Hence, in the visible range, the luminosity of UY Scuti is:

L
“ 100´0.4pM ´M@ q “ 3.3 ¨ 105 .
L@
On the other hand, according to the Stefan-Boltzmann law:
L ´ R ¯2 ´ T ¯4 ´ V ¯2{3 ´ T ¯4
“ “ .
L@ R@ T@ V@ T@
Hence, the temperature of UY Scuti is:
´ V ¯1{6 ´ L ¯1{4
@
T “ T@ “ 3350 K .
V L@
Since L@ “ 3.84 ¨ 1026 W, the energy radiated by UY Scuti is:

L “ 3.3 ¨ 105 L@ “ 1.26 ¨ 1032 W .

Due to radiation, UY Scuti looses mass at a rate of:

dmr L
“ 2 “ 1.4 ¨ 1015 kg/s .
dt c
184 Chapter 14. Exercise Solutions

Hence, the total rate at which UY Scuti looses mass is:

dm dmr {dt
“ “ 3.5 ¨ 1018 kg/s .
dt 0.004
Assuming this rate stays constant, the total lifetime of UY Scuti is:
8M@
∆τ “ “ 1.5 ¨ 105 years ,
dm{dt
which is a very short time on the cosmic scale.

8.15
The absolute magnitude of Gliese 581 can be found from its apparent magnitude and parallax,
using Eq. 8.8:
Mg “ mg ` 5 log πp2 ` 5 « 9.0 .
Since M@ “ 4.83, the ratio of the intrinsic luminosities is:
Lg
“ 10´0.4pMg ´M@ q « 1{48 .
L@
Since Gliese 581g has similar characteristics to the Earth, we require their surface temper-
ature to be approximately the same. This amounts to equating the fluxes of their respective
stars, as seen from the planets:
c
Lg L@ Lg
“ ñ ag,au “ « 0.144 .
4πa2g 4πa2C L@
The period is found using Kepler’s third law:
d
a3au
Tyears “ « 36 days .
M {M@

8.16
The distance of Altair to the Earth is d “ 1{πp2 “ 5.13 pc, i.e. da {d@ « 106 . From Eq. 8.5:
Fa
“ 10´0.4pma ´m@ q .
F@
The ratio of the fluxes can be found from the distances and intrinsic luminosities of the stars:
Fa La ´ d@ ¯2
“ .
F@ L@ da
Substituting in the first equation and isolating La {L@ :
La ´ d ¯2
a
“ ¨ 10´0.4pma ´m@ q .
L@ d@
Because Altair is a main-sequence star with a mass in the range 1.4 – 2.1M@ , looking at
Tab. 7.1, its temperature its around 8000 K. According to Stefan-Boltzmann law:
La “ 4πRa2 σTa4 ,
2 4
L@ “ 4πR@ σT@ .
Dividing the two equations, substituting for La {L@ the expression previously found, and
isolating Ra {R@ :
Ra da ´ T@ ¯2
“ ¨ 10´0.2pma ´m@ q « 1.642 .
R@ d@ Ta
 185

The ratio of the densities is:

ρa Ma ´ R@ ¯3
“ « 0.384 .
ρ@ M@ Ra
From which it follows that:
3 M@
ρa « 0.384 ¨ 3
« 437 kg/m3 .
4π R@

8.17
The composite magnitude can be found using Eq. 8.9, with n “ 3:

msys “ ´2.5 logr10´0.4m1 ` 10´0.4m2 ` 10´0.4m3 s “ 3.11 .

8.18
Let m1 be the magnitude of the larger and colder star, and m2 that of the smaller and
hotter star. The composite magnitude is smallest when both stars are visible, while it is
greatest when the hotter star is completely eclipsed. Hence, it immediately follows that
mmin “ m1 “ 4.85, while m2 is found using Eq. 8.9:

mmax “ ´2.5 log r10´0.4m1 ` 10´0.4m2 s

10´0.4m2 “ 10´0.4mmax ´ 10´0.4m1
ñ m2 “ ´2.5 log r10´0.4mmax ´ 10´0.4m1 s “ 4.16 .

Therefore, the magnitudes of the two stars are m1 “ 4.85 and m2 “ 4.16. The ration of the
maximum and minimum fluxes is:
Fmax
“ 10´0.4pmmax ´mmin q “ 2.88 .
Fmax

8.19
From Eq. 8.9, we know that the composite magnitude is given by:

ms “ ´2.5 logr10´0.4m1 ` 10´0.4m2 ` ... ` 10´0.4mn s .

In this case n “ 24, equal to the number of letters in the Greek alphabet. Furthermore, the
magnitude of the n-th star is mn “ 0.10 ¨ n, hence the previous equation becomes:

ms “ ´2.5 logr10´0.4¨0.10 ` 10´0.4¨0.20 ` ... ` 10´0.4¨2.4 s

“ ´2.5 logrp10´0.04 q ` p10´0.04 q2 ` ... ` p10´0.04 q24 s
” 1 ´ 10´0.04¨25 ı
“ ´2.5 log ´0.04
´ 1 « ´2.41 .
1 ´ 10
where we have used the fact that:
1 ´ xn`1
1 ` x ` x2 ` ... ` xn “ .
1´x
186 Chapter 14. Exercise Solutions

8.20
The brightness of the image of a star is proportional to the number of photons captured,
which in turn is proportional to the product of the flux of the star and the exposure time.
Let F1 , F2 be the fluxes of the two stars and t1 , t2 the exposure times. To obtain the same
luminosity, we need F1 t1 “ F2 t2 , i.e. F1 {F2 “ t2 {t1 . Using Eq. 8.5:
F1
“ 10´0.4pm1 ´m2 q .
F2
Substituting F1 {F2 “ t2 {t1 :
t2 “ t1 ¨ 10´0.4pm1 ´m2 q “ 1000 s .
Indeed, since the ratio of the fluxes is 1{100, the exposure time for the second star must be
100 times longer, i.e. 1000 s.

8.21
Using Eq. 8.5, we obtain the ratio of the fluxes:
F$ ´0.4pm ´m@ q
“ 10 $ .
F@
From Eq. 8.12, we have:
F$ F
“ e´τ ñ τ “ ´ ln $ .
F@ F@
Substituting the ratio of the fluxes found in the first equation, using the logarithm property
ln x “ log x{ log e, we find:
0.4
τ “ pm ´ m@ q “ 12.9 .
log e $

8.22
First, we compute the maximum distance at which a star, exploding as a supernova, would
raise the temperature on Earth beyond T “ 333 K. Assuming the Earth to be a perfect
2
black body, the power emitted in space is Pout “ 4πRC T 4 , while the incident power is
2
Pin “ Ftot πRC . Equating Pout and Pin , we find that the fourth power of the temperature
is directly proportional to the incident flux. Since the temperature on Earth was originally
T0 “ 14˝ “ 287 K, it follows that:
´ T ¯4 F `k F L{4πd2 L{L@
“ “1` “1` “1` 2 , (14.5)
T0 k k L@ {4πd2C dau
where dau is the distance of the supernova, expressed in astronomical units, and L{L@ its
luminosity, expressed as a multiple of the Sun’s luminosity. Isolating dau :
d
L{L@
dmax “ « 3.5 ¨ 105 au “ 1.7 pc . (14.6)
pT {T0 q4 ´ 1
On average, the number of stars at a distance d ă 1.7 pc is:
4 3
N“ πdmax ¨ n “ 2.88 . (14.7)
3
Since the Solar System has a lifetime of t@ “ 1010 years, and there is a supernova every
t “ 30 years among the Ngal “ 1011 stars that populate our galaxy, the probability that a
supernova wipes out life on Earth during the lifetime of the Solar System is:
N t@ 2.88 1010 2.88
p“ “ 11 “ « 0.01 . (14.8)
Ngal t 10 30 300
 187

9. Cosmic Distance Ladder

9.1
From Eq. 9.3, we obtain the distance:
1
d“ “ 3.50 pc .
πp2

We then substitute F10d {Fd “ pd{10dq2 “ 1{100 in Eq. 8.4, since the intrinsic luminosity is
the same. The magnitude of the star at a distance 10 times greater is:
1
m10d ´ md “ ´2.5 log ñ m10d “ md ´ 2.5 ¨ p´2q “ 0.34 ` 5 “ 5.34 .
100
Because m10d ă 6, the star is still visible.

9.2
Eq. 9.10 gives the ratio of the intrinsic luminosities of the two galaxies:
L1 ´v
rot,1
¯4
“ .
L2 vrot,2
Using Eq. 8.5, we obtain the ratio of their fluxes from the apparent magnitudes:
F1
“ 10´0.4pm1 ´m2 q .
F2
But the flux depends on both the intrinsic luminosity and the distance:
F1 L1 ´ d2 ¯2
“ .
F2 L2 d1
Solving the last equation for d2 {d1 , substituting L1 {L2 and F1 {F2 given by the first two
equations, we find:
d2 ´v
rot,2
¯2
“ ¨ 10´0.2pm1 ´m2 q « 7 .
d1 vrot,1

9.3
Using Eq. 8.4:

FM `∆M d2 ´ ∆d ¯
∆M “ pM ` ∆M q ´ M “ ´2.5 log “ ´2.5 log 2
“ 5 log 1 ` .
FM pd ` ∆dq d
Hence:
∆d
“ 100.2¨∆M ´ 1 « 25.9 % .
d

9.4
Since spiral galaxies are approximately circular, but the image appears elliptic, it is clear that
the galactic plane does not coincide with the observation line. Let i be the angle between the
perpendicular to the plane of the galaxy and the observation line, and let r be the radius of
the galaxy. The length of its apparent semi-major axis is a “ r, while that of its semi-minor
axis is b “ r cos i, hence cos i “ b{a. The rotational velocity of the galaxy is greater than the
observed velocity by a factor of sin i, therefore:
b
vrot “ voss { sin i “ voss { 1 ´ pb{aq2 “ 366 km/s .
188 Chapter 14. Exercise Solutions

Using the Tully-Fisher relation in the H band, we find:

MH “ ´9.5 log vrot ` 2.08 « ´22.3 .
The distance of the galaxy is:
mh ´MH
MH ´ mh “ ´5 log d ` 5 ñ d “ 10 5
`1
« 109 pc “ 1000 Mpc .
Since z is small, we can use the non-relativistic formula:
cz
zc “ H0 dMpc ñ H0 “ “ 60 km s´1 Mpc´1 .
1000 Mpc

9.5
The distance of the galaxy can be obtained from Hubble’s law:
zc
d“ “ 440 Mpc .
H
Since z ! 1, we use Eq. 8.7, neglecting the decrease in brightness due to redshift:
M “ m ` 5 ´ 5 log d « ´20 .
We can now compare the luminosity of the galaxy to that of the Sun, using Eq. 8.5:
F
“ 10´0.4pM ´M@ q « 1010 .
F@
The luminosity of this galaxy is 1010 times greater than the luminosity of the Sun. Since there
is practically no gas and dust in elliptical galaxies, we can assume there is no absorption.
Furthermore, assuming that the average mass of stars in the galaxy is equal to the mass of
the Sun, and the average luminosity is the luminosity of the Sun, it follows that the mass
of all stars in the galaxy is 1010 solar masses. Taking into account dark matter, the mass
of a galaxy is around 5 times greater than the mass of visible matter (stars, in this case).
Therefore, the final estimate of the total mass will be 5 ¨ 1010 solar masses.

9.6
From Hubble’s law (Eq. 9.13), it follows that:
v “ H0 d ,
where v is the velocity of an object and d its distance from Earth. Rearranging:
1 d distance
“ “ “ time .
H0 v velocity
Hence, by calculating the reciprocal of Hubble’s constant, we can estimate the age of the
universe. Using H0 “ 68 km s´1 Mpc´1 :
1 1 1
τ « “ “
H0 68 km s´1 Mpc´1 68{p206265 ¨ 149.6 ¨ 1012 q Mpc s´1 Mpc´1
206265 ¨ 149.6 ¨ 1012
“ s “ 4.54 ¨ 1017 s
68
4.54
“ ¨ 1017 years “ 14.4 ¨ 109 years .
86400 ¨ 365.25
Close enough to the true value of τ “ 13.7 ¨ 109 years. Since H0 is known with a certain
amount of error, i.e. H0 “ p68 ˘ 5.5q km s´1 Mpc´1 , we should take this into account when
estimating τ . The relative error on H0 , equal to 5.5{68 “ 0.081, is also the relative error on
1{H0 , hence the absolute error on τ is 14.4 ¨ 109 ¨ 0.081 “ 1.2 ¨ 109 years. Therefore:
τ “ p14.4 ˘ 1.2q ¨ 109 years ,
which contains the accepted value of 13.7 ¨ 109 years.
 189

9.7
At a distance r, the universe expands at a rate of vprq “ Hr. The mass of a spherical shell
between r and r ` dr is dm “ ρ dV “ 4πρ r2 dr. Hence, its kinetic energy is:
1 1
K“ dm v 2 “ p4πρ r2 drqpHrq2 “ 2πH 2 ρ r4 dr .
2 2
The mass inside a sphere of radius r is m “ ρV “ 4{3 πρ r3 , therefore the potential energy
of this shell is:
Gm dm G´4 ¯ 16
U “´ “´ πρ r3 p4πρ r2 drq “ ´ Gπ 2 ρ2 r4 dr .
r r 3 3
The total energy is then:
´ 8 ¯
E “ K ` U “ 2πρ H 2 ´ Gπρ r4 dr .
3
Setting E “ 0, we find the corresponding density:
8 3H 2
´ Gπρc ` H 2 “ 0 ñ ρc “ « 10´26 kg/m3 .
3 8πG
This is the so-called critical density of the universe, which can also be obtained from Ein-
stein’s theory of relativity. The critical density is the average density of matter required
for the universe to just halt its expansion, but only after an infinite time. A universe with
the critical density is said to be flat. If the density of matter in the universe is higher than
the critical density (closed universe), self-gravity slows the expansion until it halts, and ul-
timately the universe re-collapses. Instead, if the density of matter is lower than the critical
density (open universe), self-gravity is insufficient to stop the expansion, and the universe
continues to expand forever (albeit at an ever decreasing rate). The critical density cor-
responds to approximately 10 hydrogen atoms per cubic metre. Its measurement is still a
matter of debate, but experimental data suggests that our universe is flat.

9.8
The Lyman series results from the transition of an electron in an excited state (n ě 2) to the
ground state (n “ 1). Therefore, the lowest wavelength is obtained from the most energetic
transition, i.e. from n “ 8 to n “ 1. This energy difference is, by definition, the ionization
energy of hydrogen, hence:
hc
λl “ “ 91.2 nm .
Eion
For the V band, λmax “ 600.5 nm and λmin “ 501.5 nm. Using z “ ∆λ{λl , we find z “ 4.50–
5.58. Using the relativistic Doppler shift (Eq. 7.20), we have vrel “ p0.937–0.953q c, i.e. vrel “
p281–286q¨103 km/s. Applying Hubble’s law, the range of distances is 4130–4200Mpc.

9.9
The redshift parameter of the galaxy is z “ pλE ´ λ@ q{λ@ . According to Wien’s law,
the wavelengths of maximum emission are related to the temperatures by λE “ k{TE and
λ@ “ k{T@ , hence λE {λ@ “ T@ {TE . Therefore, the redshift parameter is:
λE ´ λ@ T@
z“ “ ´ 1 « 0.18 .
λ@ TE
The galaxy recedes from us with a velocity of v “ c ¨ z “ 5.4 ¨ 104 km/s. By Hubble’s law,
its distance is dg “ v{H0 « 760 Mpc. The photons emitted by the galaxy have energy
E “ hc{λ@ , while those arriving on Earth have energy E “ hc{λE , hence the ratio of the
190 Chapter 14. Exercise Solutions

emitted to the observed flux is Fg {Fo “ λE {λ@ “ T@ {TE . Additionally, we know that the
galaxy appears 1000 times fainter than  Eridani, i.e. Fo {FE “ 1{1000. We then have:
Lg Fg ´ dg ¯2 Fg Fo
“ “ 2
pdg πp,E q2
LE F E dE F o FE
5780 1
“ p760 ¨ 106 ¨ 0.311q2 « 6.6 ¨ 1013 .
4900 1000
The ratio between the luminosity of  Eridani and that of the Sun can be obtained from
their apparent magnitudes and parallaxes:
FE LE ´ d@ ¯2
mE ´ m@ “ ´2.5 log “ ´2.5 log
F@ L@ dE
LE ´ d ¯2
E ´0.4pmE ´m@ q
ñ “ ¨ 10 « 0.283 .
L@ d@
Because the galaxy mainly consists of Sun-like stars, the number of stars is equal to the ratio
between the intrinsic luminosities of the galaxy and the Sun:
Lg Lg LE
N“ “ ¨ « p6.6 ¨ 1013 q ¨ 0.283 “ 1.9 ¨ 1013 .
L@ LE L@

9.10
The ratio of the fluxes during maximum and minimum brightness can be computed from the
apparent magnitudes, using Eq. 8.5:
F1
“ 10´0.4pm1 ´m2 q .
F2
Since the star is at a constant distance from the Earth, the ratio of the fluxes is equal to the
ratio of the intrinsic luminosities, given by Stefan-Boltzmann law:
F1 L1 4πR12 σT14 F1 ´ R ¯2 ´ T ¯4
1 1
“ “ 2 4
ñ “ .
F2 L2 4πR1 σT1 F2 R2 T2
According to Wien’s law, λ1 “ k{T1 and λ2 “ k{T2 , hence T1 {T2 “ λ2 {λ1 . Substituting
T1 {T2 and F1 {F2 into the last equation:
R1 ´ λ ¯2
1
“ ¨ 10´0.2pm1 ´m2 q « 0.89 .
R2 λ2
which is the desired result.
After a period T , the surface of the star has covered a distance R2 ´ R1 , at a velocity of v.
Hence, R2 ´ R1 “ vT , and substituting R1 “ 0.89R2 , we find:
R2 p1 ` 0.89q “ vT ñ R2 “ 5.76 ¨ 109 m
ñ R1 “ 0.89R2 “ 5.12 ¨ 109 m .
Let k “ 1367 W/m2 be the solar constant and m@ “ ´26.74 the apparent magnitude of the
Sun. Then:
F2
“ 10´0.4pm2 ´m@ q ñ F2 “ 6.42 ¨ 10´10 W/m2 .
k
When the radius is maximum, the temperature of the star is:
k 0.002897
T2 “ “ K “ 4463.1 K .
λ2 649.1 ¨ 10´9
Hence, its intrinsic luminosity is:
L2 “ 4πR22 σT24 “ 9.38 ¨ 1027 W .
While the flux is: c
L2 σ
F2 “ ñ d “ R2 T22 “ 34.94 pc .
4πd2 F
 191

10. Gravitation and Kepler’s Laws

10.1
From Eq. 10.2, we can write:
GM@ GMC
g@ “ 2
, gC “ 2
,
R@ RC
where R@ , RC are the radii and M@ , MC the masses of the Sun and Earth, respectively.
Dividing the two above equations:
g@ M@ ´ RC ¯2
“ “ 27.96 .
gC MC R @

10.2
Writing the mass as a function of the radius and density, i.e. M “ 4{3 πr3 ρ, from Eq. 10.2:
4πG
g“ ρr.
3
Using the above equation to write gSW3 and gE , dividing one by the other:
gSW3 ´ρ
SW3
¯´ r
SW3
¯
“ .
gE ρE rE
Substituting ρSW3 {ρE “ 4 and rSW3 {rE “ 1{4, we finally get:
gSW3
“ 1.
gE
Hence, the gravitational acceleration is the same as Jupiter’s.

10.3
Since ρ “ 3m{p4πr3 q, the ratio of the densities can be written as:
ρC M C ´ R @ ¯3
“ .
ρ@ M@ R C
We now need to express all the terms on the RHS as a function of the four known parameters
3
θ@ , g, Tsid , s. Let us start with R@ {M@ . If θ@ is expressed in radians, then:
2R@ 2R@
θ@ « ñ dC « ,
dC θ@
where dC is the Earth-Sun distance. Substituting dC in Kepler’s third law:
8R@ 3 R@ 3 GT 2 θ@ 3
T 2 “ 4π 2 3
ñ “ .
GM@ θ@ M@ 32π 2
From the definition of s, it follows that:
180
360 ¨ s “ 2πRC ñ RC “ s.
π
Using 10.2 to express g:
GMC MC g
g“ ñ “ .
RC 2 RC 2 G
Hence:
MC MC 1 πg
“ ¨ “ .
RC 3 RC 2 RC 180 s G
Substituting everything back into the first equation:
ρC GT 2 θ@ 3 πg T 2 θ@ 3 g
“ 2
¨ “ « 3.91 .
ρ@ 32π 180 s G 5760 s π
192 Chapter 14. Exercise Solutions

10.4
The velocity in a circular orbit is given by Eq. 10.18:
c c
GM GM
vc “ “ “ 9688.4 m/s .
r RC ` H
The orbital period is given by Kepler’s third law:
c
pRC ` Hq3
T “ 2π “ 5745.4 s “ 1h 35m 45.4s .
GM

10.5
The orbital period of a geostationary satellite is equal to the period of rotation of the Earth
about its axis (Tsid “ 86164 s), hence by Kepler’s third law:

2 4π 2 ´ GM ¯1{3
Tsid “ ags 3 ñ ags “ T2 « 42150 km .
GM 4π 2
Therefore, the distance from the Earth’s surface is 35770 km. The velocity is:
d c
GM 2πGM
v“ “ 3 « 3.07 ¨ 103 m/s
ags Tsid

Geostationary satellites are often used for telecommunications, since an antenna pointed at
a fixed point in the sky is sufficient to exchange data with a satellite. However, there is a
delay of at least 35770{p3 ¨ 106 q « 0.01 seconds, so locally cables are used to transfer data.
The geostationary orbit was first proposed by Herman Potočnik in 1928 and popularized by
science fiction writer Arthur C. Clarke. In theory, to cover the entire equator, three satellites
in coplanar geostationary orbits, distant 120˝ from each other, would be sufficient.

10.6
On Earth, the sky appears to move from east to west with the same angular velocity as the
rotation of the Earth about its axis. Therefore, a satellite moves in the same direction as
the apparent rotation of the sky if its orbital period is greater than the period of rotation of
the Earth. For a geostationary satellite, the orbital period is exactly equal to the period of
rotation of the Earth, and since the angular velocity decreases with the square root of the
distance, we require the orbital radius to be greater than the radius of the geostationary orbit.
Thus, the minimum orbital radius is a “ ags , where ags is the radius of the geostationary orbit
(see previous exercise). We conclude that the minimum height from the ground is:

hmin “ ag,s ´ RC « 35770 km .

10.7
Since a is already expressed in astronomical units, we convert T in years and apply Eq. 10.14:
2
2 M 3 Tyr
Tyr “ aau ñ M “ M@ 3 “ 2.14 ¨ 1030 kg .
M@ aau
 193

10.8
The sum of the distances at perihelion and aphelion is equal to two times the semi-major axis:
dp ` da
a“ “ 16 au .
2
From Kepler’s third law:
2
?
Tyr “ a3au ñ T “ 212 years “ 64 years .

To calculate the area swept by the radius vector every year, we divide the area of the ellipse
by the orbital period (measured in years), since, according to Kepler’s second law, the area
swept by the radius vector per unit time is constant. The area of the ellipse is A “ π a b. To
find b, we first calculate c:
da ´ dp
c“ “ 15.5 au .
2
Which gives: ?
b “ a2 ´ c2 “ 3.97 au .
Hence, the area swept by the radius vector per year is:
dA πa b
“ “ 3.1 au2 {year “ 6.9 ¨ 1016 km2 {year .
dt Tyr

10.9
The energy is given by Eq. 10.26:
GmM
E“´ .
2a
The Sun-Earth distances at aphelion and perihelion are da “ ap1 ` eq and dp “ ap1 ´ eq,
respectively. Denoting with va and vp the velocities at aphelion and perihelion, applying
conservation of energy:
c c
1 2 GmM GmM GM 1 ´ e
mva ´ “´ ñ va “ ,
2 ap1 ` eq 2a a 1`e
c c
1 2 GmM GmM GM 1 ` e
mvp ´ “´ ñ va “ .
2 ap1 ´ eq 2a a 1´e
To check the correctness of our result, let us verify that the angular momentum is indeed
conserved. Using Eq. 10.12, we require that:

dp ¨ vp “ da ¨ va .
a
Substituting va and vp obtained previously, and dividing by GM {a:
c c
1`e 1´e
dp “ da ñ dp p1 ` eq “ da p1 ´ eq .
1´e 1`e
But dp “ ap1 ´ eq and da “ ap1 ` eq, hence the last equation is an identity.

10.10
The idea of this problem is to think of the trajectory of a body in free fall as the borderline
case of an elliptical orbit that tends to a straight line (when the eccentricity goes to 1). As-
suming that g is constant and using the equations of uniformly accelerated motion is incorrect
because, at the distance of r “ 105 km, g{g0 “ 1{p1 ` h{RC q2 « 0.65, so the gravitational
194 Chapter 14. Exercise Solutions

acceleration varies considerably. At the time of observation, the asteroids are at rest, so
they are located at the apogee of their orbits. At the moment of impact, neglecting Earth’s
radius, the asteroids will be in the focus of the orbit (which coincides with the perigee),
therefore they will have travelled exactly half of the orbit. The major axis is equal to the
initial distance from the centre of the Earth (d ` RC ). Hence, by Kepler’s third law:
c
rpd1 ` RC q{2s3
T1 “ π “ 1777047 s ,
GM
c
rpd2 ` RC q{2s3
T2 “ π “ 1776772 s .
GM
The time difference is ∆t “ 264 seconds. The distance along the Earth is then:

d “ v ¨ ∆t “ ωRC ∆t “ 123 km .

We have assumed that the radius of the Earth is small compared to the initial distance.
Without this assumption, we need to solve the equations of motion numerically, and we
would obtain T1 “ 1776550 s and T2 “ 1776290 s, giving ∆t “ 260 seconds. So, even though
the time intervals are both smaller by about 500 seconds, their difference does not change
substantially. The approximate solution is therefore correct up to 2 km.

10.11
In vacuum, electromagnetic radiation travels at the speed of light c, hence the distance of
perihelion and aphelion are, respectively:

dp “ p2 ¨ 60 sq ¨ c « 3.60 ¨ 1010 m “ 0.24 au ,

da “ p39 ¨ 60 sq ¨ c « 70.1 ¨ 1010 m “ 4.79 au .

The semi-major axis is then:
da ` dp
as “ “ 2.52 au .
2
The focal distance is equal to the difference between the distances at aphelion and perihelion:
da ´ dp
cs “ “ 2.28 au .
2
Hence, the eccentricity is:
cs
es ““ 0.9 .
as
The period of Seneca is given by Kepler’s third law:
b
Ts,yr “ a3s,au “ 4 years .

10.12
In a circular orbit, the centrifugal and gravitational forces are always equal, ensuring the
satellite does not accelerate in the radial direction, so that its orbital radius stays constant.
In an elliptical orbit, however, the two forces are different in general, and this causes the
body to accelerate radially, moving towards or away from the centre. Let us calculate the
radial acceleration of the body at perigee. The gravitational force is:
GmM
Fg “ ,
r2
 195

where r “ dp ` RC . The centrifugal force is:

vp2 ´ 6 ¯2 GmM
Fc “ m “ 1` ,
r 1000 r2
where we have used the fact that the velocity at perigeeais 0.6 % greater than the velocity of
a circular orbit at the same height, which is just v “ GM {r. Expanding the square and
neglecting the factor p6{1000q2 , we obtain the total force in the radial direction:
12 GmM 12 GM
Ftot “ Fc ´ Fg « ñ atot “ « 0.103 m/s2 .
1000 r2 1000 pdp ` RC q2
The satellite undergoes approximately uniformly accelerated motion, starting with zero ve-
locity at perigee, covering a distance of s “ d ´ dp “ 70 km in the radial direction. The
desired time is thus:
1 a
s “ atot t2 ñ t “ 2s{gtot « 1200 s “ 20 minutes .
2

10.13

Figure 14.7: The length of the latus rectum can be derived by applying the Pythagorean
theorem to the triangle F1 LF2 .

Solution 1. First, we find the length of the latus rectum. Looking at Fig. 14.7, let LF2 “ x.
Since F1 L ` F2 L “ 2a, it follows that F1 L “ 2a ´ F2 L. We also know that F1 F2 “ 2c “ 2ea
and OF2 “ c “ ea. Applying the Pythagorean theorem to the triangle F1 LF2 .
2 2 2
LF2 ` F1 F2 “ F1 L
x2 ` 4e2 a2 “ p2a ´ xq2
4e2 a2 “ 4a2 ´ 4ax
ñ x “ ap1 ´ e2 q .

The angle θc (in Fig. 14.7) is given by:

LF2 x 1 ´ e2
tan θc “ “ “ .
OF2 c e
Using Eq. A.7, the area of the sector of the ellipse between ´θc and θc is:
´a ¯ ?
A “ 2Apθc q “ a b arctan tan θc “ a b arctan p 1 ´ e2 {eq “ a b arccos e ,
b
196 Chapter 14. Exercise Solutions

? ?
where we have used the fact that b “ a 1 ´ e2 and arctan p 1 ´ e2 {eq “ arccos e. The total
area swept by the radius vector is then:
?
As “ A ´ 2AOLF2 “ a b arccos e ´ a2 ep1 ´ e2 q “ a bparccos e ´ e 1 ´ e2 q .

The total area of the ellipse is πab and the orbital time is T , thus Eq. 10.11 gives:
As T ?
t“ T “ parccos e ´ e 1 ´ e2 q .
Atot π
Solution 2. The length of the latus rectum can be found using Eq. 10.32, with θ “ 90˝ ,
which gives r “ ap1 ´ e2 q (as before). Using Eq. 10.33, the eccentric anomaly for the latus
rectum is:
ap1 ´ e2 q “ ap1 ´ e cos Eq ñ E “ arccos e .
The mean anomaly is then given by Eq. 10.34:
?
M “ arccos e ´ e sin parccos eq “ arccos e ´ e 1 ´ e2 .
?
where we have used sin parccos eq “ 1 ´ e2 . Since M increases linearly with time and
completes a full revolution after a period T , it follows that:
M T ?
t“ T “ parccos e ´ e 1 ´ e2 q .
π π
which is the same result obtained previously. Note that this method is generally faster.

10.14
Let af be the semi-major axis of Earth’s orbit after the catastrophe, M the initial mass of
the Sun and m the mass of the Earth. Earth’s kinetic energy immediately before and after
the catastrophe is the same, because there is no net force that changes its speed. Being
proportional to the mass of the Sun, Earth’s potential energy is instead halved after the
catastrophe. Let us examine the two cases:
Case 1, Earth in aphelion As shown in Ex. 10.9, the velocity at aphelion is obtained by
applying conservation of energy:
GM 1 ´ e
va 2 “ .
a 1`e
The sum of the kinetic and potential energies is equal to the total energy of the elliptical
orbit with semi-major axis af , and is given by Eq. 10.26:
1 GM {2m GM {2m
mva 2 ´ “´ .
2 ap1 ` eq 2af

Substituting va 2 , given by the first equation:

GM GM 1 ´ e GM
´ “ ´
4af 2a 1 ` e 2ap1 ` eq
1 1 ´ 1 1 ´ e¯
“ ´
4af 2a 1 ` e 1`e
1 2 e
“
af a1`e
1`e
ñ af “ a .
2e
 197

The period of the new orbit is given by Kepler’s third law:

4π 2 af 3
Tf 2 “ .
GM {2
While the initial period, equal to one year, was given by:
4π 2 a3
T2 “ .
GM
Dividing the two equations:
Tf,yr 2 “ 2af,au 3 ,
where the factor 2 appears because the mass of the Sun is halved after the catastrophe. Note
that af can be easily expressed in astronomical units if we divide by a the equation for af .
We then find:
Tf “ 118.8 years .

Case 2, Earth in perihelion The velocity of the Earth at perihelion is:

GM 1 ` e
vp 2 “ .
a 1´e
The energy of the Earth after the catastrophe is:
1 GM m GM m ´ 1 ` e 1 ¯
E“ mvp 2 ´ “ ´ .
2 2ap1 ´ eq 2a 1´e 1´e
From which we obtain:
GM m e
E“ .
2a 1 ´ e
Since E is greater than zero, the Earth will follow an hyperbolic orbit, and its motion will
not be periodic.

10.15
On average, a person can jump (lift their centre of mass) by 50 cm on the Earth, so their
launch velocity vl is:
1 a
mg∆h “ mvl2 ñ vl “ 2gh .
2
The escape velocity from the planet is given by Eq. 10.27:
c
2GM
vesc “ .
R
Since the density of the planet is the same as the density of the Earth:
M ´ R ¯3
“ .
MC RC
Substituting M in vesc :
c c
2GMC 2 2gR2
vesc “ R ñ vesc “ ,
RC 3 RC
2
where we have used g “ GMC {RC . Finally, setting vl “ vesc :
a
R “ hRC « 2 km .

We have retained only one significant digit, since vl was an estimate.

198 Chapter 14. Exercise Solutions

10.16
The escape velocity is given by Eq. 10.27:
c
2GM
vesc “ .
R
Assuming that, on average, the stars have the same mass as the Sun, M “ N M@ . Substi-
tuting M in the previous equation and solving for N :
2
Rvesc
N“ « 8 ¨ 104 stars .
2GM@

10.17
The Schwarzschild radius for a black hole of mass M@ is given by Eq. 10.28:
2GM@
Rs “ « 3 km .
c2
The density is:
M@ 3 ´ c2 ¯
ρm “ 3 “ “ 1.8 ¨ 1019 kg/m3 .
4{3 πRs 4πM@ 2G

10.18
By symmetry, the gravitational field of a cylinder must point in the radial direction. If we
consider a Gaussian surface with cylindrical shape whose axis coincides with that of the
planet, the only contribution to the gravitational flux is given by the lateral surface, since
the base of the cylinder is perpendicular to the gravitational field. Therefore, the total flux
of the gravitational field is:
Φ “ ´2 π l r gprq ,
where l is the height of the cylinder. From Eq. 10.4:

Φ “ ´4πGMint “ ´4π 2 G ρ R2 l .

Equating the two expressions for the flux:

2πGρR2
gprq “ .
r
In a circular orbit, the centripetal force is equal to the gravitational force:
2
c
vc,1 2πGρR2 a
2
3 GMC
m “m ñ vc,1 “ 2π G ρ R “ “ 14.72 m/s ,
r r 2 RC
3
where we have substituted ρ “ MC {p 43 πRC q and R “ RC , since the density and radius are
the same as on Earth. The angular velocity of rotation is also the same as on Earth:
vc,1 vc,1
“ ωe ñ h “ ´ RC “ 195.5 ¨ 103 km .
h ` RC ωe
The second cosmic velocity is infinite. Indeed, the gravitational field of the planet is constant,
hence it is never possible to escape from it. As you travel farther and farther away from
the planet, which is an infinitely long cylinder, imagine rescaling all distances. The infinite
cylinder would still look identical, however your distance from it would be reduced.
 199

10.19

Figure 14.8: The total gravitational field is given by the sum of the gravitational fields of
the large sphere and the small sphere, taken with negative mass.

Gravity satisfies the superposition principle, hence, in Fig. 14.8, the gravitational field at
each point in space is the sum of the gravitational field produced by the sphere of radius
a and by the sphere of radius a{4, taken with negative mass. In fact, in considering the
gravitational field of the entire sphere a, we are overestimating the total gravitational field,
since a portion of radius a{4 has been removed. Therefore, the total gravitational field can
be calculated by subtracting the contribution of the cavity or, alternatively, by adding the
contribution of the cavity, taken with negative mass. The calculation of the gravitational
field is particularly simple on the diameter AB since the two contributions point in the same
direction, and it is sufficient to subtract their modules. Let ρ be the density of the body and
let M , m be the masses of the large and small spheres, respectively. Then:
4 1
M“ π ρ a3 , m“ π ρ a3 .
3 48
Summing the contributions, for |x| ě a:

GM Gm 4πρGa3 ” 1 1 ı
gpxq “ ´ “ ´ .
x2 px ´ 5{8 aq2 3 x2 p8x ´ 5aq2
Substituting xB “ a and xA “ ´a, we find:
gB 1{9 ´ 1 169
“ “ .
gA 1{169 ´ 1 189

10.20
Since the projectile reaches high altitudes, the gravitational field is constant is not constant.
Part 1. To find the semi-major axis, we compute the total energy energy:
1 GMC m GMC m GMC m GMC m
E“ mv0 2 ´ “ ´ “´ .
2 RC 2RC RC 2RC
Comparing with Eq. 10.26, we conclude that a “ RC .
200 Chapter 14. Exercise Solutions

Part 2. We give two solutions: one uses conservation of energy and angular momentum,
the other is based on purely geometrical arguments.
Solution 1. Initially, the angle between the instantaneous velocity and the radius
vector is θ “ 30˝ . At the highest point of the orbit, the radius vector is perpendicular to the
instantaneous velocity (if it weren’t, the rocket would be ascending or descending, but then
it wouldn’t be located in the highest point of the orbit). Applying conservation of energy
and angular momentum between the launch point and the point of maximum height:
GMC m GMC m 1 ´ 2 1 ¯
´ “´ ` mv 2 ñ v 2 “ GMC ´ ,
2RC rmax 2 rmax RC
m RC v0 sin θ “ m rmax .

Squaring both sides of the last equation and substituting v 2 found in the first equation:
?
GMC ´ 3 ¯2 ´ 2 1 ¯
RC 2 “ rmax 2 ¨ GMC ´
RC 2 rmax RC
2
3 rmax
ñ RC “ 2rmax ´ ,
4 RC
which is a quadratic equation:
?
4˘ 16 ´ 12
4rmax 2 ´ 8rmax RC ` 3RC 2 ñ rmax “ RC ,
4
with solutions rmax,1 “ RC {2 (not acceptable since rmax,1 ă RC ) and rmax “ 3RC {2. The
maximum distance from the surface of the Earth is therefore hmax “ RC {2.

Figure 14.9: Projectile launched at 30˝ to the horizontal with the first cosmic velocity.

Solution 2. The sum of the distances of a point on the ellipse from the two foci is equal
to the major axis. Considering the launch point A in Fig. 14.9, we have:

AF1 ` AF2 “ 2RC ,

since a “ RC . But AF1 “ RC , hence AF2 “ AF1 “ RC . Being equidistant from the
foci, A is on the minor axis of the ellipse. The orbit of the projectile is symmetric with
 201

respect to the launch and landing points, hence the major axis passes through F1 , F2 and
is perpendicular to AB, which is the minor axis. It follows that =tAF1 “ 90˝ , =v0 At “ 30˝
and =v0 AB “ 90˝ (since AB is the minor axis), hence:

=BAF1 “ =tAF1 ´ =tAB “ =tAF1 ´ p=v0 AB ´ =v0 Atq “ 90˝ ´ p90˝ ´ 30˝ q “ 30˝ .

At the same time, =AF1 F2 “ 90˝ ´ =BAF1 “ 60˝ , therefore, by symmetry, the triangle
F1 AF2 is equilateral. The semi-focal distance is then c “ F1 F2 {2 “ RC {2. Thus, the
maximum distance from the centre of the Earth is:
1 3
rmax “ 2a ´ c “ 2RC ´ RC “ RC ,
2 2
as we found earlier.
Part 3. The angle at the centre of the Earth is α “ 2 ¨ =AF1 B “ 120˝ “ 2π{3 radians.
Hence, the range of the projectile is:
2π
s“ RC .
3

Part 4. By Kepler’s second law, the flight time is given by the ratio of the area swept by
the projectile to the total area of the elliptical orbit, multiplied by the time necessary to
travel the full orbit. To calculate the total area we need the semi-minor axis:
?
? a 3
b “ a2 ´ c2 “ RC 1 ´ 1{4 “ RC
2
?
3π
ñ Atot “ πa b “ RC 2 .
2
The area swept by the radius vector during the flight is the sum?of the area to the right of
the minor axis AB, which is just half the are of the ellipse A1 “ p 3π{4q ¨ RC 2 , and the area
of triangle F1 AB: ? ?
3 1 3
A2 “ b ¨ c “ RC ¨ RC “ RC 2 .
2 2 4
Hence, the area swept by the radius vector is:
?
3
As “ A1 ` A2 “ pπ ` 1qRC 2 .
4
The ratio of the areas is:
As π`1
R“ “ .
Atot 2π
Therefore, the flight time is:
d
RC 3
t “ R ¨ T “ pπ ` 1q « 3335.7 s “ 55m 36s .
GMC

10.21
As in the previous problem, the energy of the probe is:
1 GMC m GMC m GMC m GMC m
E“ mv0 2 ´ “ ´ “´ ,
2 RC 2RC RC 2RC
from which it follows that the semi-major axis is a “ RC . Looking at Fig. 14.10 (overleaf), it
is clear that F1 PF2 E is a square, and PE must be the minor axis of the ellipse. The solution
202 Chapter 14. Exercise Solutions

Figure 14.10: Probe launched at 45˝ to the horizontal from the north pole, landing at the
equator.

is similar to the previous problem, only with θ0 “ 45˝ . Applying conservation of energy and
angular momentum between the launching point and the point of maximum height:
1 GmM 1 GmM
mv0 2 ´ “ mv 2 ´ ,
2 RC 2 rmax
1
? RC v0 “ rmax v .
2
Squaring the second equation and substituting v 2 obtained from the first equation:
1 RC 1
´ “ 2 ´ ,
2RC 4rmax rmax
where we used v0 2 “ GM {RC . It follows that:
1 1 ´ 1 ¯
rmax 2 ´ rmax ` RC “ 0 ñ rmax “ RC 1 ˘ ? .
2RC 4 2
The only acceptable solution is for rmax ě RC , hence:
?
rmax “ RC p1 ` 1{ 2q .
?
The maximum height with respect to the surface of the Earth is hmax “ RC { 2.
To find the time of flight, we compute the total area swept by the radius vector, which is
just the sum of two contributions:
1´ 2 π ¯ R2 ´ π ¯
As “ AF1 PE ` Ahalf ellipse “ RC ` ? RC 2 “ C 1 ` ? .
2 2 2 2
From the second and third Kepler’s laws, we conclude that:
π 2
p1 `? qRC {2 4π 2 RC 3
T2 “ π
2
2 ¨
? RC
2
GMC
d
? RC 3
ñT “ 2πpπ ` 2q « 4320s “ 1h 12m .
GMC
 203

10.22
In the case of minimum speed, the asteroid initially orbits the Sun at the same distance
and in the same direction as the Earth. We can therefore assume that the initial relative
velocity between the asteroid and the Earth is zero. When the asteroid is close enough to
the Earth (at a distance smaller than the Hill sphere, about 1.5 million kilometres), it stars
accelerating towards the Earth because of their mutual gravitational attraction.
Since the radius of the Hill sphere is very large, and since the asteroid has a small but non-
zero velocity when it enters the Hill sphere, we can very well assume that the asteroid starts
at rest, from an infinite distance, and accelerates towards Earth. Then, the impact velocity
is exactly equal to the escape velocity from Earth:
c
1 GmMC 2GMC
mv 2 ´ “0 ñ v“ “ 11.2 km/s .
2 RC RC
If the asteroid hits the Earth at the equator, in the same direction as the rotation of the
Earth, the impact velocity will be approximately 0.5 km/s smaller. Hence, the minimum
velocity of an asteroid that impacts the Earth is vmin « 10.7 km/s.
In the case of the maximum velocity, the orbit of the asteroid is a parabola. In fact, since the
asteroid belongs to the Solar System, it cannot follow an hyperbolic orbit, therefore its max-
imum energy is E “ 0. The direction of the asteroid’s orbit must be opposite to the direction
of motion of the Earth around the Sun and the vertex of the parabolic orbit must coincide
with the collision point. When the asteroid is closest to the Sun, its velocity is:
c
1 GmM@ 2GM@
mv12 ´ “ 0 ñ v1 “ .
2 dC dC
a
Since the orbital velocity of the Earth is vC “ GM@ {dC , at this moment, the relative
velocity between the asteroid and the Earth is:
c c c
2GM@ GM@ ? GM@
vr “ ` “ p1 ` 2q .
dC dC dC
As before, we can assume that the asteroids starts at infinity. This time, however, the initial
velocity is vr . Applying conservation of energy:
c
1 GmMC 1 ? GM@ GMC
mv 2 ´ “ mvr2 ñ v “ p3 ` 2 2q `2 “ 72.8 km/s .
2 RC 2 dC RC
If the asteroid hits the Earth at the equator, in the opposite direction as the rotation of
the Earth, the impact velocity is 0.5 km/s greater, hence vmax « 73.3 km/s. Taking into
account the eccentricity of Earth’s orbit, we obtain an even greater velocity of impact if
the Earth is at perihelion. In this case, using dC “ ap1 ´ eq in the equation for vr , we
find vr “ 72.7 km/s. Adding in quadrature
? the contribution of the Earth’s gravitational
acceleration, we then obtain vmax “ 72.72 ` 11.22 ` 0.5 « 74 km/s.

10.23
Solution 1. The geometric solution makes use of the fact that, according to Archimedes’
theorem, the area of a parabolic segment is 4/3 of the area of the inscribed triangle. Hence,
in Fig. 14.11 (overleaf), the area of the parabolic segment, on the left of AB, is 4/3 of the
area of triangle AOB. Let p be the distance of the comet at perihelion. Since the orbit is
parabolic,
a E “ 0, and applying conservation of energy we find the velocity at perihelion
?
vp “ 2GM {p. Therefore, the angular momentum of the comet is L “ mpvp “ m 2GM p.
204 Chapter 14. Exercise Solutions

Figure 14.11: The dashed line represents the orbit of the Earth; the heavy solid line, the
parabolic orbit of the comet, with closest approach at a distance p from the Sun (S).

From Kepler’s second law, the constant rate at which the radius vector sweeps out area
is: c
dA L GM p
“ “ .
dt 2m 2
Let the x-axis be the axis of symmetry of the parabolic orbit, and let the origin coincide with
its vertex (O). The equations of the orbits of the comet and Earth are, respectively:
y 2 “ 4px , px ´ pq2 ` y 2 “ 1 .
Substituting y 2 “ 4px in the equation of the Earth’s orbit:
4px ` px ´ pq2 “ 1 ñ x2 ` 2px ` p2 ´ 1 “ 0 ,
which is a quadratic equation, with the only acceptable (|x| ă 1)asolution x “ 1 ´ p.
Substituting back into the equation of the parabola, we obtain y “ 2 pp1 ´ pq. Hence, the
?
area of the inscribed triangle is At “ yx “ 2 pp1 ´ pq3{2 , and that of the parabolic segment
?
Ap “ 8{3 pp1 ´ pq3{2 . The area swept by the radius vector is then:
8? a
As “ Ap ´ AASB “ p p1 ´ pq3{2 ´ 2 pp1 ´ pq p1 ´ 2pq
3
2a 2a
“ pp1 ´ pq r4p1 ´ pq ´ 3p1 ´ 2pqs “ pp1 ´ pq p1 ` 2pq .
3 3
The time taken to cover this area is:
a
As 2{3 pp1 ´ pq p1 ` 2pq
Tcomet “ “ a
dA{dt GM p{2
c
2 1 a 1 a
“ 2p1 ´ pq p1 ` 2pq “ 2p1 ´ pq p1 ` 2pq years ,
3 GM 3π
a
since 1{GM “ Tyr {2π, where Tyr is the duration of one year (remember we set dC “ 1).
To obtain the maximum time, we require dT {dp “ 0:
a 1 1 ` 2p
2 1´p´ ? “ 0 ñ 4p1 ´ pq “ 1 ` 2p ñ p “ 1{2 .
2 1´p
 205

Hence:
2
Tcomet, yr “ years « 77 days .
3π
Solution 2. Writing r9 “ dr{dt “ vr in Eq. 10.25, we have:
1 L2 GmM
mr9 2 `
E“ ´ .
2 2mr2 r
Since the orbit is parabolic, the energy is E “ 0. The angular momentum can be found by
considering the moment when the comet is in perihelion, at a distance p from the Sun. At
this point, r9 “ 0, hence the above equation becomes:
L2 GmM a
0“ ´ ñ L “ m 2GM p .
2mp2 p
Isolating r9 from the first equation, we find:
c ´
dr 2 GmM L2 ¯
r9 “ “ ´ .
dt m r 2mr2
The time the comet spends inside Earth’s orbit is obtained by separating variables and
integrating both sides:
ż dC ż dC
ż T {2
dr dr
c ´ ¯ “ dt ñ T “ 2 c ´ ¯,
2 GmM L2 0 2 GmM L2
m r
´ 2mr 2 m r
´ 2mr 2
p p

where dC is the radius of Earth’s orbit. But L2 “ 2GM m2 p, hence:

ż dC
c ż dC
dr 2 r
T “2 c ´ ¯ “ ? dr .
2 GmM GmM p GM p r´p
m r
´ r2
p

The above expression can be easily integrated by parts, yielding:

ż dC c
r 2 3{2 ´ 2p ¯ p
? dr “ dC 1 ` 1´ .
p r ´ p 3 d C d C

The time, as a function of p, is therefore:

? c 3 ´ c
2 2 dC 2p ¯ p
T “ 1` 1´ .
3 GM dC dC
a
Since Tyr “ 2π d3C {GM , we recover the previous solution.

10.24
Suppose we want to produce the maximum gravitational field at the origin of our coordinate
system, along the x-axis. An infinitesimal element of the planet, of volume dV and position
vector r, which makes and angle ψ with the x-axis, produces a gravitational field of:
ρ dV
dg “ G
,
r2
where ρ is the density of the planet. The component of the gravitational field along the
x-axis is dgx “ dg cos ψ, hence the specific contribution of this element is:
dgx cos ψ
“ Gρ 2 .
dV r
206 Chapter 14. Exercise Solutions

Figure 14.12: The dashed line represents the “level surfaces”; the heavy solid line, the shape
of the planet.
The contribution is the same for all parts of the planet for which the fraction cos ψ{r2 has
the same value. Such “level surfaces” (Fig. 14.12) can be described in a polar coordinate
system (with the x-axis as the polar axis) by the equation:
a
rpψq “ a cos ψ .
The larger the value of a, the farther (on average) the surface is from the centre and the
smaller its specific contribution. Imagine that the shape of a planet is described by the
above equation. Changing its surface in any way must involve moving some element from
the interior to the exterior of the planet, which
? would inevitably lower its contrition to the
gravitational field. This proves that rpψq “ a cos ψ is indeed the shape giving the greatest
gravitational field at the origin. The value of a can be related to the volume V of the planet.
Transforming the polar coordinates pr, ψq into Cartesian coordinates, we have:
x “ r cos ψ “ apcos ψq3{2 , y “ r sin ψ “ apcos ψq1{2 sin ψ .
In terms of these, the volume of the planet (as a solid of revolution) is:
ża ż0
3π 4π 3
V “ πypxq2 dx “ ´ a3 pcos ψq3{2 sin3 ψ dψ “ a .
0 2 π{2 15
Expressing a in terms of V , the equation describing the shape of the planet becomes:
c
3 15V
a
rpψq “ cos ψ .
4π
The total gravitational field at the origin is a
obtained by integrating the contribution of all
the “level shells”, with a ranging from 0 to 3 15V {p4πq. These shells have an infinitesimal
volume of dV “ dp4π{15a3 q “ 4{5πa2 da, therefore:
b
ż ż ż 3 15V c c
Gρ 4π Gρ 4
2 4 3 15V 4 3 15 ´ π ¯2{3
gx “ dgx “ dV “ πa da “ πGρ “ GM .
a2 0 a2 5 5 4π 5 4 V
2
On thea other hand, the gravitational field produced by a sphere is g “ GM {R , where
R “ 3 3V {p4πq. Therefore, the ratio of the maximum gravitational field produced by a
planet and by a spherical planet with the same mass and density is:
c c
gmax 4 3 15 ´ π ¯2{3 GM 4 3 135
“ ¯2{3 “ « 1.026 .
gsph 5 4 V 5 64
´
GM 4π{3 V

The maximum gravitational field is only 2.6% greater than that of a simple sphere.
 207

10.25

Figure 14.13: The dashed line represents the envelope; the heavy solid line, the orbit of a
piece of debris.

Since the fragments fly off with the same velocities and, at the moment of impact, they are
all at the same distance SP from the Sun, it follows that they have the same energies, hence
the major axes of their orbits have a common value of 2a. In Fig. 14.13, the heavy solid line
represents the possible orbit for a piece of debris; it passes through P and has the Sun (S)
as one of its foci. The situation has cylindrical symmetry about the line SP, therefore it is
sufficient to work in an arbitrary plane that contains it. Since, for any point on an ellipse,
the sum of the distances from the two foci is equal to the length of the major axis, it follows
that the distance of the second focus (F), as measured from P, is also fixed: it lays on a
circle (Γ1 ) with centre in P and radius r1 “ 2a ´ r.
Consider now an arbitrary point Q on the plane. If an elliptical trajectory passes through
Q, its focus must be on a circle (Γ2 ), centred in Q, with radius r2 “ 2a ´ SQ. Therefore, the
second focus (F) must belong to both Γ1 and Γ2 . Three cases are now possible:

1. if circle Γ2 intersects circle Γ1 , then point Q is on two different “eligible” ellipses, and
the two intersection points give the second foci of the two corresponding trajectories.
2. if circle Γ1 and Γ2 have no common point, then none of the possible orbits of the
debris pieces can pass through Q; hence, Q lies outside the required envelope.
3. if circles Γ1 and Γ2 are tangent (at point F1 in Fig. 14.13), only a single elliptical
trajectory passes through Q; this means that Q lies on the required envelope.

In the third case, it follows that:

PQ “ r1 ` r2 “ p2a ´ SPq ` p2a ´ SQq , (14.9)

that is:
PQ ` SQ “ 4a ´ SP “ const. (14.10)
Therefore, the points of the envelope themselves lie on an ellipse (the dashed line in Fig.
14.13), with foci at P and S. Because of the cylindrical symmetry, the three-dimensional
envelope of the orbits of the pieces of debris is a spheroid (ellipsoid of revolution), with
major axis of length 4a ´ SP, and foci at P and S.
208 Chapter 14. Exercise Solutions

11. Motion of the Planets

11.1
For a superior planet we use Eq. 11.3:
Tyr Tp
Tsyn “ “ 398.85 days .
Tp ´ Tyr
Dividing numerator and denominator by Tyr , we find:
Tp,yr 398.85
“ .
Tp,yr ´ 1 325.25
Hence:
398.85{325.25
Tp,yr “ “ 11.87 years ,
398.85{325.25 ´ 1
which is Jupiter’s orbital period

11.2
We don’t know whether this is a superior or inferior planet, hence we consider two cases.
For an inferior planet, we use Eq. 11.2:
Tyr Tp 584{325.25
“ 584 ñ Tp,yr “ “ 0.615 years .
Tyr ´ Tp 1 ` 584{325.25
For a superior planet, we use Eq. 11.3:
Tyr Tp 584{325.25
“ 584 ñ Tp,yr “ “ 2.67 years .
Tp ´ Tyr 1 ´ 584{325.25
Since there is no planet with an orbital period of 2.67 years, while the orbital period of Venus
is exactly 0.615 years, we conclude that the planet under consideration is Venus.

11.3
The orbital periods can be found using Kepler’s third law:
c c
pRC ` hi q3 pRC ` hs q3
Ti “ 2π “ 5304 s , Ts “ 2π “ 5548 s ,
GMC GMC
where i and s are the inferior and superior satellites, respectively. Using Eq. 11.2:
Ti Ts
Tsyn “ “ 120600 s “ 1d 9h 30m .
Ts ´ Ti

11.4
The maximum elongation of an inferior planet is always smaller than 90˝ . Since the elonga-
tion is greater than 90˝ in this case, it must be a superior planet.

11.5
In the evening, the Sun sets in the west. Planets that rise in the east have an angular
distance from the Sun (i.e. elongation) around 180˝ . Since the elongation of inferior planets
is always less than 90˝ , it must be a superior planet.
 209

11.6
For an inferior planet, we use Eq. 11.2, setting Tsyn “ Tp :
Tp Tyr
Tp “ ñ Tyr ´ Tp “ Tyr ñ Tp “ 0 ,
Tyr ´ Tp
which is clearly impossible. For a superior planet, Eq. 11.3 gives:
Tp Tyr
Tp “ ñ Tp ´ Tyr “ Tyr ñ Tp “ 2 Tyr .
Tp ´ Tyr
Hence, it must be a superior planet with an orbital period of two years.

11.7

Figure 14.14: Apparent motion of Mars.

The synodic period of Mars can be calculated using Eq. 11.3:

TD,yr
Tsyn,yr “ “ 2.135 years .
TD,yr ´ 1
The angle between eastern quadrature and opposition is (see Fig. 14.14):
θ “ arccos pdC {dD q “ arccos p1{dD,au q . (14.11)
Let ωr “ 2π{Tsyn be the angular velocity of Mars. Then, the time taken to arrive in eastern
quadrature starting from opposition is:
θ arccos p1{dD,au q
∆t “ “ Tsyn « 0.29 years “ 105d 22h 8m .
ωr 2π

11.8
The greatest and the smallest maximum elongations of Mercury are observed when the planet
is in perihelion and aphelion, respectively:
dp da
“ sin pmax,2 q, “ sin pmax,1 q .
dC dC
From Eq. 10.7, we have:
da ´ dp sin pmax,1 q ´ sin pmax,2 q
e“ “ “ 0.206 ,
da ` dp sin pmax,1 q ` sin pmax,2 q
where we neglected the eccentricity of Earth’s orbit, since it is 10 times smaller than Mercury’s.
210 Chapter 14. Exercise Solutions

11.9
Using Kepler’s third law (Eq. 10.14), we estimate the semi-major axis of the asteroid’s orbit:
2{3
aau “ Tyr “ 15.212{3 « 2.5 au .

From Eq. 8.5, we obtain the ratio of the maximum and minimum fluxes perceived on Earth:
Fmax
“ 10´0.4∆m “ 10 .
Fmin
The flux depends not only on the distance of the asteroid from the Sun, but also on the
distance between the observer and the asteroid. Let e be the eccentricity of the asteroid’s
orbit, R its radius and A its albedo. Then, the fluxes at perihelion and aphelion, for an
observer on Earth, are:

L@ πR2 A
Fp “ ,
4πa2 p1 ´ eq2 2πrap1 ´ eq ´ dC s2
L@ πR2 A
Fa “ ,
4πa2 p1 ` eq2 2πrap1 ` eq ´ dC s2
where we have assumed that the asteroid only reflects light on the hemisphere in the direction
of the incoming rays. A perfect sphere would reflect light isotropically, so in this case 2π
should be substituted with 4π. In reality, this coefficient depends on the surface of the
asteroid, and could be as small as π. We have also assumed that the orbit of the Earth is a
circle. This is a good assumption, since we expect the eccentricity of the orbit of the asteroid
to be much greater (as can be verified at the end). Expressing the distances in astronomical
units, taking the ratio of the above equations:
Fmax Fp ´ 1 ` e a p1 ` eq ´ 1 ¯2
au
“ “ “ 10
Fmin Fa 1 ´ e aau p1 ´ eq ´ 1
1 ` e aau p1 ` eq ´ 1 ?
ñ “ 10 .
1 ´ e aau p1 ´ eq ´ 1
Rearranging, we obtain a quadratic equation for the eccentricity:
? ? ?
e2 ap1 ´ 10q ` ep2a ´ 1qp1 ` 10q ` pa ´ 1qp1 ´ 10q “ 0 ,

which leads to e « 0.21 (an order of magnitude larger than the eccentricity of Earth’s orbit)

11.10
Because Mars is visible on the meridian when the Sun is setting, it is in eastern quadrature
and it rises at midnight. Looking at Fig. 14.14, initially the angle formed by Mars (M) with
the SO direction (as seen from S) is:

=OSM “ ´ arccos p1{dD,au q “ ´ arccos p1{1.524q “ ´49˝ .

Since the sidereal period of Mars is TD “ 1.881 years, after a year it will have covered an angle
of 360˝ {1.881 “ 191.4˝ to the west, as seen from the Sun. Hence, =OSM “ 191.4˝ ´ 49˝ “
142.4˝ . To find the angle formed with the CE direction, as seen from E, we can either apply
the sine and cosine rules to triangle SEM, or note that:
dD,au sin p=CSMq
tan p=CEMq “
dD,au cos p=CSMq ` 1
dD,au sin p180˝ ´ =OSMq
“ .
dD,au cos p180˝ ´ =OSMq ` 1
 211

Hence:
sin p180˝ ´ 142.4˝ q
=CEM “ arctan “ 22.8˝ .
cos p180˝ ´ 142.4˝ q ` 1
At the autumnal equinox, the Sun rises exactly at 6 am; therefore Mars, being further west by
22.8˝ , will rise 22.8˝ {360˝ ¨24h « 1.5h earlier, i.e. at approximately 4 : 30 am, local time.

12. Orbital Manoeuvres

12.1
Let us compute the desired
a velocity increments using Eqs. 12.2 and 12.3. In this case,
R2 {R1 “ 5.2 and v0 “ GM@ {dC “ 29.787 km/s, therefore:
„d ˆ ˙ 
1
∆v1 “ vP ´ v0 “ v0 2 1´ ´ 1 “ 8.792 km/s ,
1 ` R2 {R1
c „ d 
R1 2
∆v2 “ v0 1´ “ 5.643 km/s .
R2 1 ` R2 {R1

The transfer time can be obtained from Eq. 12.4:

c
π2
T “ pR1 ` R2 q3 “ 2.73 years .
8GM@

12.2
Summing Eqs. 12.2 and 12.3, setting α “ R2 {R1 :
„ ? 
1 2 ´ 1 ? ¯
v0 ´1` ? ´ ? ` α “ ∆vtot ,
α 1`α α
a
where v0 “ GM {R1 is the velocity of the initial circular orbit. Solving for the semi-major
axis in Eq. 12.4: c
2
3 8GM T
R1 p1 ` αq “ .
π2
Substituting R1 , thus obtained, in the first equation and simplifying:
”? ´ 1 ¯ ? c ı
2 ´ 2?2T ¯1{3
1 ` α ? ´ 1 ` 2α ´ “ ∆vtot .
α α GM π
Inserting the numerical values on the RHS:
¯ ? c
? ´ 1 2
1 ` α ? ´ 1 ` 2α ´ « 0.5 .
α α
It can be verified that α « 2. Hence:
c
1 3 8GM T 2
R1 “ « 0.01 au ,
1`α π2
from which R2 “ αR1 « 0.02 au.

12.3
It is better to fire the engine at take-off. Indeed, looking at the result derived in Pr. 12.4, it
is best to maximize the rate of fuel consumption.
212 Chapter 14. Exercise Solutions

12.4
Since the rope is in equilibrium, the centrifugal force Fc due to the rotation of the Earth is
equal to the gravitational force Fg . Let λ be the linear density of the rope. Let us consider
an infinitesimal element of rope, of length dr and mass dm “ λ dr, at a distance r from the
centre of the Earth. The gravitational force acting on this element is:
GMC λ
dFg prq “ gprq ¨ dm “ dr .
r2
The gravitational force on the rope is the sum (integral) of the contributions from all infin-
itesimal elements:
ż h`RC
GMC λdr ´ 1 1 ¯
Fg “ 2
ñ Fg “ GMC λ ´ ,
RC r RC RC ` h
where h is the total length of the rope. The centrifugal force on an infinitesimal element is:
dFc prq “ ω 2 r dm “ λω 2 r dr ,
where ω is the angular velocity of rotation of the Earth. The total centrifugal force is then:
ż RC `h
1 ´ ¯
Fc “ λω 2 r dr ñ Fc “ λω 2 2RC h ` h2 .
RC 2
Setting Fc “ Fg , we obtain:
´ h ¯2 h 2GMC
`3 ` 2 ´ 2 3 “ 0.
RC RC ω RC
2
Denoting with g “ GMC {RC the gravitational acceleration on the surface of the Earth and
solving the quadratic equation in h{RC , we find:
c
h 1” 2g ı
“ ´1` ´ 1 « 22.5 .
RC 2 ω 2 RC
The rope must be approximately 22.5 times longer than the Earth’s radius, which corresponds
to a length of 143000 km. The tension is maximum at the point where the gravitational
and centrifugal forces acting on an infinitesimal element are the same. Equating dFg and
dFc , it is clear this happens at the distance of the geostationary orbit rgs « 42000 km.
The maximum tension is then σmax {ρ « 4.8 ¨ 107 N m/kg. This is much greater than the
maximum sustainable tension of any construction material. For steel, the breaking point is
about 2.6 ¨ 105 N m/kg, for carbon it is 1.7 ¨ 106 N m/kg. One possible way to overcome this
would be using a rope with a variable cross-section.

12.5
The energetically optimal trajectory is an ellipse, with its perihelion and aphelion tangent
to the orbit of the Earth and the planet. Since the probe returned on Earth at the same
initial position, the entire flight must have taken an integer number of years. Of course, this
is impossible in the case of an inner planet, since it would take less than a year. Therefore,
we only need to consider the case of an outer planet. The time of flight is twice the value
given in Eq. 12.4:
π2
T2 “ pdC ` dp q3 . (14.12)
2GM@
a
The length of a year can be written as Tyr “ 2π d3C {GM@ . Since T “ n Tyr , were n is an
integer, it follows that:
1 dp 3{2
n “ ? p1 ` q . (14.13)
2 2 dC
 213

Writing dp,au “ dp {dC , isolating dp,au :

dp,au “ 2n2{3 ´ 1 . (14.14)
Trying every integer, we see that n “ 12 gives dp “ 9.48 au, which is exactly the distance of
Saturn from the Sun.

12.6
Writing the density as mass divided by volume:
Mp ´ Rp ¯ Mp Rp3
3
“ 2320 ` 3180 kg/m3 ñ “ 9718Rp2 ` 13320 .
4{3πRp RC Rp RC
From Eq. 10.27, the escape velocity from the planet is:
d d
2GMp ´ Rp3 ¯
vesc “ “ 2G 9718Rp2 ` 13320 . (14.15)
Rp RC
Assuming that friction with the atmosphere is negligible and that the rocket expels all its
fuel instantaneously (hence ignoring the effect of gravity, Pr. 12.4), we have:
mi ´ 1 ¯
vesc “ ∆v “ ve ln “ 4460 ¨ ln “ 14356 m/s . (14.16)
mf 1 ´ 0.96
Substituting for vesc :
d
´ Rp3 ¯
2G 9718Rp2 ` 13320 “ 14356
RC
´ R ¯2
p
´ R ¯3
p 143562
9718 ` 13320 “ 2
RC RC 2GRC
´ R ¯2 ´ R ¯3
p p
9718 ` 13320 “ 37979 ,
RC RC
which can be solved numerically, giving Rp “ 1.21RC . If the expulsion of fuel is not instant-
aneous, we need to use Eq. 15.18 (see Pr. 12.4). Assuming the burn lasts for 10 minutes, the
decrease in ∆v is approximately 0.25 km/s. This leads to Rp “ 1.19RC . The atmosphere
may reduce the ∆v by another 2 km/s, giving Rp “ 1.06RC .

13. Binary Stars

13.1
Using Eq. A.38, we compute the following Taylor expansions:
∆δ 2
cos pδ ` ∆δq “ cos δ ´ sin δ∆δ ´ cos δ ` ... ,
2
2
∆δ
sin pδ ` ∆δq “ sin δ ` cos δ∆δ ´ sin δ ` ... ,
2
2
∆α
cos ∆α “ 1 ´ ` ... .
2
In Eq. 13.1, the term in square brackets can then be written as:
´ ∆δ 2 ¯´ ∆α2 ¯ ´ ∆δ 2 ¯
cos δ cos δ ´ sin δ∆δ ´ cos δ 1´ ` sin δ sin δ ` cos δ∆δ ´ sin δ
2 2 2
∆δ 2 ∆α2 ´ 2 2 ∆δ
2¯
“1´ ´ cos δ ´ sin δ cos δ∆δ ´ cos δ .
2 2 2
214 Chapter 14. Exercise Solutions

If δ « 90˝ , the term in round brackets is an infinitesimal of second order, hence the term
containing ∆α2 can be neglected, since it is an infinitesimal of order higher than ∆δ 2 . Oth-
erwise, the terms sin δ cos δ∆δ and ´ cos2 δ ∆δ 2 {2 inside the round brackets can be neglect,
since much smaller than cos2 δ. Hence, in general:
” ∆δ 2 ∆α2 ı
d2 “ d21 ` d22 ´ 2d1 d2 1 ´ ´ cos2 δ ,
2 2
proving Eq. 13.2.

13.2
2 2
Using Eq. 9.3, we find dm “ 1{πp,m “ 25.40 pc and da “ 1{πp,a “ 25.06 pc. The differences
s
in their right ascensions and declinations are ∆α “ 78 “ 0.00567 rad and ∆δ “ 2262 “
0.0011 rad, respectively. Using Eq. 13.1:
c ” ∆δ 2 ∆α2 ı
d “ d2m ` d2a ´ 2dm da 1 ´ ´ cos2 δm « 0.351 pc « 1.14 ly
2 2
Hence, Mizar and Alcor are close enough to form a binary system, the first ever discovered.
Modern telescopes have since found that Mizar is itself a pair of binaries (Mizar Aa, Mizar
Ab, Mizar Ba, Mizar Bb), revealing what was once thought of as a single star to be four
stars orbiting each other. Alcor has been sometimes considered a fifth member of the system,
orbiting far away from the Mizar quadruplet. Recently, astronomers have discovered that
Alcor is also a binary, gravitationally bound to the Mizar system. The whole group is
therefore a sextuplet.

13.3
Using Eqs. 13.11 and 13.12, we know that, when the larger star covers the smaller entirely:

Lmin,1 R1 2 T1 4
“ .
L0 R1 T1 4 ` R2 2 T2 4
2

Instead, when the smaller star partially eclipses the larger star:

Lmin,2 pR1 2 ´ R2 2 qT1 4 ` R2 2 T2 4

“ .
L0 R1 2 T1 4 ` R2 2 T2 4
Substituting R2 “ R1 {4 and T2 “ T1 {3, we find:
Lmin,1 1296
“ ,
L0 1297
Lmin,2 1216
“ .
L0 1297
Therefore, the difference in magnitudes is:
Lmin,1 1296
∆m “ ´2.5 log “ ´2.5 log « ´0.069 .
Lmin,2 1216

13.4
The transit method consists in simultaneously monitoring the brightness of thousands of
stars, with the aim of identifying small periodic variations, which could indicate the presence
of orbiting bodies, known as extrasolar planets.
Let Rs « R@ “ 6.96 ¨ 105 km, Rr “ 0.2R@ « 1.4 ¨ 105 km, rg “ RE “ 7.0 ¨ 104 km and
rt “ RC “ 6400 km be the radii of a Sun-like star, of a red dwarf, of a gas giant and of an
 215

Earth-like planet, respectively. During a transit, the brightness of the system decreases by
a factor equal to the ratio between the cross-sectional areas of the orbiting planet and the
host star. In the case of a Sun-like star:
´ r ¯2
g
fs,g “ « 0.01 ñ 1 % ,
Rs
´ r ¯2
t
fs,t “ « 0.0000845 ñ 0.0084 % .
Rs
While, for a red dwarf:
´ r ¯2
g
fr,g “ « 0.25 ñ 25 % ,
Rn
´ r ¯2
t
fr,t “ « 0.00209 ñ 0.209 % .
Rn
The decrease in brightness due to the transit of a gas giant is about 100 times (or 5 mag-
nitudes) greater than that due to an Earth-like planet. Given the limited sensitivity of the
tools available (about 0.1 %), it is clear that gas giants are much more likely to be discovered
than Earth-like planets, which instead can only be found when orbiting a red dwarf. By
measuring the maximum variation in the brightness of the star, and estimating its radius
from the spectral class, it is possible to obtain the radius of the planet. Furthermore, using
the radial velocity method, it is possible to measure the mass of the planet, thus obtaining an
estimate of its density. The period of variation of the brightness of the star gives the time of
revolution, from which it is then possible to derive the semi-major axis of the orbit. However,
it is very rare for the orbital plane of a system to coincide with our line of sight. In fact,
if we consider an Earth-like planet orbiting around a Sun-like star, the angle of inclination
can be at most θrad ă Rs {dC « 0.00465 rad for a transit to be observable. The probability
of this happening is 2π ¨ p2θrad q{4π “ θrad , that is, only 1 in 200.

13.5
Let α “ R1 {R2 be the ratio of the radii of the larger to the smaller star and β “ T1 {T2 the
ratio of their temperatures. Using Eqs. 13.11 and 13.12, in conjunction with Eq. 8.5, we
find ∆1 and ∆2 :
Lmin, 1 α2 β 4
∆1 “ ´2.5 log “ ´2.5 log 2 ,
Lmin, 2 pα ´ 1qβ 4 ` 1
Lmin, 1 α2 β 4
∆2 “ ´2.5 log “ ´2.5 log 2 4 .
L0 α β `1
Raising both equations to exponents of 10:
α2 β 4
“ 10´0.4∆1 ,
pα2 ´ 1qβ 4 ` 1
α2 β 4
“ 10´0.4∆2 .
α2 β 4 ` 1
Solve for α2 β 4 in the second equation:
10´0.4∆2
α2 β 4 “ .
1 ´ 10´0.4∆2
Solving for β 4 in the first equation, writing it as a function of α2 β 4 :
α2 β 4
“ 10´0.4∆1
α2 β 4 ` 1 ´ β4
ñ β 4 “ 1 ´ α2 β 4 p100.4∆1 ´ 1q .
216 Chapter 14. Exercise Solutions

Hence, inserting the expression for α2 β 4 obtained earlier:

10´0.4∆2
β4 “ 1 ´ p100.4∆1 ´ 1q ñ β « 0.533 .
1 ´ 10´0.4∆2
Substituting the numerical value of β in the equation for α2 β 4 , we find α:

10´0.4∆2
α2 “ ñ α « 1.289 .
1 ´ 10´0.4p∆2 ´∆1 q
If we had not know that R1 ą R2 , T1 ă T2 we would have had to examine two cases,
according to which minimum was the deepest.

13.6

Figure 14.15: a) Initial shape of the orbits, b) smallest eccentricity for which they intersect.

The bodies must be diametrically opposite at all times, relative to their centre of mass, and
their distances must obeys the relation m dm “ 2m d2m , i.e. dm “ 2d2m . Therefore, their
orbits must have coincident apse lines and equal eccentricities. Furthermore, when one body
is at minimum distance from the centre of mass, the other must also be at minimum distance,
hence their orbits are flipped relative to the minor axis. From Fig. 14.15 b), we see that the
orbits intersect if:
1
Rp1 ` eq ` 2Rp1 ` eq “ 4R ñ e “
3

13.7
Let’s calculate the velocity of the Sun with respect to the centre of mass of the system, using
Eq. 13.6 and taking i “ 90˝ , since the orbit of the planet is on our line of sight.
„ 1{3
2πG
v1 “ m2 .
pm1 ` m2 q2 T
The frequency shift is given by the Doppler formula:
„ 1{3
v1 v1 m2 2πG
∆f “ f“ “ .
c λ λ pm1 ` m2 q2 T
In the case of the Sun-Jupiter system, we take m1 “ M@ and m2 “ ME . Since ME ! M@ ,
we can neglect the mass of Jupiter at denominator, obtaining:
„ 1{3
ME 2πG
∆f “ .
λ M@ 2 T
 217

If we observe with a telescope in the visible region of the electromagnetic spectrum, λ «

550 nm. The orbital period of Jupiter can be measured from the periodicity with which f
varies. Using T “ 11.86 years “ 3.74 ¨ 108 s:

∆f « 2.49 ¨ 107 Hz ,

which corresponds to a wavelength shift of:

λ2
∆λ “ ∆f “ 2.51 ¨ 10´5 nm .
c
Hence, the resolution needed is λ{∆λ « 2 ¨ 107 .

13.8
Optical binaries Regardless of the orientation of the system, at some point in time the
segment connecting the stars will be perpendicular to the line of sight. Since its length is
1 au, and the limiting angle is 0.12 , such a segment is visible from a distance of 10 pc. At
this distance, the apparent magnitude of each star is equal to its absolute magnitude, which
is equal to that of the Sun (M@ “ 4.83), and is more than enough for observations with an
astrograph. As a result, it will be possible to resolve all binary systems inside a sphere of
radius 10 pc. The number of such systems is:
´4 ¯
No “ n πdo 2 « 4 stars .
3
Eclipsing binaries Eclipsing systems can be registered with a photometer. Its accuracy
is 0.001 magnitudes, so we can assume that only a small partial eclipse is sufficient for
registration, i.e. the apparent disks of the stars need only touch each other slightly. In this
case, the angle between the line connecting the centres of the stars and the direction to the
observer is:
2R
γ“ “ 0.0094 rad .
dC
Such eclipses can be observed if the perpendicular to the plane of rotation of the stars lays
on a thin ring with angular width 2γ, centred on the plane perpendicular to the direction of
the observer. This ring covers an area of 2πp2γq, whilst the area of the whole sphere is 4π,
hence the probability that the stars eclipse each other is:
4πγ
pe “ “γ.
4π
In addition, the binary system must be observable with a photometer. Since it consists of
two Sun-like stars, each with absolute magnitude M “ 4.83, the total magnitude of the
system is:
M “ ´2.5 log 2 ¨ 10´0.4M “ M ´ 2.5 log 2 “ 4.83 ´ 0.75 « 4 .
The distance at which the system will have an apparent magnitude of 15 is:

M ´ m “ ´5 log de ` 5 ñ de “ 10´0.2pM ´mq`1 « 1600 pc ,

from which we obtain: ´4 ¯

Ne “ nγ πde 3 “ 1.6 ¨ 105 stars .
3
Spectroscopic binaries To determine the number of spectroscopic binaries, we find the
velocities of the stars about the centre of rotation. From Kepler’s third law (Eq. 10.20), the
orbital period is: c
d3C
T “ 2π .
2GM
218 Chapter 14. Exercise Solutions

Since each star is at a distance dC {2 from the centre of rotation, their linear velocities are:
d c
2πdC {2 dC 2GM GM
v“ “ “ .
T 2 d3C 2dC

The stars travel in opposite direction, hence their maximum relative velocity is twice as high.
If the perpendicular to the plane of rotation of the binary system is inclined by an angle i
to the direction of the observer, the observed relative speed is:

vrel “ 2v sin i .

The Doppler shift implies ∆λ{λ “ vrel {c, hence the resolution limit requires the inclination
to be greater than:
c c
2 sin i GM c 2dC
10´5 “ ñ sin i “ ¨ 10´5 “ 0.071 rad « 4˝ .
c 2dC 2 GM
As a result, the area suitable for observation covers almost the whole sphere, with the
exclusion of a patch of area 2πp2iq. Hence, the spectroscopic binary can be observed with a
probability of:
4π ´ 4πi
ps “ “ 1 ´ i “ 0.93 .
4π
The magnitude of the system must be greater than 12, hence the limiting distance is:

ds “ 10´0.2pM ´mq`1 « 400 pc .

The number of such systems is therefore:

´4 ¯
Ns “ nps πds 3 “ 2.5 ¨ 105 stars .
3
This gives an idea of how few binary systems are optical binaries. Instead, the number of
eclipsing and spectroscopic binaries are on the same order of magnitude, and are approxim-
ately 105 more frequent than optical binaries. If the stars eclipse each other, the photometer
can be used in conjunction with a spectrometer to better determine the characteristics of
the binary.
 15

Problem Solutions

1. Celestial Coordinate Systems

1.1 The seasons

The temperature of a location on the surface of the Earth is proportional to the flux incident
on it. The flux depends on both the distance to the Sun and on the inclination of the
solar rays. Let us examine which of the two effects gives rise to the greatest variation in
temperature. Let a be the semi-major axis of the Earth’s orbit and e its eccentricity. The
maximum and minimum Sun-Earth distances are:

dmax “ p1 ` eqa , dmin “ p1 ´ eqa .

Since the flux is proportional to the inverse of the distance squared, the ratio of the fluxes
at aphelion and perihelion is:
Fa ´ 1 ´ e ¯2
“ « 0.94 .
Fp 1`e
Let us now examine how the flux varies with the altitude of the Sun. If the Sun’s rays form
an angle h relative to the ground, those passing through a section of area A (perpendicular
to the direction of the rays) are distributed on Earth on a surface of area A{ sin h. The angle
h is just the altitude of the Sun. The declinations of the Sun during the summer and winter
solstices are  and ´, respectively, where  “ 23 ˝ 271 is the obliquity of the ecliptic. For an
observer in the northern hemisphere, the corresponding altitudes of the Sun are:

hs “ 90 ˝ ´ φ `  , hw “ 90 ˝ ´ φ ´  .

The flux is inversely proportional to the area on which the Sun’s rays are distributed:
Fs sin p90 ˝ ´ φ ` q
“ .
Fw sin p90 ˝ ´ φ ´ q
For intermediate latitudes (φ « 45 ˝ ), we find:
Fs
“ 2.5 .
Fw
As the Sun-Earth distance varies, the flux arriving on Earth changes at most by 6 %, while,
as the declination of the Sun changes, the flux varies by approximately 250 %, at intermediate
latitudes. Of course, the change in temperature is mitigated by the atmosphere. The Earth
is close to aphelion during the summer solstice and close to perihelion during the winter
solstice. Hence, the winter in the northern hemisphere is milder than the winter in the
southern hemisphere, while the summer is slightly warmer.

219
220 Chapter 15. Problem Solutions

1.2 Rotation of the Earth

Buildings rotate with the same angular velocity ω as the Earth. While the speed of the
Earth’s surface is ωRC cos φ, the speed of an object launched from the top of a building with
height h is ωpRC ` hq cos φ, where φ is the latitude of the place. The relative velocity is
therefore ωh cos φ and remains constant during the fall, if we neglect the curvature of the
Earth and friction with the air. Since the vertical motion
a is uniformly accelerated, h “ gt2 {2,
hence the object hitsa the ground after a time t “ 2h{g. The horizontal distance travelled
is therefore d “ ωh 2h{g cos φ. A body dropped from the Leaning Tower of Pisa (h “ 57 m,
φ “ 44˝ ) travels a horizontal distance of d « 1 cm.
The above calculation is not entirely correct, as it assumes that the horizontal speed of the
body is constant. In reality, since the object moves tangentially and the Earth is curved, the
gravitational force gradually moves away from the initial vertical, acquiring a component in
the tangential direction and decelerating the body accordingly. This correction adds a factor
of 2{3 in front of the expression previously obtained. We can show this by either considering
an inertial or an accelerating frame.
Inertial frame If the ball has moved a distance x to the right, then the westward component
of gravity is approximately gpx{RC q. To the leading order, the distance travelled horizont-
ally is x “ RC ωt cos φ, so the sideways acceleration is a “ ωgt cos φ. Integrating the last
expression, using pRC ` hqω cos φ as the initial speed, we obtain the horizontal speed:
1
vptq “ pR ` hqω cos φ ´ ωgt2 cos φ
2
Integrating again, we find the horizontal distance as a function of time:
1
xptq “ pR ` hqωt cos φ ´ ωgt3 cos φ
6
Subtracting off the distance travelled by the base of the tower:
1
∆x “ hωt cos φ ´ ωgt3 cos φ
6
a
For small deflections, we can substitute t “ 2h{g in the above equation:
d d d
2h 1 2h 2h 2 2h
∆x “ hω cos φ ´ ωg cos φ ñ ∆x “ hω cos φ
g 6 g g 3 g

Accelerating frame In general, the Coriolis force is given by Fc “ ´2m~ ω ˆ ~v . In this

case, the angle between ω ~ and ~v is π{2 ´ φ, hence the Coriolis force is directed eastward
with magnitude 2mωpgtq cos φ, where gt is the vertical speed at time t. Note that the
ball is deflected eastward, with acceleration 2ωgt, independent of which hemisphere it is in.
Integrating the last expression, we obtain the eastward speed (which is initially zero) as a
2 3
function of time vptq
a “ ωgt . Integrating again, we find a total displacement of x “ ωgt {3.
Substituting t “ 2h{g, we recover the previous result.

1.3 Foucault’s Pendulum

If the pendulum is located at the north pole, the solution is simple: for an inertial ob-
server hovering above the north pole, the pendulum’s plane of oscillation appears stationary
(because of the conservation of angular momentum), while the Earth below it rotates anti-
clockwise with a period of Tsid . Hence, for an observer at the poles, the plane of the pendulum
rotates clockwise with period T “ Tsid . On the other hand, for an observer at latitude φ, the
rotation of the Earth can be decomposed into a component pointing along the vertical line,
and another pointing along the horizon (Fig. 15.1). The component along the vertical line
 221

Figure 15.1: Components of the angular velocity of Earth’s rotation for an observer at
latitude φ.

is ωF “ ω sin φ. We are only interested in this component, since the component along the
horizon just modifies the effective gravitational acceleration experienced by the pendulum.
Hence, since ω “ 2π{Tsid , we finally get T “ Tsid { sin φ. For an observer on the equator, the
pendulum’s plane of oscillation appears stationary.

1.4 Dip of the horizon with refraction

The effective curvature of the Earth is simply diminished by the curvature of the ray: the
Earth’s curvature is 1{R, the ray’s curvature is k{R, and the difference of the curvatures
is Reff “ p1 ´ kq{R. The value of k mostly depends on the temperature gradient. It turns
out that in normal conditions, k ranges between 1{6 to 1{7 at sea level, and is smaller on
sunny days or at higher elevations. On the other hand, when there is a strong temperature
inversion, k can reach, or exceed, unity. The case k “ 1 corresponds to horizontal rays
that orbit the Earth indefinitely. Values larger than 1 correspond to ducting conditions;
if the observer is inside the duct, a pseudo-horizon appears above the astronomical one, a
remarkable phenomenon that really is observed.

2. Transformation of Coordinates

2.1 Upper and Lower Culmination

Consider a location at latitude φ and a star of declination δ.
For an inertial observer, the Earth’s axis is fixed (because of the conservation of angular mo-
mentum), and the direction of the star is also fixed, since the celestial equatorial coordinates
are independent of time. Looking at Fig. 15.2 (overleaf), we see that the altitudes of the
star at lower and upper culmination are:
hmin “ δ ´ p90˝ ´ φq “ δ ` φ ´ 90˝ ,
hmax “ δ ` p90˝ ´ φq “ 90˝ ´ φ ` δ .
where, by convention, hmax is measured from the south and hmin from the north.
Now consider the reference system of an observer on Earth. Since the direction of the Earth’s
axis remains constant, the polar star appears fixed in the sky, while all other stars circle
around it. A star with declination δ describes a circle with radius equal to the angle between
the polar star and itself, i.e. 90˝ ´ δ. The polar star forms an angle φ with the northern
horizon. Looking at Fig. 15.3 (overleaf), the altitudes of lower and upper culmination
are:
hmin “ φ ´ p90˝ ´ δq “ δ ` φ ´ 90˝ ,
hmax “ 180˝ ´ φ ´ p90˝ ´ δq “ 90˝ ´ φ ` δ .
222 Chapter 15. Problem Solutions

Figure 15.2: Inertial reference system of the Earth.

Figure 15.3: Reference system for an observer on Earth.

 223

3. Perturbation of Coordinates

3.1 Precession of the Equinoxes

Let us first calculate the contribution of the Sun. The origin of our coordinate system is the
centre of the Earth. Let R ~ be the position vector of the Sun, and ~r the position vector of an
infinitesimal element of the Earth, with mass dm. The gravitational force between the Sun
and the element is:
dF~ “ GM pR ~ ´ ~rq dm .
~ ´ ~r|3
|R
The torque about the centre of the Earth is then:

GM ~ ¯
~ ˆ ~rq dm « GM 1 ` 3 R ¨ ~r pR
´
~ “
d~τ “ ~r ˆ dF pR ~ ˆ ~rq dm ,
~ ´ ~r|3
|R R 3 R2

where we have used ~r ˆ ~r “ 0 and p1 ` xqn « 1 ` nx for x ! 1, with x “ |~r|{|R| ~ and

n “ 3. The total torque is obtained by integrating the last equation over the whole ellipsoid.
Separating the vector relation into components, we first compute τx :
ż
GM ´ rx R x ` ry R y ` rz R z ¯
τx “ 3
1`3 pry Rz ´ rz Ry q dm .
R R2
ş ş
When solving the integral, we see that all factors of the form rx dm and rx ry dm are equal
to zero. Indeed, by symmetry, there are as many elements with rx ě 0 as there are with
rx ď 0. Hence, summing (integrating) over the whole ellipsoid, every element simplifies with
ş yz plane, giving a net zero contribution. The only non-zero factors
its symmetric about the
are those of the form rx2 dm:
3GM ”ż ż ı
2 2
τx “ R y R z r y dm ´ r z dm .
R5
ş ş
The integrals ry2 dm, rz2 dm are the moments of inertia about the principal axes y and z, hence:
3GM
τx “ Ry Rz rIyy ´ Izz s .
R5
We then obtain τy and τz by cyclically permuting x, y, z:
3GM 3GM
τy “ Rz Rx rIzz ´ Ixx s , τz “ Rx Ry rIxx ´ Iyy s .
R5 R5
~ as a function of its ecliptic longitude. Looking
It is possible to write the position of the Sun R
at Fig. 15.4 (overleaf), we see that the components of R ~ are:
$
&Rx “ R sin λ cos 
’
Ry “ R cos λ
’
%
Rz “ R sin λ sin  .
The torque in the y direction is then:
3GM
τy “ sin2 λ sin  cos rIzz ´ Ixx s .
R3
Denoting with xτy y the time average of τy and with xsin2 λy the time average of the square
of the sine of the ecliptic longitude, we have:
3GM 3GM
xτy y “ sin  cos rIzz ´ Ixx sxsin2 λy ñ xτy y “ sin  cos rIzz ´ Ixx s ,
R3 2R3
224 Chapter 15. Problem Solutions

~ can be written as a function of its ecliptic longitude

Figure 15.4: The position of the Sun R
λ, which can be assumed to increase uniformly throughout a year.

where we have assumed that λ increases linearly with time, i.e. that the Earth’s orbit can
be approximated as circular, so that:

1 T
ż ´ 2π ¯ 1
xsin2 λy “ sin2 t dt “ .
T 0 T 2
During a year, the torque is maximum at the solstices and minimum at the equinoxes. From
the equation for xτy y, we see that the average torque throughout a year is equal to the average
of the torque at a solstice and at an equinox. We can also show that τx , τz are proportional
to sin λ cos λ, but since xsin λ cos λy “ 0, we have xτx y “ xτz y “ 0. From the formula for xτy y,
we see that the torque is proportional to M {R3 . Therefore, the ratio of the contribution of
the Moon and the contribution of the Sun is:
3
M$ R @ M ´ R @ ¯3
3
¨ “ $ « 2.17 .
R$ M@ M@ R$

In fact, this number is slightly different, since the orbit of the Moon is inclined by i “ 5˝ 91 «
0.09 radians with respect to the ecliptic. Therefore, the position of the Moon ranges from
 ` i to  ´ i, and it can be shown that this leads to a correction factor of 1 ´ 0.092 « 0.992.
Hence, the combined effect of the Sun and the Moon is equal to 3.15 the effect of the Sun
alone. Setting Izz “ Ip and Ixx “ Ie , the total torque acting on the Earth is:
3GM@
τtot “ 3.15 ¨ 3
sin  cos  rIp ´ Ie s .
2R@
As a result of this torque, the angular momentum vector of the Earth describes a double
cone, in which only the component of the angular momentum perpendicular to the ecliptic
changes with time. The value of this component is LK “ Ip ω sin , hence we have:
~
dL ~ ñ xτtot y “ 2π LK .
“ Ω~pr ˆ L
dt Tpr
Isolating Tpr :
3
1 4πωR@ Ip
Tpr “ ¨ 3
« 25300 years .
3.15 GM@ cos  Ip ´ Ie
This is the precession period of the Earth’s axis, with an error of about 2 %.
The value of Ip {pIp ´Ie q can be estimated from the eccentricity of the Earth. It is well known
 225

that the moments of inertia of an ellipsoid of revolution along the polar and equatorial axes are:
2 1
Ip “ M a2 , Ie “ M pa2 ` b2 q .
5 5
Hence:
Ip ´ Ie a2 ´ b2
“ .
Ie 2a2
Since the equatorial and polar radii are a “ RC “ 6378 km and b “ Rp “ 6357 km, respect-
ively; we conclude that pIp ´ Ie q{Ie « 0.00335 « 1{300.

4. Observation and Instruments

4.1 Combination of thin lenses

The image formed by the first lens is located at the same position as if the second lens were
not present. It serves as the object for the second lens, whose image is the final image of
the system. Let p be the distance of the object from the first lens and let f1 , f2 be the focal
lengths of the first and second lenses. Then, according to Eq. 4.6:
1 1 1
` “ ,
p q1 f1
where q1 is the distance of the image from the first lens. Treating the image formed by the
first lens as the object for the second lens, we see that the object distance for the second lens
must be p2 “ ´q1 (the lenses are in contact and assumed to be infinitesimally thin). Let q
be the distance of the final image from the second lens. We then have:
1 1 1
´ ` “ .
q1 q f2
Summing the two equations, q1 simplifies, and we are left with:
1 1 1 1 1 f1 f2
` “ ` “ ñ feff “ .
p q f1 f2 feff f1 ` f2

4.2 Lens maker’s equation

Assuming all angles are small in Fig. 15.5 (overleaf), we see that α, β, γ are given by:
d d d
α“ , β“ , γ“ . (15.1)
p R1 q1
The angles of incidence and refraction are θi “ α`β and θr “ β ´ γ; hence, by Snell’s law:
n1 θ1 “ n2 θ2
n1 pα ` βq “ n2 pβ ´ γq
´1 1 ¯ ´ 1 1¯
n1 d ` “ n2 d ´ .
p R1 R1 q1
Therefore, we find:
n1 n2 n2 ´ n1
` “ .
p q1 R1
Now, suppose the back of the refractive material has a radius of curvature R2 , and that the
overall lens is infinitesimally thin. As in Pr. 4.1, the image formed by R1 acts as the object
for R2 , so p2 “ ´q1 . Let q be the distance of the final image from R2 , then:
n2 n1 n1 ´ n2
´ ` “ .
q1 q R2
226 Chapter 15. Problem Solutions

Figure 15.5: Refraction of light rays originating from a point P at a distance p from a
spherical surface with radius R1 . The image forms at a distance q1 .

Summing the two above equations, we obtain:

n1 n1 ´ 1 1 ¯
` “ pn2 ´ n1 q ´ .
p q R2 R1
Setting n1 “ 1 (i.e. the index of refraction of air) and n2 “ n, we recover Eq. 4.5:
1 1 ´ 1 1 ¯
` “ pn ´ 1q ´ .
p q R2 R1

4.3 Airy Disk

The source (S) produces spherical waves with a strength aS , and is at a distance s from the
aperture element at px, yq. The wave arriving at the element is therefore:

as eiks
ψ1 prq “ .
s
The aperture, denoted by Σ, can change the amplitude or phase of the incident radiation,
and its transmission properties can be described by the aperture function, hpx, yq. Usually
h “ 0 or 1 for obstructing or open areas respectively. By Huygens’ principle, the aperture
element can now be considered to act as a source of secondary, spherical, wavelets with a
strength and phase given by:
aap 9 ψ1 hpx, yqdx dy .
Next, we need to calculate the amplitude of the secondary wavelet reaching the observation
point (P), which is a distance r away from the aperture element. This amplitude is given by:

eikr
dψp 9 aap .
r
In reality, Huygens’ principle is not exact, since the simple idea of spherical wavelets would
lead to a backward-propagating wave-front as well as a forward-propagating one, which is not
observed experimentally. To fix Huygens’ principle, we should multiply the previous equation
by the obliquity factor Kpθq, which describes the fall-off in intensity of the wavelets with
angle θ away from the forward direction. However, assuming angles are small (Fraunhofer
regime), Kpθq « 1. To calculate the total amplitude at point P, we finally sum over all
aperture elements:
as eiks eikr
ż ż
ψp 9 hpx, yq dx dy .
Σ s r
 227

Consider the diffraction pattern in a plane at a distance L from the aperture (Fig. 4.9). We
denote the coordinates of a point P in this plane by px0 , y0 q. We will assume that the source S
is a large distance behind the aperture (and centred on the aperture) so that s Ñ 8 and the
aperture is illuminated with a plane wave at normal incidence. Using the coordinate system
shown in Fig. 4.9, we find the distance r from the aperture element at px, yq to point P:

r2 “ L2 ` px ´ x0 q2 ` py ´ y0 q2
“ L2 ` x20 ` y02 ´ 2pxx0 ` yy0 q ` x2 ` y 2
´ xx0 ` yy0 x2 ` y 2 ¯
“ R2 1 ´ 2 ` ,
R2 R2
where R2 “ L2 ` x20 ` y02 . Using the binomial expansion, we obtain:
xx0 ` yy0
r «R´ ,
R
where we have assumed that the last term can be neglected, i.e. that:

x2 ` y 2
k ! π,
2R
The wave amplitude at P is then given by the integral:
ż ż ! xx0 ` yy0 )
ψp 9 hpx, yq exp ´ik dx dy .
Σ R
Changing to a polar coordinate system, we substitute dx dy with ρ dρ dφ and x, y with ρ cos φ
and ρ sin φ, respectively:
ż D{2 ż 2π " *
x0 cos φ ` y0 sin φ
ψp 9 ρ exp ´ikρ dφ dρ .
0 0 R
Since we expect the diffraction patter to be circular symmetric, it suffices to study the
intensity along a given direction, say x. Then, setting y0 “ 0 in the previous equation, and
substituting x0 “ R sin ξ, we have:
ż D{2 ż 2π
ψp 9 ρe´ikρ sin ξ cos φ dφ dρ
0 0
ż D{2
“ 2π ρJ0 pkρ sin ξq dρ
0
ż kD{2 sin ξ
2π
“ kρJ0 pkρ sin ξq dpkρ sin ξq
k2 sin2 ξ 0
2π Dk sin ξ ´ kD sin ξ ¯
“ 2 2 J1 .
k sin ξ 2 2
Which can be written as:
J1 pπD sin ξ{λq
ψp 9 ,
πD sin ξ{λ
where we have substituted k “ 2π{λ. We know that the first root of J1 pxq is at x “ 3.8317.
Hence, the first minimum in the intensity is at:
πD sin ξ 3.8317 λ
“ 3.8317 ñ ξ “
λ π D
λ
ξ « 1.2197 .
D
228 Chapter 15. Problem Solutions

As we wanted to show. Since ψp must be circular symmetric, its general form is:
a
J1 pπD sin2 ξ ` sin2 θ{λq
ψp 9 a .
πD sin2 ξ ` sin2 θ{λ

To calculate the diffraction pattern of an elliptical aperture, let us define x1 , y 1 , such that
x1 “ ρ cos φ, y 1 “ pb{aq ρ sin φ and x12 ` y 12 “ a2 {ρ2 is a circle of radius a{ρ. The amplitude
at point P“ px0 , y0 q on the diffraction screen is:
ż ż ! xx0 ` yy0 )
ψp px0 , y0 q 9 hpx, yq exp ´ik dx dy
Σ R
# +
x0 cos φ ` ab y0 sin φ b
ż a ż 2π
“ exp ´ikρ ρ dφ dρ
0 0 R a
# +
x0 cos φ ` ab y0 sin φ
ż a ż 2π
9 ρ exp ´ikρ dφ dρ
0 0 R

We can then write:

ż a ż 2π " *
´ a ¯ x0 cos φ ` y0 sin φ
ψp x 0 , y 0 9 ρ exp ´ikρ dφ dρ
b 0 0 R
a
J1 pπa x20 ` y02 {pλRqq
“ a ,
πa x20 ` y02 {pλRq
which follows from the circular symmetric solution found earlier. Therefore, substituting
pa{bq y0 Ñ y0 :
a
J1 pπa x20 ` pb{aq2 y02 {pλRqq
ψp px0 , y0 q 9 a
πa x20 ` pb{aq2 y02 {pλRq
a
J1 pπ a2 x20 ` b2 y02 {pλRqq
“ a .
πa a2 x20 ` b2 y02 {pλRq
Hence, we obtain: a
J1 pπ a2 sin2 ξ ` b2 sin2 θ{λq
ψp px0 , y0 q 9 a .
πa a2 sin2 ξ ` b2 sin2 θ{λ

5. Time System

5.1 Effect of ecliptic obliquity on the equation of time

The ecliptic longitude of the Sun is β@ “ 0˝ . Substituting β@ in the system of equations
2.12, we have: $
& sin α cos δ “ cos  sin λ
’
cos α cos δ “ cos λ
’
%
sin δ “ sin  sin λ .
Dividing the first equation by the second, discarding the first:
$
& tan α “ cos  tan λ
’
cos α cos δ “ cos λ
’
%
sin δ “ sin  sin λ .
 229

Differentiating the first equation with respect to time, assuming  is constant:

1 dα cos  dλ
“ .
cos2 α dt cos2 λ dt
Substituting cos α “ cos λ{ cos δ, obtained from the second equation:
dα cos  dλ
“ ,
dt cos2 δ dt
which is what we wanted to show. This formula shows that the angular velocity of the Sun
varies throughout the year because:

• dλ{dt is not constant, but is greatest at perihelion and smallest at aphelion;

• cos { cos2 δ varies from a minimum of cos  when δ “ 0˝ , to a maximum of 1{ cos 
when δ “ .

5.2 Equation of time

Both the energy and the angular momentum are conserved:
GmM 1
E“´ ` mv 2 , L “ m r vt “ m r ω 2 , (15.2)
r 2
where vt is the tangential velocity and ω “ dθ{dt is the angular velocity. The total velocity
v is obtained by solving the equation for the energy:
c ´
2 GmM ¯
v“ E` .
m r
In an elliptical orbit, the energy is:
GmM
E“´ .
2a
The angular momentum can be calculated by considering the instant when the body is in
perihelion:
d
2´ GmM ¯
L “ mrp vp “ mrp E`
m rp
d
2 ´ GmM GmM ¯
“ map1 ´ eq ´ `
m 2a ap1 ´ eq
c
? 2
“ mp1 ´ eq GM a ´1 `
1´e
? ?
“ m 1 ´ e2 GM a .
Thus, the angular velocity is:
dθ L ? ?
2 GM a
p1 ` e cos θq2
“ “ m 1 ´ e
dt mr2 a2 p1 ´ e2 q2
c
2
GM p1 ` e cos θq
“ ,
a3 p1 ´ e2 q3{2
as we wanted to show. The angular velocity of the Sun on the ecliptic is thus:
c
dλ GM p1 ` e cos θq2
“ .
dt a3 p1 ´ e2 q3{2
230 Chapter 15. Problem Solutions

From Pr. 5.1, the projection of this motion on the celestial equator is:
dα cos  dλ
“ .
dt cos2 δ dt
The declination of the Sun can be obtained from the ecliptic longitude, using Eq. 2.12c:

sin δ “ sin  sin λ .

Hence, we have: c
dα cos  GM r1 ` e cos pλ ´ γqs2
“ ,
dt 1 ´ sin2  sin2 λ a3 p1 ´ e2 q3{2
where we have substituted θ “ λ ´ γ. Since θ “ 0˝ corresponds to the direction of perihelion
and λ “ 0˝ to the direction of the vernal point, γ is the angular distance between the
perihelion and the vernal point. The Earth transits through perihelion on the 5th of January,
while the spring equinox is on the 21st of March, i.e. 76 days later. Hence γ “ p76{365.25q ¨
360˝ « 75˝ . Denoting with Ts “ 2πpdα{dtq´1 , the length of the true solar day is:
1 1 1 Tsid Ts
“ ´ ñ Tsol “ .
Tsol Tsid Ts Ts ´ Tsid
The difference between the length of a given day and the mean solar day is therefore Tsol ´
86400 s. For a given day, the equation of time is the sum of all these differences from the 1st
of January up to that day. We can then draw the analemma by calculating the altitude and
the azimuth of the true Sun at each day of the year, using the system of Eqs. 2.4.
In the following plots, each point represents the position of the Sun seen at a given day. The
equation of time and the analemma for an observer at latitude φ “ 45˝ , at time 12 : 00, is
shown in Figs. 15.6 a) and b), respectively. In Figs. 15.7 a), b) and c) we see the shape of
the analemma for an observer at latitude φ “ 45˝ , at observation times 10 : 00, 12 : 00 and
14 : 00, respectively. In Figs. 15.8 a), b) and c) we compare the shape of the analemma, at
time 10 : 00, for observers at latitudes 90˝ , 45˝ and 0˝ , respectively. Finally, in Figs. 15.9
a), b) and c), we see the equation of time if the eccentricity of Earth’s orbit were 10 times
smaller, the current value, and 10 times larger, respectively. You can find my code for this
exercise on the web page of the book; have fun experimenting with different cases!

Figure 15.6: a) Equation of time, b) analemma at latitude φ “ 45˝ and time 12:00.

Figure 15.7: Analemma at latitude 45˝ and times a) 10:00 , b) 12:00 and c) 14:00.
 231

Figure 15.8: Analemma at time 10:00, for observers at latitudes a) 90˝ , b) 45˝ and c) 0˝ .

Figure 15.9: The equation of time if the eccentricity were a) 10 times smaller, b) the current
value and c) 10 times larger.

6. The Moon

6.1 Meton cycle

Lunar cycles depend on the relative position between the Sun and the Moon. The position
of the Sun relative to the Earth varies with a periodicity of one tropical year (ttr “ 365.24d ).
The position of the Moon is the same, relative to the Earth, every synodic period (tsyn “
29.53d ). The length of a Meton cycle is therefore the smallest interval of time that contains an
integer number of tropical years and synodic months. For n, m integers, we must have:
x x
“ n, “ m.
29.53 365.24
Dividing the first equation by the second:
n
“ 12.36844 .
m
We try several values of m and check whether n is approximatively integer. Note that varying
m is easier than varying n, since m is smaller by one order of magnitude, and we therefore
need to try fewer cases. By inspection, we find m “ 19, which gives n “ 235.0003 « 235.
Hence, the Meton cycle is equal to 19 tropical years, or 235 synodic months.

6.2 Anomalistic month

Let ωa “ 2π{Ta and ωsid “ 2π{Tsid be the angular velocities of rotation of the apse line and
of the Moon, respectively, where Ta “ 8.8504 years “ 3232.665 days and Tsid “ 27.322 days.
Since both the apse line and the Moon rotate in the anti-clockwise direction, the relative
angular velocity is ωr “ ωsid ´ ωa . Then:
1 1 1 Ta Tsid
“ ´ ñ Tan “ « 27.554 days .
Tan Tsid Ta Ta ´ Tsid

The anomalistic month is 5h 35m longer than the sidereal month.

232 Chapter 15. Problem Solutions

6.3 Measuring the distance of the Moon

As already found in Ex. 6.5, the distances between an observer on the Earth and the Moon
at zenith or on the horizon are, respectively:
b
dz “ d$ ´ RC , dh “ d2$ ´ RC 2
.

The angular diameter of the Moon in these two cases is:

D$ D$
tan αz “ ñ αz,rad « ,
dz d$ ´ RC
D D$
tan αh “ $ ñ αh,rad « b .
dh 2 2
d$ ´ RC

Dividing the two equations:

´α
z,rad
¯2 pd$ ´ RC q2
« .
αh,rad d2$ ´ RC
2

Denoting with α “ αz,rad {αh,rad “ 237{233 and solving the second order equation for d$ ,
we finally obtain: a
1 ` 1 ´ p1 ´ α4 q
d$ “ R C « 375000 km ,
α2 ´ 1
which is 9000 km smaller than the true value.

6.4 Distance of the Moon, another method

Figure 15.10: The zenith distance is measured when the declination of the Moon is a)
minimum and b) maximum, respectively, from a place of latitude φ.

The apparent position of the Moon changes because of parallax. In Fig. 15.10 b), consider
triangle OPM. We see that =POM “ φ ´ δmax and =OPM “ 180˝ ´ zu , so it follows that
=PMO “ 180˝ ´pφ´δmax `180˝ ´zu q “ zu ´φ`δmax . Using the law of sines (Eq. A.13):

OP OM
“ .
sin p=PMOq sin p=OPMq

Substituting OP “ RC and OM “ d$ , using the reduction formulae (Appendix A.3):

d$ sin zu
“ .
RC sin pzu ´ φ ` δmax q
 233

Similarly, looking at Fig. 15.10 a), we obtain a second equation for d$ {RC . It is sufficient
to swap δmax with ´|δmin | and zu with zl in the previous formula:
d$ sin zl
“ .
RC sin pzl ´ φ ´ |δmin |q
Since the orbit of the Moon is symmetric with respect to the celestial equator, δmax “
|δmin | “ δ. The problem then reduces to solving the system of equations:
$
’ d$ “
’
’
sin zu
&R sin pzu ´ φ ` δq
C
’ d$ sin zl
’
’ “ .
RC sin pzl ´ φ ´ δq
%

Dividing the two equations, we find:

sin zu sin pzu ´ φ ` δq
“ .
sin zl sin pzl ´ φ ´ δq
Since δ is the only unknown, the above equation can be solved numerically, giving δ « 28.6˝ .
Finally, substituting the value of δ in one of the above equations, we find d$ {RC « 60.1.
Knowing that the obliquity of the ecliptic is  “ 23˝ 271 , we then obtain the angle between
the plane of the lunar orbit and the ecliptic: i “ δ ´  “ 5.15 “ 5˝ 91 .

6.5 Lunar influence

Let x$ and xC be the respective distances of the Moon and the Earth from the centre of
mass of the Earth-Moon system. Then:
MC
MC xC “ M$ x$ ñ x$ “ xC .
M$

The Earth-Moon distance is d$ “ xC ` x$ “ 3.844 ¨ 105 km. Hence, substituting x$ in the

equation for d$ , we find:
´ MC ¯ d$
d$ “ 1 ` xC ñ xC “ « 4670 km .
M$ 1 ` MC {M$

Suppose there is a first quarter Moon during the vernal equinox. Then, the centre of mass
of the Earth-Moon system starts at a distance xC behind the Earth. Since there are approx-
imately 12.5 synodic months in a year, at the end of the year there will be a last quarter
Moon. At that moment, the centre of mass will be at a distance xC in front of the Earth.
Therefore, the centre of mass will have travelled an additional distance of 2xC , compared to
the average. Since the orbital velocity of the centre of mass is constant, the time taken to
cover this additional distance is:
c
2xC dC
∆t “ “ 2xC « 310 s “ 5m 10s .
vC GM@
The next year, there is a last quarter Moon during the vernal equinox, hence the centre of
mass of the system starts at a distance xC in front of the Earth. At the end of the year there
will be a first quarter Moon, hence the centre of mass will now be at a distance xC behind
the Earth. Therefore, the centre of mass has travelled a distance smaller by 2xC , compared
to the average, and the duration of this year will be ∆t shorter. Since the year before was ∆t
longer than the average, the difference in length of the two years is 2∆t « 10m 20s .
234 Chapter 15. Problem Solutions

6.6 Longitudinal libration

Let us first find an estimate for the amplitude of longitudinal libration. The velocity of the
Moon at aphelion is (see Ex. 10.9):
c d
1 ` e GMC
v“ .
1´e a$

Therefore, the angular velocity at aphelion is:

d d
v 1`e GMC
ωriv “ “ .
p1 ´ eqa$ p1 ´ eq3 a3$

2
Since the Moon rotates with angular velocity ωrot “ GMC {a3$ about its axis (equal, on
average, to that of revolution), after a time t we will observe an additional angle of the
Moon: ˆd ˙
b
3 1`e
2A “ pωriv ´ ωrot qt “ GMC {a$ ´ 1 t.
p1 ´ eq3
Using the approximation for small e (since e “ 0.0549 ! 1), we find:
2π ” ı 4π
2A « p1 ` e{2qp1 ` 3e{2q ´ 1 t “ et .
Tsid Tsid
Since this is just an approximation, we can loosely chose t. To get rid of some numerical
factors, let us take t “ Tsid {π, in which case A « 2e “ 6˝ , approximately. This should be
considered as an order-of-magnitude estimation.
Another solution, which does not rely on a guess for t, starts with Eq. 5.9:
d
dθ GMC p1 ` e cos θq2
“ .
dt a3$ p1 ´ e2 q3{2

The relative velocity of the lunar surface as seen from Earth is:
d
´ dθ ¯ dθ GMC ” p1 ` e cos θq2 ı
« ´ ωrot “ 3 2 3{2
´1 .
dt rel dt a$ p1 ´ e q

The relative velocity is zero for:

p1 ` e cos θ0 q2 p1 ` eq3{4 ´ 1
2 3{2
´ 1 “ 0 ñ cos θ0 “ .
p1 ´ e q e
Using the approximation for small e, we get cos θ0 « 3{4 e. Since ωr “ dA{dt “ pdA{dθqpdθ{dtq:
żA ż θmax ”
p1 ´ e2 q3{2 ı
dA “ 1´ dθ .
0 0 p1 ` e cos θq2
For small e, the integral becomes:
ż cos´1 θ0 c ´3 ¯
A “ 2e cos θ dθ “ 2e 1 ´ e2 « 2e ,
0 4
to first order. Hence, we recover the previous result.
 235

7. Electromagnetic radiation

7.1 Radiation emitted by accelerated charge

Let C be the units of charge, T of time, L of length and M of mass. Writing down the
physical dimensions of the quantities under consideration:
C2T 2 L L
r0 s “ , rcs “ , rqs “ C , ras “ 2 .
L3 M T T
Our aim is to combine these quantities to get another, whose physical dimension is power:
M L2
rP s “ .
T3
Let α, β, γ and δ be the exponents with which 0 , c, q and a appear in the final formula for
P . The problem reduces to solving a system of four equations:
$
’
’ L : ´ 3α ` β ` δ “ 2 (15.3a)
’
’
& T : 2α ´ β ´ 2δ “ ´3 (15.3b)
’
’
’ M : ´ α “ 1 (15.3c)
’
% C : γ ` 2α “ 0 . (15.3d)
For instance, to get the first equation, we note that the length L appears with exponent ´3
in 0 (so we add ´3α); with exponent 1 in c (so we add β), and finally with exponent 1 in a
(so we add δ). The length appears in the expression for P with exponent 2, hence the sum
of the previous terms must equal 2, and we recover Eq. 15.3a.
From Eq. 15.3c, we immediately obtain α “ ´1 and, substituting the value of α in Eq.
15.3d, we find γ “ 2. Substituting the value of α and γ in Eqs. 15.3a and 15.3b, we have:
"
β ` δ “ ´1 (15.4a)
´ β ´ 2δ “ ´1 . (15.4b)
Summing Eqs. 15.4a and 15.4b, we find ´δ “ ´2, i.e. δ “ 2. Substituting the value of δ in
any of the two above equations, we find β “ ´3. We know that the numerical constant is
1{6π, hence we can write:
1 q 2 a2
P “
6π c3

7.2 Maxwell distribution

According to Boltzmann’s law, the velocity distribution gpvx q is proportional to:
mv 2 mv2
´ 2k xT ´ 2k xT
gpvx q 9 e B ñ gpxq “ k e B ,
where k is a constant. The probability that a particle has a velocity anywhere between ´8
and 8 is 1. Requiring gpvx q to satisfyathis condition, for a certain k, isacalled normalizing the
function. We change variable to u “ m{p2kB T qvx , hence dvx “ du 2kB T {m. Integrating
gpvx q, we find: c c
2kB T 8 ´u2
ż
2πkB T
k e du “ k ,
m ´8 m
ş8 2 ?
where we have used the Gaussian integral ´8 e´u du “ π. The above integral must
a
be equal to 1, therefore k “ m{p2πkB T q. The probability function gpvx q, and similarly
gpvy q, gpvz q, is given by:
c 2
mvx
m ´
gpvx q “ e 2kB T .
2πkB T
236 Chapter 15. Problem Solutions

Let npvx , vy , vz q be the velocity distribution function, such that npvqdvx dvy dvz is equal to
the probability that a particle’s velocity is in the range pvx , vy , vz q to pvx ` dvx , vy ` dvy , vz `
dvz q. Then:
´ m ¯3{2 mpvx2 `vy2 `vz2 q
´
npvq “ gpvx q gpvy q gpvz q “ e 2kB T
,
2πkB T
where vx2 ` vy2 ` vz2 “ v 2 . In a coordinate system where vx , vy and vz are the three axes, the
number of particles with velocities in the range v to v ` dv is given by the product of npvq
with the volume of the spherical shell with radius v and thickness dv, i.e. 4πv 2 dv:
´ mv 2
m ¯3{2 2 ´ 2k
f pvq dv “ 4πv 2 npvq dv “ 4π v e B T dv ,
2πkB T
which is Maxwell distribution law.

7.3 Natural width of emission lines

When an atom passes from the excited state to the ground state, it emits a photon with
energy E “ hf “ h{p2πq ¨ p2π{T q “ h̄ω, where ω is the angular frequency. If the time
between two successive emissions is ∆t, the atom undergoes 1{∆t transitions per second,
hence the power emitted is E{∆t “ h̄ω{∆t. As shown in Pr. 7.1, the power emitted by an
accelerated electron is on the order of P “ e2 a2 {0 c3 . The acceleration of an electron is on
the order of ω 2 r, where r is the radius of the hydrogen atom (r “ 50 ¨ 10´12 m). Equating P
and E{∆t, we then obtain:
h̄c ´ c ¯2 1
∆t “ 2 .
e ωr ω
In the case of visible radiation c{ω « 10´7 m, hence ∆t « 4 ¨ 108 ω ´1 . Using ∆E “ h̄∆ω,
and substituting ∆E in Heisenberg’s uncertainty principle, we find:
∆E ∆ω
“ « 2 ¨ 10´9 ,
E ω
hence ∆E “ h̄∆ω « 5 ¨ 10´28 J. Note that radio waves have a natural width which is ap-
proximately 1010 times smaller than for visible light, while the time between two successive
emissions is longer by 1015 . The natural width of the emission lines increases with increas-
ing pressure. Indeed, the higher the pressure, the shorter the average time between two
consecutive collisions. Collisions interrupt the emission process, shortening its characteristic
time, and increasing, according to Heisenberg’s uncertainty principle, the uncertainty in E.
Thanks to the pressure dependence of the natural width of the spectral lines, we are able to
estimate the surface gravity of stars, by simply observing their spectra. This is the basis of
the Yerkes classification system.

7.4 Doppler Broadening

As shown in Sec. 7.6, in the non-relativistic limit, the wavelength emitted by an atom with ve-
locity vx is λ “ λ0 p1´vx {cq, where λ0 is the wavelength emitted at rest. Solving for vx :
´ λ¯
vx “ c 1 ´ .
λ0
According to Boltzmann
! ) the number of atoms emitting photons between λ and λ ` dλ is
law,
proportional to exp ´ kBET , where E “ 12 mvx2 “ 12 mc2 pλ0 ´λq2 {λ20 . Hence, the intensity is:

mc2 pλ0 ´ λq2

" *
Ipλq 9 exp ´ .
2λ20 kB T
 237

The wavelength at which Ipλq is half the maximum value satisfies:

Ipλ0 ` ∆λ{2q 1
“ .
Ipλ0 q 2
Substituting for Ipλq, previously found, we obtain the desired result.
This formula is valid when the gas can be treated as ideal, therefore the temperature and
pressure should not be too high or too low. Furthermore, if the pressure is too high, the nat-
ural width of the emission lines would overwhelm the Doppler broadening. Other phenomena
that can be observed on emission lines are the Stark effect, which consists in the separation
of spectral lines caused by an electric field due to the matter in the state of plasma, or the
Zeeman effect, equivalent to the Stark effect for the magnetic field.

7.5 Wien and Stefan-Boltzmann laws

To compute the distribution of Pλ , we note that the energy contained in the interval rλ, λ`dλs
is the same as the energy is the interval rf, f `df s, where f “ c{λ. Hence, Pf df “ Pλ dλ, or:
df
Pλ “ Pf .
dλ
But df {dλ “ ´c{λ2 , and substituting f “ c{λ in Pf , we obtain:

2πhc2 1
Pλ “ .
λ5 e kBhcT λ ´ 1

The maximum of Pλ can be found by setting dPλ {dλ “ 0. Since the numerator is constant,
it is simpler to set the derivative of the denominator to zero. Denoting with x “ λT “ cT {f
and k “ hc{kB “ 0.0143878 mm K, we have:
hc hc khcx
5λ4 pe kB x ´ 1q ´ λ4 e B “0
kB x
ñ ek{x p5 ´ k{xq “ 5 .

Solving this equation numerically, we find x “ 0.201405 k, hence λT “ 2.89777 mm K, which

is Wien’s law. The maximum of Pf can be found setting dPf {f “ 0. As before, using
x “ λT “ cT {f and k “ hc{kB “ 0.0143878 mm K:

f2 ” ´ k ¯ k kı
k
3 ex ´ 1 ´ ex “ 0.
pe ´ 1q2
x x

The term inside the brackets must be equal to zero. Thus, we need to numerically solve:
´ k ¯ xk
3´ e “ 3,
x
which gives x “ 0.354429 k as the only real solution. Hence, the maximum of Pf is at
λT “ 5.09945 mm K. Integrating the power over all frequencies:

f3
ż8
2πh 8
ż
P “ Pf df “ 2 hf
df .
0 c 0 e kB T ´ 1

Using the substitution u “ h{pkB T qf , with df “ kB T {h du:

2πh ´ kB T ¯4 8 u3
ż
P “ 2 u
du .
c h 0 e ´1
238 Chapter 15. Problem Solutions

The last integral is just π 4 {15:

2π 5 kB
4
P “ T4 .
15c h3
2

This equation can be written in the form:

2π 5 kB
4
P “ σT 4 , where σ“ .
15c2 h3

7.6 Ultraviolet catastrophe

Figure 15.11: Rayleigh-Jeans and Planck laws for a body at the temperature of the Sun
(T “ 5778 K). In the ultraviolet part of the electromagnetic spectrum, Planck’s function
decreases after having reached a maximum, while Rayleigh-Jeans increases to infinity.

In the previous problem, we showed that:

2πhc2 1
Pλ “ .
λ5 e kBhcT λ ´ 1
hc
For large λ, hc{pkB T λq tends to zero, hence e kB T λ tends to 1 ` hc{pkB T λq. Therefore:

2πhc2 kB T λ 2πckB T
Pλ “ “ . (15.5)
λ5 hc λ4
However, for λ Ñ 0, P Ñ 8. Of course, a body cannot emit an infinite amount of energy. It
is for this reason that classical physics, when applied to the study of the black body, leads to
the so-called “ultraviolet catastrophe”, being this the part of the electromagnetic spectrum
where the Rayleigh-Jeans law significantly deviates from the true distribution.

7.7 Why do stars emit energy?

Backing our arguments with data directly measurable on Earth, let us show that the first
two models cannot be correct, while the third one is reasonable. Let k “ 1367 W/m2 be the
solar constant and L@ “ 4πd2C ¨ k « 3.8 ¨ 1026 W the power emitted by the Sun.

Hypothesis 1 An asteroid arriving from infinity has an initial energy of zero. Applying
conservation of energy, we find the kinetic energy of the asteroid when it hits the surface of
the Sun:
GmM@
K `U “0 ñ K “ .
R@
 239

The power produced in this way is:

∆K GM@ ∆m
P “ “ .
∆t R@ ∆t
Since P “ L@ , the mass of the Sun increases at the rate of:
∆m R@
“ L@ .
∆t GM@
Hence, every year, the mass of the Sun would increase by:
R@
∆myr “ 365 ¨ 86400 ¨ L@ « 6.3 ¨ 1022 kg .
GM@
The duration of a year and the semi-major axis of the Earth depend on the mass of the Sun.
Let da, dM and dT be the variations in the semi-major axis, the mass of the Sun and the
length of a year, after a time interval dt. At t “ 0, the semi-major axis of the Earth is a, the
mass of the Sun is M and the length of a year is T . Therefore, at t “ 0, the kinetic energy
of the Earth is:
GmM
K“ .
2a
When t “ dt, the mass of the Sun increases by dM , thus the potential energy of the Earth
decreases, but the kinetic energy remains the same (similarly to Ex. 10.14). Therefore, the
total energy is now:
GmpM ` dM q GmM GmM ” ´ dM ¯ı GmM
E“´ ` “ 1´2 1` «´ ,
a 2a 2a M 2ap1 ´ 2 dM {M q
where we have used the approximation p1 ` xq « 1{p1 ´ xq for x “ dM {M ! 1. The total
energy can be written as E “ ´GmM {2pa ` daq, where a ` da is the new semi-major axis.
Comparing this expression for E with the above equation:
´ dM ¯ da dM
a ` da “ a 1 ´ 2 ñ “ ´2 .
M a M
According to Kepler’s third law, the new orbital period (T ` dT ) is:

4π 2
pT ` dT q2 “ pa ` daq3
GpM ` dM q
4π 2 3 ´ dM ¯´ da ¯
T 2 ` 2T dT « a 1´ 1`3
GM M a
´ dM ¯´ dM ¯
T 2 ` 2T dT 2
«T 1´ 1´6
M M
2 dM
2T dT « ´7 T
M
dT 7 dM
«´ .
T 2 M
We can then find by how much the period changes in a year:
7T
∆Tyear “ ´ ∆myear « ´1.4 s .
2M
240 Chapter 15. Problem Solutions

Hence, if the Sun got its energy solely from the impact of asteroids, the orbital period of the
Earth would decrease by around 1.4 s every year. Integrating the above equation:
ż T ptq
7 M ptq dM
ż
dT
“´
T0 T 2 M0 M
T ptq 7 M ptq
ln “ ´ ln
T0 2 M0
´ M ptq ¯´7{2
T ptq “ T0
M0
´ ∆myr ¯´7{2
T ptq « T0 1 ` tyr .
M0
After just 6.9 million years, the duration of a year would halve. Of course, this is not observed.

Hypothesis 2 In Pr. 10.4 we obtained the potential energy of a star:

3GM 2
U “´ .
5R
According to the contraction hypothesis, when R decreases, the potential energy decreases,
thus the star must radiate the energy lost into space. In fact, according to the Virial theorem
(Sec. 10.6), the magnitude of the kinetic energy is always equal to half the magnitude of the
potential energy. Hence, while the star contracts, only half of the potential energy is emitted
as radiation, while the rest goes to increasing the thermal velocity of atoms inside the star.
The total energy that the star can emit according to this process is:

3GM 2
Etot “ « 1.14 ¨ 1041 J .
10R
Assuming the solar luminosity stays roughly constant, the lifetime of the Sun would be:
Etot
tlife “ « 9.5 ¨ 106 years .
L@
If this hypothesis were true, the Sun would only be able to radiate energy for a little more
than 9 million years. However, we have ample evidence that shows the Solar System has
existed for more than 5 billion years, hence this hypothesis must also be wrong. At the time
when stars were believed to emit energy by this process, one of the objections to Darwin’s
theory of evolution was that it could only take place over a couple billion years, whereas the
Solar System was thought to be much younger than that.
It is interesting to note that, if this hypothesis were true, the mean temperature of the Sun
would increase every year by:
3 ∆T
nR “ L@
2 ∆t
3 M@ ∆T
kB “ L@
2 mp ∆t
∆T 2L@ mp
ñ “ « 0.5 K{year ,
∆t 3M@ kB
where mp is the mass of a proton (we assumed that the Sun is entirely made up of hydrogen
and neglected the mass of an electron). Again, this variation is not measured.

Hypothesis 3 According to the process of nuclear fission, 2 protons and 2 neutrons are
needed to form a helium atom. However, the mass of reagents and products is not the same,
 241

but it is smaller by 0.72 %, compared to the mass of a helium atom. This mass difference is
directly converted into energy, according to the equation E “ mc2 . Assuming that around
10 % of the total mass of the Sun can be converted into helium through this reaction, the
total energy the Sun can emit is:
Etot “ M@ ¨ 0.1 ¨ 0.0072 ¨ c2 “ 1.3 ¨ 1044 J .
Hence, its lifetime is:
Etot
tlife “ « 10 ¨ 109 years .
L@
Which sounds plausible.

7.8 Inside a star

Hydrostatic equilibrium In Sec. 10.2, we find that the gravitational acceleration at a distance
r from the centre of a star is:
Gmprq
gprq “ .
r2
Consider a cylindrical element, at a distance r from the centre of the star, with base area
dA, thickness dr and mass dm “ ρ dA dr. The gravitational force acting on it is:
Gmprq
Fg “ gprq dm “ ρ dA dr .
r2
The force due to the radiation pressure is always directed radially and is opposite to the force
of gravity. Let dP be the pressure difference between the top and bottom of the cylinder.
The cylinder experiences an outwards force of:
Fp “ dP dA .
Equating the two forces:
Gmprq dP Gmprqρ
dP dA “ ρ dA dr ñ “ .
r2 dr r2
Mass distribution Consider a spherical shell of radius r and thickness dr, with mass dm “
p4πr2 drqρ. Then:
dm
“ 4πr2 ρ .
dr
Energy production The contribution to the total luminosity from an infinitesimal mass
element dm is dLr “  dm. The mass between r and r ` dr is dm “ 4πρr2 dr, hence:
dLr
“ 4πr2 ρ .
dr
Temperature Consider a cylindrical element between r and r ` dr, with base area dA. The
radiation pressure is directly proportional to the incident flux:
Fr Lr
dPr “ pkρ drq “ kρ dr ,
c 4πr2 c
where Lr is the luminosity of the star at a distance r from the centre, which can be calculated
using the result from the previous part. At the same time, the pressure in a photon gas is
P “ aT 4 {3. If dT is the temperature difference between the bottom and the top of the
cylinder, then the pressure difference is:
´1 ¯ 4
dPr “ ´d aT 4 “ ´ aT 3 dT .
3 3
Equating the two expressions:
Lr 4 dT 3 kρ Lr
kρ dr “ ´ aT 3 dT ñ “´
4πr2 c 3 dr 4ac T 3 4πr2
242 Chapter 15. Problem Solutions

8. Flux and magnitude

8.1 Olbert paradox

Let L be the average luminosity of a star and n the average number density of stars in the
universe. The number of stars between r and r ` dr is:
dNr “ n ¨ 4πr2 dr .
Hence, the total luminosity produced by stars in this spherical shell is:
dLr “ L ¨ dNr “ 4πnLr2 dr .
Therefore, the flux arriving on Earth is:
dLr
dFr “ “ 4πnL dr .
4πr2
As we see, the flux does not depend on the distance of the spherical shell from the Earth.
Thus, the fluxes arriving on Earth from spherical shells of radii R and r (with r ‰ R) are the
same. Since there are an infinite number of these shells, the flux arriving on Earth should be
infinite, therefore the night sky should be as bright as daylight! This reasoning is also valid
if we assume that stars are grouped into galaxies, since we only used the average luminosity
and density of stars (which is certainly non-zero). Olbert believed that the answer to this
paradox is that space is opaque. At that time, thermodynamics had yet to be fully developed,
but today we know that, if this hypothesis were true, the interstellar medium would absorb
part of the light emitted by stars, re-emitting it in the form of black body radiation: the
energy arriving on Earth would be the same as if space were completely transparent. In
reality, the speed of light is finite and, because light from distant regions of the universe has
not yet had time to travel to the Earth, the observable universe is finite. At the same time,
the universe expands according to Hubble’s law, so radiation from distant stars is fainter,
due to the Doppler effect.

8.2 Earth’s temperature

Assuming the Sun to be a perfect black body, the flux arriving on Earth is given by the solar
constant:
L@ σR@ 2 T@ 4
k“ 2 “ .
4πdC dC 2
To obtain the power incident on the Earth (Li ), we multiply the flux by Earth’s cross-
sectional area (πRC 2 ):
σR@ 2 T@ 4
Li “ πRC 2 .
dC 2
The reflected power is proportional to the albedo of the Earth, according to Lr “ AC Li .
The difference between the incident and reflected power is equal to the power (P ) absorbed
by the Earth:
σR@ 2 T@ 4
P “ Li ´ Lr “ Li p1 ´ AC q “ πRC 2 p1 ´ AC q .
dC 2
In a stationary regime (at constant temperature) the energy absorbed is equal to the energy
emitted, according to the black body radiation formula:
P “ 4πσRC 2 TC 4 .
Solving for TC 4 , we find: c
R@
TC “ T@ p1 ´ AC q1{4 . (15.6)
2dC
 243

The temperature is independent on the radius of the planet. Inserting the numerical values
and assuming Earth has an albedo of AC “ 0.31:
TC “ 254 ˝ K “ ´19 ˝ C .
This is lower by about 30˝ C, compared to the average temperature on Earth. In fact, we
neglected the role of the Earth’s atmosphere.

8.3 Earth’s temperature, another model

For the Sun, the wavelength of maximum emission is λmax,s “ 501.5 nm, which corresponds
to visible light. The Earth, as shown in Pr. 8.2, has an average temperature of around
250 K, hence its wavelength of maximum emission should be around λmax,t “ 11500 nm,
which corresponds to the infra-red part of the electromagnetic spectrum. Since ei ă ev , this
phenomenon raises the Earth’s temperature.
Let us model the Earth as surrounded by a thin sheet of atmosphere, with a radius of about
RC , and constant temperature Tatm . The radiation that enters the atmosphere is:
2 4
2 σR@ T@ 2
Leff “ p1 ´ Aatm qkπRC “ p1 ´ Aatm q 2
πRC ,
dC
where k is the solar constant. The atmosphere directly absorbs a luminosity of Leff p1 ´ ev q,
and transmits the remaining Leff ev to the surface. The energy arriving on the surface and
then reflected is Leff ev As , while the energy absorbed is Leff ev p1 ´ As q. The energy reflected
by the surface and absorbed by the atmosphere is Leff ev As p1 ´ ev q, the remaining Leff ev 2 As
being lost in space. Energy is not only emitted by the Sun, but also by the atmosphere and
the surface of the Earth. The atmosphere absorbs a fraction equal to p1 ´ ei q of the thermal
2 4
energy emitted by the Earth, which is 4πRC σTC . On the other hand, the Earth absorbs
2 4
half of the energy emitted by the atmosphere, which is 8πRC Tatm (taking into account both
inner and outer surfaces). Therefore, the conditions for the atmosphere and the Earth to be
in thermal equilibrium are:
# 2 4 2 4
Leff rp1 ´ ev q ` As ev p1 ´ ev qs ` 4πσRC TC p1 ´ ei q “ 8πσRC Tatm
2 4 2 4
Leff ev p1 ´ As q ` 4πσRC Tatm “ 4πσRC TC .
2 4
Isolating 4πσRC TC in the second equation, and substituting into the first:
2 4
Leff rp1 ´ ev q ` As ev p1 ´ ev q ` ev p1 ´ As qp1 ´ ei qs “ 4πσRC Tatm r1 ` ei s .
2 4
Isolating 4πRC σTatm :
2 4 Leff r1 ´ As ev 2 ´ ev ei ` As ev ei s
4πRC σTatm “ .
1 ` ei
2 4 4
Substituting 4πRC σTatm into the second equation of the system, and isolating TC :
4 Leff 1 ` ev ´ As ev p1 ` ev q
TC “ 2
.
4πRC σ 1 ` ei
Inserting the expression for Leff and taking the fourth root of both sides:
c
R@ ” 1 ` ev ı1{4
TC “ T@ p1 ´ Aatm qp1 ´ As ev q ¨ .
2dC 1 ` ei
Inserting numerical values, we find TC « 283 K“ 10˝ C. In reality, the Sun and the Earth do
not emit at only one wavelength, hence the effective ev , ei could be slightly different. At the
same time, we have assumed that the Earth and atmosphere only reflect visible light. This
model predicts that the temperature of the Earth increases with decreasing ei (greenhouse
effect), to a maximum temperature of Tmax « 300 K = 27˝ C, when ei goes to zero and the
albedo of the Earth is reduced, due to the melting of the ice, to As “ 0.06. The change in
Aatm is, instead, harder to estimate. In any case, it wouldn’t be such a friendly world!
244 Chapter 15. Problem Solutions

9. Cosmic Distance Ladder

9.1 The Astronomical Unit

As explained in Sec. 11.1, the distance of Venus from the Sun can be obtained from its
maximum elongation:
dB
sin max “ ñ dB “ 0.723 dC .
dC
The angle between?the two images of Venus is ∆φ “ 232 ˘ 62 , while the parallax baseline is
b “ 2RC sin 60˝ “ 3RC . Hence, the distance (d) of Venus from the Earth is:
b
∆φrad “ ñ d “ p9.91 ˘ 2.5q ¨ 1010 m .
d
We can determine the astronomical unit by applying the Pythagorean theorem:

d2 “ d2C ´ d2B “ p1 ´ 0.7232 qd2C

(15.7)
dC « d ¨ 1.4475 “ p1.43 ˘ 0.35q ¨ 1011 m .

9.2 Luminosity-radius relationship

Differentiating Stefan-Boltzmann law (L “ 4πR2 σT 4 ), we find:

dL “ 8πRσT 4 dR ` 16πR2 σT 3 dT
´ dR dT ¯
“ 4πR2 σT 4 2 `4 .
R T
Dividing by L:
dL dR dT
“2 `4 .
L R T
Writing V “ 4{3πR3 in the adiabatic relationship, we find T R3pγ´1q “ const, which differ-
entiated gives:

dT R3pγ´1q ` 3pγ ´ 1qT R3pγ´1q´1 dR “ 0

dT dR
ñ “ ´3pγ ´ 1q .
T R
Substituting this expression for dT {T in the equation for dL{L, we find:
dL dR ” dR ı
“2 ` 4 ´ 3pγ ´ 1q
L R R
dR dR
“2 ´ 12pγ ´ 1q .
R R
Hence, using γ “ 5{3:
dL dR
“ ´6 .
L R
When the star expands, dR is positive, thus the above equation tells us that dL must be
negative, i.e. the luminosity decreases.

9.3 Period-density relationship

The pressure at a distance r from the centre of the star can be obtained using the equation
of hydrostatic equilibrium (Pr. 7.8):
dP Gmprqρ Gρ ´ 4 3 ¯ 4
“´ “´ 2 ¨ πr ρ “ ´ πGρ2 r ,
dr r r 3 3
 245

which can be easily integrated assuming that the pressure on the surface of the star is zero:
ż0 żR
4
dP “ ´ πGρ2 r
P prq r 3
2
´P prq “ ´ πGρ2 pR2 ´ r2 q
3
2
ñ P prq “ πGρ2 pR2 ´ r2 q .
3
The velocity of sound is therefore:
c c
γP 2
v“ “ γπGρpR2 ´ r2 q .
ρ 3
The period of pulsation is:
żR żR żR
dr dr 2 dr
T “2 «2 b “ b a .
0 v 0 2 2 2
γπGρpR ´ r q 2
γπGρ 0 pR 2 ´ r2 q
3 3

The last integral can be solved by substituting r “ R sin θ and dr “ R cos θ dθ. The lower
and upper limits become θi “ arcsin p0{Rq “ 0 and θf “ arcsin pR{Rq “ π{2, so we have:
ż π{2 ż π{2
R cos θ dθ π
a “ dθ “ .
0
2 2
R p1 ´ sin θq 0 2
Hence, substituting in the equation for the period:
c
2 π 3π
T “ b ¨ ñT “ .
2
γπGρ 2 2γGρ
3

This shows that the period of pulsation is inversely proportional to the square root of the
density. Indeed, the period of pulsation is greater for supergiants than for white dwarfs.
Taking M “ 5M@ and R “ 50R@ for a typical Cepheid, we find T “ 10 days, which is close
to the observed period.

10. Gravitation and Kepler’s Laws

10.1 Planets as point masses

In Sec. 10.2 we stated that a sphere behaves in the same way as a point mass, with the
same mass, placed at its centre. The easiest way to prove this statement is by computing
the potential of a spherical shell, since the force can be obtained by differentiating the
potential, according to F “ ´dU prq{dr. It is sufficient to show that a spherical shell obeys
this statement, since a solid sphere can be regarded as a collection of shells.
Let R be the radius of the spherical shell, r the distance between the centre of the shell and
a point P at its exterior, d the distance between P and a circular annulus constituting the
shell, and θ the angle between d and r (Fig. 15.12). The distance d can be written as a
function of r, R and θ, by applying the law of cosines (Eq. A.14) to triangle APO:
?
d “ R2 ` r2 ´ 2rR cos θ .
The area of the circular annulus between θ and θ ` dθ is equal to the circumference times
the height: p2πR sin θq ¨ pRdθq. Let σ “ M {p4πR2 q be the surface density. The potential
energy at point P due to the circular annulus between θ and θ ` dθ is:
GmσpRdθqp2πR sin θq
dU “ ´ .
d
246 Chapter 15. Problem Solutions

Figure 15.12: The potential at a point P, at a distance r from the centre of a spherical shell
of radius R, can be calculated by sectioning the shell into infinitesimal circular annuli and
summing (integrating) all of these contributions to the potential.

To obtain the potential energy of the spherical shell, we need to sum (integrate) the contri-
bution of all circular annuli:
żπ
2πσGR2 m sin θ
U prq “ ´ ? dθ
0 R2 ` r2 ´ 2rR cos θ (15.8)
2πσGRm ? 2
”
2
ıπ
“´ R ` r ´ 2rR cos θ .
r 0

We now consider two cases. If r ě R:

2πσGRm “ ‰ 4πR2 σGm GM m
U prq “ ´ pr ` Rq ´ pr ´ Rq “ ´ “´ ,
r r r
which is the potential due to a point mass, with the same mass of the shell, and placed at
its centre. If r ă R:
2πσGRm “ ‰ 4πR2 σGm GM m
U prq “ ´ pr ` Rq ´ pR ´ rq “ ´ “´ ,
r R R
which is independent of r.
By differentiating the potential U prq with respect to r, we find F prq in both cases:
$
&´ GM m , if r ě R
F prq “ r2
%0, if r ă R .
By symmetry, F prq is entirely radial. Since a solid sphere is made up of many spherical
shells, and since each spherical shell can be treated as a material point, it follows that a
solid sphere produces a gravitational field, on its exterior, equal to the gravitational field
produced by a material point, with the same mass, and placed at its centre. This result is
valid as long as the density is spherically symmetry.
The force experienced inside a spherical shell is zero. Therefore, when calculating the force
inside a solid sphere, at a distance r from its centre, it is sufficient to consider the mass
inside a spherical surface with radius r (the mass outside does not contribute to the total
gravitational field).

10.2 Using Gauss’ theorem

As explained in Sec. 10.2, for r ě RC , we can treat the Earth as a point mass, and use Eq.
10.1 in conjunction with Newton’s second law to obtain:
MC
gprq “ G .
r2
 247

For the Earth, MC “ 5.972 ¨ 1024 kg and RC “ 6.378 ¨ 106 m, hence:

Nm2 5.972 ¨ 1024 kg
gpRC q “ 6.674 ¨ 10´11 ¨ “ 9.81 m/s2 ,
kg2 6.371 ¨ 106 m2
which is the well-known value for the gravitational acceleration on the surface of the Earth.
To compute the gravitational acceleration inside the Earth, at a distance r from its centre,
we only need to take into account the mass mprq inside a sphere of radius r:
4 4
mprq “ πρ r3 , MC “ 3
πρRC .
3 3
Dividing the two equations:
r3
mprq “ MC 3
.
RC
This is the mass responsible for the gravitational acceleration at a distance r. Using Eq.
10.1 in conjunction with Newton’s second law, we find, for r ă RC :
GMC
gprq “ 3
r,
RC
where the gravitational field is directed towards the centre of the body. To sum up:
$ GM
C
& r2 , if r ě RC
’
’
gprq “
GM
% 3 C r, if r ă RC .
’
’
RC
We note that the gravitational field at the centre of the Earth is zero, therefore an object
would be in (stable) equilibrium at that point.

10.3 Deep tunnels

A body being transported on a circular orbit to the antipode, must cover half the orbit.
Hence, the time taken is: d
3
R$
t1 “ π .
GM$
We now calculate the time taken to reach the antipode through a tunnel. In Pr. 10.2, we
found that the gravitational field inside a spherical body is:

~ “ ´ GM$ ~r ,
gprq 3
R$

where the minus sign has been introduced because the gravitational field points towards the
centre of the body, in direction opposite the position vector ~r. Consider, by analogy, the
motion of a body of mass m, given by Hooke’s law:

~ “ ´ k r.
~ “ ´k~r ñ aprq
F
m
where k is the spring constant. The above equation can be written as:
~ “ ´ω 2 r ,
aprq

where ω is the angular frequency. In this case, the period of oscillation is:
2π
T “ .
ω
248 Chapter 15. Problem Solutions

The equation of motion of a body falling through the tunnel and one attached to a spring
are the same, if we set ω 2 “ GM$ {R$ 3
. The time taken to reach the antipode, i.e. to cover
half a full oscillation, is: d
3
π R$
t2 “ “π .
ω GM$
We note that t1 “ t2 , hence it takes the same time to transfer a load using a circular orbit
or a tunnel. For a more accurate calculation, we would need to consider the centrifugal force
due to the rotation of the Moon about its axis, which slightly increases the transfer period.
You can verify that the transfer time is the same for any straight tunnel connecting two
points on the surface of the Moon (try it!).

10.4 Potential energy of a spherical body

Imagine we assemble a spherical body by successively compounding spherical shell. The
potential energy of the whole body is then the sum of the potential energy of each of these
shells. The potential energy of a shell of radius r, thickness dr and mass dm “ p4πr2 drqρ,
which is placed on the surface of a body of mass mprq and radius r, is:
Gmprq dm
dU “ ´ .
r
Assuming the density is constant, mprq “ M r3 {R3 . We then obtain:

4πGM ρ r4
dU “ ´ dr .
R3
Integrating both sides to find the total potential energy:
żU żR
4πGM ρr4 dr
dU “ ´
0 0 R3
4πGM ρR2
ñU “´ .
5
Finally, substituting ρ “ 3M {p4πR3 q:

3GM 2
U “´ . (15.9)
5R

Eq. 15.9 gives the potential energy of a body of mass M , radius R and constant density.

10.5 Escape velocity

Without loss of generality, let us assume that m1 is initially at rest and that m2 has an
initial velocity of ~v . We want to find the minimum velocity needed for both masses to travel
indefinitely far away from each other. In Sec. 10.5, we assumed that m1 remains stationary
while m2 is escaping. However, it is clear that m1 will move in the same direction as m2 ,
because of the reciprocal gravitational attraction. Looking at Fig. 15.13 a), at the point of
maximum distance, the relative velocity between the two bodies is zero, therefore they will
both travel with the same velocity v~1 .
The key idea is to consider the reference system of the centre of mass of m1 and m2 . In this
system, the velocity of the centre of mass if zero, by definition, and the total momentum is
always zero, hence the bodies will move in opposite directions, as shown in Fig. 15.13 b).
The maximum distance is attained when both bodies are stationary, i.e. when their final
kinetic energy is zero in the CM frame.
In the lab frame, the centre of mass moves with velocity:
 249

Figure 15.13: Initial and final configurations in the reference system of a) the laboratory and
b) the centre of mass.
m1 ¨ 0 ` m2~v m2
~
u“ “ ~v .
m1 ` m2 m1 ` m2
Hence, in the CM frame, m2 and m1 move with velocities:
m1
v~2 “ ~v ´ ~
u“ ~v ,
m1 ` m2
m2
v~1 “ ´~ u“´ ~v .
m1 ` m2
The initial kinetic energy in the frame of the CM is thus:
1 1 1 m1 m2 2
KCM,i “ m1 v1 2 ` m1 v1 2 “ v .
2 2 2 m1 ` m2
At the end, both the kinetic and potential energy are zero in the CM frame, hence Ef,CM “ 0.
Applying conservation of energy, Ei,CM “ Ef,CM , which implies KCM,i `Ui “ 0, thus:
1 m1 m2 Gm1 m2
ve 2 “
2 m1 ` m2 d
c
2Gpm1 ` m2 q
ñ ve “ .
d
In the limit m1 " m2 , the above equation reduces to Eq. 10.27.

10.6 Roche limit

Let us consider a central body with mass M and radius R, and a secondary body with
mass m and radius r, which is at a distance x from the central body (Fig. 15.14). Let A
and B be the points of the secondary body which are farther and closer to the central body,
respectively. The gravitational field is the vector sum of the gravitational fields of the central
body gc and of the secondary body gs . Taking the magnitudes:
GM Gm
gA “ gc ` gs “ ` 2 ,
px ` rq2 r
GM Gm
gB “ gc ´ gs “ ´ 2 .
px ´ rq2 r
We require the relative acceleration between A and B to be zero, since the body is rigid:
gA ´ gB “ 0
GM GM Gm
ñ ´ “2 2 .
px ´ rq2 px ` rq2 r
250 Chapter 15. Problem Solutions

Figure 15.14: The rigid body-condition is satisfied if the gravitational accelerations in A and
B are the same.

Multiplying both sides by pr ` xq2 pr ´ xq2 :

´ x4 ¯
2rx “ α r2 ` 2 ´ 2x2 ,
r
where α “ m{M . When M " m, we expect the body to disintegrate at a distance x much
greater than its radius r. Hence, on the RHS of the above equation, we only retain the term
x4 {r2 :
c
x4 3 2M
2rx “ α 2 ñ x “ r .
r m
The ratio of the masses can also be written as a function of the densities and radii of the
two bodies:
M
4
πρc R3 ρc ´ R ¯3
“ 34 “ .
m 3
πρs r 3 ρs r
Hence:
c
2ρc
x“R3 . (15.10)
ρs
This is the Roche limit for a non-rotating body. If the rotation of the body about its axis
cannot be neglected, we need to add the contribution of the centrifugal force:
GM Gm
gA “ gc ` gs ´ acf “ ` 2 ´ ω2 r ,
px ` rq2 r
GM Gm
gB “ gc ´ gs ` acf “ ´ 2 ` ω2 r .
px ´ rq2 r
Equating the accelerations and using the approximation x " r, we find, following a similar
reasoning as before:
d
2ρc
xrot “ R 3 3ω 2
. (15.11)
ρs ´ 4πG

As expected, the centrifugal force increases the distance at which the body disintegrates. In
the case ω “ 0, Eq. 15.11 reduces to Eq. 15.10. Denoting with xs and xr the Roche limits for
a stationary and a rotating body, respectively, in the case of the Earth-Sun system we find:
xe,s « 556330 km xe,r « 556980 km .
While, for the Earth-Moon system:
xm,s « 9485 km xm,r « 9485 km .
 251

We see that rotation can be neglected in both cases (remember that the approximation x " r
likely introduces an error comparable to the correction due to rotation). If a secondary body
revolves about a central body at a distance equal to its Roche limit, and its rotation is
synchronous with its revolution:
GM 4 G
ω2 “ “ πR3 pc ¨ 3 .
x3 3 x
Substituting ω in 15.11, we find:
d
2ρc
x“R3 3 ,
ρs ´ pc R
x3

which, after some manipulation, leads to the synchronous Roche limit:

c
3ρc
xsyn “ R 3 . (15.12)
ρs

10.7 Jeans Limit

The force of gravity tends to compress the nebula, while the gas pressure tends to expand it.
In equilibrium, these two forces are equal. Instead of working with forces however, it is more
convenient to think in terms of energy. We can equal the difference in potential energy of
the nebula, following an infinitesimal contraction dr, with the work done by the gas during
this contraction. Equating the differences in energy and work is just another way of equating
the forces. As we found in Pr. 10.4, the potential energy of the nebula is:

3GM 2
U “´ .
5r
Differentiating the above equation:

3GM 2
dU “ dr .
5r2
The work done by the gas during the contraction is:

dW “ P dV .

Assuming the gas is ideal, we know that P “ nRT {V , where V “ 4{3 πr3 . Substituting
dV “ 4πr2 dr, we find:
3nRT
dW “ dr .
r
Equating dU and dW :
GM 2
“ nRT .
5r
Let µ be the molar mass of the gas, we then have n “ M {µ. Furthermore, M “ 4{3πρr3 ,
where ρ is the density of the nebula, hence r “ r3M {p4πρqs1{3 . Substituting the expressions
for n and r in the last equation, isolating M :
c
3 ´ RT ¯3{2
MJ “ 5 . (15.13)
4πρ Gµ
252 Chapter 15. Problem Solutions

10.8 Hill sphere and Lagrangian points L1 , L2

Figure 15.15: Positions of the Lagrangian points.

The Lagrangian points are five points in the Solar System whose positions relative to the
Earth and Sun are fixed (Fig. 15.15). It can be shown that only L4 and L5 are stable points,
while L1 , L2 and L3 are all unstable. Nonetheless, L1 is very convenient for observing the
Sun, and many probes, such as the SOHO (Solar and Heliospheric Observatory) and the
ACE (Advanced Composition Explorer ), have taken advantage of it. Instead, L2 , placed on
the opposite side of the Earth relative to the Sun and shielded from the solar radiation, is
used for studying deep space, especially in the infra-red (which requires low temperatures).
Among the best known satellites in L2 we find Gaia, and the future James Webb telescope.
Now, let us calculate the position of L1 and L2 . Let m be the mass of the Earth, M the mass
of the Sun, r the distance of the Earth from the Sun and x the distance of the points L1 and
L2 from the Earth. For now, we don’t make any assumptions about the m{M ratio. The
two bodies orbit around the common centre of mass, which is at a distance rm{pM ` mqsr
from the centre of the Sun. The distance rL1 between L1 and the centre of mass corresponds
to the radius of its orbit:
m M
r L1 “ r ´ x ´ r“ r ´ x.
M `m M `m
Equating the gravitational and centrifugal forces in L1 :
GM Gm
ω 2 r L1 “ ´ 2 ,
pr ´ xq2 x
 253

where ω 2 “ GpM ` mq{r3 is the square of the angular velocity of revolution of the Earth.
Multiplying by pr ´ xq2 {GM :
´ 1 x ¯´ x ¯2 ´ x ¯2
p1 ` αq ´ 1´ “1´α 1´
1`α r r r
´ x x2 x3 x4 x5 ¯ x3 x4 x5
ñα 1´2 ` 2 ` 3 ´2 4 ` 5 “3 3 ´3 4 ` 5 ,
r r r r r r r r
where α “ m{M « 3 ¨ 10´6 . Since r " x (this hypothesis can be verified at the end, but
appears reasonable given that m{M ! 1), in approximating the above equation we only
retain terms of zero order in x{r on the LHS (ignoring x{r, x2 {r2 , x3 {r3 and so on), and
only the third order term on the RHS. Thus:

3x3 {r3 “ α .

It follows that: c
3
m
x“r . (15.14)
3M
For the Earth-Sun system, we obtain x “ 149.6 ¨ 107 m “ 0.01 ua. Therefore, L1 is at a
distance from the Earth of one hundredth of an astronomical unit, in the direction of the
Sun. Hence, the assumption r " x was indeed valid. With a similar reasoning, we can show
that the distance of L2 from Earth is also equal to x (in the direction opposite the Sun).
If, instead, we numerically solve the fifth order equation, we find

xL1 “ 1.4914 ¨ 109 m ,

xL2 “ 1.5014 ¨ 109 m .

As we expect, the error is on the order of 1{100 (i.e. the size of x{r). The points L1 and
L2 define a sphere around the Earth such that all the bodies inside it gravitationally belong
to the Earth, while the bodies outside it belong to the Sun. If we take into account the
eccentricity of the Earth’s orbit, we need to replace r Ñ ap1 ´ eq in Eq. 15.14. Therefore, a
more general formula for the radius of the Hill sphere is:
c
m
rH “ p1 ´ eqa 3 . (15.15)
3M

For the above formula to hold, the mass of the secondary body should be much smaller
than the mass of the primary body, so that the approximations m{M, x{r ! 1 are valid.
While this is certainly true for the Earth-Sun system, it is not a very good approximation
for the Earth-Moon system (indeed the Hill sphere of the Moon is about 1/6 of the Earth-
Moon distance). In reality, there are other phenomena (such as radiation pressure and the
Yarkovsky effect), that make a body at the edge of the Hill sphere unstable. Therefore,
bodies are stable, in the long term, only if their distance is less than half of the Hill sphere.

10.9 Position of Lagrangian point L3

Let m be the mass of the Earth, M the mass of the Sun, r the Earth-Sun distance and
x the distance of L3 from the point on Earth’s orbit opposite the Earth. Looking at Fig.
15.16 (overleaf), we find the distances L3 -centre of mass, L3 -Sun and L3 -Earth, respect-
254 Chapter 15. Problem Solutions

Figure 15.16: Schematics showing the orbits of the Sun and the Earth (small and large
circles, respectively).

ively:
m M
dL3 ,c “ r ´ r`x“ r ` x,
M `m M `m
m M ´m
dL3 ,s “ r ´ 2 r`x“ r ` x,
M `m M `m
2M
dL3 ,e “ dL3 ,s ` r “ r ` x.
M `m
Equating the centrifugal and gravitational forces at L3 , we find:
GM Gm
ω 2 dL3 ,c “ ` 2 ,
d2L3 ,s dL3 ,e

where ω 2 “ GpM ` mq{r3 is the square of the angular velocity of revolution of the Earth.
Denoting by α “ m{M « 3 ¨ 10´6 , similarly to the previous problem, it is possible to neglect
the terms of order greater than one in x{r and α. After some manipulation, we get:
17 ´ 53 ¯ x
α“ 3` α .
4 4 r
But αx{r is an infinitesimal of second order, hence, to be consistent with the previous
approximation, we neglect it. We finally obtain:
17
x“ αr. (15.16)
12
Therefore, the point L3 is at a distance x « 636 km away from the point on Earth’s orbit
opposite the Earth. It can be shown that L3 is a point of unstable equilibrium, therefore
there aren’t any bodies that orbit the Sun at that point. We can also write down the position
of L3 with respect to the centre of mass of the system:
M 17 ´ 5 ¯
dL3 ,c “ r ` x « p1 ´ αqr ` αr “ 1 ` α r. (15.17)
M `m 12 12

10.10 The three-body problem and Lagrangian points L4 , L5

Let us place the origin of our coordinate system in the centre of mass of m1 , m2 and m3 .
Then:
m1 r~1 ` m2 r~2 ` m3 r~3 “ 0 .
 255

Figure 15.17: The three-body problem has a simple solution if the bodies form an equilateral
triangle.

The forces acting on the first body are:

Gm1 m2
F~2,1 “ ´ pr~1 ´ r~2 q ,
a31,2
Gm1 m3
F~3,1 “ ´ pr~1 ´ r~3 q .
a31,3

The centripetal force is:

F~c “ m1 ω 2 ¨ r~1 .
Equating the centripetal and gravitational forces, we find:
Gm1 m2 Gm1 m3
pr~1 ´ r~2 q ` pr~1 ´ r~3 q ´ m1 ω 2 ¨ r~1 “ 0 .
a31,2 a31,3

Solving for r~3 in the CM equation, we find r~3 “ ´pm1 {m3 qr~1 ´ pm2 {m3 qr~2 . Substituting in
the above equation:
Gm1 m2 Gm1 m3 ” ´ m1 ¯ m2 ı
3
pr~1 ´ r~2 q ` 3
r~1 1 ` ` r~2 ´ m1 ω 2 ¨ r~1 “ 0
a1,2 a1,3 m3 m3
” m
2 m3 ´ m1 ¯ ω 2 ı ” m
2 m2 ı
ñ r~1 3
` 3
1` ´ “ r~2 3
´ .
a1,2 a1,3 m3 G a1,3 a1,2 3
The three bodies cannot be aligned, hence r~1 and r~2 cannot be on the same line. For the
above equation to be satisfied, the coefficients of r~1 and r~2 should be both zero. We therefore
obtain two equations:
m2 m2
´ “ 0,
a1,3 3 a1,2 3
m2 m3 ´ m1 ¯ ω 2
` 1` ´ “ 0.
a1,2 3 a1,3 3 m3 G
The first equation tells us that a1,2 “ a1,3 but, as we could have chosen any of the three
bodies at the beginning, by symmetry we must also have a1,2 “ a1,3 “ a2,3 “ a. Now, the
second equation gives:
G GM
ω 2 “ 3 pm1 ` m2 ` m3 q “ 3 ,
a a
256 Chapter 15. Problem Solutions

where M is the total mass of the system. We have shown that the three bodies all move with
the same orbital period and that they are placed on the vertices of an equilateral triangle
(Fig. 15.17). In the Solar System, the points L5 and L4 are at the same distance from the
Sun as the Earth, and are placed 60˝ before and after the Earth, respectively.
Analysing the energy of the system in L4 and L5 , we discover that it is a local maximum.
Therefore, even a small disturbance would move the body away from the Lagrangian points
and, consequently, there would be no stable orbits around L4 and L5 . However, in the Solar
System, we observe a large amount of so-called Trojans asteroids distributed in two oblong
regions around the points L4 and L5 of Jupiter. The reason why these points can host stable
orbits is due to the Coriolis force in the rotating frame of the Sun. This force is large enough
to sufficiently curve the trajectory of an asteroid that moves at moderate speeds, forcing it
to stay close to L4 or L5 .

10.11 The fate of abandoned satellites

Part 1 Let r be the radius of the orbit of the satellite. Its total energy is:
GmM
Etot “ ´ .
2r
Since the orbit is circular, xKy “ K and xU y “ U . According to the Virial theorem:
1 GmM
K“´ U “ .
2 2r
Friction with air reduces the total energy of the satellite. The above equations shows that,
as r decreases, K increases: it seems like friction increases the velocity of the satellite. How-
ever, we have not taken into account the gravitational force. In a perfectly circular orbit, the
gravitational force is perpendicular to the instantaneous velocity, hence it does no work. In
this case, the orbit is not perfectly circular, and the force of gravity has a small but non-zero
tangential component. In the next part, we will show that the value of this component is
exactly two times the friction force, so that the total force acting on the body is equal to
the friction force, in magnitude, but opposite in direction.

Part 2 The work done by the friction force is dWa “ F~a ¨ ds, ~ where ds ~ “ ~v dt and v is the
velocity of the satellite. The satellite looses energy due to friction, according to dEtot “ dWa .
Hence:
~ .
dEtot “ F~a ¨ ds
Note that F~a is directed opposite to ds,~ hence the scalar product is negative, and the energy
decreases. The kinetic energy is K “ 12 mv 2 . Differentiating, we find:

dv dt ~ “ pF~g ` F~a q ¨ ds
~
dK “ mv dv “ mv ds “ ma ds “ F~tot ¨ ds
dt ds
~ .
ñ dK “ pF~g ` F~a q ¨ ds

Applying the Virial theorem dEtot “ ´dK. Substituting the expressions obtained previously
for dEtot and dK, we have:

F~a “ ´pF~g ` F~a q ñ F~g “ ´2F~a .

The tangential component of the gravitational force is therefore equal to two times the
friction force, and has opposite direction. Hence, the total force acting on the body in the
tangential direction is F~tot “ F~g ` F~a “ 2F~a ´ F~a “ F~a , equal to the friction force, but
opposite in direction. That is why the satellite accelerates!
 257

Part 3 Differentiating the kinetic energy with respect to the radius (r) of the orbit:
dK GM m
“´ .
dr 2r2
If the radius of the orbit decreases, dr is negative and dK is positive, i.e. the velocity
increases. After every revolution, the radius of the orbit decreases by dr “ 100 m. The work
done by the force acting on the satellite is:
¿ ¿
dW “ F~tot ¨ ds ~ “ Fa ds “ Fa p2πrq .

The friction force is Fa “ cρv 2 . The velocity in the circular orbit is v 2 “ GM {r, hence:
´ GM ¯
dW “ cρ ¨ p2πrq ñ dW “ 2πcρGM .
r
This work is equal to the increase in kinetic energy after one revolution. Therefore:
GM m
2cπρGM “ ´ dr .
2r2
Taking the magnitude of dr and isolating ρ, we find an expression for the density:
m
ρ“ |dr| « 4 ¨ 10´10 kg/m3 .
4cπr2

10.12 Kepler’s equation

In the usual notation, the equation of an ellipse is:
x2 y2
2
` 2 “ 1.
a b
Placing the origin at the rightmost focus, we perform the translation x Ñ x ` c:
px ` cq2 y2
2
` 2 “1
a b
Consider a plane polar coordinate system, with x “ r cos θ and y “ r sin θ. Substituting in
the previous equation:
pr cos θ ` cq2 r2 sin2 θ
2
` “1
? a b2
Since c “ a2 ´ b2 and c “ a e, we have b2 “ a2 p1 ´ e2 q:
r2 sin2 θ
pr cos θ ` a eq2 ` “ a2
1 ´ e2
´ 1 ´ cos2 θ ¯
r2 cos2 θ ` ` a2 e2 ` 2a er cos θ “ a2
1 ´ e2
1 ´ e2 cos2 θ
r2 ` 2ae r cos θ ´ a2 p1 ´ e2 q “ 0 .
1 ´ e2
Solving the quadratic equation, we find:
1 ´ e2 ” a
2 e2 r 2 cos2 θ ` a2 p1 ´ e2 cos2 θq
ı
r“ ´ a e cos θ ˘ a
1 ´ e2 cos2 θ
1 ´ e2 ” ı
“ ´ a e cos θ ˘ a
1 ´ e2 cos2 θ
ap1 ´ e2 q ” ı
“ 1 ˘ e cos θ
p1 ´ e cos θqp1 ` e cos θq
ap1 ´ e2 q
“ .
1 ˘ e cos θ
258 Chapter 15. Problem Solutions

Since the origin is in the rightmost focus, rp0˝ q “ ap1 ´ eq, we chose the positive sign in the
previous equation. We then find:

ap1 ´ e2 q
rpθq “ ,
1 ` e cos θ
as we wanted to show.
Let H be the intersection of the line passing through EP and the major axis. Looking at
Fig. 10.8, we have:
2 2 2
FP “ PH ` FH .
Since an ellipse is just the scaled version of a circle, by factors a and b along the major and
minor axis, respectively:

r2 “ pb sin Eq2 ` pa cos E ´ eaq2

“ a2 p1 ´ e2 q sin2 E ` a2 cos2 E ` e2 a2 ´ 2ea2 cos E
“ a2 ` e2 a2 p1 ´ sin2 Eq ´ 2ea2 cos E
“ a2 ` e2 a2 cos2 E ´ 2ea2 cos E
“ a2 p1 ´ e cos Eq2 .

Hence:
r “ ap1 ´ e cos Eq ,
as we wanted to show.
To derive Kepler’s equation, we differentiate the above formula with respect to time:

r9 “ ae sin E E9 .

On the other hand, r is related to the true anomaly by:

ap1 ´ e2 q 1 1 ` e cos θ
rpθq “ ñ “ .
1 ` e cos θ r ap1 ´ e2 q
Differentiating the above equation with respect to time:

r9 e sin θ θ9
2
“ .
r ap1 ´ e2 q
From Kepler’s second law, we know that the radius vector sweeps out equal areas in equal
times. Multiplying Eq. 5.9 by r2 , setting θ “ 0˝ :
?
r2 θ9 “ ωa2 1 ´ e2 ,
a
where ω “ GM {a3 is the mean angular velocity of revolution. Substituting θ9 in the
previous equation:
ae sin θ
r9 “ ω ? .
1 ´ e2
9
Comparing with r9 “ ae sin E E:
sin θ
E9 “ ω ? .
1 ´ e2 sin E
Since an ellipse is just the scaled version of a circle, by factors a and b along the major and
minor axis, respectively; PH “ b EH, which can be written as:
?
r sin θ “ a 1 ´ e2 sin E .
 259

9 we then have:
Comparing with the equation for E,

rE9 “ ω a .

Since r “ ap1 ´ e cos Eq, this becomes:

p1 ´ e cos EqE9 “ ω .

Integration over time yields:

E ´ e sin E “ ωpt ´ t0 q .
If t0 is interpreted as the time elapsed since the body passed through perihelion, then M “
ωpt ´ t0 q. Hence:
M “ E ´ e sin E ,
which is Kepler’s equation.

10.13 Relaxation time

Figure 15.18: If two stars interact weakly, their motion is approximately a straight line,
and only the component of the velocity perpendicular to the direction of motion is changed
significantly.

If the impact parameter is large enough, the incoming star travels approximately on a straight
line with velocity v, as shown in Fig. 15.18. The distance between the two stars is r “ b{ cos θ,
hence the gravitational force along the y direction is:
Gm2
Fy “ Fg cos θ “ cos3 θ .
b2
After an infinitesimal time dt, the variation of linear momentum in the y direction is dpy “
Fy dt “ pFy {vq dx. But tan θ “ x{b, which differentiated is 1{ cos2 θ dθ “ dx{b. We then find
dpy “ Gm2 {pvbq cos θ dθ. Integrating between ´π{2 and π{2:
ż py,f ż π{2
Gm2
dpy “ cos θ dθ
py,i ´π{2 vb

2Gm2
ñ ∆p “ .
vb
The variation of the velocity along the y direction is ∆vy “ ∆py {m. Clearly, this approxim-
ation is valid only in the case ∆vy ! v, while for ∆vy Á v the interaction is substantial and
should be computed using the concepts presented in Appendix B.
We want to understand if stars for which ∆vy Á v give rise to a greater perturbation com-
pared to stars for which ∆vy ! v. In the first case, the interaction is significant, but, at the
same time, it is rare that two stars come very close to each other. In the second case, the
260 Chapter 15. Problem Solutions

interaction is weak, but the stars can be far away from each other.
Significant interactions happen when the impact parameter is smaller than the critical value bc :
2Gm Gm
Á v ñ b À bc « 2 ,
vb v
where we ignore numerical factors, since this is just an estimate. Therefore, a significant
interaction takes place only if the incoming star passes through an area of σ « b2c « G2 m2 {v 4 .
Let n be the number density of stars. In a volume σ dx there are nσ dx stars, hence, every
second, a number nσ dx{dt “ nσv of stars pass through this volume. On average, to register
the passage of a star, we need to wait a time of 1 “ tgc nσv ñ tgc “ 1{pnσvq. Let R be the
dimension of the system and N the total number of stars. Then:

1 R3 v 3 N R3 v 3
tgc « « 2 2
« 2 2 .
nσv N pG m q G M

where we have used n « N {R3 , and M “ N m is the total mass of the system. If the system
is ina
gravitational equilibrium, the velocity (v) can be found from the Virial theorem, i.e.
v “ GM {R. It follows that:
´ R ¯´ R2 v 4 ¯ R
tgc « N «N .
v G2 M 2 v
This is the time scale in the case of significant interactions between stars.
If the interaction is weak, b " bc . Following a collision, the velocity of the star changes by
dvy « Gm{pvbq. Since stars are equally likely to come from any direction, we expect the
mean velocity to be constant. The star therefore traces a chaotic path about the orbit it
would have in the case of a continuous mass distribution. The mean square speed produced
by these interaction in an interval rb, b ` dbs and time ∆t is:
´ Gm ¯2
xdvy2 y “ p2πb dbqpv∆tqn .
vb
Integrating over all possible impact parameters, we find the total mean square speed:
ż b2 ´ G2 m2 ¯ 2πnG2 m2 ´b ¯
2
xdvy2 ytot « ∆t p2πb dbqpvnq 2 v2
“ ∆t ln .
b1 b v b 1

We can take b2 « R and b1 « bc « Gm{v 2 . Therefore b2 {b1 « Rv 2 {Gm “ N pRv 2 {GM q « N

if the system is in Virial equilibrium. The system significantly deviates from the model of
continuous mass distribution on the time scale for which xdvy2 ytot « v 2 . Isolating ∆t from
the above equation, it follows that:

v3 ´ N ¯´ R ¯ tgc
∆tgc « « « ,
2πG2 m2 n ln N ln N v ln N
which is shorter than the scale time tgc . This is the relaxation time. a
For a galaxy, N « 1011 , R « 1020 m and m « 1030 kg, hence R{v « R3 {GN m « 1015 s «
109 years, therefore tgc « 1017 years. This is much larger than the age of the universe,
therefore galaxies are considered to be “relaxed”, and can be approximated with a continuous
mass distribution. a
For a star cluster, N « 106 , R « 1017 m and m « 1030 kg, thus R{v « R3 {GN m « 1013 s «
105 years, and tgc « 5 ¨ 109 years. The age of the universe is on the same order of magnitude,
therefore stellar interactions are important, and effectively randomize the velocities of all
stars in the cluster. It is for this reason that star clusters are spherical.
 261

11. Motion of the Planets

11.1 Great opposition

We neglect the precession of Mars’ apse line. The sidereal period of Mars is PD “ 1.88082 years,
while the synodic period can be calculated using Eq. 11.3:
PD,yr
Psyn,yr “ “ 2.1353 years .
PD,yr ´ 1

At the end of a synodic period, Mars will have covered an angle corresponding to Psyn {Psid “
1.13531 of a full circle, and will be displaced by 0.13531 of a full circle, relative to the
starting position. The time taken to return to the same point with respect to its apse line
is 1{0.13531 “ 7.39067 synodic periods, which corresponds to 7.39067 ¨ 2.1353 “ 15.78 years.
Therefore, great oppositions happen approximately once every 15 ´ 17 years. A more precise
cycle is five times longer than the previous one, with a periodicity of 78.906 « 79 years.

12. Orbital Manoeuvres

12.1 Circular orbits

The most effective way to put a satellite in an equatorial orbit around the Earth is to launch
it from a location on the equator, in the same direction as the rotation of the Earth, from
west to est. Indeed, the Earth completes a rotation about its axis every 86164 seconds, thus
its angular velocity of rotation is ω “ 2π{T “ 7.29 ¨ 105 rad/s. The equatorial radius is
RC “ 6378 km, therefore a satellite launched in this way has a an initial velocity of ωRC “
465.1 m/s. The velocity of a circular orbit at sea level is v “ GM {RC « 7.9 km/s, hence
the minimum ∆v required to put a satellite in circular orbit is approximately 7.4 km/s.
To insert a satellite in a polar orbit, perpendicular to the equator, it is best to launch it from
a location near the poles, so that the component of the velocity along the equator (which
will have to be cancelled) is the smallest possible. The minimum ∆v required is simply equal
to the velocity of a circular orbit at sea level, which is 7.9 km/s.
To escape Earth’s a gravitational field, the velocity must be equal to the escape velocity from
the Earth, i.e v “ 2GM {RC « 11.2 km/s. If the satellite is launched from the equator, in
the same direction as the rotation of the Earth, the ∆v is reduced to 10.7 km/s.
In reality, the ∆v should be a bit larger than the previously stated limits. Of course, it is
not possible to put a body in circular orbit at sea level, but we must first reach a height of
a few hundred kilometres above the surface. This requires an additional ∆v to contrast the
gravitational force (which adds about a 1 km/s), while the velocity of the new circular orbit
is essentially the same, since v “ GM {pRC ` hq and h ! RC . In addition, the friction
force with the atmosphere (1.5 km/s) also tends to slow down the satellite. Therefore, it is
reasonable to assume we need a further 3 km/s above the previously stated limits.

12.2 Changing the plane of the orbit

The manoeuvre should be carried out in the ascending or descending node, depending on
whether the desired final orbit is above or below the starting orbit. Let v0 be the velocity
of the circular orbit. Then, the projections of v~0 on the plane of the final orbit and on the
perpendicular to it are:
vik “ v0 cos i , viK “ v0 sin i .
The final velocity must be:
vfk “ v0 , vfK “ 0 .
262 Chapter 15. Problem Solutions

Hence, the necessary ∆v is:

∆vk “ vfk ´ vik “ v0 p1 ´ cos iq ,
∆vK “ vfK ´ viK “ ´v0 sin i .
Adding the two components in quadrature:
b b
∆vtot “ ∆vk2 ` ∆vK 2
“ v02 p1 ´ cos iq2 ` v02 sin2 i
a ? ?
“ v0 1 ` cos2 i ´ 2 cos i ` sin2 i “ 2v0 1 ´ cos i .
Using the half-angle formula in Eq. A.22, we find:
´i¯
∆vtot “ 2v0 sin .
2
Usually, satellites are directly launched with the desired orbital inclination, in order to avoid
wasting fuel with this orbital manoeuvre. A satellite in a circular orbit, at a height of
h “ 200 km above the surface, would require ∆v “ 270 m/s for a change of 1˝ in its orbital
inclination. If the expulsion speed is 4.5 km/s and the mass of the satellite is 1000 kg, we
need an additional ∆m “ 60 kg of fuel for this manoeuvre. However, to bring additional fuel
in orbit, we also require more fuel to lift the satellite. Suppose ∆vorb « 9 km/s is needed to
place the satellite into orbit (taking into account friction with air and the effect of gravity).
Then:
mi
∆vorb “ ve ln .
mf
If an additional mass of fuel ∆m must be left at the end, the initial mass must be greater
by an amount ∆M :
mi ` ∆M
∆vorb “ ve ln .
mf ` ∆m
Dividing the above equations, we find:
mi mi ` ∆M mi
ln “ ln ñ ∆M “ ∆m ,
mf mf ` ∆m mf

where mi {mf “ evorb {ve « 7. Hence, the additional mass of fuel needed to launch the
satellite is:
mi
∆M “ ∆m “ 420 kg ,
mf
approximately 40 % of the mass of the satellite itself.

12.3 Third cosmic velocity

The trajectory of the probe escaping the Solar System can be divided into two parts.
First, the satellite moves far away from Earth, to enter the sphere of influence of the Sun.
The energy needed to achieve this is approximately equal to the energy required to escape the
gravitational field of the Earth entirely. Indeed, the gravitational potential energy is inversely
proportional to the distance, thus the error made in assuming that the final distance is infinite
rather than the Hill sphere of the Earth (about 1.5 million kilometres) is not significant.
Denoting with vi the initial speed of the probe and with vf its speed upon entering the
sphere of influence of the Sun, applying energy conservation:
c
1 2 GmMC 1 2 2GMC
mvi ´ “ mvf ñ vf “ vi2 ´ .
2 RC 2 RC
In the second part, the gravitational field of the Earth can be neglected. Now, the probe
is at a distance dC from the Sun and is moving with a velocity of vf ` vC in the frame of
 263

the Sun. To find the minimum velocity required to escape the Solar System, we require the
speed of the probe to be zero at infinity. Therefore, the probe must travel in a parabolic
orbit, with energy E “ 0:
1 GmM@
mpvf ` vC q2 ´ “ 0.
2 dC
Solving for vf : c
? GM@
vf “ p 2 ´ 1q .
dC
Substituting vf in the first equation and solving for vi , we find:
c
2GMC ? M@
vi “ ` p3 ´ 2 2q « 16.6 km/s .
RC dC
If the satellite is launched from the equator, the required velocity is smaller by about
0.5 km/s, therefore the third cosmic velocity is approximately v « 16 km/s. Notice that
the energy scales with the square of the velocity. Since the escape velocity from Earth is
ve “ 11.2 km/s, and from the Sun is vf “ ? 12.3 km/s, a quicker way to find vi is by adding
in quadrature the two contributions: vi “ 11.22 ` 12.32 ´ 0.5 « 16 km/s.
To collide with the Sun, the probe first needs to escape Earth’s gravitational field, which
requires a velocity of 11.2 km/s. After this, the probe would orbit the Sun, at the same
distance and with the same velocity as the Earth. To reach the centre, the angular mo-
mentum of the probe needs to be zero (see Pr. 12.5), hence it is necessary to cancel all of its
tangential velocity, which requires a ∆v equal to Earth’s orbital velocity. To minimize this
∆v, we chose to launch the probe when the Earth is at aphelion, its velocity being:
c
GmM GmM 1 1 ´ e GM
´ “´ ` mva2 ñ va “ “ 29.3 km/s .
2a a 2 1`e a
Since energy
? scales with the square of the velocity, the speed needed to collide with the Sun
is vc “ 11.22 ` 29.32 ´ 0.5 « 31 km/s. Hence, escaping the Sun requires less energy than
colliding with it!
Notice that, in both cases, we must deliver at take off all the ∆v the probe needs for
both parts of its motion. By doing so, we take advantage of the fact that the energy is
proportional to the square of the velocity (Oberth effect). In fact, if we only gave the
probe enough energy to escape the Earth first, and then fired the engine again to escape
? Sun, the total ∆v needed would have been 11.2 ` 12.3 ´ 0.5 “ 23 km/s, instead of
the
11.22 ` 12.32 ´ 0.5 « 16 km/s.

12.4 Rocket in a gravitational field

While the linear momentum is conserved for a rocket moving in free space, this is not true
when the rocket is in a gravitational field of strength g (for instance, it is escaping the Earth),
since its momentum changes by dp “ ´mg dt after a time dt. Looking at Fig. 12.1:
dp “ pm ` dmqpv ` dvq ´ p´dmq ¨ pve ´ vq ´ mv “ ´mg dt .
Simplifying, we obtain:
m dv ` ve dm “ ´mg dt .
Denoting with r “ ´dm{dt the rate of exhaustion, it follows that dt “ ´dm{r (for r ‰ 0),
and substituting in the above equation:
´v g¯
e
dv “ ´ ´ dm .
m r
If r is independent of time, we can easily integrate both sides, to get:
mi g
∆v “ ve ln ´ pmi ´ mf q . (15.18)
mf r
264 Chapter 15. Problem Solutions

If the gravitational field goes to zero, or the rate of fuel consumption goes to infinity, the
previous equation reduces to Eq.12.1, which describes a rocket moving in empty space.
Therefore, the gravitational field reduces the effective ∆v a rocket can deliver. This is the
effect of gravity mentioned in Pr. 12.1. If ve {m “ g{r, then dv “ 0, therefore the rocket
hovers at a constant height. If r “ 0, we must revert back to the equation expressing the
conservation of linear momentum to deduce that dv{dt “ ´g, as expected.

12.5 Reaching the centre

We assume the potential has the form:
α
U “´ .
rk
We want to find out what k needs to be for the body to reach the centre, with non-zero
angular momentum (L ě 0). The body reaches the centre if the radial velocity, for r Ñ 0, is
positive, i.e. if the radial kinetic energy is positive in this limit:
1 L2
mr9 2 “ E ´ U prq ´ ą 0.
2 2mr2
It follows that r can take any arbitrary small value only if:
L2 L2
rr2 U prqsrÑ0 ă ´ ñ rr´k`2 srÑ0 ą .
2m 2mα
Let us distinguish three cases:

• if k ą 2, then 8 ą L2 {2mα is always true, therefore the body can reach the centre;
?
• if k “ 2, then 1 ą L2 {2mα and the body reaches the centre only if L ă 2mα;
• if k ă 2, the body never reaches the centre if L ‰ 0.

The last case corresponds to the gravitational potential (k “ 1). This is the reason why is
necessary to cancel the tangential speed of the probe (so that L “ 0) in order to reach the
centre of the Sun in Pr. 12.3.

12.6 Gravitational slingshot

Let v0 be the velocity of the probe with respect to the Sun once it has left the gravitational
field of the Earth. Let dC be the radius of the Earth’s orbit and dp “ xdC the radius of the
orbit of the planet which is used as a gravitational slingshot. Let vt and vr be the tangential
and radial velocities of the probe when it is at a distance dp from the Sun, with mass M
(Fig. 15.19). Applying conservation of energy and angular momentum:
1 GM m 1 GM m
mv0 2 ´ “ mpvt 2 ` vr 2 q ´
2 dC 2 xdC
v0
mdC v0 “ mxdC vt ñ vt “ .
x
Substituting vt , obtained from the second equation into the first, and isolating vr :
c ´
1¯ GM ´ 1¯
vr “ v0 2 1 ´ 2 ´ 2 1´ .
x dC x
The velocities of the Earth and of the planet are, respectively:
GmM vt 2 GM
2
“m ñ vC 2 “ ,
dC dC dC
GmM vp 2 1
“ m ñ vp 2 “ vC 2 .
d2C x2 xdC x
 265

Hence, vr can be written as:

c ´ 1¯ ´ 1¯
vr “ v0 2 1 ´ 2 ´ 2vC 1 ´ .
x x
Requiring that the term under the square root is positive, we have a condition on v0 :
c
v0 2x
ě gpxq “ .
vC 1`x
If v0 {vC were smaller than gpxq, the probe wouldn’t have enough energy to reach the planet.
Let us now consider the reference system of the planet. We assume that the gravitational
field of the Sun can be neglected close to the planet. Let vrel be the relative velocity between
the probe and the planet when the probe is just entering its sphere of influence:
a
vrel “ pvt ´ vp q2 ` vr 2 .

The probe will travel on a hyperbolic orbit, with speed at infinity vrel and some angle of
deviation (Fig. 15.20). In the reference system of the planet, it appears not much has
happened; however, in the frame of the Sun, the velocity of the probe has increased. In the
best case, the probe is deviated in such a way that its velocity, upon exiting the sphere of
influence of the planet, is directed along the velocity of the planet. Therefore, in the frame
of the Sun, the final velocity of the probe is vf “ vrel ` vp?. If this velocity is greater than the
escape velocity from the Sun at that distance, which? is 2vp , then the probe will leave the
Solar System. The condition is therefore vrel ě p 2 ´ 1qvp . Substituting vrel as a function
of vt , vp and vr , we obtain:
d
?
v0 1 1 8
ě ? ` 3
` 2 ´ ” f pxq .
vC x 3 x x
We therefore have two constraints on the velocity v0 . The first ensures the probe does indeed
reach the planet, while the second is needed for the planet to escape the Solar System.

Figure 15.19: Applying conservation of energy and angular momentum, it is possible to find
the components of the velocity of the probe when it reaches the other planet.
266 Chapter 15. Problem Solutions

Figure 15.20: In the reference system of the planet, the probe travels on an hyperbolic orbit,
with velocity v~rel at infinite distance. The planet deflects the path of the probe, therefore
changing its velocity in the reference system of the Sun.

? functions on the same graph, we see that f pxq is always greater than gpxq for
Plotting both
x ă 2 `? 8 (Fig. 15.21). At this point, the two functions have the same value, and, for
x ą 2 ` 8, gpxq is greater than f pxq. The minimum velocity is therefore:
$ d
?
’
’ 1 1 8 ?
&? ` 2 for 1 ď x ď 2 ` 8
’
v0 3
` ´
“ cx 3 x x
vC ’
’ 2x ?
’
% for x ą 2 ` 8 .
1`x
Solving f pxq1 “ 0, we find the planet that gives the lowest possible velocity v0 :
b
´ ? ¯
x “ 9 ` 81 ´ 24 8 {8 “ 1.58 .

Surprisingly, this value is very close to the Sun-Mars distance (1.52 au). Therefore, Mars is
the best planet to use as a gravitational slingshot.

Figure 15.21: The velocity must be greater than the maximum of f pxq and gpxq.
Appendix

267
 A

Mathematics

A.1 Vectors

The simplest physical quantity one can imagine is the scalar, specified only by
its magnitude, a number complete with units in which it is measured. Examples
of scalar quantities are temperature, energy, time and density.
A vector is a quantity that requires both a magnitude, and a direction in space
to specify it completely; we may think of it as an arrow in space. Examples of
vectors are velocity, acceleration, force, linear and angular momentum.

Addition and Subtraction

There are two ways to sum vectors. With the parallelogram method, we trans-
late the first vector in such a way that its tail coincides with the tail of the
second vector. As you can see in Fig. A.1 a), the vector sum is the diagonal of
the parallelogram formed by these two vectors. With the head-to-tail method,
we translate the first vector in such a way that its tail coincides with the head
of the second vector. In Fig. A.1 b), the resultant vector is the vector whose
head and tail are the head and tail of the first and second vector, respectively.
The sum of two vectors is associative, i.e. ~a ` ~b “ ~b ` ~a. Subtraction is similar
to addition, since ~a ´ ~b “ ~a ` p´~bq, where ´~b is the vector with the same
magnitude as ~b, but opposite direction.

Figure A.1: Vector addition using: a) parallelogram and b) head-to-tail methods.

268
A.1. Vectors 269

Figure A.2: Multiplication of a vector ~a by a scalar λ.

Multiplication by a scalar
Not to be confused with the scalar product, the multiplication of a vector ~a
by a scalar λ returns a vector with magnitude |λ||~a| (where |~a| denotes the
magnitude of ~a), and the same direction as ~a if λ ą 0, otherwise opposite
direction (Fig. A.2). The following properties hold:
pλµq~a “ λpµ~aq “ µpλ~aq ,
λp~a ` ~bq “ λ~a ` λ~b ,
pλ ` µq~a “ λ~a ` µ~a .

Basis vectors and components

Given three distinct and non-coplanar vectors u~1 , u~2 , u~3 , it is possible to write
any other vector ~a in space as:
~a “ a1 u~1 ` a2 u~2 ` a3 u~3 .
The three vectors u~1 , u~2 , u~3 are said to form a basis, and the scalars a1 , a2 , a3
are called the components of the vector ~a with respect to the basis. We say that
the vector has been resolved into components. The basis is called orthogonal
if u~1 , u~2 , u~3 are all perpendicular to each other; orthonormal if, additionally,
all three vectors have unit modulus. An example of an orthonormal basis is
the Cartesian basis, specified by three unit vectors î, ĵ, k̂ which point along the
x, y, z axes, respectively. We can therefore write ~a as:
~a “ ax î ` ay ĵ ` az k̂ , (A.1)
where ax , ay , az are the projections of ~a on the axes x, y, z, respectively. Using
a Cartesian basis, it is easy to write down vector addition:
~a ` ~b “ pax î ` ay ĵ ` az k̂q ` pbx î ` by ĵ ` bz k̂q
“ pax ` bx qî ` pay ` by qĵ ` paz ` bz qk̂ .
Multiplication of ~a by a scalar can be written as:
λ~a “ λpax î ` ay ĵ ` az k̂q “ pλax qî ` pλay qĵ ` pλaz qk̂ .
270 Appendix A. Mathematics

Magnitude of a vector
The magnitude of a vector ~a, denoted by |~a|, can be written in terms of its
Cartesian components as:
b
|~a| “ a2x ` a2y ` a2z , (A.2)

which can be viewed as the Pythagorean theorem in three dimensions. A vector

whose magnitude equals unity is called a unit vector. The unit vector in the
direction ~a is usually denoted by â, and may be evaluated as:
~a
â “ . (A.3)
|~a|

Scalar product

Figure A.3: The scalar product ~a ¨ ~b equals the magnitude of ~a multiplied by the projection
of ~b onto ~a, or vice versa.

The scalar product of two vectors ~a and ~b, denoted by ~a ¨~b, is the scalar:

~a ¨ ~b “ |~a||~b| cos θ , (A.4)

where θ is the angle between the two vectors, placed tail to tail or head to
head. Thus, the scalar product ~a ¨ ~b equals the magnitude of ~a multiplied by
the projection of ~b onto ~a, or vice versa (Fig. A.3). If ~a is perpendicular to ~b,
their scalar product is zero. If ~a and ~b are parallel to each other and point in
the same direction, their scalar product is |~a||~b|, otherwise it is ´|~a||~b|, if their
direction is opposite. An example of scalar product is the work W “ F~ ¨~r.

Vector product
The vector product of ~a and ~b, denoted by ~a ˆ ~b, is the vector perpendicular
to both ~a and ~b, with magnitude:

~a ˆ ~b “ |~a||~b| sin θ , (A.5)

A.2. Conic Sections 271

where θ is the angle between the two vectors, placed tail to tail or head to
head (Fig. A.4). The direction can be found with the right hand rule: if the
thumb is pointed in the direction of the first vector, the index in the direction
of the second vector, then the middle finger gives the direction of their vector
product. If we exchange the order of the two vectors, their vector product
changes sign, as you can verify with the right hand rule. Therefore, the vector
product is anti-commutative: ~a ˆ ~b “ ´~b ˆ ~a . From its definition, we see that
the vector product has the very useful property that if ~a ˆ ~b “ 0, then either
~a is parallel or anti-parallel to ~b (unless either of them is zero). We also note
that ~a ˆ ~a “ 0. An example of vector product is the torque ~τ “ F~ ˆ ~r.

Figure A.4: The vector product of ~a and ~b, denoted by ~a ˆ ~b, is the vector perpendicular to
both ~a and ~b, with magnitude equal to the area A of the parallelogram formed by the two
vectors. The direction can be found with the right hand rule.

A.2 Conic Sections

Figure A.5: a) Conic sections, b) shape of the corresponding orbits.

272 Appendix A. Mathematics

As the name suggests, conic sections are curves obtained from the intersection
of a circular cone with a plane (Fig. A.5).

Circle
The circle is the locus of points in the plane that are equidistant from a fixed
point, called the centre. Let r be the radius of the circle, i.e. the constant dis-
tance of any point P“ px, yq on the circle to the centre C“ p0, 0q. Then:

x2 ` y 2 “ r 2 .

Ellipse
The ellipse is the locus of points in the plane such that the sum of the distances
from two fixed points, called foci, is constant. Consider a Cartesian coordinate
system and let the position of the two foci be F1 “ p´c, 0q and F2 “ pc, 0q.
Take a point P“ px, yq on the ellipse, and let 2a be the sum of the distances
of P from the two foci. Then:
a a
2a “ px ` cq2 ` y 2 ` px ´ cq2 ` y 2 .

Moving the second square root to the LHS and squaring both sides:
a
p 2a ´ px ´ cq2 ` y 2 q2 “ px ` cq2 ` y 2
a
4a2 ` px ´ cq2 ` y 2 ´ 4a px ´ cq2 ` y 2 “ px ` cq2 ` y 2
a
4a2 ´ 4xc “ 4a px ´ cq2 ` y 2
c a
a ´ x “ px ´ cq2 ` y 2 .
a
Squaring again both sides:
´ c ¯2
a2 ` x2 ´ 2xc “ x2 ` c2 ´ 2xc ` y 2
a
´ c ¯2
a2 ´ c2 “ ´x2 ` x2 ` y 2
a
” ´ c ¯2 ı
a2 ´ c2 “ x2 1 ´ ` y2 .
a
For simplicity, let us define b2 “ a2 ´ c2 . The above equation becomes:
b2
x2 ` y 2 “ b2 .
a2
Dividing both sides by b2 , we obtain the canonical equation of an ellipse:

x2 y 2
` 2 “1 . (A.6)
a2 b
A.2. Conic Sections 273

Figure A.6: a) Schematics of an ellipse, b) the degree of flattening along the semi-major axis
increases with increasing eccentricity.

Here, a is the semi-major axis, b the semi-minor axis and c the semi-focal
distance, as shown in Fig. A.6 a). The ratio between the semi-focal distance
and the semi-major axis is called the eccentricity of the ellipse, and is denoted
by e:
c
e“ with 0 ă e ă 1 .
a
The eccentricity is always contained in the range 0 to 1, and measures the
degree of flattening along the major axis. As you can see in Fig. A.6 b),
if a “ b, then e “ 0, and the ellipse reduces to a circle. With increasing
eccentricity, the ellipse is more and more flattened along the major axis until,
in the degenerate case e Ñ 1, it becomes a segment connecting the two foci.
The area of a sector on an ellipse between the major axis and a segment forming
an angle θ with it is:
1 ´a ¯
Apθq “ a b arctan tan θ . (A.7)
2 b
Which is easy to prove by thinking of an ellipse as a circle scaled by factors
a and b along the x and y axes, respectively. The total area of an ellipse is:

A “ πa b . (A.8)

Parabola
The parabola is the locus of points in the plane that are equidistant from a
fixed line, called directrix, and a fixed point, called focus (Fig. A.7, overleaf).
The line passing through the focus and perpendicular to the directrix is the
axis of symmetry of the parabola.
Consider a Cartesian coordinate system and let F“ pp, 0q be the position of the
274 Appendix A. Mathematics

Figure A.7: Parabola

focus and x “ ´p the equation defining the directrix. Taking a point P“ px, yq
on the parabola, we have:
a
px ´ pq2 ` y 2 “ x ` p .

Squaring both sides:

x2 ` p2 ´ 2xp ` y 2 “ x2 ` p2 ` 2px

Hence, the canonical equation for a parabola, with focus in F“ pp, 0q and
directrix x “ ´p, is:
y 2 “ 4px . (A.9)
An important property of the parabola is that any incident ray parallel to its
axis of symmetry is reflected in its focus.

Hyperbola
The hyperbola is the locus of points in the plane such that the difference in the
distances from two fixed points, called foci, is constant (Fig. A.8). Consider
a Cartesian coordinate system and let F1 “ p´c, 0q and F2 “ pc, 0q be the
position of the two foci. Take a point P=px, yq on the hyperbola, and let 2a
be the difference of the distances from the two foci. Then:
ˇa a ˇ
2a “ ˇ px ` cq2 ` y 2 ´ px ´ cq2 ` y 2 ˇ ,

where we have taken the modulus of the difference since we don’t know, a
priori, which distance is greater. Moving the second square root to the LHS,
A.2. Conic Sections 275

Figure A.8: Hyperbola

squaring both sides, and considering both positive and negative cases:
a
p2a ˘ px ´ cq2 ` y 2 q2 “ px ` cq2 ` y 2
a
4a2 ` px ´ cq2 ` y 2 ˘ 4a px ´ cq2 ` y 2 “ px ` cq2 ` y 2
a
˘4a px ´ cq2 ` y 2 “ ´4a2 ` 4xc
a c
˘ px ´ cq2 ` y 2 “ x ´ a .
a
Squaring both sides again:
´ c ¯2
a2 ` x2 ´ 2xc “ x2 ` c2 ´ 2xc ` y 2
a
´ c ¯2
x2 ´ x2 ´ y 2 “ c2 ´ a2
a
”´ c ¯2 ı
x2 ´ 1 ´ y 2 “ c2 ´ a2 .
a
For simplicity, define b2 “ c2 ´ a2 (note that c and a are swapped compared
to the ellipse, since c ą a). The above equation becomes:

b2
x2 ´ y 2 “ b2 .
a2
Dividing both sides by b2 , we finally obtain the canonical equation for the
hyperbola:
x2 y 2
´ 2 “1 . (A.10)
a2 b
The hyperbola has eccentricity e ą 1, since c ą a. The asymptotes are de-
scribed by:
b
y“˘ .
a
276 Appendix A. Mathematics

A.3 Plane Trigonometry

Measuring Angles
The radian is the SI unit for measuring angles. Given a circle of radius r, the
radian is defined as the angle at the centre subtended by an arc of length r.
Therefore, an arc of length s subtends an angle of:
s
αrad “ .
r
A full circle, measured in radians, is equal to the circumference divided by the
radius, i.e. αc, rad “ 2πr{r “ 2π. On the other hand, a full circle measured
in degrees is αc, deg “ 360˝ , hence the conversion from degrees to radians is:

αrad αdeg π
“ ˝
ñ αrad “ αdeg . (A.11)
2π 360 180
Similarly, the steradian is the SI unit for measuring solid angles. Given a
sphere of radius r, the steradian is numerically equal to the solid angle at the
centre subtended by an area of r2 on the sphere. The solid angle subtended
by an element of area A, at a distance r, is thus:
A
ωsr “ 2 .
r
Since the area of a sphere is 4πr2 , a full sphere measures 4π steradians.

Sine, cosine and tangent

Trigonometry is useful in physics to describe periodic motion. Given a right
triangle (Fig. A.9), we define three trigonometric functions:
• the sine of an angle, denoted by “sin”, is the ratio of the leg opposite the
angle to the hypotenuse;
• the cosine of an angle, denoted by “cos”, is the ratio of the leg adjacent
the angle to the hypotenuse;
• the tangent of an angle, denoted by “tan”, is the ratio of the leg opposite
the angle to the adjacent leg.

Figure A.9: Definition of sine, cosine and tangent.

A.3. Plane Trigonometry 277

Hence, we can write:

b “ a sin β , c “ a sin γ ;
b “ a cos γ , c “ a cos β ;
b “ c tan β , c “ b tan γ .

Pythagorean identity
The sum of the sine squared and cosine squared of any angle is equal to unity.
Indeed, looking at Fig. A.9, the lengths of the opposite and adjacent legs to
β are equal to b “ a sin β and c “ a cos β. According to the Pythagorean
theorem, a2 “ b2 ` c2 . Hence, substituting for b and c, dividing both sides by
a2 , we find:
sin2 β ` cos2 β “ 1 . (A.12)

Law of Sines

Figure A.10: Proof of the law of sines.

Consider a generic triangle with sides of length a, b, c and angle opposite those
sides of α, β, γ, respectively. Let R be the radius of the circumference that
circumscribes the triangle. The law of sines states that:

a b c
“ “ “ 2R . (A.13)
sin α sin β sin γ

Let us prove that c{ sin γ “ 2R; the other equations follow by symmetry.
Looking at Fig. A.10, let H be the height of triangle ABO relative to the
base AB. Since the angle at the circumference is half the angle at the centre
278 Appendix A. Mathematics

subtended by the same arc, we have =AOB “ 2γ. Because triangles AHO
and HBO are both right triangles in H, AO “ BO “ R, and since they share
OH, it follows that they are congruent, hence =AOH “ γ. Then, AH “ c{2,
but we also know that AH “ R sin γ, hence c{ sin γ “ 2R, as we wanted to
show.

Law of Cosines

Figure A.11: Proof of the law of cosines.

The law of cosines allows us to compute the length of one side of a triangle,
knowing the length of the other two and the angle between them. In particular:

a2 “ b2 ` c2 ´ 2b c cos α . (A.14)
By symmetry, we can substitute a Ñ b, b Ñ c, c Ñ a and α Ñ β to obtain b2 “
c2 ` a2 ´ 2c a cos β and, with a similar transformation, c2 “ a2 ` b2 ´ 2a b cos γ.
Swapping every index with the next (imagine they are placed on a circle) is
called performing a cyclic permutation. We only prove Eq. A.14; the others
follow by cyclic permutation.
Looking at Fig. A.11, we draw the height relative to the base AC and we call
H the point of intersection. We have:

AH “ c cos α , BH “ c sin α .

It follows that HC “ b´c cos α. Applying the Pythagorean theorem to triangle

HCB:
2 2
a2 “ BH ` HC “ pc sin αq2 ` pb ´ c cos αq2
“ c2 psin2 β ` cos2 βq ` b2 ´ 2b c cos α .

Using the Pythagorean identity:

a2 “ b2 ` c2 ´ 2b c cos α ,

which is what we wanted to prove.

A.3. Plane Trigonometry 279

Reduction Formulae
In this section we look for a simple way to memorize the value of trigonometric
functions for some common angles. Since the sine and cosine have a period
of 2π (i.e. a full circle), we only need to know the value of trigonometric
functions for angles in the interval r0, 2πs. We can do even better: from the
trigonometric function of an angle α in r0, π{4s, we would like to find the
trigonometric functions for:
π 3
˘ α, π ˘ α, π ˘ α, ´α .
2 2
The strategy is to visualize the angles on the trigonometric circle (Fig. A.12),
applying rotations or reflections in order to reduce the problem to that of an
angle in the interval r0, π{4s. We then find:
` ˘ ` ˘
• for π2 ˘ α, sin π2 ˘ α “ cos α , cos π2 ˘ α “ ¯ sin α ;
` ˘ ` ˘
• for π ˘ α, sin π ˘ α “ ¯ sin α , cos π ˘ α “ ´ cos α ;
` ˘ ` ˘
• for 32 π ˘ α, sin 32 π ˘ α “ ´ cos α , cos 32 π ˘ α “ ¯ sin α ;
• for ´α, sin p´αq “ ´ sin α , cos p´αq “ cos α .

Figure A.12: By visualizing the angles on the trigonometric circle, and applying rotations
or reflections, it is possible to reduce the problem of finding the trigonometric function of an
angle in r0, 2πs, to that of an angle in r0, π{4s.
280 Appendix A. Mathematics

Looking at Fig. A.12, let us compute, for instance, the sine, cosine and tangent
of π ´ α. Clearly, the two angles have the same sine, hence sin pπ ´ αq “ sin α.
Instead, the cosine has opposite sign (because the corresponding triangles are in
the first and second quadrant, respectively), hence cos pπ ´ αq “ ´ cos α. The
tangent is equal to the ratio of sine and cosine, therefore tan pπ ´ αq “ ´ tan α.
With a similar reasoning, it is possible to obtain all the above relations.
For completeness, we give the trigonometric values of some common angles in
the interval r0, π{4s:
?
˝ 3 1 1
cos 30 “ , sin 30˝ “ , tan 30˝ “ ? ;
2 2 3
1 1
cos 45˝ “ ? , sin 45˝ “ ? , tan 45˝ “ 1 .
2 2
Using the reduction formula for π{2 ´ α, and taking α “ 30˝ , we can then
compute the trigonometric functions for 60˝ :
?
˝ 1 ˝ 3 ?
cos 60 “ , sin 60 “ , tan 60˝ “ 3 .
2 2

Trigonometric Identities
Below, we give some of the most useful trigonometric identities, which should
be remembered.

Addition Formulae:
` ˘
sin α ˘ β “ sin α cos β ˘ cos α cos β , (A.15)

` ˘
cos α ˘ β “ cos α cos β ¯ sin α sin β , (A.16)

` ˘ tan α ˘ tan β
tan α ˘ β “ . (A.17)
1 ¯ tan α tan β
Double Angle Formulae: Setting α “ β in the addition formulae:

sin 2α “ 2 sin α cos α , (A.18)

cos 2α “ cos2 α ´ sin2 α , (A.19)

2 tan α
tan 2α “ . (A.20)
1 ´ tan2 α
A.4. Spherical Trigonometry 281

Half-angle Formulae: Using the Pythagorean identity in Eq. A.19, setting

α Ñ α{2: c
α 1 ´ cos α
sin “ , (A.21)
2 2
c
α 1 ` cos α
cos “ . (A.22)
2 2

A.4 Spherical Trigonometry

Figure A.13

Let ABC be a triangle on a sphere of unit radius, with sides of length a, b, c

and angles opposite the respective sides of A, B and C (Fig. A.13).

Spherical Law of Cosines

The spherical law of cosines states that:

cos c “ cos a cos b ` sin a sin b cos C , (A.23)

and the corresponding cyclic permutations. In the case C “ 90 ˝ , we obtain

the Pythagorean theorem on a sphere:

cos c “ cos a cos b . (A.24)

If the triangle is small, i.e. a, b, c ! 1, we can use the approximation cos x «

1 ´ x2 {2 for x ! 1 (see Sec. A.6), to obtain the Pythagorean theorem on a
282 Appendix A. Mathematics

plane:

c2 ´ a2 ¯´ b2 ¯
1´ « 1´ 1´
2 2 2
c2 a2 b2
1´ «1´ ´
2 2 2
2 2 2
ñc «a `b ,

where we have neglected the term a2 b2 {4, since it is an infinitesimal of higher

order.

Spherical Law of Sines

The spherical law of sines states that:

sin a sin b sin c

“ “ . (A.25)
sin A sin B sin C

If a, b, c ! 1, we can use the approximation sin x « x for x ! 1, to obtain the

law of sines on a plane.
If the sphere has radius r, it is sufficient to substitute a Ñ a{r, b Ñ b{r and
c Ñ c{r in the previous formulae.

Spherical Excess
The sum of the angles of a spherical triangle is always greater than 180˝ . The
quantity:
E “ A ` B ` C ´ 180˝ , (A.26)
is called the spherical excess. It is not a constant, but depends on the triangle.
The area of a spherical triangle is related to the spherical excess in a very
simple way:
A “ Er2 (A.27)
A.5. Special Functions 283

A.5 Special Functions

Exponentials
Let us consider a function of the form f pxq “ ax , where a ą 0, called the
exponential function. We can distinguish three cases:
• a “ 1, then f pxq “ 1 for every x;
• a ą 1, as in Fig. A.14 a);
• 0 ă a ă 1, as in Fig. A.14 b).
The exponential function has some important properties:
• f p0q “ 1 for every a. Indeed a0 “ 1 for every a;
• f pxq ą 0 for every x. Indeed a ą 0 implies ax ą 0.
In the case a “ e, where e is Euler’s number, we obtain the natural exponential
function. The functions f pxq “ ex and f pxq “ e´x are shown in Fig. A.14 c).

Logarithms
The logarithmic function has the form f pxq “ loga x, where a ą 0 and a ‰ 1.
The meaning of this notation is the following:

f pxq “ loga x if and only if af pxq “ x . (A.28)

Clearly, the case a “ 1 is not valid because 1f pxq is always equal to unity. Since
a ą 0, it follows that af pxq ą 0 for every x, i.e. the function is only defined for
a positive argument (the domain is x ą 0).
The graphs of the logarithmic function for a ą 1 and 0 ă a ă 1 are shown in
Figs. A.14 d) and e), respectively. As shown in Fig. A.14 f), the logarithmic
function is simply a reflection of the exponential function about the line f pxq “
x. In other words, the two functions are related by an exchange of axes: x
becomes f pxq, and f pxq becomes x.
The logarithmic function has some important properties, which can be proven
by using A.28 and reducing the problem to one involving exponentials:
• logarithm of a product: loga px ¨ yq “ loga x ` loga y;
• exponent rule: loga pxy q “ y loga x;
• logarithm of a fraction: loga px{yq “ loga x ´ loga y;
• change of basis: loga x “ logy x{ logy a.
284 Appendix A. Mathematics

Figure A.14: a), b) and c) exponential functions; d), e) logarithmic functions. In f) we see
that the logarithmic function is simply a reflection of the exponential function about the line
f pxq “ x.
A.6. Derivatives 285

A.6 Derivatives

In physics and astronomy, differential calculus is of fundamental importance.

In this section we introduce the concept of a derivative in a straightforward
and intuitive manner.
Let f pxq be a function with real argument. Consider two points P1 “ px1 , f px1 qq
and P2 “ px1 ` ∆x, f px1 ` ∆xqq on f pxq. The line passing through P1 and
P2 is a secant of f pxq but, as we make ∆x smaller, it becomes closer to the
line tangent to f pxq in P1 . The angular coefficient of this line is ∆f {∆x, with
∆f “ f px ` ∆xq ´ f pxq. Thus, for ∆x infinitesimally small, ∆f {∆x tends to
the angular coefficient of the line tangent to f pxq in P1 . Instead of using the
notation ∆ for a generic variation, we will use the letter d to denote an infinites-
imal variation. Hence, ∆x becomes dx and ∆f becomes df “ f px ` dxq ´ f pxq.
In this limit, the angular coefficient of the line tangent to f pxq in x, which we
denote with f 1 pxq, is called the first derivative of f pxq in x, and is given by:

f px ` ∆xq ´ f pxq df
f 1 pxq “ lim “ . (A.29)
∆xÑ0 ∆x dx
Qualitatively, the derivative measures the rate of variation of a function, as x
is changed. The greater the derivative, the stronger the growth is; the smaller
the derivative, the faster the decrease.
From the definition of the derivative, it is clear that the value of ∆f for a small
(but non-infinitesimal) displacement ∆x is:
df
∆f « ∆x .
dx
The accuracy of the above relation increases as ∆x becomes smaller. For ∆x
that goes to zero, we can substitute ∆x and ∆f with dx and df , respectively:

df
df “ dx . (A.30)
dx

In the same way, we can define the second derivative of a function f pxq with
respect to x, which we denote by f 2 pxq, as the angular coefficient of the line
tangent to f 1 pxq in x:

f 1 px ` ∆xq ´ f 1 pxq df 1 pxq

f 2 pxq “ lim “ . (A.31)
∆xÑ0 ∆x dx
A physical example of a second derivative is the acceleration, equal to the
derivative of velocity w.r.t. time, with velocity being the derivative of distance
w.r.t time. Hence, acceleration is the second derivative of distance w.r.t. time.
Since we often consider derivatives w.r.t. time, it is convenient to define x, 9 x
:
286 Appendix A. Mathematics

as a short-hand notation for dx{dt and d2 x{dt2 , respectively.

We can then define the n-th derivative of a function as:
f pn´1q px ` ∆xq ´ f pn´1q pxq
f pnq pxq “ lim . (A.32)
∆xÑ0 ∆x
In this notation, we have f 1 pxq “ f p1q pxq, f 2 pxq “ f p2q pxq, etc. and, formally,
f p0q pxq “ f pxq.

Basic derivatives
All derivatives can be obtained by solving the limit A.29. In most cases, the
limit can be solved by simply thinking what quantities can be neglected as ∆x
goes to zero. Let us calculate the derivative with respect to x of f pxq “ x2 :

df pxq f px ` ∆xq ´ f pxq Function Derivative

f 1 pxq “ “ lim
dx ∆xÑ0 ∆x k 0
px ` ∆xq2 ´ x2
“ lim xα αxα´1
∆xÑ0 ∆x
2x∆x ` p∆xq2 ex ex
“ lim
∆xÑ0 ∆x log x 1{x
“ lim p2x ` ∆xq . sin x cos x
∆xÑ0
cos x ´ sin x
As ∆x goes to zero, 2x ` ∆x is just 2x:
tan x 1{ cos2 x
1 ?
f pxq “ 2x .
arcsin x 1{ 1 ´ x2
?
It is useful to memorize the basic derivatives in arccos x ´1{ 1 ´ x2
Tab. A.1. Using the techniques explained in the
arctan x 1{p1 ` x2 q
next section, they can be used to find the deriv-
atives of more complex functions. Table A.1

Product Rule
We want to calculate the derivative of a function f pxq which can be written
as the product of two or more functions: f pxq “ upxq vpxq. For example,
if f pxq “ x3 sin x, we see that upxq “ x3 and vpxq “ sin x. Of course, the
separation is not unique: we could have chosen upxq “ x2 , vpxq “ x sin x or
even upxq “ x4 tan x, vpxq “ cos x{x. However, the point of separating f pxq is
to find two functions, upxq and vpxq, whose derivatives are simpler to calculate.
Ultimately, we would like to trace back the calculation of the derivative of any
function to the basic derivatives. Since f pxq “ upxq vpxq, it follows that:

f px ` ∆xq ´ f pxq “ upx ` ∆xqvpx ` ∆xq ´ upxqvpxq

“ upx ` ∆xqrvpx ` ∆xq ´ vpxqs ` rupx ` ∆xq ´ upxqsvpxq .
A.6. Derivatives 287

From the definition A.29:

df f px ` ∆xq ´ f pxq
“ lim
dx ∆xÑ0 ∆x
! ” vpx ` ∆xq ´ vpxq ı ” upx ` ∆xq ´ upxq ı )
“ lim upx ` ∆xq ` vpxq .
∆xÑ0 ∆x ∆x
As ∆x becomes infinitesimally small, the terms in the square brackets become
du{dx and dv{dx, respectively, while upx`∆xq becomes upxq. Therefore:

df d dvpxq dupxq
“ rupxqvpxqs “ upxq ` vpxq ,
dx dx dx dx
which can be written as:

f 1 “ puvq1 “ uv 1 ` u1 v (A.33)

This equation is known as the product rule. Hence, the initial problem is
reduced to:
d 3 d d 3
px sin xq “ x3 psin xq ` px q sin x “ x3 cos x ` 3x2 sin x .
dx dx dx
We can extend the product rule to three or more separations:
d du
f pxq “ upxqvpxqwpxq ñ f 1 pxq “ u pvwq ` vu ,
dx dx
where we have used Eq. A.33, applied to u and vw. Using again Eq. A.33 to
expand pvwq1 :
d dw dv du
puvwq “ uv `u w` vw ,
dx dx dx dx
which can be written as:

puvwq1 “ uvw1 ` uv 1 w ` u1 vw . (A.34)

It is clear that the same reasoning can be extended to the product of any
number of terms.

Chain Rule
Consider a function of a function f pupxqq. For example, if f pxq “ p3 ` x2 q3 ,
we can take upxq “ 3 ` x2 , so that f pxq “ upxq3 . If ∆f , ∆u and ∆x are finite,
we can write:
∆f ∆f puq ∆upxq
“ .
∆x ∆upxq ∆x
288 Appendix A. Mathematics

When the above quantities become infinitesimally small, we can write:

df df puq dupxq
“ . (A.35)
dx dupxq dx

The above equation is known as the chain rule, and can be used to compute
the derivative of a function of a function. We can now solve the example given
at the beginning of the section:

df df puq du ” d 3 ı ” d ı
“ “ pu q ¨ p3 ` x2 q “ 3u2 ¨ 2x “ 6xp3 ` x2 q2 .
dx du dx du dx
In a similar way, we can use the chain rule to compute the derivative of the
reciprocal of a function. Let f pxq “ 1{vpxq “ vpxq´1 . Applying Eq. A.35:

df 1 dv
“´ 2 . (A.36)
dx v dx

Quotient Rule
Using the product rule and the chain rule (Eqs. A.33, A.35), it is easy to find
the derivative of the quotient of two functions:
´ u ¯1 ´ 1 ¯1 ´1¯ ´ v 1 ¯ u1
f1 “ “u ` u1 “u ´ 2 ` .
v v v v v
The last equation can be written as:
´ u ¯1 vu1 ´ uv 1
f1 “ “ , (A.37)
v v2

which is often referred to as the quotient rule. For example, the derivative
of f pxq “ sin x{x can be found by taking u “ sin x and v “ x, from which
u1 “ cos x and v 1 “ 1. Applying A.37:
x cos x ´ sin x cos x sin x
f 1 pxq “ 2
“ ´ 2 .
x x x

Special Points
We have seen that the derivative measures the rate of change of a function f pxq
in the neighbourhood of a point. If the increment of f pxq in the neighbourhood
of x0 is zero, we call this a stationary point. Stationary points can be divided
into three categories.
A.6. Derivatives 289

Figure A.15: In B, D and E the first derivative is zero: these are called stationary points. In
particular, B and D are points of minimum and maximum, respectively (collectively known
as turning points), while E is a point of inflection.

Referring to Fig. A.15, B is a point of minimum, since the function increases

in both directions moving away from that point; D is a point of maximum,
since the function decreases in both directions away from that point; finally, E
is an inflection point, as the function increases on one side of E and decreases
on the other. Note that B and D are not absolute maxima or minima, but,
rather, local maxima and minima. From a mathematical point of view, we
can distinguish maxima, minima and inflection points by studying the second
derivative of the function at those points. By definition, all stationary points
are such that df {dx “ 0. Looking at the minimum B, we see that df pxq{dx is
negative in A and positive in C, passing through zero in B. Therefore, in the
neighbourhood of B, the first derivative is increasing, hence the second deriv-
ative (rate of change of the first derivative) is positive. Similarly, the second
derivative must be negative in D. It is less obvious that in E the second deriv-
ative is zero. Intuitively, we note that, on the left of E, the function is concave
upwards, therefore the first derivative is increasing and the second derivative
is positive. Similarly, on the right of E, the function is concave downwards,
therefore the second derivative is negative. Since the second derivative varies
continuously, it must be zero in E.
To sum up, stationary points are characterized by df {dx “ 0 and:
• d2 f {dx2 ą 0, for a minimum;
• d2 f {dx2 ă 0, for a maximum;
• d2 f {dx2 “ 0 for an inflection point, and d2 f {dx2 changes sign through
the point.
290 Appendix A. Mathematics

Taylor Series
The Taylor series allows us to approximate a function, in the neighbourhood of
a point, with a polynomial of degree n. The general form of the Taylor series is:

f 2 px0 q 2 f 3 px0 q 3
f px ` x0 q “ f px0 q ` f 1 px0 qx ` x ` x ` .. (A.38)
2! 3!
The validity of the above formula can be checked with the following reasoning.
Taking the first derivative with respect to x and setting x “ 0, we obtain f 1 px0 q
on both sides. Indeed, the derivative of the LHS is f 1 px ` x0 q, which, in x “ 0,
is equal to f 1 px0 q. The derivative of the RHS is f 1 px0 q ` f 2 px0 qx ` .., hence, in
x “ 0, the only non-zero term is f 1 px0 q. Similarly, taking the second derivative
and setting x “ 0, we obtain f 2 pxq on both sides, and so on. Since both sides
are equal in x “ 0, and their n-th derivatives in x “ 0 are the same for every
n, it follows that they are the same function. Some examples of Taylor series
and their first order approximations are given in Tab. A.2.

Function Taylor series For x ! 1

1
1´x 1 ` x ` x2 ` x3 ` ... 1`x
ln p1 ` xq x ´ x2 {2 ` x3 {3 ´ ... x
ex 1`x` x2 {2! ` x3 {3! ` ... 1`x
sin x x ´ x3 {3! ` x5 {5! ´ ... x
cos x 1 ´ x2 {2! ` x4 {4! ´ ... 1
1 ` nx ` n2 x2 ` n3 x3 ` ...
` ˘ ` ˘
p1 ` xqn 1 ` nx
Table A.2

To obtain the first formula in the table, we need to compute the n´th deriv-
atives of 1{p1 ´ xq at x0 “ 0. Using the chain rule with upxq “ 1 ´ x:
1
f pxq “ ñ f p0q “ 1 ,
1´x
df du 1
f 1 pxq “ “ ñ f 1 p0q “ 1 ,
du dx p1 ´ xq2
df 1 pxq 1 2
f 2 pxq “ “ ´2 3
¨ p´1q “ ñ f 2 p0q “ 2 ,
dx p1 ´ xq p1 ´ xq3
6
f 3 pxq “ ñ f 3 p0q “ 6 ,
p1 ´ xq3
... ,
n!
f n pxq “ ñ f 1 p0q “ n! .
p1 ´ xqn
A.7. Integrals 291

Substituting the derivatives in Eq. A.38, with x0 “ 0:

f 2 p0q f n p0q n
f pxq “ f p0q ` f 1 p0qx ` ` ... ` x ` ...
2! n!
2 6 n!
“ 1 ` x ` ` ` ... ` xn ` ...
2! 3! n!
1
ñ “ 1 ` x ` x ` x3 ` ... ,
2
1´x
as we wanted to show.
We define the order of a power series as the minimum degree of the independent
variable. For instance, the order of x2 is 2, while the order of x2 ` 3x is 1.
The Taylor series allows us to approximate a function with a power series. For
x ! 1, we can often truncate the Taylor series to first order, since xn ! x for
n ą 1. For example, the first order approximation of 1{p1 ´ xq is obtained by
neglecting all terms of degree greater than 1, hence 1{p1 ´ xq « 1 ` x.

A.7 Integrals

Figure A.16: The shaded area under the curve is the integral of f pxq between the extremes
şb
a and b, and is denoted by a f pxq dx.

The notion of an integral as the area under a curve will already be familiar
to many. In Fig. A.16, where the function f pxq is drawn as a solid line, the
shaded area represents the quantity denoted by:
żb
I“ f pxq dx . (A.39)
a

This expression is known as the definite integral of f pxq between the lower
and upper limits x “ a and x “ b, respectively. Here, f pxq is called the
integrand.
292 Appendix A. Mathematics

Integration from basic principles

According to Riemann’s definition, the integral of a function f pxq between the
extremes a and b is the limit of the sum of the areas of the rectangles with
height f pxi q and width pxi ´ xi´1 q, for a “ x0 ă x1 ... ă xi ă ... xn “ b, as n
goes to infinity. Hence:
n
ÿ
S “ lim f pxi qpxi ´ xi´1 q . (A.40)
nÑ8
i“1

In this limit, with the constraint that the length of the interval rxi´1 , xi s goes
to zero (i.e. to dx), S converges, for a continuous function in a finite interval,
to the definite integral of f pxq. The symbol representing the integral (see Eq.
A.39) is a reminder of the S, which stands for summation.

Properties of the integral

Let a ă b ă c be three real numbers. The following properties hold:
• reversing the limits of integration changes the sign of the definite integral:
żb ża
f pxq dx “ ´ f pxq dx ;
a b

• if the upper and lower limits are the same, the integral is zero:
ża
f pxq dx “ 0 ;
a

• the definite integral of f pxq over the interval ra, cs is equal to the sum of
the integrals over the intervals ra, bs and rb, cs:
żc żb żc
f pxqdx “ f pxq dx ` f pxq dx ;
a a b

• the definite integral of the sum of two functions is equal to the sum of
the integrals of these functions:
żb żb żb
pf pxq ` gpxqq dx “ f pxq dx ` gpxq dx ;
a a a

• a constant factor can be moved outside the integral sign:

żb żb
c ¨ f pxq dx “ c f pxq dx .
a a
A.7. Integrals 293

How to solve integrals

Let us define the indefinite integral F pxq of f pxq as:
żx
F pxq “ f px1 q dx1 , (A.41)
x0

where x1 has been used to distinguish the independent variable from the ex-
treme of integration. Using the properties above, we find:
żb
f pxq dx “ F paq ´ F pbq . (A.42)
a

If we know F pxq, it is then possible to compute the definite integral of f pxq.

But how do we find F pxq?

Fundamental theorem of calculus

Figure A.17: F pxq is the indefinite integral, equal to the area bounded by f pxq between
rx0 , xs. Considering a small additional stripe between x and x ` dx, with width dx and area
f pxq dx, the change in F pxq is f pxq dx. Therefore, dF pxq “ f pxq dx, i.e. F 1 pxq “ f pxq. For
this reason, F pxq is also called the anti-derivative of f pxq.

The fundamental theorem of calculus states that F 1 pxq “ f pxq, i.e. the de-
rivative of F pxq is equal to f pxq. Since the derivative of a constant factor is
zero, if follows that F pxq is defined up to an additive constant. Integration
and differentiation are therefore inverse operations. This theorem is easy to
visualize graphically: for an infinitesimal variation dx, the increment dF pxq of
F pxq is equal to the area of the rectangle between x and x ` dx, with base dx
and height f pxq, i.e. dF pxq “ f pxqdx (Fig. A.17). By the definition of the
derivative, F 1 pxq “ dF pxq{dx. Hence dF pxq “ f pxq dx implies F 1 pxq “ f pxq,
as we wanted to show. Now that we know how to find F pxq, it is sufficient
to apply Eq. A.42 to find the definite integral of f pxq, as we will see in the
following examples.
294 Appendix A. Mathematics

şa
Let us calculate, for instance, the integral 0 x dx. We want to find the function
F pxq whose derivative is f pxq “ x. It is easy to see that F pxq “ x2 {2 ` c satis-
fies the required condition. We then use Eq. A.42 and calculate the difference
of F pxq, evaluated at the extremes:
ża
a2
x dx “ F paq ´ F p0q “ .
0 2
şa 2
As another example, let us compute the integral 0 2xe´x . It is easy to see
2 2
that F pxq “ ´e´x ` c is the anti-derivative of 2xe´x . Hence:
ża
2 2
2xe´x “ F paq ´ F p0q “ ´e´a ` 1 .
0

At times, we may want to compute an integral without specifying the extremes

of integration. This can be useful, for instance, to examine how the integral it-
self varies as the extremes are changed. In this case, it suffices to find F pxq and
write down Eq. A.41. Notice that the solution will carry an arbitrary constant,
which we call constant of integration.
For the previous examples:
ż f pxq F pxq
x dx “ x2 {2 ` c ,
ż k kx
2 2
2xe´x dx “ ´e´x ` c . n
x , n ‰ ´1 x n`1 {pn ` 1q
1{x ln x
Here, the extremes of integration are im-
plicitly x0 and x, where x0 is an arbitrary ex ex
constant related to c by F px0 q “ c. In sin x ´ cos x
Tab. A.3, we present some common indef-
cos x sin x
inite integrals, where the constant of integ- ?
ration has been omitted for clarity. These 1{ 1 ´ x2 arcsin x
?
are the basis for computing more complex 1{ 1 ´ x2 ´ arccos x
integrals, using the rules explained in the
1{1 ` x2 arctan x
next section.
Table A.3
Integration by substitution
The indefinite integral of simple integrands can be found immediately by in-
spection, as shown in the previous section. For more complex functions, we
can use some strategies to simplify the calculation.
If the integral does not appear to be elementary, the first strategy we can try is
integration by substitution. The idea is to define a new variable that simplifies
the integral,
ş possibly reducing it to an elementary one. For example, if we
are given f pxq dx, we may want to define a new variable u “ upxq, which
A.7. Integrals 295

ş
transforms the starting integral into one of the form gpuq du. To find du in
terms of dx, we use Eq. A.30. It is important to note that once the variable
has been changed and the new integral has been obtained, no reference to x
or dx should remain. Choosing the new variable is not always easy: below we
see some examples in order of difficulty.
ż
cosp2x ` 7q dx .

The choice upxq “ 2x ` 7 looks promising. Using Eq. A.30 to compute du:

du
du “ dx ñ du “ 2 dx .
dx
Substituting in the initial integral:
ż ż
1
cosp2x ` 7q dx “ cos u du ,
2
which is a basic integral:
ż
1 1
cos u du “ sin u ` c .
2 2

Reverting back to the expression for upxq, we then obtain:

ż
1
cosp2x ` 7q dx “ sin p2x ` 7q ` c .
2

Let us now calculate: ż

3
dx .
xpln xq2
After some trials, we see that the substitution upxq “ ln x will simplify the
integral. Then, du “ dx{x, and substituting in the integral:
ż ż ż
3 1 dx 1
2
dx “ 3 2
“3 du ,
xpln xq pln xq x u2

which is an elementary integral:

u´1
ż ż
1 ´2 3
3 2
du “ 3 u du “ 3 ` c “ ´ ` c.
u ´1 u
We can now substitute u in terms of x:
ż
3 3
2
dx “ ´ ` c.
xpln xq ln x
296 Appendix A. Mathematics

Integration by parts
Another strategy to simplify integrals is integration by parts. Using Eq. A.33
for the derivative of a product:
d dv du
puvq “ u ` v.
dx dx dx
Isolating the first term on the RHS:
dv d du
u “ puvq ´ v.
dx dx dx
Integrating both sides w.r.t. x:
ż ż
dv du
u dx “ uv ´ v dx ,
dx dx
which can be written as:
ż ż
1
uv dx “ uv ´ u1 v dx . (A.43)

The above equation is the rule of integration by parts. To integrate by parts,

we write the integrand as a product of two functions u and v 1 , so that the
integral of u1 v is easier to calculate compared to the initial integral. As an
example, let us calculate: ż
x cos x dx .

Substitution fails here, so we try integration by parts. Taking dv{dx “ x and

u “ cos x, we find v “ x2 {2 and du{dx “ ´ sin x. Applying Eq. A.43:

x2
ż ż
1
x cos x dx “ cos x ` x2 sin x dx .
2 2
This integral looks more complicated than the starting integral, so we try
another choice for u and v. If we take u “ x and dv{dx “ cos x, we have
du{dx “ 1 and v “ sin x, and it follows that:
ż ż
x cos x dx “ x sin x ´ sin x dx .

This is an elementary integral:

ż
x cos x dx “ x sin x ` cos x ` c .
 B

Kepler’s Laws

B.1 Solving the equation of motion

The velocity of a body orbiting around a centre of attraction can be decom-

posed into radial and tangential components:
dr
vr “ “ r9 ,
dt
dθ
vt “ rω “ r “ rθ9 .
dt
The energy can be written as:
1 1 GmM
E “ mr9 2 ` mr2 θ92 ´ . (B.1)
2 2 r
Angular momentum is also conserved:

L “ mr2 θ9 . (B.2)

We want an expression in which the only unknown is r. Hence, we isolate θ9 in

Eq. B.2, and substitute it in Eq. B.1:

1 L2 GmM
E “ mr9 2 ` 2
´ . (B.3)
2 2mr r
Isolating r:
9 c c
2 L2 GmM
r9 “ E´ ` . (B.4)
m 2mr2 r
From the above equation we can obtain rptq. Since we are interested in the
shape of the orbit, we want to find rpθq. Using the chain rule (Eq. A.35):

dr dr dθ dr L dr
r9 “ “ “ θ9 “ .
dt dθ dt dθ mr2 dθ

297
298 Appendix B. Kepler’s Laws

Hence, substituting in B.4 and reordering:

?
dr 2m
b “ dθ . (B.5)
r2 E ´ L2
` GmM L
2mr2 r

Since this equation is valid for any r and θ, we can imagine summing all
such equations separated by infinitesimal increments dθ. This sum is just an
integral: ż ż ?
dr 2m
b “ dθ . (B.6)
r2 E ´ 2mr L2
` GmM L
2 r

Now, it comes down to solve two integrals. The one on the RHS is straight-
forward: ż ? ?
2m 2m
dθ “ θ ` c1 .
L L
To solve the integral on the LHS, we change variable to x “ 1{r, so that
dx “ ´1{r2 dr: ż
1 dx
´? b .
E 2
1 ´ L x2 ` GmM x
2Em E

Completing the square at the denominator:

ż
1 dx
´ ? b
2 m3 M 2
.
E G L GM m
am
p1 ` 2L2 E q ´ p ?2Em x´ 2
L 2E q

Using the substitution:

c c
G2 m3 M 2 L GM m m
1` cos y “ ? x´ ,
2L2 E 2Em L 2E
c
G2 m3 M 2 L
1` sin y dy “ ´ ? dx ,
2L2 E 2Em
the previous integral simplifies to:
? ż ? ? „ ? L x ´ GM m a m 
2m 2m 2m 2Em L 2E
dy “ y ` c2 “ arccos b ` c2 .
L L L 1` G2 m3 M 2
2L2 E

The axes are often chosen so that the integration constant is zero. Then, Eq.
B.6 becomes:
?L 1
´ GM m
am
2Em r L 2E
b “ cos θ .
2 3 2
1 ` G 2Lm M
2E
B.2. The orbits 299

Isolating r:
L2
GM m2
r“ b .
2L2 E
1` G2 m3 M 2
` 1 cos θ
If we denote with:
L2
r0 “ , (B.7)
GM m2
c
2E
e“ 1` r0 . (B.8)
GmM
The equation for rpθq can be written in a simpler form:

r0
r“ . (B.9)
1 ` e cos θ

B.2 The orbits

Starting from Eq. B.9, it is possible to determine the type of orbits described
by the planets. In any case, Eq. B.9 represents a conic section. Setting
dr{dθ “ 0 in Eq. B.9, it is possible to obtain the maximum and minimum
distances of the body from the centre of attraction:
$ r
0
& if e ă 1
rmax “ p1 ´ eq (B.10)
%8 if e ě 1 .

r0
rmin “ . (B.11)
1`e
a
a Eq. B.9 in Cartesian coordinates, using r “
We can write x2 ` y 2 and
cos θ “ x{p x2 ` y 2 q:

x2 ` y 2 “ r02 ´ 2r0 e x ` e2 x2 . (B.12)

Circle (e “ 0)
In the case e “ 0, Eq. B.9 becomes r “ r0 , hence the planet describes a
circular orbit with radius r0 . From Eq. B.8 it follows that:
GmM
E“´ .
2r0
We see that the energy is negative.
300 Appendix B. Kepler’s Laws

Ellipse (0 ă e ă 1)
From Eq. B.8, it follows that:
GmM
´ ă E ă 0.
2r0
Hence, the energy is negative, but it is greater than the energy of a circular
orbit with radius r0 . Completing the square for x in Eq. B.12, we obtain:

r0 e 2
px ` 1´e2
q y2 r0 r0
` “ 1 with a “ and b “ ? . (B.13)
a2 b2 1 ´ e2 1 ´ e2
This is the equation of an ellipse with semi-major axis a, semi-minor
? axis b
and centre in p´r0 e{p1 ´ e2 q, 0q. The semi-focal distance is c “ a2 ´ b2 “
r0 e{p1 ´ e2 q, hence one focus is in the origin. Now
a that we know the semi-
major axis a, we can write the eccentricity as e “ 1 ´ r0 {a. Substituting e
in Eq. B.8, we find:
c c
r0 2E
1´ “ 1` r0
a GmM
GmM
ñE“´ ,
2a
which is the well-known expression for the energy of an elliptical orbit.

Parabola (e “ 1)
From Eq. B.8, it follows that E “ 0. Hence, the velocity of the body tends
to zero when its distance from the centre of attraction tends to infinity. The
minimum and maximum distances are rmin “ r0 {p1 ` eq and rmax “ 8, re-
spectively. In this case, Eq. B.12 becomes:

y 2 “ r02 ´ 2r0 x “ ´2r0 px ´ r0 {2q . (B.14)

This is the equation of a parabola with vertex in pr0 {2, 0q and focal length
r0 {2. Hence, the focal point is in the origin.

Hyperbola (e ą 1)
From Eq. B.8, it follows that E ą 0. The body can escape a from the central
potential with a non-zero velocity at infinity, given by v “ 2E{m. Again,
rmin “ r0 {p1 ` eq and rmax “ 8. Completing the square for x, Eq. B.12
becomes:
r0 e 2
px ´ e2 ´1
q y2 r0 r0
2
´ 2
“ 1 with a “ 2 and b “ ? . (B.15)
a b e ´1 e2 ´ 1
B.3. Proof of Kepler’s third law 301

This is the equation for a hyperbola with centre (point of intersection

? of the
asymptotes) at pr0 e{pe2 ´ 1q, 0q. The semi-focal length is c “ a2 ` b2 “
r0 e{pe2 ´ 1q. The impact parameter is the minimum distance from the centre
of the potential that the body would reach if it were to travel on a straight line
from infinity. It can be shown that the impact parameter is equal to b.

B.3 Proof of Kepler’s third law

In Ch. 10 we proved Kepler’s third law only in the case of a circular orbit.
We are now in a position to prove Kepler’s third law for elliptical orbits. As
shown in Sec. “The Second Law”:
dA L
“ .
dt 2m
Integrating over a period:
LT
Atot “ ,
2m
where?T is the orbital period. The area of an ellipse is Atot “ πab, where
b “ a 1 ´ e2 . Substituting and squaring both sides:
” L2 ı T2
π 2 a4 “ . (B.16)
mp1 ´ e2 q 4m

Using L2 “ GM m2 r0 , the term in square brackets becomes GmM r0 {p1 ´ e2 q.

Since a “ r0 {p1 ´ e2 q, this can also be written as GmM a:

T2
π 2 a4 “ pGmM aq
4m
4π 2 3
ñ T2 “ a ,
GM
which is what we wanted to prove.
 C

Virial Theorem

To prove the Viral theorem, consider the quantity:

ÿ
G“ p~i ¨ r~i ,
i

where the summation extends over all the particles in the system. The total
time derivative of this quantity is:
dG ÿ 9 ÿ
“ r~i ¨ p~i ` p~9i ¨ r~i . (C.1)
dt i i

The first term can be transformed into:

ÿ ÿ ÿ
r~9i ¨ p~i “ mi r~9i ¨ r~9i “ mi ¨ vi2 “ 2K .
i i i

Since force equals rate of change of momentum, the second term is:
ÿ ÿ
p~9i ¨ r~i “ F~i ¨ r~i .
i i

Then, Eq. C.1 reduces to:

dG ÿ
“ 2K ` F~i ¨ r~i .
dt i

The time average of the last equation, over an interval τ , is obtained by integ-
rating both sides from 0 to τ , and dividing by τ :
żτ
1 dG dG ÿ
“ “ 2K ` F~i ¨ r~i ,
τ 0 dt dt i

or: ÿ 1” ı
2K ` F~i ¨ r~i “ Gpτ q ´ Gp0q . (C.2)
i
τ
302
 303

If the motion is periodic, i.e. all coordinates repeat after a certain time, the
term Gpτ q ´ Gp0q is bound, while τ can be taken arbitrarily large. Therefore,
the right hand side of Eq. C.2 vanishes. It then follows that:
1ÿ~
K“´ Fi ¨ r~i . (C.3)
2 i

Now, for a potential of the form U “ ´αr´n , the force is:

dU U
F “´ “ ´nαr´n´1 “ n .
dr r
Hence, the right hand side of Eq. C.3 is just ´nU {2, and we obtain the Virial
equation for a general potential U “ ´αr´n :

n
K“´ U .
2

For the gravitational potential, n “ 1, hence K “ ´U {2.

 D

Tables and constants

Sun
Mean radius 695475 km
Mass 1.99 ¨ 1030 kg
Surface temperature 5778 K
Absolute magnitude `4.83
Apparent magnitude ´26.74
Distance to the galactic centre 2700 ly
Spectral class G2V

Moon

Mean radius 1738 km

Mass 7.35 ¨ 1022 kg
Semi-major axis 3.844 ¨ 106 km
Sidereal period 27.322d
Orbital eccentricity 0.0549
Orbital inclination 5.145˝
Apparent magnitude (full) ´12.74

304
 305

Symbols of celestial bodies

Name Symbol
Sun @
Mercury A
Venus B
Earth C
Moon $
Mars D
Jupiter E
Saturn F
Uranus G
Neptune H

Solar System
Mercury Venus Earth Mars Jupiter Saturn Uranus Neptune
Mean
2440 6052 6371 3397 71493 60267 25557 24766
radius (km)
Mass (kg) 3.30 ¨1023 4.87 ¨1024 5.97 ¨1024 6.42 ¨1023 1.90 ¨1027 5.69 ¨1026 8.68 ¨1025 1.02 ¨1026
Semi-major
0.3871 0.7282 1 1.5237 5.2033 9.5826 19.2184 30.11
axis (au)
Orbital
period 0.240846 0.615198 1 1.88082 11.862 29.4571 84.0205 164.8
(years)
Orbital
0.2056 0.0068 0.0167 0.0934 0.0484 0.0542 0.0472 0.0086
eccentricity

Physical constants
Name Symbol Value Units
Speed of light c 2.9979 ¨ 108 m s´1
Universal gravitational constant G 6.67 ¨ 10´11 N m2 kg´2
Gravitational acceleration g 9.81 m s´2
Astronomical unit au 149.6 ¨ 1011 m
Parsec pc 206265 au
Planck’s constant h 6.63 ¨ 1034 Js
Stefan-Boltzmann constant σ 5.67 ¨ 10´8 W m2 K4
Boltzmann constant kB 1.38 ¨ 1023 J K1
´1
Hubble constant H0 67.8 kms Mpc´1
 Suggested resources

The following texts are suggested for an in-depth study of some of the topics
presented in this book.

Theory

• Fundamental Astronomy, Hannu et al.

This book presents many topics in a clear and balanced way, highlighting
their physical basis. It requires a slightly more advanced knowledge of
mathematics and physics.
• An Introduction to Modern Astrophysics, Carroll B., Ostlie D.
This book contains all the astronomy you would need for a bachelor’s
degree. However, compared to the previous one, it can be dispersed and
difficult to follow.
• Introduction to Classical Mechanics, David Morin.
One of the best books on classical physics. With accessible explanations
and hundreds of problems, this book is ideal for developing your problem
solving skills.
• Concepts in Thermal Physics, Stephen and Katherine Blundell.
This book covers the basics of thermal physics, thermodynamics and
statistical mechanics, with applications in stellar astrophysics, condensed
matter physics and climate change.
• Cosmology and Astrophysics Through Problems, Padmanabhan.
This book is very advanced and requires good physics knowledge. Very
recommended for those who want to study beyond the international com-
petition and prepare for university.

Observation

• Stellarium.
This software is one of the best resources to study the sky, interactively.
• Sky and Telescope’s Pocket Sky Atlas, Roger W. Sinnott.
It may be useful to have a reference on paper, by combining Stellarium
with an atlas.
306
 Index

aberration, 26 chain rule, 297

absorption, 41, 93 chromatic aberration, 43
achromatic, 43 circumpolar, 3, 16
action and reaction, 114 CMBR, 74
active optics, 41 color index, 92
adaptive optics, 41 conjunction, 137
albedo, 86 coordinate system, 2
almucantar, 5 cosine theorem, 278
altitude, 3, 5, 14, 17 culmination, 4, 17
angular momentum, 119, 126, 297 lower, 4, 16
angular resolution, 36 upper, 4, 15
annual parallax, 89 cycle of eccentricity, 30
anti-meridian, 52
aperture ratio, 40 daylight saving time, 52
aphelion, 48, 118 de Broglie, 69
arcsecond, 91, 101, 103 declination, 7, 14, 17, 20, 24, 45
ascending node, 7 descending node, 7
astronomical unit, 121 diffraction, 38
atmospheric refraction, 16, 17 dip of the horizon, 4
Aufbau principle, 75 dispersion, 41
autumnal equinox, 25 distinguishable particles, 70
azimuth, 5, 14 Doppler effect, 76
double angle formulae, 280
basis vector, 269 draconic month, 63
binary
astrometric, 149 eccentricity, 30, 48, 118
contact, 148 eclipse, 59
distant, 148 lunar, 60
neighbouring, 148 solar, 62
photometric, 152 eclipse seasons, 59
spectroscopic, 150 ecliptic, 7, 8, 20, 23, 48, 59
visual, 148 effective potential, 127
black body, 72, 74 electromagnetic spectrum, 68, 87
curve, 72 ellipse, 118
radiation, 72 elongation, 137
black hole, 129 emissivity, 74
bolometric correction, 92 energy, 124, 127, 128, 297
Boltzmann distribution law, 70 kinetic, 124
bremmstrahlung, 74 potential, 124
equation of time, 50
calendar equatorial bulge, 23
Gregorian, 53 equinox, 20
Julian, 53 equipartition theorem, 71
cardinal points, 4 escape velocity, 71, 128, 248
Carnot, 278 Euler’s number, 283
celestial equator, 6, 7, 48 event horizon, 129
central force, 119 exeligmos, 64
centre of mass, 123 exponential, 283
centripetal force, 122 extinction, 92, 93, 104
Cepheid, 105 eyepiece, 35
307
308 Index

filed of view, 42 longitude, 3

first contact, 61 ecliptic, 8
first point of Aries, 7 galactic, 9
first quarter, 57 luminosity class, 79
flux, 86 lunar phase, 57
focal lunation, 57
length, 35
plane, 35 magnification, 38
point, 118 maximum, 39
Fraunhofer diffraction, 38 minimum, 39
Fraunhofer lines, 75 magnitude, 86, 87, 268, 270
full moon, 57, 59 absolute, 89, 104
apparent, 89, 104
Gaia, 103 bolometric, 87, 91
Gauss, 117, 247 composite system, 90
gaussian surface, 116 integrated, 91
gravitational acceleration, 116 visual, 91
great circle, 4 Maxwell distribution law, 70
Greenwich Mean Time, 51 mean quadratic velocity, 70
meridian, 2, 45
Harvard stellar classification, 79 Meton cycle, 66
Hill sphere, 253 molecular mass, 71
Hipparcos, 103 momentum, 69
Hohmann transfer orbit, 143 month
horizon anomalistic, 66
apparent, 4, 27 draconic, 58
astronomical, 4, 17 sidereal, 57
geometrical, 4 synodic, 57
hour angle, 6, 7, 14, 15, 17 Moon, 56
hour circle, 6
Hubble, 106, 108 nadir, 4, 6
constant, 108 new moon, 59
law, 108 nodal precession, 58
hydrogen 21 cm, 9, 106 node line, 59
nutation, 26
impulse, 144
inertial reference system, 114 Oberth effect, 143
interstellar medium, 92, 104 objective, 35
intrinsic brightness, 89, 104 obliquity of the ecliptic, 24, 48
isolated system, 126 opacity, 93, 94
isotropy, 86 opposition, 137, 139
optical depth, 92, 104
Kepler, 118, 297 orbit, 118, 127
first law, 118 bound
second law, 119 circular, 127, 299
third law, 121 elliptical, 127, 299
Kirchoff, 72 unbound
hyperbolic, 128, 300
last contact, 61 parabolic, 128, 300
last quarter, 57
latitude, 3, 14, 17 pair of cones, 23
ecliptic, 8 parallax, 27, 89, 99, 124
galactic, 9 annual, 102
law of cosines, 281 diurnal, 101
law of sines, 277, 281 spectroscopic, 104
libration, 64 parallax baseline, 99
light year, 103 parallel, 2
local time, 51 parallel of declination, 6
logarithm, 283 parsec, 89, 103
Index 309

penumbra, 61 spectral class, 104

perihelion, 48, 118 spectral type, 92
period spectrum, 72–74
sidereal, 139 absorption, 74
synodic, 139 continuous, 73, 74
photon, 68 emission, 74
Planck constant, 68 speed of light, 129
Planck’s law, 74 spherical trigonometry, 281
planet standard candle, 105
inferior, 137 Stefan-Boltzmann, 86
superior, 137 constant, 73
planetary atmosphere, 71, 128 law, 73, 88
Pogson law, 87 steradian, 276
point mass, 114 surface brightness, 91
Pole Star, 23 surface temperature, 79, 88
precession of the equinoxes, 23 system
prime meridian, 3 ecliptic, 8, 19
prime vertical, 4 equatorial, 6
principal vertical, 4 first or HA-dec, 6, 14
Ptolemaic system, 139 second or RA-dec, 7, 19
galactic, 9
quadrature, 138, 139 horizontal, 4, 14
quantization, 68
quantization of energy, 74 tangential velocity, 297
telescope, 32
radial velocity, 297
Galileo’s refracting, 35
radial velocity method, 214
Newton’s reflecting, 35
radian, 276
temperature, 86, 104
Rayleigh scattering, 93
tidal force, 249
Rayleigh’s criterion, 38
time dilation, 76
recession velocity, 108
time zone, 51, 52
redshift, 108
torque, 119
reduction formulae, 279
transit, 124
relative speed, 76
transit method, 214
retrograde, 137
transmittance, 94
right ascension, 7, 17, 24, 48
triangulation, 99
Roche limit, 249
trigonometric function, 276
rocket equation, 142
Tully-Fisher relation, 106
RR Lyrae, 106
Tycho Brahe, 118
Saros, 64
Saros cycle, 63 UBV, 92
scalar product, 117, 124, 270 umbra, 61
scattering, 93 universal law of gravitation, 114
Schwarzschild radius, 129 Universal Time, 51
seeing, 41
semi-major axis, 118, 121 vector product, 119, 270
semi-minor axis, 118 vernal equinox, 7, 23, 25, 45
shadow, 61 virial theorem, 129
sidereal
time, 7, 17, 45 Wien’s law, 72
year, 52
Yarkovsky effect, 253
sidereal day, 45
year, 58
singularity, 129
draconic, 58
Snell’s Law, 27
sidereal, 52
solar day
tropical, 8, 53, 58
mean, 45, 48
Yerkes stellar classification, 80
true, 45, 48
solar time, 45
zenith, 4, 6
solstice, 7
zodiac, 24
special relativity, 76