(Undergraduate Lecture Notes in Physics) Giuseppe Severino (Auth.) - The Structure and Evolution of The Sun-Springer International Publishing (2017)

Undergraduate Lecture Notes in Physics
Giuseppe Severino
The Structure
and Evolution
of the Sun
Undergraduate Lecture Notes in Physics (ULNP) publishes authoritative texts covering
topics throughout pure and applied physics. Each title in the series is suitable as a basis for
undergraduate instruction, typically containing practice problems, worked examples, chapter
summaries, and suggestions for further reading.
ULNP titles must provide at least one of the following:

• An exceptionally clear and concise treatment of a standard undergraduate subject.
• A solid undergraduate-level introduction to a graduate, advanced, or non-standard subject.
• A novel perspective or an unusual approach to teaching a subject.
ULNP especially encourages new, original, and idiosyncratic approaches to physics teaching
at the undergraduate level.
The purpose of ULNP is to provide intriguing, absorbing books that will continue to be the
reader's preferred reference throughout their academic career.
Series editors
Neil Ashby
University of Colorado, Boulder, CO, USA
William Brantley
Department of Physics, Furman University, Greenville, SC, USA
Matthew Deady
Physics Program, Bard College, Annandale-on-Hudson, NY, USA
Michael Fowler
Department of Physics, University of Virginia, Charlottesville, VA, USA
Morten Hjorth-Jensen
Department of Physics, University of Oslo, Oslo, Norway
Michael Inglis
SUNY Suffolk County Community College, Long Island, NY, USA
Heinz Klose
Humboldt University, Oldenburg, Niedersachsen, Germany
More information about this series at http://www.springer.com/series/8917

Giuseppe Severino
The Structure and Evolution

of the Sun
123
Giuseppe Severino
Napoli
Italy
ISSN 2192-4791 ISSN 2192-4805 (electronic)

ISBN 978-3-319-64960-3 ISBN 978-3-319-64961-0 (eBook)
DOI 10.1007/978-3-319-64961-0
Library of Congress Control Number: 2017951430
© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
The solar disk—Aton—lights up with its rays ending by hands the family of
Pharaoh Akhenaton, who in turn holds out his hands to him and makes an offering.
Egypt, about 1350 B.C.
To my family
and, in particular, to Albina,
“mon enfant, ma soeur”,
for our journey over the rainbow together.
Preface
After studying the Sun for 41 years (since receiving my degree in 1974), I have
finally realized my dream of using my experience in this field to write a book. Why
a book on the Sun? The Sun is our closest star and without it, life on Earth would
cease to exist. Even though man has studied the Sun for centuries, there is still
much to learn. At present there has been a “drop” in interest in solar physics;
scientists saying that we know all there is to know, just like the ancient Greeks said
to Socrates, but like them, today’s scientists would be wrong. Our technology alone
makes gaining insight into our Sun easier, and with motivated astronomers taking
an interest in solar physics, our knowledge will grow and our lives on Earth will
benefit.
The primary objective of the book is to illustrate the structure and evolution
of the Sun, making the reader capable of constructing a simplified model of it. The
final chapter is devoted to solar evolution. This selection of topics leaves solar
activity phenomena out of the book. This is certainly an important loss, since solar
activity, with its influence on the Earth, is the most researched field in solar physics
today. However, the study of the solar structure and evolution represents a natural
first step for a student moving from physics to astrophysics. He/she can understand
how the Sun and stars basically work and, also, learn new concepts and methods,
which are largely used in astronomy, in a coherent manner, that is, organized in one
homogeneous subject and not scattered in many different topics.
Initially, I thought “The Sun for almost everyone” as a possible title of the book.
This title was inspired by the book “Physics for everyone” by Landau and
Kitaigorodskij, published by MIR, which I recently started reading again, redis-
covering its great value, many years after my first reading. However, I realized that
I could not write a book, which is intended to be at the university level, using only
the four operations and minimizing the formulas, as the Russians did. Thus, I had to
give up the initial title, even though I retained the aim of helping the reader as much
as possible. I added a preliminary chapter with the basic physics to read the book.
I tried to make the involved mathematics as clear as possible. In particular, I wrote
all the steps necessary to get the formulas, so that the reader can proceed without
recurring to pen and paper to perform further calculations. Moreover, instead of
ix
x Preface
referring to a preceding formula, I decided to rewrite the formula so as not to

constrain the reader to browse the book losing the thread of the argument, but, on
the contrary, in order to make reading the book as smooth as possible, like a novel.
I would like to express all my gratitude to the many persons I feel indebted to.
Marcella Marconi helped me in improving many chapters of the book with her
extensive comments. Scilla Degli Innocenti provided me with the unpublished solar
evolutionary models by Senesi et al. Moreover, she, Vincenzo Andretta and Lucio
Crivellari critically read through particular chapters. Stuart M. Jefferies, Colleen
Jefferies, Carmen De Dominicis and Giovanna Monaco revised the English of a
number of chapters. Massimo Della Valle contributed comments on a specific topic.
Massimo Capaccioli, Elvira Covino, Juan Alcalà and John Leibacher critically read
through preliminary parts of the book. Finally, the Editor of the series
Undergraduate Lecture Notes in Physics, prof. Mike Inglis, helped me to see my
original project in the right perspective.
This book is dedicated to my family and, in particular, to my wife, Albina, who
has been to me like the Sun to the Earth.
Napoli, Italy Giuseppe Severino

June 2017
Contents
1 Some Basic Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Basic Structure of Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Elementary Particles and Fundamental Forces . . . . . . . . . . 2
1.1.2 Atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1.3 Units of Measurement and Exponential Notation . . . . . . . 6
1.1.4 Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.5 Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.1.6 Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Thermodynamic Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.1 The Energy Distribution of Gas Particles . . . . . . . . . . . . . 12
1.2.2 Maxwell Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.2.3 Boltzmann Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.2.4 Saha Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.2.5 The Indefinite Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.2.6 The Definite Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.3 The Nature of Light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.3.1 Waves, Photons and the Electromagnetic Spectrum. . . . . . 35
1.3.2 The Black Body . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.4 The Nuclear Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.4.1 The Missing Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.4.2 The Mass–Energy Relation . . . . . . . . . . . . . . . . . . . . . . . . 46
1.4.3 The Nuclear Fission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
1.4.4 Nuclear Power Stations . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.4.5 Radioactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.4.6 The Nuclear Fusion in the Sun . . . . . . . . . . . . . . . . . . . . . 50
1.4.7 Further Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2 Introduction to the Sun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.1 Our Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
xi
xii Contents
2.1.2 The Steady Sun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

2.1.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.2 The Characteristic Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.2.1 The Distance from Earth . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.2.2 The Radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.2.3 The Mass. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.2.4 The Luminosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.2.5 The Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3 The Photosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.1 The Physical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.1.1 Physical Constrains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.1.2 Hydrostatic Equilibrium. . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.1.3 Scale Height . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.1.4 Pressure Stratification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.1.5 Plane Parallel Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.1.6 Optical Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.1.7 Flux Formation Depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.1.8 Photospheric Opacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.1.9 Hydrogen Ionization Equilibrium . . . . . . . . . . . . . . . . . . . 84
3.1.10 Electric Charge and Mass Conservation . . . . . . . . . . . . . . 86
3.1.11 Evaluating Density and Pressure . . . . . . . . . . . . . . . . . . . . 88
3.2 A Model of the Photosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.2.1 Radiation Intensity and Flux . . . . . . . . . . . . . . . . . . . . . . . 90
3.2.2 The Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.2.3 The Origin of Continua and Lines . . . . . . . . . . . . . . . . . . . 93
3.2.4 Equation of Radiation Transfer . . . . . . . . . . . . . . . . . . . . . 95
3.2.5 Solution of the Transfer Equation . . . . . . . . . . . . . . . . . . . 97
3.2.6 The Eddington-Barbier Approximation . . . . . . . . . . . . . . . 99
3.2.7 LTE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.2.8 Center to Limb Intensity . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.2.9 The Temperature Variation . . . . . . . . . . . . . . . . . . . . . . . . 102
3.2.10 Temperature as Function of Height . . . . . . . . . . . . . . . . . . 106
3.2.11 Complete Model Photosphere . . . . . . . . . . . . . . . . . . . . . . 109
3.2.12 Accurate Models of Photosphere . . . . . . . . . . . . . . . . . . . . 115
3.2.13 The Chemical Composition . . . . . . . . . . . . . . . . . . . . . . . . 120
3.2.14 Differences Between Photosphere and the Earth’s
Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 124
3.2.15 Further Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 125
4 The Convection Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.1 Glimpses of the Solar Interior. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Contents xiii
4.1.1 Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

4.1.2 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.1.3 Opacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.2 The Radiative Temperature Gradient . . . . . . . . . . . . . . . . . . . . . . . 132
4.2.1 The Diffusion Approximation . . . . . . . . . . . . . . . . . . . . . . 132
4.3 The Convective Temperature Gradient . . . . . . . . . . . . . . . . . . . . . . 136
4.3.1 Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.3.2 The Convective Temperature Gradient . . . . . . . . . . . . . . . 138
4.4 The Convection Zone Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.4.1 Convective Stability Criterion . . . . . . . . . . . . . . . . . . . . . . 144
4.4.2 A Model of Convection Zone . . . . . . . . . . . . . . . . . . . . . . 151
4.4.3 The Extension of the Convection Zone . . . . . . . . . . . . . . . 163
4.4.4 Checking the Hydrostatic Equilibrium . . . . . . . . . . . . . . . . 165
5 The Radiation Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.1 The Physical Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.1.1 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
5.1.2 The Opacity Power Laws . . . . . . . . . . . . . . . . . . . . . . . . . 172
5.1.3 Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
5.1.4 State Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
5.1.5 Mass Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
5.2 The Radiation Zone Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
6 The Core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
6.1 Cross Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
6.1.1 Solar Nuclear Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . 189
6.1.2 The First Reaction of the ppI Chain . . . . . . . . . . . . . . . . . 192
6.1.3 De Broglie Wavelength . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
6.1.4 Wave Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.1.5 The Uncertainty Principle . . . . . . . . . . . . . . . . . . . . . . . . . 199
6.1.6 Tunnel Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
6.1.7 Cross Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
6.2 Reaction Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
6.2.1 The Average over the Maxwell Distribution . . . . . . . . . . . 204
6.2.2 The Gamow Peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
6.2.3 Lifetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
6.2.4 Energy Production Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
6.3 The Core Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
6.3.1 Luminosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
6.3.2 Energy Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
6.3.3 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
xiv Contents
6.3.4 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

7 Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
7.1 Pre-main Sequence Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
7.1.1 Onset of Gravitational Collapse . . . . . . . . . . . . . . . . . . . . . 227
7.1.2 The Contribution of Core-Collapse Supernovae . . . . . . . . 229
7.1.3 Jeans Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
7.1.4 Isothermal Collapse and Fragmentation . . . . . . . . . . . . . . . 233
7.1.5 Adiabatic Collapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
7.1.6 ProtoSun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
7.1.7 Sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
7.1.8 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
7.1.9 Contraction Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.1.10 The Hayashi Track . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
7.1.11 Virial Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
7.2 Main Sequence Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
7.2.1 Stability of the Solar Equilibrium . . . . . . . . . . . . . . . . . . . 245
7.2.2 Equations of Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
7.2.3 Evolution Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
7.2.4 Main Sequence Evolution . . . . . . . . . . . . . . . . . . . . . . . . . 250
7.2.5 MS Evolution and Life on the Earth . . . . . . . . . . . . . . . . . 252
7.3 Post-main Sequence Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
7.3.1 Hydrogen Shell Fusion and Red Giant Phase . . . . . . . . . . 256
7.3.2 Mass Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
7.3.3 Degenerate Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
7.3.4 Helium Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
7.3.5 Thermal Instability and Helium Flash . . . . . . . . . . . . . . . . 260
7.3.6 Horizontal Branch and Asymptotic Giant Branch . . . . . . . 261
7.3.7 Planetary Nebula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
7.3.8 White Dwarf and Evolution Last Phases . . . . . . . . . . . . . . 262
List of Exercises and Answers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Chapter 1
Some Basic Physics
Abstract Before discussing the structure and history of our nearest star, the Sun,
we will review the underlying basic physics. This will help us to better appreciate
the subject. We start with a description of the fundamental structure of matter in
order to focus on the constituent particles and working forces which are present at
very different spatial scales. Despite their apparent complexity, only a limited
number of elementary particles and fundamental forces are needed to explain the
Sun as well as the rest of the Universe. Then, because the Sun is a huge mass of gas
which can be modeled using simple physical laws that apply to a perfect gas in
thermodynamic equilibrium, we introduce the equation of state, relating gas tem-
perature, density and pressure, and the Maxwell distribution, which determines the
number of gas particles (or molecules) having kinetic energies in a specific range as
function of the gas temperature. Actually, solar gas molecules have internal degrees
of freedom, that is, they not only exchange kinetic energy through elastic collisions,
but also undergo excitation and ionization processes, where further energy
exchanges occur. These two types of processes can also be treated in equilibrium
conditions, leading us to the Boltzmann and Saha distributions. The Boltzmann
equation determines the ratio of the numbers of particles occupying two subsequent
excitation states as a function of the excitation energy difference and gas temper-
ature. The Saha equation gives the ratio of the numbers of particles in two sub-
sequent ionization states as a function of the ionization energy, gas temperature and
electron density.
Excitation and ionization processes are also related to light emission, and the Sun is
our primary source of light. Consequently, a basic knowledge of the nature and
behavior of light is indispensable. For our purposes, the light particles, or photons,
can be described again by an equilibrium distribution over their energies; this is
called the black body distribution. Finally, we will take a panoramic look at the
energy produced by the nuclear fusion and fission of atomic nuclei. Nuclear fusion
is the basic energy source for the Sun and stars, while nuclear fission is currently
exploited to supply the increasing energy needs of human life.
© Springer International Publishing AG 2017 1

G. Severino, The Structure and Evolution of the Sun, Undergraduate Lecture
Notes in Physics, DOI 10.1007/978-3-319-64961-0_1
2 1 Some Basic Physics
1.1 Basic Structure of Matter
1.1.1 Elementary Particles and Fundamental Forces
What is the microscopic structure of our world, including animals and mankind?
The answer to this question requires finding elementary particles that, having no
internal structure, represent the fundamental building blocks from which the
Universe is made. Research into this question has led to the discovery of many
supposed elementary particles and the situation was a real zoo till the so-called
Standard Model arrived and put this field of research in order. The Standard Model
was developed in the early 1970s as a collaborative effort between scientists around
the world. According to this model, matter is composed of 12 elementary particles
which, interacting through 4 fundamental forces, allow for the formation of all the
other particles. These basic particles are 6 quarks (up, down, charm, strange, top
and bottom) and 6 leptons (electron, electronic neutrino, muon, muonic neutrino,
tauon and tauonic neutrino). Moreover, there are further elementary particles that,
in the analogy of building blocks, are the cement (or glue) that transfer, or mediate,
the fundamental forces. The 4 fundamental forces are gravity, electromagnetism,
weak interaction and strong interaction. Only the first two forces, gravity and
electromagnetism, are familiar to us. This is because they have an infinite range of
action and, therefore, are important at the macroscopic level. The other two forces,
weak interaction and strong interaction, have very short ranges and become dom-
inant at the microscopic level of atomic nuclei and elementary particles. The par-
ticles that transfer fundamental forces (the glue) are: photons for electromagnetism,
W and Z bosons for weak interaction, and gluons for strong interaction.
Therefore, there are 16 fundamental particles included in the Standard Model, as
summarized in Fig. 1.1. Actually, in addition to these 16 particles we have to add two
more particles, namely the Higgs boson, which was predicted long ago by the British
physicist Peter Higgs (born in 1929) for explaining why the W and Z bosons have
mass and which was only recently detected, and the graviton, which was proposed as
a mediator of gravity by theory but has not yet been found experimentally.
Two particles exchange a force if they are provided with the property that
characterizes that force. So, matter undergoing gravity requires mass; for electro-
magnetism the characteristic property is the electric charge, which we know to be
of two types: positive or negative. For the weak interaction the property is the
flavor, which can be of 12 different types, as many as the number of quarks and
leptons. For the strong interaction the typical property is the color, which can be of
three different types, blue, red and green. Obviously, flavor and color, as well as the
quarks’ names, are words inferred from the common use, but which gain a new
meaning in connection with elementary particles. To complete the picture, we have
to say that for each of the 12 matter elementary particles an antiparticle exists
which is characterized by the inversion of one or more of the particle’s charac-
teristic properties. For example, if we denote −e as the electron negative electric
1.1 Basic Structure of Matter 3
Fig. 1.1 Elementary particles included in the standard model. Each particle is represented by a
circle. Violet circles refer to the six quarks, green circles to the six leptons, and pink circles to the
four force mediators. Finally, on the right, the mosaic is completed by the higgs boson (in yellow)
and the graviton (in black dashes)
Fig. 1.2 Quark structure of proton (left) and neutron (right). A proton consists of 2 up quarks and
1 down quark while a neutron consists of 2 down quarks and 1 up quark. Assuming the modulus of
the electron charge, e, as a unit, a down quark has a negative charge of −1/3 e and an up quark has
a positive charge of +2/3 e. Therefore, the charge of a proton is +2/3 e +2/3 e −1/3 e = +e, and the
neutron charge is −1/3 e −1/3 e +2/3 e = 0
charge, its antiparticle, which is called positron, has the same mass but positive
electric charge +e, that is opposite to the electron charge.
We can now construct any other particle and the entire structure of matter using
the elementary particles and fundamental forces included in the Standard Model. As
an example, three quarks, two up and one down, linked up by strong interaction, form
a proton, which has a positive electric charge +e, equal and opposite to the electron
charge. Another example is three quarks, two down and one up, form a neutron,
which is instead without charge and, hence, neutral, as its name says (Fig. 1.2).
1.1.2 Atoms
Protons and neutrons, linked by strong interaction, form the nuclei of atoms.
A cloud of electrons in orbit around the nucleus under the action of the electro-
magnetic force between the positive nucleus and the negative electrons then
completes the atomic structure. The number of protons, each with charge +e, is
equal to the number of electrons, each with charge −e, so that the atom is elec-
trically neutral.
The word atom comes from a Greek word meaning indivisible. We know now
that this is not true, atoms have an internal structure, one which deserves a more
detailed description. All atoms are made up of a massive nucleus and a number of
lighter electrons orbiting around the nucleus. This begs the question, why are
electrons bound to a nucleus without falling into it? If we want an object, for
example a bucket, to rotate around us, we need a suitable velocity to set it in
motion, and a rope to keep it in orbit. In the case of the solar system, the gravi-
tational force is equivalent to the rope: it constrains the Earth and the other planets
to rotate around the Sun, each one with its own orbital velocity. Unlike the solar
system, which is governed by gravity, the stability in the atom is brought by the
electric force since the nucleus and electrons have opposite charges: positive for the
nucleus and negative for the electrons, and, hence, they attract each other.
To make things easier, let us consider one electron in orbit around a proton in an
atom of hydrogen, noting that the processes we are about to describe occur in a
similar way in an atom with many electrons. Above we compared an electron in
orbit around an atomic nucleus to a tiny solar system or to a bucket on a rope which
a man uses to make it rotate around him. To tell the truth, these types of analogy
cannot represent all the different types of atomic behavior and, in fact, the atomic
dimension is the threshold of physics with new characteristics in respect to those
ones we are used to in our macroscopic world. This new physics is named quantum
mechanics.
The first big surprise is that for the electron bound to the nucleus only a discrete
series of orbits is allowed. This was first proposed by the Danish scientist Niels
Bohr (1885–1962) in 1913 (Fig. 1.3).
In each orbit, or better still, state, since the motion of the electron around the
nucleus is actually more complex than the simple circles of the Bohr model, the
electron has a definite velocity and energy and also a definite mean distance from
the nucleus. Every time the electron passes from one state to another, an event
called transition, its energy changes. If the atom is isolated, the only possible
transition is towards a state with lower energy. The energy lost by the electron in
making such a transition is emitted as light. All states, having higher energy than
the ground state, are called excited states. The state of minimum energy is called the
ground state.
If the atom is not isolated, but instead is embedded in a gas composed of other
atoms and particles as well as light, other types of transitions are possible in
addition to the simple decay toward the ground state. So, for instance, a collision
Fig. 1.3 Bohr’s model of the Hydrogen atom. The nucleus consists of one proton (red central
circle) and there is just one electron orbiting around it in one of a discrete series of allowed circular
orbits. Here only the first four orbits are plotted and the electron is represented by the blue circle in
the innermost orbit, or ground state. Each orbit is distinguished by a value of the integer number n,
called the principal quantum number, which determines the energy of the electron inside the atom
according to the formula. En = −EH/n2, with n = 1, 2, 3, …∞, where the symbol ∞ denotes
infinity. The minus sign means that the electron requires the energy −En to break loose from the
nucleus and move freely. In particular, for n = 1, −E1 = EH is the energy necessary to ionize
hydrogen from the ground state
with another gas particle, which is then accelerated, can make the atom decay from
an excited state without the emission of light. Moreover, transitions from a lower
energy state to a higher energy state may occur. In this case, the energy necessary
for the electron to jump is provided either by light when a photon is absorbed, or by
another gas particle, which slows down in a collision. In particular, the electron can
be excited to the point of losing its link to the nucleus. In this situation, one says
that the atom is ionized and the electron can move freely with a velocity and an
energy that can span a continuous range (no longer only a discrete set).
Exercise 1.1 Write down the energy balance for the following transitions occurring
to a hydrogen atom:
1. A gas particle with kinetic energy mv21/2 hits a hydrogen atom which is in the
ground state. After collision, the particle has kinetic energy mv22/2 with v1 > v2
(the symbol > means greater than) and the hydrogen atom is in the excited state
with n = 2.
2. A hydrogen atom decays from the excited state with n = 3 to the one with n = 2
and emits a photon with energy Ep.
3. A hydrogen atom in the ground state is ionized by a photon with energy Ep.
After collision, the free electron has velocity v.
We called the atom a microscopic solar system, but what really is the atomic
size? The size of an atom is extremely small, of the order of the length resulting
from dividing 1 m in 10 billions of equal parts. Atomic nuclei are even smaller,
having sizes of the order of one million billion smaller than one meter.
1.1.3 Units of Measurement and Exponential Notation
Speaking of very small numbers, like those we encounter in the microscopic world
of atoms and nuclei, or of very large numbers, like, for instance, the astronomical
unit (symbol AU), which is the mean distance from Earth to the Sun and amounts to
nearly 150 billion meters, it is worthwhile to comment shortly on the units of
measurement and exponential notation. Our usual units of measure, like the meter
in our case, go naturally with the sizes we encounter in the common experiences of
a terrestrial laboratory. However, when they are used to describe the sizes we
encounter at the microscopic level, the required numbers rapidly become very
small. Analogously, passing to the Sun or to the Universe, the required numbers
become extremely large. A simple way of denoting a large number is based on the
property that each multiplication of a number by 10 yields the addition of a 0 in the
result. For instance, starting with not a large number, the product 31010 = 300 is
written, in what is called the exponential notation, as 3 102 = 300, where 2
indicates the number of times that 10 enters as a factor in the product whose result is
100.
Returning back to the Sun–Earth distance we can now write
1 AU ¼ 150 billion meters ¼ 150,000,000,000 m ¼ 15 1010 m:
For numbers smaller than 1, we use the division by 10 for their representation,
taking as exponent of the power of 10 the number of divisions with the sign
designated as minus; for example, 1 hundredth = 0.01 = 1/100 = 1/102 = 10−2.
Therefore, we can write now that atoms have sizes of the order of
1 m over 10 billion ¼ 1=10,000,000,000 m ¼ 1=1010 m ¼ 1010 m;
and atomic nuclei have sizes of the order of
1 m over 1 million billion ¼ 1=1,000,000,000,000,000 m ¼ 1=1015 m ¼ 1015 m:
Different powers of 10 can also be denoted with the initial letter of a suitable
prefix, as listed in Table 1.1. So, for instance, m is the symbol for the prefix milli,
which represents 10−3; hence
Table 1.1 Power of 10 with their corresponding symbol letters and prefixes
1012 T Tera
109 G Giga
106 M Mega
103 k Kilo
10−2 c Centi
10−3 m Milli
10−6 l Micro
10−9 n Nano
10−12 p Pico
Notice that the symbol for micro (one millionth) is l, a Greek letter which is read mu
103 m ¼ 1 millimeter ¼ 1 mm:
Moreover, using the symbol G for the prefix Giga, we can write
1 AU ¼ 15 1010 m ¼ 150 109 m ¼ 150 Gm:
Thanks to the exponential notation we can easily represent very small and very
large numbers referring to the same unit of measurement and without resorting to a
plethora of ad hoc units.
Exercise 1.2 Ångström, symbol Å, and Fermi, symbol Fm, are length units used in
atomic spectroscopy and nuclear physics. Since 1 Å = 10−10 m and 1 Fm = 10−15 m,
find their values in nm and pm, respectively (Anders Jonas Ångström was a Swedish
physicist, 1814–1874, and Enrico Fermi an Italian physicist, 1901–1954).
1.1.4 Elements
The atomic structure, consisting of a nucleus with protons and neutrons, and
electrons in orbit around it, is the same for all the known chemical elements. What
changes passing from one element to another is the number of protons, which is
called the atomic number and is denoted by the letter Z. As Z increases the number
of neutrons increases, which helps to stabilize the nucleus, and also the number of
electrons increases so as to keep the atom neutral. The atomic number determines
what element we are dealing with. For instance, the chemical element with only one
proton (Z = 1) is hydrogen, the element with two protons (Z = 2) is helium, and so
on. A stable nucleus is in its minimum (negative) energy state where it will remain
unless it receives a suitable amount of energy. There are 81 different types of stable
nuclei (Table 1.2).
The stable nuclei with 20 or less protons (Z 20) are characterized by a
number of neutrons almost equal to that of protons. For Z > 20 the number of
neutrons for a stable nucleus is greater than Z. Atomic nuclei with the same number
of protons and different numbers of neutrons are called isotopes. So, for example,
Table 1.2 The known stable elements.

Z Element Symbol Isotopes Z Element Symbol Isotopes
1 Hydrogen H 2 43 [Technetium] [Tc] –
2 Helium He 2 44 Ruthenium Ru 7
3 Lithium Li 2 45 Rhodium Rh 1
4 Beryllium Be 1 46 Palladium Pd 6
5 Boron B 2 47 Silver Ag 2
6 Carbon C 2 48 Cadmium Cd 8
7 Nitrogen N 2 49 Indium In 2
8 Oxygen O 3 50 Tin Sn 10
9 Fluorine F 2 51 Antimony Sb 2
10 Neon Ne 3 52 Tellurium Te 8
11 Sodium Na 1 53 Iodine I 1
12 Magnesium Mg 3 34 Xenon Xe 9
13 Aluminum Al 1 55 Cesium Cs 1
14 Silicon Si 3 56 Barium Ba 7
15 Phosphorus P 2 57 Lanthanum La 2
16 Sulfur S 4 58 Cerium Ce 4
17 Chlorine Cl 2 59 Praseodymium Pr 1
18 Argon Ar 3 60 Neodymium Nd 7
19 Potassium K 3 61 [Promethium] [Pm] –
20 Calcium Ca 6 62 Samarium Sm 7
21 Scandium Sc 2 63 Europium Eu 2
22 Titanium Ti 5 64 Gadolinium Gd 7
23 Vanadium V 2 65 Terbium Tb 1
24 Chromium Cr 4 66 Dysprosium Dy 7
25 Manganese Mn 1 67 Holmium Hb 1
26 Iron Fe 4 68 Erbium Er 6
27 Cobalt Co 1 69 Thallium Tm 1
28 Nickel Ni 5 70 Ytterbium Yb 7
29 Copper Cu 2 71 Lutetium Lu 2
30 Zinc Zn 5 72 Hafnium Hf 6
31 Gallium Ga 2 73 Tantalum Ta 2
32 Germanic Ge 5 74 Tungsten W 5
33 Arsenic As 1 75 Rhenium Re 2
34 Selenium Se 6 76 Osmium Os 7
35 Bromine Br 2 77 Iridium Ir 2
36 Krypton Kr 6 78 Platinum Pt 6
37 Rubidium Rb 2 79 Gold Au 1
38 Strontium Sr 4 80 Mercury Hg 7
39 Yttrium Y 1 81 Thallium Tl 2
40 Zirconium Zr 5 82 Lead Pb 4
(continued)
Table 1.2 (continued)

Z Element Symbol Isotopes Z Element Symbol Isotopes
41 Niobium Nb 1 83 Bismuth Bs 1
42 Molybdenum Mb 7
For each element columns report from left to right: atomic number Z, name, chemical symbol and
number of stable isotopes. The elements Z = 43 and Z = 61 have no stable isotopes
there are two stable isotopes with Z = 1: hydrogen, with one proton, and deuterium,
whose nucleus contains a proton and a neutron. With Z = 2 we find two stable
isotopes: tritium, with 2 protons and one neutron, and helium, which has 2 protons
and 2 neutrons.
1.1.5 Molecules
Electromagnetism is the force at the base of chemical bonds that allows atoms of
the same element or of different elements to mutually combine and form larger
particles, called molecules. For example, a molecule of water is obtained by the
combination of two hydrogen atoms with one oxygen atom (Z = 8).
Molecules are the components that repeat themselves in the structure of the
different materials of our world. So, a macroscopic piece of matter contains a large
number of molecules of the same type; for instance, one gram of water contains
3.3 1022 molecules of water.
1.1.6 Gases
The fact that a material is a solid, liquid or gas, each of which is called a state of
aggregation, depends primarily on the temperature which controls the relative
distances and the motion of its molecules, called thermal agitation. With increasing
thermal agitation, in fact, the mean distance between molecules increases and,
hence, mutual interaction is reduced. In particular, the molecular components in a
solid are relatively very close to each other and oscillate around positions fixed in
space. The molecules of a liquid interact with each other less strongly than in a
solid, so they cannot maintain fixed relative positions and, at the macroscopic level,
a liquid can flow and take the shape of its container. The average distance between
molecules in a gas is much larger than the size of the molecules themselves and the
characteristic intermolecular distances in solids and liquids. So, a gas is made up by
molecules where each one is in rectilinear motion with constant velocity, until it
collides with another molecule, and at the macroscopic level, a gas is also able to fill
the full volume of its container. In general, when considering two molecules that
collide as an isolated system, that is, with no interaction from the outside, the
collision preserves the total energy of the two molecules while it allows the
transformation of a part of the initial kinetic energy into internal energy of the
molecules. The conservation of total energy occurs, for example, in the collision of
a bullet with a target, where the full kinetic energy of the bullet is transformed into
thermal energy of the target and of the bullet itself. If, however, in collisions
between gas molecules the total kinetic energy of the colliding molecules remains
unchanged, as occurs for instance in the case of collisions between billiard balls,
then the collisions are called elastic collisions and the gas is said perfect or ideal.
In a perfect gas there is a simple relationship between pressure, density and
temperature. A perfect gas exerts a pressure P which is directly proportional to its
temperature T and to the number of molecules per unit of volume N. This behavior
is summarized in the gas law or equation of state, which is
P ¼ NkB T:
Here the product kBT measures the most probable value of the kinetic energy of a
molecule, as we will see later, and kB is a constant, named after the Austrian
scientist Ludwig Boltzmann (1844–1906). The value of the Boltzmann constant
depends on the units of the measure used to define energy and temperature; in
particular, when we measure energy in erg (g cm2 s−2) and temperature in
absolute degree or Kelvin (after William Thomson, Lord Kelvin, Scots-Irish
physicist, 1824–1907, symbol K), it results
kB ¼ 1:38062 1016 erg K1 :
Exercise 1.3 Show that 1 kWh = 3.6 1013 erg, where kWh means
kilo-Watt-hour and Wh is the usual unit for the energy consumption of household
appliances. Then, find the value of the Boltzmann constant in kWh K−1. Note that
James Watt was a Scottish engineer (1736−1819) and James Prescott Joule was an
English physicist (1818−1889).
Exercise 1.4 ◦C is the symbol for centigrade or Celsius degree (named after the
Swedish astronomer Anders Celsius, 1701–1744). The relation between K and ◦C
degrees is such that 0 K = −273.15 ◦C. What kind of problems could be implied by
using ◦C instead of K degrees in determining the value of kB?
The gas density, which is usually indicated by the Greek letter q (read rho), is
defined by the ratio of the mass of a gas to its volume. If we denote with m the mass
of a molecule, the gas density is also the product of the particle mass and its number
density, i.e.
q ¼ mN:
In terms of the density q the equation of state can be written
P ¼ NkB T ¼ mNkB T=m ¼ qkB T=m:
As we will examine in detail in the following chapters, the Sun is a huge mass of
gas, mostly hydrogen, whose behavior can often be approximated with that of a
perfect gas. In the Sun’s case the gas is composed of a mixture of atoms of different
elements; hence, we can take the mass m as the average mass of the different
components. The averaged mass m is usually measured in the unit of the atomic
mass mu, which we will come back to later. The value of m in this unit is denoted by
the Greek letter l (read mu) and is called the mean molecular weight. So, we can
write
m ¼ lmu :
The atomic mass unit differs by less than 1% from the hydrogen mass and in
grams is
mu ¼ 1:660538921 1024 g:
The mean molecular weight is a number which takes values that are not too
different from l = 1, which corresponds to a perfect gas made only with hydrogen.
The equation of state then becomes
P ¼ q kB T=m ¼ qkB T=ðlmu Þ
and, finally,
P ¼ qRT=l:
The constant
R ¼ kB =mu ¼ 1:38062 1016 erg K1 =1:660538921 1024 g

¼ 8:314472 107 erg K1 g1 ;
is called the universal gas constant.

After having briefly described the structure of matter from quarks to the
macroscopic level (a summary of the relevant length scales is given in Table 1.3),
let us continue, in detail, with the conditions in which matter and, in particular, a
gas, can exist. Here the basic concept is that of thermodynamic equilibrium.
Table 1.3 Length scales from the level of elementary particles to the solar system
Nucleus 10−15 m
Atom 10−10 m
Molecule 10−9 m
Standard meter 1m
Earth-Sun mean distance (astronomical unit, AU) 1.496 1011 m
1.2 Thermodynamic Equilibrium
1.2.1 The Energy Distribution of Gas Particles
The macroscopic properties of a gas in thermodynamic equilibrium (whose

abbreviation is TE) are constant in time and uniform in space. At the microscopic
level thermodynamic equilibrium corresponds to the most probable configuration of
the system of molecules from which the gas is composed. The particles of an ideal
gas interact only through elastic collisions, in which the conservation of the total
amounts of both kinetic energy and momentum allows the exchange of kinetic
energy and momentum between colliding particles. Incidentally, we remember that
a particle with mass m and vector velocity v, has a momentum p, which is the
product of mass and velocity
p ¼ mv;
while its kinetic energy is
1
E ¼ mv2 ;
2
where v is the modulus of v (Fig. 1.4).

Exercise 1.5 Show that applying the Pythagorean theorem twice gives the relation
between the modulus of a vector and its orthogonal coordinates.
Since elastic collisions allow kinetic energy exchanges between molecules, the
microscopic configurations of a perfect gas can differ from each other due to a
different distribution of the possible values of energy among molecules. In other
words, a specific microscopic configuration is characterized by well-defined particle
numbers corresponding to all possible energy ranges (Fig. 1.5). In particular, the
most probable configuration of an ideal gas in TE corresponds to the distribution of
kinetic energy, or equivalently velocity modulus, which is designated after the
Scottish physicist James Clerk Maxwell (1831–1879).
1.2 Thermodynamic Equilibrium 13
Fig. 1.4 The velocity v of a particle (brown sphere) is a vector quantity (red arrow) that is
determined giving modulus and direction, or, equivalently, the three coordinates in a Cartesian
orthogonal system. The coordinates vx, vy and vz are the projections of v on the Cartesian axes and,
hence, the edges of the parallelepiped (in dashes) whose diagonal has a length equal to the
modulus v. Pythagorean theorem (after Pythagoras, Greek philosopher, about 570–495 BC)
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
implies that v ¼ v2x þ v2y þ v2z
Fig. 1.5 Energy distribution for gas particles. We start plotting a set of equal energy intervals,
[Ei, Ei + DE], on the x-axis. On the y-axis we report the probability density DNi/(NDE), where, for
each energy interval DNi is the number of gas particles that have energy in that interval and N is the
total number of particles. Because N is usually a very large number, the ratio DNi/N represents the
probability to find a particle within the energy interval [Ei, Ei + DE]. Assuming that the kinetic
energy of gas particles is a continuous variable, if we make the interval DE approach zero, the
probability density DNi/(NDE) approaches the energy distribution, which is a continuous function
(red curve)
1.2.2 Maxwell Distribution
Consider a gas in thermodynamic equilibrium, characterized by a temperature

T which is unvarying in time and space. Denote by dN/N the fraction of the number
of particles having energy within the values E and E + dE, over the total number of
particles N. Note that the letter d denotes a differential, thus, dN and dE, as dvl
which is considered later, represent infinitesimal intervals of number density,
energy and velocity respectively, and dN/N represents the probability of finding a
particle in the energy interval [E, E + dE]. Let us begin with the simple,
one-dimensional case of the distribution of the velocity component along a given
direction, vl. The probability of finding a particle with velocity within the interval
[vl, vl +dvl] is
pffiffiffi 2
dN=N ¼ 1= pvp eðvl =vp Þ dvl
with
vp ¼ ð2kB T=mÞ1=2 :
As we will see later, vp is the most probable value that the velocity modulus of a
gas particle, with mass m, can take.
The probability dN/N is equal to the velocity interval dvl multiplied by the
velocity function,
pffiffiffi ðv =v Þ2
1= pv p e l p ;
which is called a Gaussian function, after the German mathematician Carl Friedrich
Gauss (1777–1855). The base of the Gaussian is the irrational number
e ¼ 2:7182818 . . .;
which is called the Euler’s number after Leonhard Euler (Swiss scientist, 1707–
1783). As shown in Fig. 1.6, the one-dimensional velocity distribution has the
maximum value at zero velocity and is symmetric around zero, since positive and
negative velocity components have equal probabilities of occurring. The resulting
shape is a peak centered at zero velocity and vanishing for large velocities
(vl < −3vp and vl >3vp). For vl = ±vp, the probability of finding particles reduces by
a factor e with respect to its maximum value. The quantity 2vp defines the char-
acteristic peak width.
To determine the Maxwell distribution for the velocity modulus, which is what is
commonly used because the square of the velocity modulus is proportional to the
kinetic energy, we calculate the probability, dN/N, of finding a particle with its three
Fig. 1.6 The one-dimensional velocity distribution. The velocity vl in units of the most probable
velocity vp is on the x axis and the quantity dN vp/(Ndvl) on the y axis, where dN/N is the number
of particles over the total number in the velocity interval dvl. The resulting velocity distribution
function is a Gaussian. Then, while particles can have any positive or negative velocity, their
number is strongly centered at small velocities, with moduli |vl| < vp (< means less than).
Departing from the maximum value Pm at zero velocity, the number of particles drops sharply and
symmetrically for both positive and negative velocities; in particular, it reduces to Pm/e when
velocity equals ±vp, and is much less than 1 for vl < −3 vp and vl > 3 vp
velocity components within the intervals [vx, vx + dvx], [vy, vy + dvy] and [vz, vz +
dvz]. With the assumption that the velocity components are independent of each
other, we simply multiply the one-dimensional probabilities along each axis, that is,
pffiffiffi 2 pffiffiffi 2 pffiffiffi 2
dN=N ¼ 1=ð pvp Þ eðvx =vp Þ dvx 1= pvp eðvy =vp Þ dvy 1= pvp eðvz =vp Þ dvz ;
from which
pffiffiffi 3
dN=N ¼ 1= pvp eðvx þ vy þ vz Þ=vp Þ dvx dvy dvz :
2 2 2 2
The product of the velocity intervals, dvx dvy dvz, is the volume of an infinitesimal
cube in the space of velocities in Cartesian orthogonal coordinates (Fig. 1.7, panel a).
Inside this elementary volume, the velocity vector v is approximately constant; we
need to replace this volume with an elementary volume, where only the velocity
modulus is approximately constant. The squared modulus v is defined as
v2 ¼ v2x þ v2y þ v2z
and this expression represents a sphere with radius v in the space (vx, vy, vz), Then, it
is natural to consider the velocity space in spherical coordinates (Fig. 1.7, panel b).
Fig. 1.7 The velocity space in Cartesian orthogonal coordinates (panel a) and in spherical
coordinates (panel b). In Cartesian coordinates the velocity vector v has as components the
projections vx, vy and vz onto the three reference axes. In spherical coordinates the velocity vector
has as coordinates the modulus v = √v2x + v2y + v2z , the azimuth u and the inclination h. The
elementary volumes, where the velocity vector is approximately constant to v, are plotted in red for
both the coordinate systems
Here, the elementary volume, where the velocity vector is approximately constant
to v, is the piece of spherical corona whose edges are the infinitesimal increment of
the modulus dv, and the arches vdh and vsenhdu that belong to the sphere with
radius v and correspond to the infinitesimal increments of the inclination h and
azimuth u, respectively. The volume of this piece of corona is the product of the
edges, that is
dVv ¼ dv v dh v senh du ¼ v2 senh dv dh du:
The elementary volume, where the velocity modulus is approximately constant

to v while u and h can take any value, is the spherical corona of radius v and
infinitesimal width dv, whose volume is equal to the sphere’s surface multiplied by
the corona’s width
4pv2 dv:
Replacing the infinitesimal cube with the elementary spherical corona, the
equilibrium distribution for the velocity modulus, or the Maxwell distribution, is
pffiffiffi 3 2 2
dN=N ¼ 1= pvp ev =vp 4pv2 dv:
The Maxwell distribution is represented in Fig. 1.8. It contains a squared

velocity multiplied by a Gaussian function; the former increases with increasing
modulus while the latter decreases. Therefore, the Maxwell distribution reaches a
Fig. 1.8 The Maxwell distribution. The velocity modulus v in unit of the most probable velocity
vp is on the x axis and the quantity dN vp/(N dv) on the y axis, where dN/N is the number of
particles over the total number in the velocity modulus interval [v, v + dv]. The Maxwell function
has its maximum value Pm at v = vp. The number of particles vanishes at v = 0, since no particle is
standing motionless, and progressively reduces for v > vp, but with a tail which is still significant
at high velocity, for v > 3vp
maximum value at a velocity that differs from zero and coincides with the most
probable velocity vp. In addition, the distribution is asymmetric and vanishes much
more quickly for low speeds than it does for high speeds.
Worked Exercise 1.6 The maximum point of the Maxwell distribution.
To determine the velocity where the Maxwell distribution peaks, we set its
derivative to zero (see Worked Exercise 1.7 for the concept of derivative and a
Table with derivatives of elementary functions). Then, neglecting constant
quantities
2
0 ¼ d v ev = vp =dv:
2 2
Expressing the derivative of a product, we get

2 2
0 ¼ dv2 =dv ev =vp þ v2 d ev =vp =dv:
2 2
Moreover, developing the derivative of a function of a function, we get

2 2
0 ¼ 2vev =vp þ v ev =vp d v =vp
2 2 2 2 2
2 2
¼ 2vev =vp þ v 2v=vp ev =vp
2 2 2 2

¼ 2vev =vp 1 v =vp :
2 2 2 2
Because for any finite, non zero velocity the term 2vev = vp never vanishes, it is
2 2
sufficient to search for solutions where

2 2
1 v =vp ¼ 0:
Then
2 2
v =vp ¼ 1;
that is
v ¼ vp ¼ ð2kB T=mÞ1=2 :
Therefore, the maximum value of the Maxwell distribution occurs at the most
probable velocity vp.
Exercise end.
Worked Exercise 1.7 Derivatives.
Consider a continuous function f of an independent variable x and two points
x and x + Dx, which differ by an increment Dx of the independent variable x.
The derivative, denoted by f′ or df/dx, of f at x is the limit of the incremental ratio,
that is, the ratio of the function increment, Df(x) = f(x + Dx) – f(x), and the
increment Dx, as Dx approaches 0. Thus
f 0 ð xÞ ¼ df ð xÞ=dx ¼ lim ½f ðx þ DxÞf ð xÞ=Dx ¼ lim Df ð xÞ=Dx

Dx!0 Dx!0
The derivative df(x)/dx represents the slope of the curve y = f(x) plotted in a
Cartesian coordinate system with axes x and y (Fig. 1.9).
The derivatives of the elementary functions relevant for this text are listed in
Table 1.4.
The notation df(x)/dx emphasizes that the derivative is the ratio of two
infinitesimal increments. An infinitesimal increment of f(x) is called differential and
is the product of the derivative of f and the infinitesimal increment of the inde-
pendent variable, i.e.
Fig. 1.9 Geometric interpretation of the derivative of a function f(x). As Dx approaches 0, the
incremental ratio Df(x)/Dx = BC/AC = tan h approaches the derivative df(x)/dx and the direction of
the segment AB approaches the tangent t to the curve y = f(x) at the point A. Thus, the derivative df
(x)/dx is equal to the slope of the curve y = f(x) at A
Table 1.4 Derivatives of Name f(x) df(x)/dx

elementary functions
Constant c 0
Power xn nxn−1
Exponential ex ex
Natural logarithm ln x 1/x
Decimal logarithm log x 1/(xln10)
Sine sin x cos x
Cosine cos x −sin x
Tangent tan x 1/cos2x
Arcsine arcsin x 1/√(1-x2)
Arccosine arccos x −1/√(1-x2)
Arctangent arctan x 1/(1 + x2)
Sum f+g f ′ + g′
Product fg f′ g +f g′
Function of function f(g(x)) (df′/dg)g′(x)
df ð xÞ ¼ f 0 ð xÞdx:
In this book we will also use differentials of a function of two variables f(x,y),
which have the form
df ðx; yÞ ¼ ½@f ðx; yÞ=@xdx þ ½@f ðx; yÞ=@ydy
where ∂f(x,y)/∂x is the x derivative of f at y = constant and ∂f(x,y)/∂y is the

y derivative of f at x = constant.
Exercise end.
As shown in Exercise 1.8, the Maxwell distribution in terms of kinetic energy
takes the form
dN=N ¼ 2p1=2 Ep3=2 eE=Ep E 1=2 dE:
Worked Exercise 1.8 The Maxwell distribution in terms of energy.

Since kinetic energy and velocity modulus are related by
E ¼ mv2 =2;
it results
v2 ¼ 2E=m
and
v ¼ ð2E=mÞ1=2 :
Then, the infinitesimal variations of velocity modulus and energy are related by
h i
dv ¼ d ð2E=mÞ1=2 ¼ d ð2E=mÞ1=2 =dE dE
¼ ð1=2Þð2E=mÞ1=21 2dE=m ¼ ð2mE Þ1=2 dE:
Denoting with Ep the energy corresponding to the most probable velocity vp, we
have
Ep ¼ mv2p =2 ¼ mð2kB T=mÞ=2 ¼ kB T:
This result confirms that the most probable energy equals kBT. Note that it is
possible to show that the average particle energy, <E>, is equal to
\E [ ¼ 3 kB T=2 ¼ 3Ep =2:
Replacing the velocity modulus with the energy, the Maxwell distribution
becomes
pffiffiffi 3 2 2
dN=N ¼ 1= pvp ev =vp 4pv2 dv
h
pffiffiffi 1=2 i3 E = E
¼ 1= p 2Ep =m e p
4pð2E=mÞð2mE Þ1=2 dE:
Thus
dN=N ¼ 2ð3=2 þ 31=2Þ pð3=2 þ 1Þ Ep3=2 mð3=21 1=2Þ eðE=Ep Þ E ð11=2Þ dE
and, finally, the distribution of the kinetic energy at equilibrium is
dN=N ¼ 2p1=2 Ep3=2 eE = Ep E 1=2 dE:
Exercise end.
Solar gas, unlike an ideal gas, is composed of particles having not only kinetic
energy but also internal degrees of freedom. This is because the atoms can be
excited and ionized and can undergo the corresponding inverse processes of
de–excitation and recombination. These processes occur through inelastic collisions
with other particles, including photons. We expect thermodynamic equilibrium to
hold for excitation and ionization processes when the direct processes of excitation
and ionization are perfectly balanced by the inverse processes. Under these con-
ditions, we apply two further equilibrium distributions for the excitation and ion-
ization energies in a solar gas. These distributions are named after Boltzmann and
the Indian astrophysicist Meghnad Saha (1893–1956), respectively.
Let us discuss first how the gas atoms are distributed over their excited energy
levels when thermodynamic equilibrium holds.
1.2.3 Boltzmann Distribution
In TE the fraction of atoms in the excited state with energy Ei over the total number,
Ni/N, follows the Boltzmann distribution that is
Ni =N ¼ gi eEi =ðkB TÞ =U;
where the energy level, labeled by the index i, can take all the possible integer
values from 0, corresponding to the ground level, to ∞, corresponding to the
ionization in the continuum. Note that here Ei denotes the energy necessary to
excite the atom from its ground state to the excited state of index i; thus, E0 = 0 and
E∞ is the ionization energy (this is different from the energies of the Bohr hydrogen
model in Fig. 1.3). gi is the statistical weight of the level, that is the number of
different states, or sublevels, having the same energy Ei. U is the partition function,
that is, the sum of all the statistical weights, each one multiplied by the corre-
sponding value of the Boltzmann exponential, i.e.
X
U¼ gi eEi =ðkB TÞ ;
i
P
where the symbol i indicates the summation over all energy levels.
Exercise 1.9 Show that, for large temperatures, the partition functions are simply
the sum of all level statistical weights. Since statistical weights are greater or equal
to 1, this would imply a divergence in the sum. However, electrons are not allowed
to populate energy levels very close to continuum because of interactions with
neighboring gas particles.
The Boltzmann distribution shows that, at a fixed temperature T, the populations
of excited levels decrease exponentially with increasing energy Ei; this confirms the
expectation that it is easier to find atoms with electrons in the lower energy levels.
Vice versa, fixing the excitation energy Ei and neglecting the dependence of U on
temperature that is generally weak, the populations of excited levels increase with
increasing temperature; this can be understood because with increasing temperature
the mean kinetic energy of atoms increases fostering the excitation of higher energy
levels.
From the form just written for the Boltzmann distribution we can infer the ratio
between the populations of two excited levels having energies and statistical
weights Ei, gi and Ej, gj, respectively. In fact, the fraction Nj/N is
Nj =N ¼ gj eEj =ðkB TÞ =U
and the ratio Ni/ Nj is

h i h i
Ni =Nj ¼ ðNi =N Þ= Nj =N ¼ gi eEi =ðkB TÞ =U = gj eEj =ðkB TÞ =U ;
from which
Ni =Nj ¼ gi eðEi Ej Þ=ðkB TÞ =gj ;
where Ei − Ej corresponds to the energy jump between the two levels of indices
i and j.
Figure 1.10 shows the variation with temperature of the relative populations of
each level for a gas of hydrogen atoms having only the ground and first excited
levels. While the sum of the two relative populations is obviously constant to 1, the
relative population of atoms in the ground level equals 1 for T = 0 and at low
temperatures, for T < 20,000 K, the atoms are substantially all in the ground level.
Above this the relative population decreases with increasing temperature
approaching the ratio of the statistical weights, g0/(g0 + g1) = 0.2, as T approaches
∞. At the same time the relative population of the excited level is zero for T = 0
and increases with increasing temperature approaching the ratio of the statistical
Fig. 1.10 Variation with temperature of the Boltzmann equation for hydrogen atoms with only
two levels corresponding to the ground and first excited levels. The fraction of atoms in the ground
level over total is defined as N0 =N ¼ N0 =ðN0 þ N1 Þ ¼ 1=ð1 þ N1 =N0 Þ and is represented by the
black line. The fraction of atoms in the excited level is defined as N1 =N ¼ N1 =ðN0 þ N1 Þ ¼
ðN1 =N0 Þ=ð1 þ N1 =N0 Þ and is represented by the red line. The ratio between hydrogen in the
excited and ground levels is given by the Boltzmann distribution, that is N1 =N0 ¼
g1 eðE1 E0 Þ=ðkB TÞ =g0 : The two horizontal thin lines refer to the excitation fractions equal to 0.2
and 0.8, which correspond to the ratios of the level statistical weights to their sum
weights, g1/(g0 + g1) = 0.8, as T approaches ∞. Therefore, at high temperatures,

that is for T >200,000 K, the atoms are substantially distributed over the levels
according to their statistical weights.
Exercise 1.10 The statistical weights of hydrogen energy levels are given by
gn ¼ n 2 ;
where n is the principal quantum number. Verify the values for the ratios of the
level statistical weights to their sum given in the text.
In Chap. 3 we will justify the shape of the Boltzmann equation by an analogy
based on the equation of the hydrostatic equilibrium, which is introduced in that
chapter. Next we look at how atoms are distributed among the different ionization
states in conditions of thermodynamic equilibrium.
1.2.4 Saha Distribution
In TE the ionization processes of a chemical species are in equilibrium with the

inverse processes of recombination. This equilibrium is represented by the relation
E $ Eþ þ e
where E indicates an atom or ion and E+ indicates the same chemical species in the
next ionization state so that one electron e has been released. The double arrow
indicates that ionization and recombination processes occur in both directions. In
particular, in TE the product of the number densities of the dissociated species, N+,
and electron number density, Ne, divided by the number density of the ionizing
species, N, is a function only of the equilibrium temperature and satisfies the Saha
equation, that is
N þ Ne =N ¼ uðT Þ;
where the Saha function u(T) is

3=2 E =ðk TÞ
u ¼ ðU þ =U Þ2 2pme kB T=h2 e I B :
Here
me ¼ 9:10956 1028 g
is the electron mass and
h ¼ 6:62620 1027 erg s
is the Planck constant, named after the German physicist Max Planck (1858–1947).
U and U+ are the partition functions, or statistical weights, for the two species
before and after ionization, respectively. The factor 2 expresses the two possible
orientations of the electron spin, that is the intrinsic rotation motion of the electron
around itself, and EI is the ionization potential, that is, the energy which is required
for ionizing a species from its ground state to the ground state of its next ion.
As the Saha function is given by the Boltzmann exponential, e EI =ðkB TÞ , multi-
plied by the power of the electron most probable energy (kBT)3/2, ionization is
reduced by increasing the ionization potential and increased by increasing the
temperature.
Figure 1.11 shows the fractions, over total population, of neutral and ionized
hydrogen, as function of temperature for a pure hydrogen gas in TE. These ion-
ization fractions are computed in the Exercise 1.11.
Fig. 1.11 Temperature variation of the Saha equation for hydrogen. The black line represents
the fraction of neutral atoms over total N/Nt and the red line refers to the fraction of ionized atoms
N+/Nt
Worked Exercise 1.11 Ionization fractions of a hydrogen gas in TE.

Denoting with Nt the total number of hydrogen nuclei per unit of volume, the
fraction of neutral atoms over total is defined as
N=Nt ¼ N=ðN þ N þ Þ ¼ 1=ð1 þ N þ =N Þ
and the fraction of ionized atoms is defined as
N þ =Nt ¼ N þ =ðN þ N þ Þ ¼ðN þ =N Þ=ð1 þ N þ =N Þ:
The ratio of ions to neutral atoms N+/N is inferred from the Saha equation, that is
N þ Ne =N ¼ u;
where the electron and proton number densities are equals in a gas of only
hydrogen. Therefore
N þ ¼ Ne
and the Saha equation becomes
N 2þ =N ¼ u:
Inserting the total number of hydrogen nuclei, Nt, which is constant, we obtain

N 2þ =Nt2 Nt2 =N ¼ u
and in terms of the ionization fractions

N 2þ =Nt2 ðNt =N Þ ¼ u=Nt
Considering the neutral fraction as an unknown quantity, we write
N=Nt ¼ x
and the ionized fraction is
N þ =Nt ¼ðNt N Þ=Nt ¼Nt =Nt N=Nt ¼ 1 x:
Then, the Saha equation becomes
ð1 xÞ2 =x ¼ u=Nt :
Multiplying by x and expanding the squared term we get
1 2x þ x2 ¼ xu=Nt :
Grouping the terms in x we have
x2 ð2 þ u=Nt Þx þ 1 ¼ 0:
The two solutions to this quadratic equation are

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
x ¼ ð2 þ u=Nt Þ ð2 þ u=Nt Þ2 4 2:
The only solution having physical meaning is the one with the minus sign before
the square root. This is because as T approaches ∞ this solution implies, correctly,
that the neutral fraction approaches 0. The fractions of neutral and ionized
hydrogen, as function of temperature inferred from this solution, are plotted in
Fig. 1.11.
Exercise end.
The fraction of neutral atoms equals 1 for T = 0 and remains close to 1 at low
temperatures meaning that hydrogen is substantially neutral for T < 10,000 K.
Then the neutral fraction decreases with increasing temperature, approaching 0 as
T approaches ∞.
Simultaneously with this, the ion fraction vanishes at T = 0 and increases with
increasing temperature approaching 1 as T approaches ∞. At high temperatures, i.e.
for T > 20,000 K, hydrogen atoms are substantially all ionized.
The combined application of the Boltzmann and Saha equations allows us to
calculate the fraction of atoms in a particular excitation level of a specified ion-
ization state. As we will see later, this type of information is important for inter-
preting the opacity as well as the emission of the Sun and stars. Thus, for example
in the simple case of a two level hydrogen atom, the fraction of neutral hydrogen in
the excited level can be written
N1 =Nt ¼ ðN1 =N Þ ðN=Nt Þ:
Then, the fraction of excited hydrogen over neutral hydrogen, N1/N, is obtained
by the Boltzmann equation, as shown in the caption of Fig. 1.10, and the fraction of
all hydrogen which is neutral, N/Nt, is inferred from the Saha equation, as described
in the Worked Exercise 1.11. The fraction of all hydrogen which is neutral and
excited is plotted as a function of temperature in Fig. 1.12. This fraction is clearly
peaked around T = 16,000 K. In fact, at temperatures slightly lower than this one,
the number of excited hydrogen atoms drops, and, at slightly higher temperatures, it
is the number of neutral hydrogen that is strongly reduced.
Worked Exercise 1.12 The degeneracy of a free electron.
Energy states of a free electron are a continuum. Given a small (infinitesimal)
increment dv of the velocity modulus v, the degeneracy of a free electron with
velocity modulus between v and and v + dv is the number of states that the electron
can access in the three dimensional velocity space. As discussed in relation to the
Fig. 1.12 Temperature variation of the combined Saha and Boltzmann equations for the fraction
of neutral hydrogen in the excited level over total hydrogen
Maxwellian distribution, an electron with speed within v and v + dv, is represented

in the velocity space by any point of the spherical corona with radius v and
infinitesimal width dv; this corona has volume equal to the sphere’s surface 4pv2
multiplied by the corona’s width dv (Fig. 1.7, panel b). Then, the electron occupies
the volume dVv given by
dVv ¼ 4pv2 dv:
This result has to be modified to take into account the constraints imposed by the
Heisenberg uncertainty principle and the Pauli exclusion principle, both established
by quantum mechanics (Werner Karl Heisenberg was a German physicist, 1901–1976
and Wolfgang Pauli an Austrian physicist, 1900–1958). We will discuss again the
uncertainty principle in Chap. 6, and we disclose here only what is strictly necessary.
The uncertainty principle does not allow the electron to continuously span the
velocity space. In fact, this principle requires that the product of the variations of
the momentum components, dpx, dpy, dpz, and the variations of the position.
components, dx, dy, dz, cannot be lower than h3, where h is the Planck constant,
that is
dx dy dz dpx dpy dpz h3 :
Note that for the constant on the right side we have chosen h3 instead of (h/2p)3
to be consistent with the degeneracy computed according to quantum statistics.
Since the momentum components are the products of the electron mass me and the
corresponding velocity components, we can write in terms of velocity
dpx dpy dpz ¼ m3e dvx dvy dvz ;
where dvx dvy dvz is an elementary volume of the velocity space. Therefore, the
uncertainty principle states that an electron, having velocity within v and v + dv,
can occupy a minimum volume in the velocity space, dVm v , such that
dx dy dz m3e dVvm ¼ h3 ;
from which

dVvm ¼ h3 = dx dy dz m3e :
Because of the exclusion principle, the volume dVm v can actually host two
electrons with opposite spins, so that the volume at disposal to a single electron is
given by half this value, that is

dVvm ¼ h3 = 2dx dy dz m3e :
One electron, sitting where the electron number density equals Ne electrons per
volume unit, has an elementary volume available in the position space, dx dy dz,
such that
Ne dx dy dz ¼ 1
and, hence,
dx dy dz ¼ 1=Ne :
Therefore, the electron with velocity within v and v + dv can occupy a minimum
volume in the velocity space equal to
dVvm ¼ h3 =ð2dx dy dz m3e Þ ¼ h3 Ne =ð2m3e Þ:
In conclusion, this minimum volume represents the size of a velocity sublevel

and, hence, the free electron degeneracy is given by the ratio of the volume dVv to
the minimum volume dVm v , that is

dge ðvÞ ¼ dVv =dVvm ¼ 4pv2 dv2m3e = h3 Ne ¼ 8pm3e v2 dv= h3 Ne :
Exercise end.
Worked Exercise 1.13 The Saha distribution.
We can justify the shape of the Saha equation starting from the Boltzmann
equation written for a process which leads an atom, initially in its ground state, to
dissociate itself to the next ion, again in its ground state, and one free electron
having velocity modulus within the interval [v, v + dv]. The energy required by the
process allows the atom to make the energy jump between the ground states of the
two next ions, and, moreover, feeds the kinetic energy of the free electron.
Therefore, the necessary energy is the sum of the ionization potential, EI, and the
kinetic energy of the electron, mev2/2. Referring to the Boltzmann distribution for
the populations of the atomic energy levels, we expect that the ratio of the ion
number in the ground level with one free electron having velocity within v and
v + dv, dN+0 (v), and the atom population in the ground level, N0, can be written as

dN0þ ðvÞ=N0 ¼ dg0þ ðvÞ=g0 e ðEI þ me v =2Þ=ðkB TÞ :
2
Here g0 is the statistical weight of the species which dissociates itself and dg+0 (v)
is the statistical weight of the ionized state. The interpretation of the latter weight
deserves particular attention for the presence of two particles. First, we can write
this statistical weight as comprised of the statistical weight of the ionized species in
the ground state, g+0 , multiplied by the statistical weight, dge(v), of the electron
having velocity within v and v + dv, that is
dg0þ ðvÞ ¼ g0þ dge ðvÞ:
Exercise 1.12 shows that the degeneracy of a free electron is

dge ðvÞ ¼ 8pm3e v2 dv= h3 Ne :
Inserting this result in the Boltzmann equation we find

dN0þ ðvÞ=N0 ¼ 8pm3e v2 g0þ eðEI þ me v =2Þ=ðkBT TÞ
2
dv= g0 h3 Ne :
Up to now the final state of the ionization process involves one electron with a
well-determined value of the velocity modulus. To allow for all electron speeds it is
sufficient to sum up all possible velocity moduli v within 0 and ∞. Since the Planck
constant is a very small number, we can assume that v varies with continuity and
replace the sum with an integral (see Worked Exercise 1.4 for the concept of
integral and a Table with some common indefinite integrals).
Z1 Z1h
i
N0þ =N0 8pm3e v2 g0þ eðEI þ me v =2Þ=ðkB TÞ
2
¼ dN0 þ ðvÞ=N0 ¼ = g0 h3 Ne dv;
0 0
where N+0 is the ion number in the ground state per volume unit. Removing the
quantities that are independent of velocity from the integral, we get
h Z1
i
N0þ ðvÞ=N0 8pm3e g0þ eEI =ðkB TÞ = v2 e m e v =ð2kB TÞ
2
¼ 3
g0 h Ne dv:
0
Using the substitution
x ¼ ½me =ð2kB T Þ1=2 v;
which implies
dx ¼ ½me =ð2kB T Þ1=2 dv;
the integral becomes
Z1 Z1
2 me v2 =ð2kB TÞ
x2 ½me =ð2kB T Þ1 ex ½me =ð2kB T Þ1=2 dx
2
ve dv ¼
0 0
Z1
3=2
x2 ex dx:
2
¼ ½me =ð2kB T Þ
0
Since
Z1 .
x2 ex dx ¼ p1=2 4;
2
we then obtain
. Z1

N0þ Ne =N0 ¼ 8pm3e g0þ e EI =ðkB TÞ g0 h3 Ne v2 eme v =ð2kB TÞ dv
2
0
. 3=2 Z1

8pm3e g0þ e EI =ðkB TÞ x2 ex dx
2
¼ 3
g0 h N e me =ð2kB T Þ
0
. 3=2

¼ 8pm3e g0þ e EI =ðkB TÞ g0 h3 N e me =ð2kB T Þ p1=2 =22

¼ 232 p1 þ 1=2 m33=2
e ð2kB T Þ3=2 g0þ eEI =ðkB TÞ = g0 h3 Ne

¼ 2ð2pme kB TÞ3=2 g0þ e EI =ðkB TÞ = g0 h3 Ne :
Therefore, the Saha equation becomes

3=2 E =ðk TÞ
N0þ Ne =N0 ¼ g0þ =g0 2 2p me kB T=h2 e I B :
We can infer the form with the total populations of the two next ionization states,
N and N+, and with the relative partition functions, U and U+, by means of the
Boltzmann equations for the ground levels of both the atom and ion. In fact,
because
N0 =N ¼ g0 =U
and
N0þ =N þ ¼ g0þ =U þ ;
then,
N ¼ N0 U=g0
and
N þ ¼ N0þ U þ =g0þ :
Therefore

N þ Ne =N ¼ N0þ U þ =g0þ Ne =ðN0 U=g0 Þ

¼ N0þ Ne =N0 g0 =g0þ ð U þ =U Þ
3=2 E =ðk TÞ
¼ ðg0þ =g0 Þ2 2p me kB T=h2 e I B g0 =g0þ ðU þ =U Þ:
Simplifying the statistical weights of the ground levels we obtain the Saha
equation for the total populations of the two next ions
N þ Ne =N ¼ ðU þ =U Þ2ð2pme kB T=h2 Þ3=2 e EI =ðk B TÞ :
Exercise end.
Worked Exercise 1.14 Integrals.
1.2.5 The Indefinite Integral
A function F(x) is called a primitive of a function f (x) when f(x) is the derivative of
F(x), that is
f ð xÞ ¼ d F ð xÞ=dx:
Finding a primitive is the inverse operation to find a derivative. Since the

derivative of a constant is 0, it is immediately apparent that given a primitive there
is an infinite number of other primitives that differ from each other by a constant
value c,
f ð xÞ ¼ d F ð xÞ=dx ¼ d ½F ð xÞ þ c=dx
The comprehensive expression of all the primitives of a given function is called

indefinite integral and is denoted by the symbol
Z
f ð xÞdx ¼ F ð xÞ þ c:
The integrals of elementary functions of interest for this book are listed in
Table 1.5.
1.2.6 The Definite Integral
The definite integral of a function f(x) over the x interval [a, b] is denoted by
Table 1.5 Indefinite integrals of elementary functions

R
Name f (x) f ð xÞdx
Constant c cx
Power xn xn þ 1 =ðn þ 1Þ
x
Exponential e ex
ax
xe ðx=a 1=a2 Þeax
2 ax
xe ðx2 =a 2x=a2 þ 2=a3 Þeax
Reciprocal 1/x ln x
Sine sin x −cos x
Cosine cos x sin x
1/cos2x tan x
p
1= ð1 x2 Þ arcsin x
p
1= ð1 x2 Þ arccos x
1=ð1 þ x2 Þ arctan x
Sum f+g R
f ð xÞdx þ R gð xÞdx
R
Integral by parts f g0 f g f 0 gdx
Note that many results listed here can be obtained from Table 1.4 by simply exchanging the order of the second
and third columns. Integration by parts is the inverse operation of derivative of a product. An arbitrary constant
can be added to the functions listed in column 3
Zb
f ð xÞdx
a
and in a Cartesian coordinate system xy it represents the area delimited by the curve
y = f(x), the x axis and the two straight lines x = a and x = b (Fig. 1.13).
The defined integral of a function f(t) over the interval [a, x], where the upper
limit x is variable, that is
Zx
f ðtÞdt
a
is a primitive of the function f. Thus

2 3
Zx
d4 f ðtÞdt5 ¼ f ð xÞ:
a
Denoting with F(x) + c the indefinite integral of f(x), we get
Zx
f ðtÞdt ¼ F ð xÞ þ c:
a
Fig. 1.13 Geometric interpretation of a definite integral. The definite integral R ba f(x)dx is the area
delimited by the curve y = f(x) (red curve), the x axis and the two straight lines x = a and
x = b. Given n points, a = x1 < x2 <…xi <…xn = b, each differing from the next by Dx, the
definite integral can be approximated by the sum of the areas of the rectangles where Dx is the base
and f(xi) is the height. As n increases and Dx decreases, the total area of the rectangles approaches
Rb P
the definite integral, that is, a f ð xÞdx ¼ limDx!0; i!1 i f ðxi ÞDx : If F(x) is a primitive of f, then dF
P P
(x)/dx = f(x) and we can write limDx!0; i!1 i f ðxi ÞDx ¼ limDx!0;i!1 i ½dF ðxi Þ=dxDx ¼
P R
b
limDx!0;i!1 i DF ðxi Þ ¼ F ðbÞ F ðaÞ. Thus, we get f ð xÞdx ¼ F ðbÞ F ðaÞ:
a
Since the integral vanishes if the upper limit is equal to the lower limit, that is
x = a, it results
Za
f ðtÞdt ¼ F ðaÞ þ c ¼ 0
a
and, hence,
c ¼ F ðaÞ:
Thus, in general we get
Zx
f ðtÞdt ¼ F ð xÞ F ðaÞ
a
and, taking x = b,
Zb
f ð xÞdx ¼ F ðbÞ F ðaÞ:
a
Therefore, the definite integral of a function f(x) between a and b is equal to the
difference F(b)−F(a), where F(x) is a primitive of the function f.
Exercise end.
Up to this point the energy distribution for photons was left out of our study of
the TE conditions. To approach this, let us first discuss in detail the nature of light.
1.3 The Nature of Light
1.3.1 Waves, Photons and the Electromagnetic Spectrum
The Sun is our primary source of light and the mechanism of light generation is
strongly related to the structure of matter that we have just discussed. We know that
atoms can emit and absorb light, and these processes correspond to electron tran-
sitions between two of their possible states with different energies around the
nucleus. But, what is the nature of light? Actually, not all the light emitted by atoms
is visible to our eyes and, since it is the electromagnetic force that binds electrons to
nuclei, the expression electromagnetic radiation is the most precise definition for
both visible and invisible light. Most of the observed behavior of electromagnetic
radiation can be described in terms of wave properties. This phenomenon belongs to
our common experience, for example, in the case of sea and sound waves. The
pattern of crests and troughs that the sea surface (in wave motion) shows close to an
indented coastline can appear complex. However, this apparent complexity is the
result of a simple structure. This pattern can be considered the superposition of a
large number of monochromatic elementary waves, like the sinusoidal wave in
Fig. 1.14, which is a simple one–dimensional example. While all monochromatic
waves move with the same velocity c, each monochromatic wave is characterized by
a different wavelength k (k is a Greek letter to be read lambda), which is the distance
between two adjacent crests (or troughs). The wave frequency m (m is a Greek letter to
be read nu) is the number of crests passing a fixed point, for example a buoy along
beach, in a fixed time; for example, one second. This means an elementary wave
moves one wavelength in a time equal to the inverse of its frequency, so that the
wave velocity c equals the product of wavelength and frequency, that is
c ¼ km:
Like a wave, light can be reflected, for instance by a mirror, diffracted, that is
displaced, for example passing from air to water, and, can interfere with other light.
In particular, in the interference of two monochromatic waves having the same
wavelength, the waves add algebraically so that there is a maximum strengthening
Fig. 1.14 The characteristics of a one–dimensional, monochromatic, sinusoidal wave.

A sinusoidal wave of unit maximum amplitude has the mathematical shape y ¼ sin h; where h
is the wave phase (panel a). The phase h varies with the time t and with the distance x according to
h ¼ 2pðmt x=kÞ þ u; where m is the frequency, k the wavelength and u the phase constant.
Therefore, the wave amplitude at a fixed point and at a fixed time depends on the corresponding
value of the phase; in particular, for x = t = 0, the wave has amplitude y ¼ sin u: The points where
the amplitude is constant in time are characterized by a constant phase, that is, h ¼ constant and,
since it is u = constant, 2pðmt x=kÞ ¼ constant and mt x=k ¼ constant; so that x=k ¼
mt þ constant and, finally x ¼ kmt þ constant: Therefore, the points with equal phases cover a
distance x = km in the time t = 1 s, that is, they move with velocity km ¼ c; called the phase
velocity. Panel b shows the spatial variation of the wave for t = 0, that is y ¼ sinð2px=k þ uÞ;
with u = p/2 and distance measured in units of wavelength
if the two waves have the same phase, that is their crests overlap as well as their
troughs do; on the contrary, there is a minimum when the two waves have opposite
phases and, hence, the crests of a wave and the troughs of the other overlap.
A light wave moves in vacuum with a velocity c very close to 3 105 km s−1.
This is a large value for a velocity! For example, it allows the solar light reflected
by the Moon to reach us on the Earth in a bit more than 1 s, and the light directly
emitted by the Sun to arrive in just over 8 min. The velocity of light in a vacuum is
considered the upper limit velocity in nature by the theory of special relativity
proposed in 1905 by the German physicist Albert Einstein (1879–1955).
Light can have a wide range of wavelengths, called the electromagnetic spectrum
(Table 1.6). Because the light velocity is constant, the frequency, m = c/k, decreases.
with increasing wavelength. So, in the first row of the Table, at one extreme of the
electromagnetic spectrum we find very short wavelengths and very high frequencies:
this is the domain of the c rays (c is a Greek letter to be read gamma), with
k * 10−13 m and m * 1021 Hz. Here, we use the symbol * to mean about equal
to, and 1 Hz = 1 Hertz is defined as the frequency of 1 cycle per second and is
named after the German physicist Heinrich Rudolf Hertz (1857–1894). Then, with
1.3 The Nature of Light 37
Table 1.6 The Radiation Wavelength (m) Frequency (Hz)

electromagnetic spectrum
c rays 10−13 1021
−10
X rays 10 (0.1 nm) 1018
−8
Ultraviolet (UV) 10 1016
−7
Visible 6 10 (0.6 lm) 1014
−4
Infrared 10 1013
−2
Microwaves 10 (1 cm) 1010 (10 GHz)
Broadcast TV 1 108
Broadcast radio 5 10 (0.5 km)
2
106 (1 MHz)
Radiofrequencies 105 (0.1 Mm) 104 (10 kHz)
For each type of radiation a typical wavelength and frequency are
listed
gradually increasing wavelength and decreasing frequency, we find the X rays,

ultraviolet and visible radiation. The visible part of the spectrum spans the range
between 0.4 and 0.8 lm and only here can the electromagnetic radiation be detected
by our eyes. The increase of k, or the decrease of m, in the visible is perceived by our
eyes as a variation of the light color, which passes from blue to red.
Beyond the visible range, at longer wavelengths, we find the infrared and then
the radio waves. Among these longer wavelengths we note the microwaves, with
k * 1 cm and m * 10 GHz, which are produced, for example, in a common home
oven. Finally, at the low frequency extreme of the electromagnetic spectrum, there
are the radiofrequencies, with k * 0.1 Mm and m * 10 kHz.
The description of the electromagnetic radiation in terms of waves is actually only
a part of the story. As quantum mechanics tells us, the full story is that every object
shows the properties typical of both a particle and a wave, even though this dualism
does not usually appear in macroscopic bodies. That is, light can also be described in
terms of particles. In effect, the particles composing light are the photons that the
Standard Model considers the mediators of the electromagnetic force.
The kinetic energy of a particle in motion is proportional to the product of mass
and squared velocity, that is
E ¼ mv2 =2:
A photon has no mass and its energy is directly proportional to the frequency of
light and inversely proportional to wavelength, that is
E ¼ hm ¼ hc=k;
where h is the Planck constant.

Returning back to the atomic transitions, it appears that we are justified in
describing light emissions and absorptions in terms of collisions between atoms and
photons. In particular, the wavelength of an emitted or absorbed photon corre-
sponds perfectly to the electron energy jump in an atomic transition that can involve
both bound and free states This is another basic piece of information for interpreting
solar and stellar spectra.
After having explained some important features about the nature of electro-
magnetic radiation, we can now justify the shape of the energy distribution for
photons, assuming that they are in equilibrium with the matter’s atoms, whose
equilibrium distributions we have already examined.
1.3.2 The Black Body
A body radiating in conditions of thermodynamic equilibrium is called a black

body. In order to picture what a black body is, let’s think of a wood-burning oven
but with extreme features. The external walls are perfectly insulated and the energy
losses from the entrance are negligible, so that practically a black body is a cavity at
a temperature that is uniform in space and constant in time. Its internal walls absorb
all the radiation incident on them, without reflections. At the same time, the walls
reemit radiation, which is more intense in the cavity as the temperature is higher.
The large number of emission and absorption processes guaranties that the photons
inside the cavity are distributed in energy, and, hence, in frequency, in a fashion
depending only on the equilibrium temperature. At the microscopic level this
temperature corresponds to a most probable energy equal to kBT, which is the same
for both the radiating atoms of the walls and the radiated photons.
At any point P, and at any time t, the amount of light energy flowing in the
direction s through the surface dS, orthogonal to s, in the time interval dt, in the
solid angle dx around s and in the wavelength band dk, is given by the product
dEk ¼ Ik dS dt dk dx:
The quantity Ik is called the radiation intensity and represents the energy radi-
ated from a point in a direction, per unit normal area, time, wavelength and solid
angle (Fig. 1.15, panel a). In the case of a black body, the radiation intensity is a
characteristic function, called the Planck function, and is denoted by Bk. In order to
find the shape of Bk, we exploit the fact that the TE conditions imply the detailed
balance of the microscopic processes and, hence, the number of a certain type of
transition equals the number of the inverse transitions. In particular, here we use the
balance of radiation emission and absorption.
Let us consider an elementary volume inside a black body
dV ¼ dS ds;
such that the width ds is lying along s and the surface dS is orthogonal to
s (Fig. 1.15, panel b). The energy, dEk, emitted by dV in the time interval dt, in the
wavelength range dk and in the solid angle dx around s, is proportional to the
Fig. 1.15 Panel a: characteristics of the radiation intensity. The radiation intensity, Ik, is the
energy radiated per unit normal area, time, wavelength and solid angle. Thus, in any point P and at
any time t, the amount of light energy that flows in the direction s through the surface dS
orthogonal to s, in the time interval dt, in the solid angle dx around s and in the wavelength
interval dk, is dEk ¼ Ik dS dt dk dx: Panel b: detailed balance for the black body radiation.
Radiation emissions and absorptions occur at an equal rate within a volume dV = Sds along a
direction s inside a black body, which implies jk = kkBk
product dV dt dk dx by a proportionality coefficient, jk, called the emission

coefficient, which is a measure of the local emission probability. Thus, we get
dEke ¼ jk dV dt dk dx ¼ jk dS ds dt dk dx:
We denote with kkds the fraction of the radiation intensity which is absorbed
when covering the distance ds along s, in the same wavelength and time intervals
and in the same solid angle, where kk is called the absorption coefficient. The
absorbed intensity is then given by kkdsBk and the corresponding absorbed energy
is
dEka ¼ kk Bk dS ds dt dk dx:
Because of equilibrium, the emitted and absorbed energies are equal, that is
jk dS ds dt dk dx ¼ kk Bk dS ds dt dk dx:
To simplify the equality we cancel out the common factors on the left and right
hand sides to give
j k ¼ kk B k :
Then, the Planck function is also the ratio of the emission to the absorption
coefficient in equilibrium conditions, that is
Fig. 1.16 Two—level atom and allowed transitions. The arrows between the two levels represent
the electron energy jumps and are in red and yellow for collisional and radiative transitions,
respectively. In the order from left to right the indicated processes are: (1) Collisional excitation, in
which an incident gas particle gives the atom a part of its kinetic energy triggering excitation.
(2) Radiative excitation or absorption, in which electron excitation occurs at the expense of
radiation energy with the absorption of a photon. (3) Collisional de-excitation, in which an excited
atom collides with a gas particle and the electron decays to the ground level releasing energy to
increase the kinetic energy of the incident particle. (4) Spontaneous emission, in which the energy
of the decaying electron feeds an emitted photon. (5) Stimulated emission, represented with a
double arrow in the figure, in which the ambient radiation field catalyzes the emission of a photon
having the same direction and frequency of the photon that acts as stimulus. We note that the
stimulated emission is the basic process of the laser effect
Bk ¼ jk =kk :
For the sake of simplicity, in order to find an expression for the emission and
absorption coefficients we assume that atoms have only two discrete energy levels
(Fig. 1.16). As Einstein first recognized, neglecting collisional transitions, an
electron can decay from the excited level to the ground level either spontaneously,
after a short lifetime in the excited state, or through triggering by the ambient
radiation field. This provides two possible emission processes: spontaneous emis-
sion and stimulated emission and the total emission coefficient is the sum
jk ¼ jsp
k þ jk :
st
The spontaneous emission coefficient jsp k is the product of several factors. The
first three factors are: the number of excited atoms, N2, the probability that the
spontaneous emission occurs, A21, and the energy of the emitted photon, hc/k. A21 is
called the Einstein coefficient for spontaneous emission. The effective value of A21
is characteristic of each particular atomic transition and has to be computed with the
methods of quantum mechanics. The spontaneous emission coefficient contains the

factor 1/(4p), because all directions within the full solid angle, which encompasses
R
dx = 4p radians, have the same emission probability. The last factor is the so–
called profile, which is a function of wavelength, wk (w is a Greek letter to be read
psi) and which represents the probability that the sublevels of the excited level are
occupied. Therefore, the complete form of jsp k is
jsp
k ¼ N2 A21 ðhc=kÞ½1=ð4pÞwk :
The formula for the stimulated emission jstk is analogous except for we have to
include the radiation field Bk, which acts as a catalyst, and use the Einstein coef-
ficient for stimulated emission B21 in the place of A21. Therefore,
jstk ¼ N2 B21 ðhc=kÞ½1=ð4pÞwk Bk :
So the total emission coefficient is the sum
jk ¼ jsp
k þ jk ¼ N2 A21 ðhc=kÞ½1=ð4pÞwk þ N2 B21 ðhc=kÞ½1=ð4pÞwk Bk :
st
The expression for the absorption coefficient kk is analogous to that for the
spontaneous emission but with N1 replacing N2 and the Einstein coefficient for
absorption B12 replacing A21, that is
kk ¼ N1 B12 ðhc=kÞ½1=ð4pÞwk :
Note that all the three emission and absorption profiles have been taken to be
equal to the spontaneous emission profile wk, which is not always true. By inserting
the formulas for the emission and absorption coefficients in the relation Bk = jk/kk,
as shown in the Worked Exercise 1.15, we can infer that the Planck function has the
form

Bk ¼ 2hc2 =k5 =½ ehc =ðkkB TÞ 1:
Worked Exercise 1.15 Determine the Planck function.

Inserting the formulas for the emission and absorption coefficients in the Planck
function, we can simplify by removing the term (hc/k)[1/(4p)]wk, which is in
common to numerator and denominator, to obtain
Bk ¼ jk =kk ¼ ðN2 A21 þ N2 B21 Bk Þ=ðN1 B12 Þ:
Gathering the terms in Bk we obtain
N1 B12 Bk ¼ N2 A21 þ N2 B21 Bk ;

and
N1 B12 Bk N2 B21 Bk ¼ N2 A21 :
Factoring out the common term Bk, we get
ðN1 B12 N2 B21 ÞBk ¼ N2 A21 ;
thus, Bk is
Bk ¼ N2 A21 =ðN1 B12 N2 B21 Þ:
Here we factor out the ratio A21/B21 in the numerator, and the ratios B12/B21 and
N1/N2, in the denominator, which are all known quantities, i.e.
Bk ¼ ðN2 A21 =B21 Þ=½N1 B12 =B21 N2 ¼ ðA21 =B21 Þ=½N1 B12 =ðN2 B21 Þ 1:
The ratios of the Einstein coefficients are
A21 =B21 ¼ 2hc2 =k5
and
B12 =B21 ¼ g2 =g1 :
The Planck function is then given by

Bk ¼ ðA21 =B21 Þ=½N1 B12 =ðN2 B21 Þ 1 ¼ 2hc2 =k5 =½ N1 g2 =ðN2 g1 Þ 1:
Moreover, the Boltzmann distribution yields the following expression for the
ratio of the populations of the two atomic levels
N1 =N2 ¼ ðg1 =g2 Þ eð E1 E2 Þ=ðkB TÞ ¼ ðg1 =g2 Þ ehc=ðkkB TÞ ;
where to highlight the dependence on wavelength we have replaced the excita-

tion energy E2 − E1 with the energy of the corresponding emitted photon
E = hm = hc/k. Therefore, we have
N1 g2 =ðN2 g1 Þ ¼ ehc=ðkkB TÞ ;
and, finally, the Planck function is

Bk ¼ 2hc2 =k5 =½ ehc =ðkkB TÞ 1:
Exercise end.
Here we recall that h is the Planck constant and kB is the Boltzmann constant.
Fig. 1.17 The Planck function for two temperature values. With increasing temperature the
wavelength of the curve’s maximum (marked by a vertical thin line) moves from km = 0.58 lm in
the yellow at T = 5000 K, to km = 0.4833 lm in the blue at T = 6000 K, according to the Wien
law km = 0.29/T cm. Note that at the extremes of the spectrum, the Planck function vanishes,
decreasing exponentially at low energies, for k approaching 0 (Rayleigh-Jeans approximation to
Bk), and as k−4 at high energies, for k approaching ∞ (Wien approximation)
The Planck function is plotted in Fig. 1.17 for two temperature values. We can
see that the maximum value of the Planck function increases with increasing
temperature of the black body. The wavelength km, where the black body emission
Bk is a maximum satisfies the Wien’s displacement law (Wilhelm Wien, German
physicist, 1864–1928), according to which the temperature T and the wavelength
km are inversely proportional, that is
km ¼ 0:20 hc=ðkB T Þ ¼ 0:20 6:62 1027 erg s 2:99 1010 cm s1 =ð1:38 1016 erg K1 =TÞ
¼ 0:29=T cm K1 ;
with T in K degrees Therefore, as we expect, the maximum of the Planck function

moves towards higher energies with increasing temperature and, hence, towards
smaller wavelengths.
Exercise 1.16 Derive the Wien’s displacement law. Hint. Use x = hc/(kkBT) as
independent variable in the Planck function and neglect 1 in the term ex −1 of the
final equation.
Exercise 1.17 Derive the Wien and Rayleigh-Jeans formulas, which are the
approximations to the Planck function at the long and short wavelength ends of the
spectrum, respectively (see Fig. 1.7). John William Strutt, 3rd Baron of Rayleigh
(1842–1919), and Sir James Jeans (1877–1946) are two English physicists who
contributed to radiation theory.
The Planck function is for the photon energies the analog of the Maxwell dis-
tribution for the gas particle kinetic energies. However, the Maxwellian function is
a probability density and is normalized to 1, that is, its integral over all particle
speeds is unity. On the contrary, the Planck function represents a radiation intensity
and its integral over all wavelengths, denoted with B, is given by
B ¼ rT 4 =p:
Here, r = 5.6696 10−5 erg cm−2 s−1 K−4 is called the Stefan-Boltzmann
constant (Joseph Stefan, Slovene physicist, 1835–1893). Moreover, it is possible to
show that B is related to the full amount of energy which the black body emits per
unit time and surface, that is, the flux F, by
F ¼ pB:
Then, inserting the value of B, we find the black body emission flux
F ¼ rT 4 ;
which is called the Stefan-Boltzmann law.

Worked Exercise 1.18 The Planck function in unit of frequency.
The Planck function in unit of frequency takes the form
Bm ¼ 2hv3 c2 ðehm=kB T 1Þ1 :
This expression can be obtained from Bk considering the frequency and wave-
length bands corresponding to the same amount of radiation intensity, that is
Bm dv ¼ Bk dk:
Since
m ¼ c=k;
then,

dm ¼ d ðc=kÞ ¼ c=k2 d k;
and, considering the moduli, because here we are interested only in the band width,

jdmj ¼ c=k2 jd kj:
Therefore,
Bm ¼ Bk dk=dv ¼ Bk k2 =c;
and the Planck function in unit of frequency is obtained by replacing k = c/v in Bk

on the right hand side of the equality.
Exercise end.
This concludes our introduction to physics. Next we discuss nuclear energy,
which is of fundamental importance for the structure and evolution of the Sun and
stars.
1.4 The Nuclear Energy
1.4.1 The Missing Mass
A huge energy resource is hidden in the nuclei of atoms that has been feeding the
Sun and stars for billions of years. However, it has only been in the last century that
man has started to harness this energy and understand its use. To know more about
nuclear energy, let us pay further attention to atomic nuclei and, in particular, to
their stability.
Nuclei are composed of two types of particles, protons and neutrons, which have
roughly equal masses of about 10−24 g. Due to the small size the atomic mass unit,
mu, has been introduced. The mass mu is one-twelfth part of the carbon 12 nucleus,
which is the most abundant stable isotope of this element and contains six protons
and six neutrons. As said earlier, with regard to the state equation of a perfect gas,
the atomic mass unit is
mu ¼ 1:66027 1024 g:
A proton and a neutron have masses equal to 1.007823 mu and 1.008665 mu,
respectively, that is, they are slightly heavier than the atomic mass unit. Deuterium
is the hydrogen isotope with a nucleus, called deuteron, made by a proton–neutron
pair. The deuteron has a mass equal to 2.01410 mu, which is 0.002388 mu smaller
than the sum of the masses of the proton and neutron taken separately. This is a
surprising characteristic of all stable nuclei: the mass of a stable nucleus is always
smaller than the sum of the masses of its nucleons (i.e. protons and neutrons).
Where does the missing mass go? To answer this question let us consider the
formation of a deuteron from one neutron and one proton which are initially sep-
arated. This process occurs with a collision of the two particles and the final result is
not only the formation of one deuteron but also the emission of a photon with the
frequency of the c rays. Thus, in the collision process, a certain amount of mass
disappears while a certain amount of energy appears. Einstein showed that this
coincidence is not exceptional, because, as a general law, there is a precise relation

between mass and energy.
1.4.2 The Mass–Energy Relation
According to the Einstein’s assumption, to every body with mass m is associated an

energy E equal to the product of its mass and the squared light velocity, that is
E ¼ mc2 :
Returning back to the formation of a deuteron and using the Einstein formula for
calculating the energy corresponding to the missing mass of 0.002388 mu, we find
the value
E ¼ mc2 ¼ 2:39 103 1:66 1024 g ð2:99 1010 cm s1 Þ2 ¼ 3:55 106 erg:
It is worthwhile to express this amount of energy in eV, a symbol that means

electron Volt and denotes an energy unit commonly used in nuclear physics. Note
that here e is the modulus of the electron charge and V is the unit of voltage, or
electric potential difference, so named after the Italian scientist Alessandro Volta
(1745−1827). As Exercise 1.19 shows
1 eV ¼ 1:6022 1012 erg;
then
E ¼ 3:55 106 erg=ð1:60 1012 erg=eVÞ ¼ 2:22 106 eV ¼ 2:22 MeV;
where MeV is short for Mega−electron−Volt.

Worked Exercise 1.19 The eV energy unit.
The electron charge is −e = −1.6022 10−19 C, where C is the symbol for
Coulomb, the electric charge unit called after the French scientist Charles Augustin
de Coulomb (1736–1806). We recall that the electric potential difference is the
work done by the electric force per unit electric charge. In particular, 1 V = 1 J/
1C = 107 erg/C. Thus, to get over a negative potential difference of −1 V, an
electron requires an energy of −e(−1 V) = 1.6022 10−19 C 107 erg/C =
1.6022 10−12 erg. Therefore,
1 eV ¼ 1:6022 1012 erg:
Exercise end
The energy value corresponding to the missing mass of a deuteron is equal to the
energy of the emitted c ray, which we can compute from the light frequency with
1.4 The Nuclear Energy 47
the formula E = hm. We note that, for the mass–energy balance, the formation of a
deuteron from a neutron–proton pair is a process analogous to the formation of a
hydrogen atom from an electron–proton pair. In fact, the latter process implies the
emission of a photon with an energy corresponding to the difference between the
sum of the proton and electron masses and the hydrogen atom mass. However,
since the atomic bond is weaker than the nuclear one, the missing mass in the
hydrogen atom is very small and, in fact, the emitted photon, at a frequency of
about 90 nm in the ultraviolet, has an energy more than 105 times smaller than the
energy of the c ray that deuteron emits.
Exercise 1.20 Note that 1 MeV represents a very small amount of energy with
respect to those we encounter in our daily experience, and show that
1 MeV = 4.45056 10−20 kWh (see also Exercise 1.3).
Exercise 1.21 Knowing that the ionization potential of hydrogen is
EH = 13.59 eV, calculate the missing mass in the hydrogen atom.
1.4.3 The Nuclear Fission
Because its mass is always smaller than the total mass of its nucleons, a stable
nucleus has a missing mass that corresponds, through the Einstein’s formula
E = mc2, to the energy released when the nucleus formed. Vice versa, if we want to
separate a stable nucleus into its components, we need to supply the energy cor-
responding to the amount of missing mass that has to reappear.
The energy required for breaking the bond among neutrons and protons inside
the nucleus is called the binding energy. The nuclear binding energy represents an
important resource that the Sun and stars, and now also man, exploit. There are two
types of processes to release this energy: fission of a nucleus in two or more lighter
nuclei, and fusion of a number of nuclei in a heavier nucleus. In order to understand
when these processes can occur, it is necessary to consider the variation of the
binding energy per nucleon with the mass number (Fig. 1.18). Here the mass
number of a nucleus, denoted by the letter A, is the full number of its nucleons,
hence A = Z + N, where Z is the number of protons and N is the number of
neutrons present in the nucleus.
The figure shows that the binding energy per nucleon increases rapidly starting
from deuterium, which has mass number A = 2, till it reaches a maximum of about
9 MeV at the position of iron, which has symbol Fe, proton number Z = 26 and
mass number A = 56 for the most common of its stable isotopes. For the elements
heavier than iron, the binding energy per nucleon shows a gradual and continuous
decrease with mass number.
When a reaction produces a higher binding energy per nucleon with respect to
the initial conditions, the excess binding energy can be released. Therefore, in order
to produce energy, elements lighter than iron can enter into a fusion reaction whilst
those heavier than iron can be involved in a fission reaction. In fact, fusion creates
Fig. 1.18 The mean binding energy per nucleon for the stable nuclei, in MeV, as a function of the
mass number A. The curve is inferred from the semi–empirical formula for nuclear masses
proposed by the German physicist Carl Friedrich von Weizsäcker (1912–2007) E ¼ a1 A þ
a2 A2=3 þ a3 Z 2 =A1=3 þ a4 ðZ A=2Þ2 A þ a5 =A1=2 with a1 ¼ 15:21MeV; a2 ¼ 17:18 MeV; a3 ¼
0:686 MeV; a4 ¼ 91:478 MeV; a5 = ± 11. MeV; this term occurs only for the nuclei having odd
Z and N, for which the sign + holds, and for the nuclei with even Z and N, for which the
sign – holds. Note that some points clearly depart from the average behavior of stable nuclei, like
the He4 nucleus (a particle with N = 4) whose nucleons have an unusually strong bond
higher mass number nuclei while fission does the opposite, thus, in both cases, the
products of the reactions can get close to the maximum of the binding energy per
nucleon.
The year 1945 marked the end of the World War II and, at the same time, the
beginning of the Atomic Age with the detonation of the first two atomic bombs at
Hiroshima and Nagasaki. These two bombs, improperly referred to as atomic,
exploited the nuclear fission of uranium and plutonium, generated by neutrons.
Uranium fission occurs according to the following preferential channel
n þ 235 U !139Ba þ 95 Kr þ 2n:
This formula describes the process in which a neutron, denoted by the letter n,
breaks a uranium nucleus U, with mass number A = 235, into two lighter nuclei,
barium, Ba, with mass number A = 139, and krypton, Kr, with mass number A = 95,
producing two free neutrons. The binding energy released in this reaction equals
about 200 MeV, that is 200 MeV 4.45 10−20 kWh/MeV = 9 10−18 kWh.
Clearly, this amount of energy is very small. It is possible, however, to make the
neutrons produced by the fission trigger other fission reactions generating what is
called a chain reaction. Here, all the energy that is available, for example, from one
kg of uranium 235, can be released in a very short time. Because in this amount of
matter there are 2.5 1024 atoms of uranium, this results in the sudden production of
about 2.5 1024 9 10−18 kWh = 2 107 kWh. This is roughly equal to the
energy released by the first atomic bomb and corresponds to roughly the amount of
electric energy which all of Italy consumed, on average, in half an hour during 2007.
1.4.4 Nuclear Power Stations
Controlling the physical conditions inside a reactor, which is where fission reactions
occur, we see that on average only one of the produced neutrons triggers a new
reaction; therefore, the released energy, which remains constant in time, does not
make an explosion and can be stored for commercial usage. Nuclear power stations,
which produce energy for civil usage, are based on controlled chain fission. Since
the last century the technology for using the energy produced by nuclear fission has
vastly improved, however, the issue of security and radioactive waste disposal still
remains a big problem.
1.4.5 Radioactivity
Radioactive materials are capable of emitting different types of particles, namely:

(1) helium nuclei, 4He, composed of two protons and two neutrons, also called a
particles (a radiation; a is a Greek letter to be read alpha), (2) electrons
(b radioactivity; b is a Greek letter to be read beta) and, finally, (3) light as c rays.
All radioactive materials, which exist in nature or are man–made, have unstable
atomic nuclei. A natural radioactive element is uranium, whose isotope 235, formed
by 92 protons and 143 neutrons, takes a very long time to decay (of order of a
billion years). As it breaks down it continues to produce other unstable elements
until it transforms into a lead stable isotope with Z = 82 and A = 207. Radioactive
materials emit radiation whose energy can be transferred to the bodies it impacts: in
particular, when radiation hits a living organism, it can ionize its molecules and,
hence, cause serious health problems. To measure the biological effects of a dose of
absorbed radiation, taking into account that the danger differs among the three types
of radioactivity, the unit of measure is the Sievert, or Sv (after the Swedish medical
physicist Rolf Maximilian Sievert, 1896−1966). For example, a chest radiography
(medical X rays) absorbs an amount of about 1.4 10−4 Sv, less than a tenth of the
amount of 2.4 10−3 Sv absorbed each year in average due to natural radioactivity.
Italian law limits the amount of radiation a person can be exposed to per year; this
limit equals the mean level of the natural radioactivity increased by 10−3 Sv for
common people and by 2 10−2 Sv for those working with radioactive isotopes.
The effects of 1 Sv on health can result in weak, temporary alterations of

hemoglobin. Amounts within 2 and 5 Sv can cause nausea, hair loss, and hemor-
rhages. 4 Sv can be lethal in the 50% of cases without suitable therapy. With more
than 6 Sv survival is improbable.
Two recent and serious accidents involving nuclear power stations occurred at
Chernobyl, Ukraine, in 1986, and at Fukushima, Japan, in 2011. The radiation level
at Fukushima, after the meltdown of three of the plant’s six reactors, reached 0.4 Sv
within an hour, a rate of 0.4 Sv/h. Three days later, the level had decreased to
2 10−2 Sv/h. At Chernobyl, immediately after the explosion, one measured
300 Sv/h close to the core of the reactor and 10–15 Sv/h in the area of the involved
unit, but somewhat farther from the core; this value was at least 25 times higher
than the value of 0.4 Sv/h recorded at Fukushima in an area about at the same
distance from the reactor.
1.4.6 The Nuclear Fusion in the Sun
Nuclear fusion is the mechanism that generates the energy in the Sun and stars. The
first fusion process we can imagine is the one allowing the formation of the helium
nucleus (or a particle), the second stable element in order of weight. There are
various channels to obtain the fusion of hydrogen into helium, however the most
common reactions realizing this process in the Sun constitute the first branch of the
so called proton–proton chain (known with the abbreviation ppI)
p þ p !2H þ e þ þ me þ 1:442 MeV

2
H þ p !3He þ c þ 5:494 MeV
3
He þ 3 He !4He þ 2p þ 12:86 MeV:
In the first reaction two protons, p, merge into a nucleus of deuterium, 2H, with
the emission of a positron, e+, the electron antiparticle, and an electronic neutrino,
me. The energy indicated on the right is the sum of the averaged kinetic energies of
the particles produced in the reaction and corresponds to the mass difference
between the involved particles before and after the collision; in particular, the
neutrino has, on average, an energy equal to 0.265 MeV. The positron quickly
annihilates with an electron, its antiparticle, releasing two c rays, that is
e þ þ e ! 2c:
In the second reaction the deuteron interacts with a third proton to form tritium,
that is the helium isotope with only one neutron, or 3He, releasing both kinetic
energy and c radiation. Tritium, in turn, combines with a second nucleus of 3He to
form an a particle, 4He, and releases two protons and further kinetic energy.
A complete cycle requires that the first two reactions occur twice in order that the
two nuclei of 3He necessary for the third reaction can form. In this way, in one cycle
4 protons transform themselves into one helium nucleus and a total amount of
26.732 MeV is released as kinetic energy of the gas particles and c rays. This
energy equals the squared light velocity multiplied by a mass equal to about 2.8%
of the proton mass that is missing in the a particle with respect to the sum of the
masses of the four protons (and two electrons) producing it.
Exercise 1.22 Verify that according to the mass—energy relation the total energy
release of the hydrogen fusion into helium is about 2.8% of the proton mass.
A characteristic of the pp chain is the difficulty of triggering it, which, as we will
see in detail in the chapter about the solar core, has two problematic issues. First,
two protons must be very close before they merge but, both having positive
charges, naturally repulse each other. Second, the weak interaction process, by
which a proton decays into a neutron, is a particularly slow reaction. The difficulty
of triggering is a major disadvantage of the fusion process compared to the fission
process, as far as the controlled utilization of nuclear energy is concerned. An
advantage of the fusion, however, is that it can produce a large amount of energy. In
fact, the energy released per nucleon by fusion is about 27 MeV divided by the 4
involved protons, that is 6.6 MeV/n. This value is, in particular, nearly 10 times
higher than the energy produced in the fission of uranium, which equals 200 MeV
divided by the involved 235 nucleons, that is, about 0.8 MeV/n. Moreover, con-
tinuing to list the pros and cons, we expect that nuclear fusion stations would be
safer and have smaller waste problems than the fission stations.
The possibility of generating high temperatures and, hence, large kinetic ener-
gies of the particles in a hydrogen gas by means of the preliminary explosion of a
fission bomb, has resulted in the development of “hydrogen bombs”, or H–bombs.
These bombs are capable of releasing fusion energy in an explosive manner.
Because controlling the extreme conditions that fusion reactions require for a
sufficiently long time is a difficult task, man has not yet managed to realize nuclear
fusion stations. However, based on recent progress in this research area, it is rea-
sonable to foresee that this goal can be reached in the not too distant future.
In addition to the hydrogen fusion other processes of nuclear fusion are possible,
which, starting from the light elements, yield nuclei of heavier elements up to iron
(Z = 26), which is very close to the maximum value of the binding energy per
nucleon. Triggering these processes requires overcoming the electric repulsion due
to an increasing number of protons and, hence, higher and higher temperatures of
the gas are needed.
Nowadays, only the Sun and stars are able to exploit the energy of nuclear fusion
in a controlled way and, thanks to that, they can live for millions, maybe billions, of
years, as we will see in the following chapters.
Finally, note that the values of the physical quantities introduced in the text are
listed together in Table 1.7.
Table 1.7 Physical quantities

Name Symbol Value
Boltzmann constant kB 1.38062 10−16 erg K−1
Planck constant h 6.62620 10−27 erg s
Speed of light c 2.997930 1010 cm s−1
Stefan-Boltzmann constant r 5.66956 10−5 erg cm−2 K−4 s−1
Gravitation constant G 6.67384 10−8 dyn cm2 g−2
Gas constant R 8.314472 107 erg K−1g−1
Electron charge e 1.6022 10−19 C
Atomic mass unit mu 1.66027 10−24 g
Proton mass mp 1.672661 10−24 g
Neutron mass mn 1.674556 10−24 g
Electron mass me 9.10956 10−28 g
Hydrogen mass mH 1.67352 10−24 g
Bohr radius of H atom a0 5.29172 10−9 cm
Hydrogen ionization potential EH 13.59 eV
Electron Volt eV 1.6022 10−12 erg
pp cycle energy output Eppc 26.732 MeV
For each quantity the name, symbol and value are given. The units of measurement for the length,
mass and time are centimeter (cm), gram (g) and second (s), respectively; these units are the base
of the so called CGS system. Dyn is the symbol for the force unit dyne (1 dyn = 1 g cm s−2), so
called from the Greek word dύmali1 (dynamis) meaning power, force
1.4.7 Further Exercises
Exercise 1.23 The Bohr formula E = −EH/n2 represents the total energy of an
electron rotating about a hydrogen nucleus. Show that the potential energy of the
electron, –e2/r, is in modulus twice its kinetic energy mv2/2. Then, knowing that
e = 4.80 10−10 esu and EH = 2.17 10−11 erg, estimate the order of magnitude
of the atomic size by computing the radius of the smallest hydrogen orbit, which
corresponds to n = 1 (the ground state). Note that the esu is the electrostatic unit of
charge that must be used instead of the Coulomb when the electron potential energy
is written in the form –e2/r (The esu unit of electric charge is defined in the Worked
Exercise 6.4).
Exercise 1.24 Knowing that in 1 g of water there are 3.3 1022 molecules,
estimate the size, lm, of a water molecule.
Exercise 1.25 Estimate the Avogadro’s number, NA, that is the number of mole-
cules contained in a mole of material. By definition a mole is an amount of material
which equals in grams the material molecular weight. Realize that the Avogadro’s
number is independent of material.
Exercise 1.26 Estimate the air mean molecular weight la. Air is a mixture whose
main components are nitrogen for 78%, and oxygen for 21%.
Exercise 1.27 Estimate the most probable energy and velocity of an air molecule
when the air temperature is T = 15 ◦C = 15 + 273 K = 288 K.
Exercise 1.28 If the air has density q = 1.3 mg cm−3 and temperature T = 15◦ C,
estimate the air pressure. The resulting pressure is close to 1.013250 106 dyn cm−3,
the value that defines 1 atm (atm is the symbol for the pressure unit called atmosphere).
R p=2 R 2p
Exercise 1.29 Show that h¼p=2 u¼0 ðv2 senhÞ dh du ¼ 4pv2 dv.
Exercise 1.30 Show that for the general case of several ionization states the
fraction of all ions that is in a particular state can be obtained by repeated appli-
cations of the Saha equation.
Exercise 1.31 From the formula E = hm, evaluate the photon energies corre-
sponding to the frequencies listed in Table 1.6.
Exercise 1.32 Verify that the photon emitted in the formation of deuteron is a c ray
while the photon emitted in the hydrogen recombination occurs in the UV.
Chapter 2
Introduction to the Sun
Abstract The Sun has a basic impact on human life; thus, during our existence
man has viewed our star in many different ways. From the scientific point of view,
the only one of concern here, the Sun is an analog of the Rosetta stone for stellar
studies, thanks to its closeness to the Earth. This book is limited to the introduction
of the structure and evolution of our star and neglects the solar magnetic field and
related phenomena. As a consequence, we treat the Sun as a spherically symmetric
star, with an atmosphere made only of the photosphere, which is constant for most
of its lifetime. Even with this restriction we can address many fundamental ques-
tions which concern the global physical parameters of the Sun, the present model of
its internal structure, its energy source and the way the energy flows outwards and,
finally, the evolution of our star from its formation to its shutdown. This intro-
ductory chapter contains two sections, the first summarizes our approach and the
second describes the basic solar data.
2.1 Our Approach
2.1.1 Motivation
No one can avoid forming a mental image of the Sun without it being strongly
connected to our life. Naturally, this idea depends on a person’s cultural level and
power of imagination. For example, under the guise of irrationality and fear, the
Sun may become a mysterious source of benign and malignant influences whose
position in the sky affects people and their lives. Actually, this belief is called
astrology and is still thriving today. Even in the past, the Sun was identified with a
god who had to be propitiated with a suitable cult, as did primitive religions. From
a scientific point of view, the Sun is interesting to many research fields in addition

56 2 Introduction to the Sun
to solar physics. The Sun greatly affects the Earth and our lives here and in space.
The influence of the Sun is first realized by its light radiation which it constantly
emits and which supports life on the Earth, and secondly, to a lesser extent, by the
intriguing phenomena of solar magnetic activity. The Sun represents an extraor-
dinary laboratory of physics, where we can observe phenomena needing extreme
conditions of temperature and density, and sizes that cannot be reached in terrestrial
labs. Although there are billions of stars in the Universe that could, in principle,
also serve as such laboratories, the Sun is the only star that is close enough for us to
observe in detail. Hence, the Sun is a basic reference point for the study of the
physical processes occurring in the other stars.
Fig. 2.1 The images of the Sun at the website of the SOHO space mission (http://sohowww.
nascom.nasa.gov/data/realtime-images.html) for the day September 9th, 2013. From top left to
right, the first four images represent the solar disk observed at different wavelengths in the extreme
ultraviolet by the instrument EIT. Because the level of the solar atmosphere that we observe
depends on the opacity of the solar gas and the opacity changes with wavelength, as we will see in
detail in Chap. 3, these images correspond to four different levels in the solar transition region and
corona, which are the outer parts of the solar atmosphere. Each atmospheric level has a different
temperature, passing in the order from about 70 thousand K to 1, 1.5 and 2 million K degrees. The
colors (blue, green, yellow and red), selected to represent the different temperatures, are arbitrary
since they correspond physically to visible wavelengths and not to the ultraviolet wavelengths of
observation. In the last two images of the second row, which have been obtained by the coron-
agraph in white light LASCO and are again represented with arbitrary colors, the corona is seen
extending far beyond the solar disk. The visible solar disk is indicated by a white circumference in
the images and was suitably occulted in the instrument to reach the same effect of a total eclipse.
The red image extends outwards to 8.4 million km from the Sun’s center, the blue to 45 million
km. The photosphere is represented in the first image of the second row, which is the only one
whose color has a real meaning. The magnetic field at the photospheric level is reported in the
following image, where black and white denote opposite magnetic polarities
2.1 Our Approach 57
2.1.2 The Steady Sun
Nowadays the reader can easily access images of the solar atmosphere with high
spatial resolution and at different wavelengths, which are provided by solar
observations with ground based and space borne instruments. The images in
Fig. 2.1 represent the solar atmosphere at different levels.
The photosphere is the base of the atmosphere and corresponds to the visible
solar disk (first two images in the second row). The outer layers are the chromo-
sphere, transition region and corona in order from inwards to outwards (first row
and last two images in the second row). Apparently the photosphere is uniform and
spherically symmetric, but it may locally present regions with a strong magnetic
field, sunspots, where the field can reach values of the order of 3000 Gauss. The
number and size of sunspots varies during a magnetic activity cycle, which is about
11 years long. The outer atmosphere reaches very high temperatures, extends
outwards over the solar disk and is not at all spherically symmetric. The charac-
teristics of the outer atmosphere are connected with the magnetic field, which,
starting from its roots in the photosphere spreads outwards and plays a dominant
role in determining the structure of these layers. Moreover, the Sun has a bipolar
global magnetic field of few Gauss, similar to the Earth’s mean magnetic field. We
expect that the magnetic field cannot easily penetrate too deep inside the Sun,
because the interior is highly ionized and electrically conductive, so that any change
of the field induces a current that will oppose to it, according to the Lenz’s law
(Heinrich Lenz was a Russian physicist, 1804−1865).
The description and explanation of solar global structure and evolution, which is
our goal, does not require us to go into the detail (even though fascinating), of the
solar magnetic field and related phenomena. Therefore, we restrict ourselves to
study the so-called quiet, or better, steady Sun. This is an ideal star having a perfect
spherical symmetry, a visible surface, the photosphere, without spots, and is free
from chromosphere, transition region and corona. The steady Sun is in uniform and
stable equilibrium and constantly emits the same amount of energy per unit of time
that is created inside its core by nuclear fusion reactions.
2.1.3 Methods
We can begin to understand how the steady Sun works once we start thinking about
a number of basic questions. In a similar approach to how a child tries to understand
a toy he or she is attracted to, by taking it apart and analyzing a piece at a time, we
will use basic questions to guide us in dismantling the solar machine into suitable
parts that we can more easily understand.
Wherever we live on the Earth, we enjoy solar light 6 months a year, and we all
have a healthy natural instinct to not directly observe by eye the solar disk, which is
so bright that it can seriously damage our retina. But, how much light does the Sun
actually emit? And, more generally, what are the characteristic data for the Sun, like
size, mass and temperature, and how are they derived? The answers to these
questions are given later in this chapter. Details of the chemical elements forming
the Sun are given in Chap. 3.
The Sun is continuously releasing energy into the region of space called the
Heliosphere, and as such is the primary energy source for our Earth and the other
bodies that are gravitationally linked to it. However, the primary use of solar energy
is to keep the structure of the whole Sun in its present extraordinary equilibrium.
So, naturally the second basic question refers to the global structure that charac-
terizes the Sun at present. The answer to this question requires determining the
ways in which energy is produced in the solar core and flows outwards. Energy
generation and transfer in turn are directly related to the temperature and density
stratification of the solar gas. Modeling the structure of the present Sun is the main
goal of this book and makes up the next four chapters, each one devoted to a deeper
layer, starting at the photosphere and going down to the convection zone, radiation
zone and core. We will construct an analytical, layer by layer model of the Sun,
which emphasizes the basic physical concepts and approximates the structure of the
so called solar standard model, that is a current and accurate solar numerical model.
However large it may be, we expect that the source of solar energy will eventually
become exhausted. Then, our third and last question, which is discussed in the seventh
chapter, concerns the life of the Sun, or, as astronomers say, its evolution, starting from
its formation, passing through the present stage, till the end of its activity.
Before scouting the structure and evolution of the Sun in the next chapters, in the
following section we describe some basic solar properties.
2.2 The Characteristic Data
By characteristic data of the Sun we mean, first of all, its distance from the Earth,
and then after this its global parameters like: radius, mass, luminosity and surface
temperature. Table 2.1 summarizes the values of the solar data. Although the values
are given to the numerical accuracy with which we know them at present, only the
first few significant figures will often be sufficient for us.
Table 2.1 Characteristic data of the Sun

Sun–Earth distance AU 1.495978707 1013 cm
Radius RS 6.960 1010 cm
Mass MS 1.9889 1033 g
Mean density qS 1.408 g cm−3
Surface gravity gS 2.7398 104 cm s−2
Luminosity LS 3.844 1033 erg s−1
Effective temperature Te 5778 K
For each quantity the name, symbol and value in CGS units are given
2.2 The Characteristic Data 59
The Sun’s distance from the Earth, radius and mass are determined by suitable
measurements and using Kepler’s laws that describe the motions of the planets
around the Sun (Johannes Kepler was a German scientist, 1571, 1630).
2.2.1 The Distance from Earth
The Earth’s orbit around the Sun is an ellipse with the Sun’s position coinciding
with one of the foci. Since the ellipse deviates only slightly from a circle (see
Exercise 2.1), we are satisfied with the average distance of the Sun from the Earth.
The Sun–Earth mean distance is about 150 million km. The astronomical unit,
symbol AU, is a unit of length derived from this distance which is of convenient use
in astronomy, in particular to express the distances in the solar system. Originally
based on the measurement of the maximum and minimum distances from the Earth
to the Sun, corresponding to the points in the orbit called aphelion and perihelion
respectively, in 2012 the astronomical unit has been conventionally fixed to the
exact value
1 AU ¼ 1:495978707 1013 cm ¼ 149:5978707 Gm:
The light covers one AU in a time t such that
c ¼ AU=t:
Then, since the light has a velocity of
c ¼ 2:997950 1010 cm=s ¼ 0:2997950 Gm= s;
we get
t ¼ AU=c ¼ 150 Gm=0:300 Gm=s ¼ 500 s;
that is 8 min and 20 s.

Exercise 2.1 Denoting with a and p the distances that the Earth’s center has from
the Sun’s center at the aphelion and perihelion, respectively and assuming that
(a + p)/2 = 1 AU and (a − p)/(a + p) = 0.016, find the values of a and p in AU.
Worked Exercise 2.2 Measuring the Sun–Earth distance.
The method is a combination of an astronomical triangulation and a radar
measurement of distance. The distance d from the planet Venus to the Earth can be
determined from the time tr that a radio wave impulse needs to reach Venus, be
reflected by its surface and come back to Earth, as d = ctr, where c is the light
velocity. This implies that in the triangle EVS, whose vertices are the positions in
the sky of the Earth (E), Venus (V) and Sun (S), the length of the side EV = d is
Fig. 2.2 The astronomical

triangle used to determine the
Sun–Earth distance. At time t,
the Earth (E), Venus (V) and
Sun (S) are aligned. Two
revolution periods of Venus
later, at time t + 2PV, the
Earth lies at the lower vertex
of the triangle VSE. From
observations and the radar
measurement of the Venus–
Earth distance (d), the
astronomical unit (AU) can be
determined
known (Fig. 2.2). The angle VES can be observed and the angle VSE is derived from
the revolution period of Venus PV = 0.615 years. In fact, the two positions of the
Earth in the figure have been chosen to correspond respectively to the time t of a
transit of Venus over the Sun, when the Earth, Venus and Sun are aligned, and the
time t + 2PV, two revolution periods later, when Venus is in the same position and
the Earth is ahead on its orbit of the angle VSE. Because the time
2PV = 20.615 = 1.230 y is elapsed, it is VSE = 360°(1.230−1) = 82.8°. According
to the sine rule, AU/sin(EVS) = d/sin(VSE) with EVS = 180◦ − (VES + VSE) and
VES = 38.3° from observation. Thus, the astronomical unit is given by AU = dsin
(EVS)/sin(VSE), where d = 174 Gm from the radio measurement of distance.
Exercise end.
Exercise 2.3 Discuss the approximations which are implied in the solution of
Exercise 2.2.
Exercise 2.4 Why the radar measurement of distance is not used directly with the
Sun?
2.2.2 The Radius
The solar radius, RS, is determined from the astronomical unit and the angular
diameter aS of the Sun. Because the Sun’s size is small with respect to its distance
from the Earth, the angle under which we see it, expressed in radians, is approxi-
mated by the ratio of the Sun’s diameter to distance (Fig. 2.3). Then
Fig. 2.3 The solar radius is

more than 200 times smaller
than the astronomical unit (in
the figure it is only 1/6). This
justifies that the solar
diameter 2RS is approximated
with the arc AB (red) that the
Sun occupies on a circle of
radius 1AU centered at the
Earth. Therefore, the angular
diameter is aS = 2RS/AU
aS ¼ 2RS =AU
and
RS ¼ AU aS =2:
The Sun’s disk in the sky as seen from the Earth has a diameter spanning in
average an angle of a bit more than half degree, 0.53° or 1920″ in seconds of arc.
The same angle is covered, for instance, by a football having a diameter of about
22 cm, placed at a distance of 23.78 m, or by a coin of 1 €, diameter 23 mm, at a
distance of 2.486 m. To determine the solar radius, RS, we have to transform the
angular diameter of the Sun from degrees into radians. Since the degree to radian
conversion’s ratio is the same for all angles and a flat angle measures
p = 3.14159… in radians (p is a Greek letter to be read pi) and 180° in degrees, we
can write
aSðradianÞ ¼ aSðdegreeÞ p=180 :
Therefore, the solar radius is
RS ¼ AU aS =2 ¼ AU aSðdegreeÞ p=ð180 2Þ
¼ 150 Gm 0:53 3:14=360 ¼ 0:696 Gm
In particular, with this value of the solar radius, one second of arc on the Sun
corresponds to
2RS =1920 ¼ 2 0:696 109 m=1:92 103 ¼ 0:725 106 m ¼ 725 km:
As a matter of comparison, the mean radius of the Earth is 6371 km, about 110
times smaller than the solar radius, and one second of arc in km on the Sun
corresponds to about the motorway distance between Naples and Venice, in Italy.
Worked Exercise 2.5 Measuring the solar angular diameter.
The solar disk has not a perfect sharp edge, like a solid sphere has. The apparent
extension of the solar disk is related to the amount of emitted radiation and depends
upon the radiation wavelength, as the images shown in Fig. 2.1 confirm. When light
is restricted to the total emission in the visible spectrum, the edge is rather sharp
because the emission at the solar limb decreases to about 15% in only 500 km.
Therefore, the extremes of a solar diameter are defined as the points where the
visible radiation is reduced to a factor 1/e of its value at the disk center. There are
further issues concerning the way radiation is observed to vanish towards the limb.
The instrumental apparatus and, in the case of ground based observations, the
Earth’s atmosphere, can degrade solar images. A significant part of these effects can
be modeled by means of the so called point spread function, which accounts for the
broadened and smoothed image of a single source point (Fig. 2.4).
Fig. 2.4 The point spread function. An imaging system, which may include the Earth’s
atmosphere and telescope, degrades the observed image of a light source. The figure sketches the
one dimensional case and the degradation effects are restricted to depression and broadening of the
observed intensity. A single source point has an intensity profile (left panel) much thinner than the
observed intensity profile (right panel), the ratio of their widths at half maximum being 1 to 10.
Thus, the observed intensity profile is a direct approximation of the point spread function of the
observing system
Fig. 2.5 The Lorentzian function LðrÞ ¼ c=½pðr2 þ c2 Þ for c = 1. This function is normalized,
þ1
i.e. R 1 LðrÞdr ¼ 1: The horizontal segment labeled c is the half width of the central peak at the
level where the Lorentzian function is half of its maximum value. For rc and r c the
Lorentzian function scales as 1/r2
If r is the coordinate along a solar diameter with r = 0 at disk center, I(r) is the
radiation intensity, and P(r) the point spread function, the observed intensity Iobs(r)
is the integral
Zþ 1
Iobs ðr Þ ¼ I ðr 0 ÞPðr 0 r Þdr 0
1
The integral is the convolution of the intensity and point spread function.
Because of this operation the observed intensity at the point r is the summation of
the intensities from all the points r′ in the neighborhood of r where the point spread
function does not vanish. This implies that, even if the solar disk would have a step
like edge, instrumental and possible atmospheric point spread functions make the
observed intensity not vanish outside the limb (Exercise 2.7).
Exercise end.
Exercise 2.6 The Lorentzian function.
The normalized Lorentzian function is defined as

Lðr Þ ¼ c= p r 2 þ c2
where c is a characteristic parameter (Fig. 2.5. Hendrik Lorentz was a Dutch

physicist, 1893–1928). This function can be used as a point spread function for the
seeing, which is the word to design the degradation that the Earth’s atmosphere
Fig. 2.6 The observed radiation intensity in the visible wavelength range along a solar diameter is
modeled by the convolution of a rectangular intensity profile and a normalized Lorentzian point
spread function. The result is plotted for two values of the Lorentzian half width c corresponding
to the values of n = 1920 (black curve) and n = 192 (red curve). Here n = 2RS /c, where RS is the
solar radius. Thus, n = 1920, that is the value of the solar angular diameter in arseconds,
corresponds to c = 725 km, or 1′′, and n = 192 corresponds to c = 7250 km, or 10′′
induces on ground based observations. Find the maximum value of the Lorentzian
function and show that c is the half width at half maximum. Verify that this function
is normalized, i.e. its integral between r = −∞ and r = +∞ is 1 [Hint Make use of
d arctan r/dr = 1/[(r2 + 1)]].
Exercise 2.7 Determine the intensity profile resulting by a convolution of a rect-
angular intensity profile (i.e. I(r) = 0 for r < −RS and r > RS, and I(r) = I0 for
−RS < r < RS) and a Lorentzian point spread function with half width c (Fig. 2.6).
Exercise 2.8 The normalized Gaussian function has the form GðrÞ ¼
pffiffiffi
½1=ðr pÞer =r : Show that a Lorentzian spread function with half width c pro-
2 2
duces a larger solar angular diameter than a Gaussian point spread function with the
parameter r equal to the Lorentzian half width c.
2.2.3 The Mass
The solar mass MS can be derived from the Newton’s law of gravitation. The
gravity force provides the centripetal acceleration of the Earth orbiting around the
Sun. Then, if m and v denote respectively the mass and speed of the Earth in a
circular orbit of radius a, we can write
GmMs mv2
¼
a2 a
and then
GMs
¼ v2 :
a
The orbital speed is the length of the circular orbit divided by the orbital period P
2pa
v¼ :
P
Therefore

GMS 2pa 2
¼ v2 ¼
a P
and
GMS a3
¼ 2:
4p2 P
This equation approximates the exact expression of Kepler’s third law, where
a is the semi-major axis of the elliptical orbit and the sum MS þ m replaces MS ; this
approximation is good enough since the Earth orbit is nearly circular and MS m
as we shall see soon. Thus, the solar mass is (with a in cm and P in seconds to be
consistent with the units of G given in Table 1.5)
3
4p2 a3 4 3:142 ð1:5 1013 cmÞ
MS ¼ ¼
GP2 6:67 108 dyn cm2 g2 ð365 24 60 60 sÞ2
¼ 2:00 1033 g
An accurate value for the mass of the Sun is MS = 1.9889 1033 g. The Sun has
about 300 thousand times as much mass as the Earth whose mass is 5.976 1027 g.
Exercise 2.9 Show that the centripetal acceleration of a body moving along a
2
circular orbit of radius a with constant speed v is given by va .
The radius and mass allow the evaluation of the mean density, qS, and surface
gravity, gS. Since the Sun is a sphere of gas that is held together by the gravitational
attraction of its component particles, its density is stratified and increases passing
from the surface to progressively inner layers, as we will see in the next chapter. As
a first step we can compute the mean density, which is the ratio of the solar mass to
the volume, that is
.h 3 i
qS ¼ MS 4pR3S =3 ¼ 2:00 1033 g 4 3:14 6:96 1010 cm =3
¼ 1:41 g cm3 :
The solar mean density is nearly one and half the density of water and about a
quarter of the mean density of the Earth.
The surface gravity is the acceleration impressed to a mass m which is on the
Sun’s surface, at 1 RS from center. The surface gravity is inferred from the gravi-
tational attraction that the Sun exercises on the mass m, which by virtue of the
second law of dynamics is

F ¼ gS m ¼ G MS m R2S ;
where G = 6.67 10−8 dyn cm2 g−2 is the gravitational constant.

Therefore, it results
gS ¼ G MS =R2S
¼ 6:67 108 dyn cm2 g2 2:00 1033 g= ð6:96 1010 cmÞ2
¼ 2:74 104 cm s2 :
The gravity at the surface of the Sun is more than 27 times the gravity at the
surface of the Earth, where it reaches 9.78 102 cm s−2.
Exercise 2.10 A satellite is in orbit around the Sun. Find the distance r at which the
satellite would remain immobile relative to the solar surface, assuming that the Sun
rotates around itself in 27 days. Notice that the rotation of the solar surface is
actually differential, with the rotation period that is shorter at the equator (about
24 days) and longer at the poles (almost 38 days).
2.2.4 The Luminosity
The solar luminosity, which is the energy per second that the Sun emits into space
as electromagnetic radiation, has the value
LS ¼ 3:844 1033 erg s1 :
To appreciate how large this number is, look at the Worked Exercise 2.11 which
shows that the solar constant, CS, that is the flux of solar energy at a distance of one
astronomical unit, is given by
CS ¼ 1:4 kW m2 :
Worked Exercise 2.11 The solar constant.

Let us start to transform the unit of measure of luminosity from erg/s into kW.
Since
1 kW ¼ 103 W ¼ 103 J s1 ¼ 103 107 erg s1 ¼ 1010 erg s1 ;
it is
1 erg s1 ¼ 1010 kW
and
LS ¼ 3:844 1033 erg s1 ¼ 3:844 1023 kW:
The solar luminosity at the distance of 1 AU is distributed over the sphere

centered in the Sun and with radius equal to 1 AU (Fig. 2.7). Then, the solar
constant, CS, is given by the luminosity divided by the surface of this sphere
.h 2 i
CS ¼ LS = 4pAU 2 ¼ 3:84 1023 kW 4 3:14 1:50 1011 m
¼ 1:4 kW m2 :
Note that the solar radiation flux is LS/(4pR2S) at the solar surface, and is
LS/(4pAU2) at the radial distance of 1 AU, then it drops off with the square of the
radial distance r from the Sun, as 1/r2. This scaling rule occurs when radiation
undergoes neither absorption nor emission along its path.
Fig. 2.7 The energy radiated

by the Sun to the distance of
one astronomical unit is
distributed over the sphere
centered in the Sun and with
radius equal to 1 AU. The
amount of solar energy that
reaches the Earth is
approximated by the radiation
directed toward the Earth’s
section (represented by the
vertical segment). Note that in
reality the Earth’s radius is
much smaller than the
astronomical unit and Sun’s
radius with respect to what the
figure suggests
Finally, we consider the energy P that reaches the whole Earth in one second.
Because solar rays hit a portion of the terrestrial surface, which is approximated by
a section of the Earth, this power equals the solar constant multiplied by the area of
the Earth’s section S. The Earth’s radius is RT = 6371 km, hence
S ¼ pR2T ¼ 3:14 ð6371 kmÞ2 ¼ 1:3 108 km2 :
Thus, we get for P
P ¼ CS S ¼ 1:4 kW m2 1:3 108 km2

¼ 1:4 kW m2 1:3 1014 m2 ¼ 1:7 1014 kW:
The Sun in one hour delivers to the Earth 1.7 1014 kWh, that is about 10
million times as much energy as the first atomic bomb released in 1945.
Exercise end.
Exercise 2.12 Show that the flux that reaches the top of the Earth’s atmosphere
fluctuates around the value of the solar constant by about 6% during a year due to
the Earth’s varying distance from the Sun.
The solar constant includes all types of solar electromagnetic radiation, not just
the visible light. It is measured at the top of the Earth’s atmosphere and the mea-
surement is adjusted to the distance of one AU scaling the flux by the inverse square
of the distance (see Exercise 2.11).
The solar constant does not remain completely constant but typically varies by
few parts per thousand or less because of solar magnetic activity. Fluctuations occur
over a period of 11 years, due to the solar magnetic cycle, and over smaller
timescales, from month to day, due to the rotation and evolution of solar magnetic
regions.
2.2.5 The Temperature
In order to determine a temperature we approximate the Sun with a black body. The
assumption of thermodynamic equilibrium is justified because the solar luminosity
is a negligible amount of energy lost with respect to the total available energy and,
moreover, the Sun is the seat of a large number of light emission and absorption
processes. In this way the emission spectrum is close to the Planck function and we
can infer a temperature from the Stefan–Boltzmann law. The outward radiation flux
F from a unit surface element of a black body at temperature T is
F ¼ rT 4 ;
where r = 5.66956 10−5 erg cm−2 K−4 s−1 is the Stefan–Boltzmann constant. In
the case of the Sun, the emitted flux is given by the ratio of the luminosity to the
area of the solar surface and, hence, is computed in terms of the luminosity and
radius as

FS ¼ LS = 4pR2S :
Therefore, the Stefan–Boltzmann law allows us to define a particular tempera-

ture of the Sun, which is called the effective temperature, Te, and is given by

FS ¼ rTe4 ¼ LS 4pR2S :
Thus

Te4 ¼ FS =r ¼ LS 4pR2S r
and
4
Te ¼ LS 4pR2S r
n .h 2 io4
¼ 3:844 1033 erg s1 4 3:1415 6:96 1010 cm 5:66956 105 erg cm2 K4 s1
¼ 5778 K:
The effective temperature is the surface temperature of the Sun since it is a

characteristic of the visible solar disk, or photosphere, while, as we will see later,
the internal layers of the Sun have temperatures much higher than the value of the
effective temperature. The structure of the photosphere is discussed in the next
chapter.
A characteristic, which is related to the surface temperature of a star, is its color.
We know that the color of an incandescent solid body changes with the temperature
to which it is heated, passing, with increasing temperature, from red to orange,
yellow, white and, finally, blue. The Sun is a star of yellow color. Note that the
Fig. 2.8 The sensitivity of

the human eye to light. The
response to color depends on
the level of the external light.
The two curves give the
fraction of light which is
perceived at each wavelength
in low (left curve) and high
(right curve) light conditions.
The maximum sensitivity is
set to unity
body’s color that we perceive depends not only on the amount of its emission at
different wavelengths, but also on the sensitivity of the human eye, which acts as a
light filter. The curves representing our standard sensitivity to light of different
wavelengths have the shapes of bells covering the visible spectrum with their
maxima in the yellow green wavelength range (Fig. 2.8).
Exercise 2.13 Find the distance from the Earth at which a satellite in orbit is
immobile relative to the ground.
Exercise 2.14 Evaluate the gravitational potential energy of the Earth in orbit
around the Sun.
Exercise 2.15 Find the orbital velocity of the Earth around the Sun.
Exercise 2.16 Determine the escape velocity from the solar system at the position
of the Earth. This escape velocity is the minimum velocity necessary to overcome
the solar attraction and leave the solar system.
Exercise 2.17 Show that the escape velocity from the Sun is a factor 3.76 greater
than the escape velocity from the Earth.
Exercise 2.18 The semi-major axes of the orbits for Venus, Mars, Jupiter and
Saturn are 0.723, 1.524, 5.204, and 9.580 AU, respectively. Find the corresponding
orbital periods from the Kepler’s third law.
Exercise 2.19 Show that at a height h, which is small relative to the Earth’s radius,
the gravitational potential energy of a body with mass m can be written E = mgEh,
where gE is the Earth surface gravity acceleration and the Earth’s surface is taken as
the zero energy level.
Exercise 2.20 A meteorite with mass m = 1 kg hits the Earth’s atmosphere with a
velocity of 40 km/s at a height of 10 km and starts to burn. If all the energy of the
meteorite is transformed into red light at a wavelength k = 0.65 lm, how much
energy and how many photons are produced?
Exercise 2.21 An incandescent solid body, which is approximated with a black
body, emits most of his light in the red at k = 0.65 lm. What is its temperature?
How much energy is required to shift the body’s emission peak to the yellow at
k = 0.57 lm, if the mass is m = 10 kg and the specific heat is Cs = 4 106
erg g−1K−1?
Exercise 2.22 Knowing that the Moon’s rotation period around the Earth is
27.3 days, estimate the Moon-Earth distance.
Chapter 3
The Photosphere
Abstract The visible solar disk, or photosphere, is the external layer of the Sun
where most of its light emerges from. The first section of this chapter completes the
thermodynamic description of the photosphere by computing values for pressure
and density in addition to the effective temperature. While the effective temperature
was straightforwardly derived from the total emergent flux, the determination of
pressure and density is a more complex issue. We need to both treat the ionization
equilibrium of the negative hydrogen, because photoionizations of this ion are the
main opacity source in the photosphere, and assess the mechanical equilibrium of
the solar gas that determines the pressure and density variation with height.
A photon mean free path is the average depth from where radiation emerges from
the photosphere. The pressure scale height is the distance over which the gas
pressure and density drop outwards by a factor equal to e. The photon mean free
path and pressure scale height are comparable and both can be used to measure the
thickness of the photosphere. As a consequence, there is an implicit relation
between the thermodynamic parameters of the photosphere that, together with the
gas equation of state, allows their determination.
The photosphere is so thin that it can be modeled as a plane parallel atmosphere.
In the second section we look at a semi-empirical, approximate model of the
photosphere. The temperature variation as function of height is determined from the
solar disk limb darkening observed at visible wavelengths. To understand this we
derive the radiation transfer equation and discuss its solution, introducing the
Eddington-Barbier approximation and the LTE assumption. We then compare the
approximated structure of the photosphere with an accurate numerical model in
order to better understand pros and cons of our approach. Finally, we present a
detailed chemical composition and compare the resulting photospheric data with
those of the Earth’s atmosphere.

72 3 The Photosphere
3.1 The Physical Parameters
3.1.1 Physical Constrains
The effective temperature of the Sun’s visible surface of 5778 K has been derived
from the total emergent solar flux and, hence, is characteristic of the external layer
of the Sun where most of its light emerges from. This layer is named photosphere,
from two Greek words meaning sphere of light, and constitutes the visible solar
disk. In order to describe the physical conditions existing in the photosphere we
need to establish a value not only for temperature but also for pressure and density.
We know that the pressure P, density q and temperature T are related by the
equation of state for a perfect gas, which can be written as
P ¼ Rq T=l;
where R = 8.314472 107 erg K−1 g−1 is the gas constant and l is the mean
molecular weight. The Sun is a mixture of chemical elements among which
hydrogen represents more than 90% of the total number of atoms. Helium con-
tributes almost 10% of the total number of atoms, while the abundance of all the
other elements together, which astronomers call generically metals, is of the order
of 0.1%.
To compute l in the photosphere we can neglect metals, because they are
relatively few, and electrons, because only metals with low ionization potential are
largely ionized. Thus, we approximate the gas with a mixture of only neutral
hydrogen and helium in the ratio of 10 to 1 by number. The atomic weight of these
11 particles is 10 + 4 = 14 in unit of atomic mass, and, hence, the molecular weight
of the mixture is 14/11 = 1.27. Actually the molecular weight is close to 1.27 in
most of the photosphere except for the deepest layers where hydrogen ionization is
no longer negligible.
The equation of state alone is not sufficient to find the pressure from the effective
temperature. In fact, as we will see later, we cannot use the mean density of about
1.4 g cm−3, because gravity strongly stratifies the solar gas and the mean density is
much larger than the surface density. A further constraint on the photospheric
parameters can be obtained by determining the thickness of the photosphere. How
thick the photosphere is can be stated in two different ways. On the one hand, the
photosphere is the layer where most of the observed solar light comes from, thus its
thickness depends on the depth where the emergent photons originate. On the other
hand, both the density and pressure drop exponentially passing from bottom to top
so we may define the thickness of the photosphere as the distance over which
pressure and density decrease by a factor e.
Let us begin with the pressure variation with height. The gas stratification is
determined by the condition of hydrostatic equilibrium.
3.1 The Physical Parameters 73
3.1.2 Hydrostatic Equilibrium
We are not surprised that the size of the solar disk does not change with time. From
a physical point of view the solar radius could not remain constant without the
mechanical equilibrium of the forces that act on each part of the Sun. Because the
Sun is a gas, this equilibrium goes under the name of hydrostatic equilibrium. Two
forces that play a role in the hydrostatic equilibrium are gravity and pressure
gradient. Pressure gradient is often called the buoyancy force and this recalls
immediately to us the Archimedes’ principle that states that a body immersed in a
fluid receives an upward push equal to the weight of the fluid that it displaced
(Archimedes was a Greek scientist, about 287–212 BC). Archimedes’ principle is
the base for the concept of hydrostatic equilibrium. In fact, when a body is partially
or fully immersed in a fluid on Earth—and everyone has the direct experience of his
body in the water—the parts of its surface that are submersed are subject to the fluid
pressure. Pressure is larger on the deeper submerged parts so that the component of
the pressure force directed upwards overcomes the component directed downwards.
Thus, the resulting pressure force is directed upwards and produces the buoyancy
that sets itself against the body’s weight. Naturally, if we replace the body’s sub-
merged part with further fluid equal to the surrounding medium, this fluid remains
in mechanical equilibrium because its weight is balanced by buoyancy.
The same principle ensures the mechanical equilibrium of the solar gas: each
macroscopic volume element does not move because its weight is perfectly bal-
anced by the resultant of the pressure forces acting on its surface. In particular, let
us consider a volume element of gas that has the shape of a parallelepiped with its
base A perpendicular to the radial direction, denoted by z, and which belongs to a
layer of thickness dz (Fig. 3.1). Note that the z-axis has the opposite direction
of gravity and the origin deep in the photosphere, thus we call z height. dz is
infinitesimal, as discussed later, and A is small enough to make the parallelepiped
horizontally homogeneous because of the assumption of spherical symmetry
(Sect. 2.1 and later in this section).
The pressure force acting on this gas element is equal to the product of the
surface A and the difference, dP, between the pressure on the upper face, P(z + dz),
and the one on the base, P(z), and is directed toward the outer atmosphere. The
element has a volume dV = Adz, a mass dm = q dV, and, hence, a weight gS dm,
where gS is the surface gravity of the Sun, which is directed inwards.
Therefore, the balance of the weight and pressure forces can be written
½Pðz þ dzÞ PðzÞA ¼ dP A ¼ gS dm ¼ gS qAdz:
Dividing both sides of the equality by the element volume, we get the equation
of hydrostatic equilibrium, i.e.
Fig. 3.1 The hydrostatic equilibrium of the photosphere. A gas parallelepiped with height
oriented along the solar radius, which coincides with the z axis, is in equilibrium along this
direction. Then, the weight force, −gSdm, directed inwards, is balanced by the force created by the
pressure difference between the two horizontal faces of the element, dP A, directed outwards.
The element is also horizontally in equilibrium, because the pressure forces on the opposite side
faces (not indicated in the figure) are equal in magnitude and opposite in direction
dP=dz ¼ gS q;
where dP/dz is the pressure gradient with height inside the photosphere. The minus
sign indicates that the pressure diminishes outwards.
The equation of hydrostatic equilibrium is the first example of the equations that
we will encounter in the following which express the different types of equilibrium
characterizing the present structure of the Sun. Similar to pressure in the case of the
hydrostatic equilibrium, we cannot explicitly express the relevant physical quan-
tities, nevertheless we can constrain their variations. Moreover, the variations, or
differences, are computed for very small, that is, infinitesimal variations of the
variable upon which the quantity depends. Infinitesimal differences are often
approximated with finite differences; for instance the height increment is in finite
terms dz ¼ z2 z1 , where z1 and z2 are two heights close enough that the corre-
sponding variations of the gas properties can be considered small with respect to
their local values.
3.1.3 Scale Height
The pressure gradient is the ratio between the pressure and height differences and
thus the right hand side of the equation of hydrostatic equilibrium, −gS q, has the
dimensions of pressure divided by height. This implies that by replacing the density
with the pressure derived from the equation of state
q ¼ lP=ðRTÞ;
the hydrostatic equilibrium becomes
dP=dz ¼ gS lP=ðRTÞ;
and it is a height the quantity
H ¼ RT=ðgS lÞ:
H is called the pressure scale height and the pressure gradient is the ratio
between the pressure itself and the scale height with the sign minus, i.e.
dP=dz ¼ P=H:
Taking the effective temperature as a typical photospheric value, and l = 1.27 as

the mean molecular weight, the scale height has the value
H ¼ RT=ðgS lÞ ¼ 8:31 107 erg K1 g1 5778 K=ð2:74 104 cms2 1:27Þ
¼ 1:37 107 cm ¼ 137 km:
To understand the meaning of the pressure scale height we need to find the
pressure stratification that the hydrostatic equilibrium determines.
3.1.4 Pressure Stratification
In order to determine the pressure variation with height we assume that the scale
height is constant. Then, separating the two variables, pressure and height,
dP=P ¼ dz=H;
we can integrate the differential equation between the level z0, that we take as a
reference, and any level z, where the pressures are P(z0) and P(z) respectively, i.e.
Zz Zz
dP=P ¼ dz=H:
z0 z0
On the right hand side the constant H goes out of the integral, which then is
simply the height increment
Zz
dz ¼ z z0 :
z0
Since d ðlnPÞ=dz ¼ 1=P, where ln denotes the natural logarithm, the left hand
side is
Zz
dP=P ¼ ½ln Pzz0 ¼ lnPðzÞln Pðz0 Þ
z0
and, hence, the equation becomes
lnPðzÞln Pðz0 Þ ¼ ðz z0 Þ=H:
Because the difference of two logarithms is equal to the logarithm of the ratio of
their arguments, it is
ln½PðzÞ=Pðz0 Þ ¼ ðz z0 Þ=H:
Because the natural logarithm is the inverse function of the exponential, we get
PðzÞ=Pðz0 Þ ¼ eðzz0 Þ=H
and
PðzÞ ¼ Pðz0 Þeðzz0 Þ=H :
Therefore, with a constant scale height the pressure decreases exponentially with
height. In particular, when we move upwards by one scale height, that is
z z0 ¼ H, we obtain
PðzÞ ¼ Pðz0 þ HÞ ¼ Pðz0 Þe H=H ¼ Pðz0 Þe1 ¼ Pðz0 Þ=e;
and the pressure drops by a factor e. Vice versa, moving inwards by one scale
height, i.e. z z0 ¼ H, the pressure increases by a factor e, in fact
PðzÞ ¼ Pðz0 HÞ ¼ Pðz0 Þe H=H ¼ Pðz0 Þe1 ¼ Pðz0 Þe:
Worked Exercise 3.1 Density stratification.

Under the assumption of constant temperature and mean molecular weight the
density also decreases exponentially with height and has the same scale height as
the pressure. In fact, from the equation of state the density at height z is
qðzÞ ¼ lPðzÞ=ðRTÞ:
Using the solution of the hydrostatic equilibrium for the pressure, the density
becomes
qðzÞ ¼ lPðz0 Þe ðz z0 Þ=H =ðRTÞ:
Since
lPðz0 Þ=ðRTÞ ¼ qðz0 Þ;
we get
qðzÞ ¼ qðz0 Þe ðz z0 Þ=H :
Exercise end.
Exercise 3.2 Show that, with constant temperature and mean molecular weight,
dq=dz ¼ q=H: The pressure decreases by a factor 10 moving outwards by a
distance z – z0 = H/log e = 137 km/0.434 = 315 km, where log denotes the deci-
mal logarithm (Exercise 3.3).
Exercise 3.3 Verify that pressure decreases by a factor 10 when moving outwards
by a distance z − z0 = H/log e.
3.1.5 Plane Parallel Model
If we assume H/log e = 315 km as its full width, the photosphere spans about
5 10−4 of the solar radius, RS = 6.96 105 km. As a consequence, the curvature
of the photosphere is negligible on distances much larger than one pressure scale
height, and the photosphere can be represented by a plane parallel model atmo-
sphere (Fig. 3.2). The plane parallel layers are orthogonal to the direction of gravity
and horizontally homogeneous because of the spherical symmetry. Moreover,
passing from the observation of the disk center to that of the limb corresponds to
follow lines of sight that are more and more inclined with respect to the vertical.
Worked Exercise 3.4. The Boltzmann distribution derived from the hydrostatic
equilibrium.
We can justify the Boltzmann distribution for the populations of the atomic
energy levels through the analogy with hydrostatic equilibrium. The density vari-
ation with height z for a gas at uniform temperature T inside a field of gravity with
constant acceleration g is
Fig. 3.2 Panel a Unlike the other stars, the Sun is close enough to the Earth that it is seen as a disk
and not like a single point. The light that comes from different points of the disk and reaches an
observer on the Earth actually follows different directions. However, the Sun is also sufficiently far
that these directions can be approximated with a set of parallel lines of sight (plotted with
horizontal arrows). Therefore, when we observe the disk center, the line of sight is normal to the
surface and parallel to the radius and, when we observe the limb, the line of sight is parallel to the
surface and perpendicular to the radius. Panel b Due to its small thickness with respect to the solar
radius, the photosphere can be described by a model of plane parallel layers that are orthogonal to
the direction of gravity. The reference axis is the vertical z axis which is oriented towards the
observer. Lines of sight, that are progressively more inclined with respect to the vertical, represent
observations moving from the disk center to the limb. Thus, the inclination of the line of sight
relative to the vertical in the plane parallel model corresponds to the angle h between the line of
sight and the solar radius
qðzÞ=qðz0 Þ ¼ eðzz0 Þ=H
where the scale height is
H ¼ RT=ðglÞ ¼ kB T=ðgmÞ;
where m = lmu is the mean molecular mass. The density q and the number of
molecules per unit volume N are simply related by
q ¼ mN;
then
qðzÞ=qðz0 Þ ¼ NðzÞ=Nðz0 Þ
and
NðzÞ=Nðz0 Þ ¼ emgðzz0 Þ=ðkB TÞ :
Therefore, the ratio of the number densities at two different heights z and z0
exponentially increases with temperature, but decreases with Ez – Ez0 = mg(z – z0),
which is the difference between the gravitational potential energies at these levels.
In other words, gravity produces a distribution of gas particles over potential
energies analogous to the Boltzmann distribution over atomic excitation energies, in

the particular case that the energy levels are not degenerate, so that the statistical
weights are all unity.
Exercise end.
Having determined the thickness of the photosphere from the hydrostatic pres-
sure stratification, we now consider the depth in the photosphere where most of the
observed solar light comes from.
3.1.6 Optical Thickness
Let us restrict ourselves to a single line of sight and consider the photons emerging
from the photosphere along this direction. Stating the depth in the photosphere
where this radiation intensity originates is analogous to stating how far in the fog
we can see the light of a lighthouse. In the case of the lighthouse, we know that the
maximum distance to be seen is smaller if the fog is thick while it is larger when the
fog is thin. Similarly, the depth where the photospheric light is formed depends on
how opaque the solar gas is. For this reason, the concept of optical depth, or optical
thickness, that measures the distance between two points according to the opacity of
the interposed material, comes in very useful. The optical thickness, sk (s is a Greek
letter to be read tau), is given by the product of the geometrical distance, d, along
the line of sight, by its opacity, kk, which we take simply constant for the time
being, i.e.
sk ¼ kk d:
The label k indicates that the opacity and, hence, the optical thickness depend in
general on the light wavelength. In this way two geometrical distances that are
equal but characterized by different opacities, will have different optical thick-
nesses; in particular, the optical thickness is larger where the atmosphere is more
opaque (Fig. 3.3).
At the microscopic level, the gas opacity is due to photon absorption by gas
atoms. In order to highlight the physical parameters that determine the opacity, we
will consider the simple case of a sphere of radius r moving through a gas of
N fixed, point-like particles per unit volume. In this simple model the photon—
atom collisions, in which light absorption occurs, are represented by the collisions
of the sphere with the gas particles, and a collision occurs whenever the sphere
encounters a point-like particle at a distance equal or less than its radius
r. Therefore, the useful area for a collision is given by the sphere’s cross-section,
p r2, and is usually denoted by the Greek letter r (name sigma), i.e.
Fig. 3.3 A lighthouse is at geometrical distance d from an observer (a). The lighthouse has an
optical distance s1 = k1d from the observer in a medium with constant opacity k1 (b). The
lighthouse has a larger optical distance s2 = k2d from the observer when the interposed medium is
more opaque, k2 > k1 (c)
Fig. 3.4 The collisions of a

moving sphere of radius
r with the fixed, point-like
particles of a gas with number
density N. The particles that
the sphere can impact within
the path L are those (in green)
contained in the cylinder
whose height is L and whose
base is the cross section of the
sphere, r = p r2, called the
cross section of the process
r ¼ p r2 :
The number of the collisions occurring while the sphere covers the distance L is
equal to the number of the point-like particles that are contained inside the cylinder
of length L whose base is the sphere cross section, i.e. N p r2 L (Fig. 3.4).
Likewise, the number of photon absorptions occurring on a distance L is
Nrk L;
where N represents the number of atoms or ions per unit volume capable of
absorbing radiation in a particular transition at wavelength k, and the cross section
rk is interpreted in a statistical way as the probability that absorption occurs.
Because the opacity at a point, kk, results from all the absorptions occurring in its
neighborhood and is defined as the number of absorptions per unit length, we get
kk ¼ Nrk L=L ¼ Nrk :
Thus, the opacity is the product of the number density of the absorbers in a
particular transition and the transition probability given by the cross section.
The photon mean free path, which is defined as the distance that photons can
travel in average before they are absorbed, is a useful concept, correlated with the
opacity. In the case of Fig. 3.4, the mean free path, l, corresponds to the average
distance between two subsequent collisions, and is given by the total distance
covered by the sphere divided by the corresponding number of collisions, that is
l ¼ L=ðNrk LÞ ¼ 1=ðNrk Þ:
Therefore, the mean free path decreases proportionally when the gas density or
the cross section increase.
Combining the expressions for the opacity and photon mean free path, we find
that they are reciprocal quantities
lk ¼ 1=N rk ¼ 1=kk :
This result is important. We can insert it in the definition of optical depth and
obtain
sk ¼ kk lk ¼ kk ð1=kk Þ ¼ 1:
Thus, the optical depth along a line of sight corresponding to one photon mean
free path is unity. Note that opacity is the inverse of a length and, hence, the optical
thickness is dimensionless.
One mean free path, being the distance that a photon covers in average before it
is absorbed, also represents the mean distance where the photons leaving the
photosphere come from and, hence, is a measure of the geometrical thickness of the
photosphere. Equivalently, optical depth unity is an estimate for the optical
thickness of the photosphere.
3.1.7 Flux Formation Depth
The last paragraph answered the question about the formation depth of the radiation
intensity emitted by the photosphere. In many circumstances, however, we deal not
with the rate of energy that is emitted per unit area along a single line of sight,
which is the emitted intensity, but with the total rate of energy that is emitted per
unit area, or emission flux. Two examples of emission flux are the solar constant,
discussed in Chap. 2, and the light that an observer on the Earth receives from stars.
We turn back to the definitions of radiation intensity and flux at the beginning of
Sect. 3.2. Thus, we introduce a mean free path, lf, and an associated optical depth,
sf, that respectively define the geometrical and optical formation depths for the
radiation flux. We will see later that the optical formation depth for flux is equal to
2/3, hence, omitting the wavelength dependence for sake of simplicity
sf ¼ klf ¼ 2=3;
where k is a suitable constant opacity.

Because the optical depth where the radiation intensity originates is unity, we
have
si ¼ kli ¼ 1;
where li denotes the mean free path for the radiation intensity. Comparing sf and si,
we find that the flux and intensity mean free paths are related by
lf ¼ 2li =3:
Worked Exercise 3.5 On the flux formation optical depth.

In Fig. 3.5 we adopted a plane parallel slab to represent the photosphere.
The vertical z axis has z = 0 at the bottom and z = Z at the surface. A point P at
height z has distances from the surface equal to Z − z along the vertical and to
S − s along a line of sight inclined by an angle h. The distances S − s and Z − z are
related by
S s ¼ ðZ zÞ=cos h ¼ ðZ zÞ=l;
where we let l = cos h (l is a Greek letter, name mu).

Let us take for S − s one mean free path for the intensity. Then, the corre-
sponding optical thickness is
sk ¼ kk ðS sÞ ¼ kk li ¼ 1
along the line of sight, and

Fig. 3.5 The distance of a point P inside the photosphere to the surface is PA = Z – z along the
vertical and PB = S – s along a line of sight inclined by an angle h with respect to the vertical.
Because the triangle PAB is rectangle in A, the hypotenuse PB, the side PA and the angle h are
related by PA ¼ PBcos h; hence ðZ zÞ ¼ ðS sÞcos h ¼ ðS sÞl:
sk z ¼ kk ðZ zÞ ¼ kk ðS sÞl; ¼ kk li l ¼ l;
along the vertical. The vertical optical depth skz, where the intensity at different
inclinations h is formed, is proportional to l = cos h and, hence, decreases from 1
to 0 as the line of sight changes from vertical (h = 0◦) to horizontal (h = 90◦). The
photons that contribute to the emergent radiation flux follow all the possible
inclinations between h = 0◦ and h = 90◦. Therefore, the radiation flux originates in
average from a vertical optical depth that is smaller than 1. We will see later that
this optical depth is equal to 2/3 and corresponds to a level which is about 50 km
above the level at unit optical depth.
Exercise End.
We are now in the position to state a further physical condition that, together
with the equation of state, correlates the photospheric thermodynamic parameters.
The mean formation optical depth of the solar radiation flux is given by
sf ¼ klf ¼ 2=3;
where lf is the flux mean free path. We require that lf equals the pressure scale
height H, because both quantities are estimates of the geometrical thickness of the
photosphere. Thus,
lf ¼ H
and we get
2=3 ¼ sf ¼ k H:
This equation implicitly correlates the values of the temperature, pressure and
density at optical depth 2/3. We already know the expression for the scale height in
terms of the temperature, thus we have to examine the way the opacity in the visible
wavelength range depends upon the temperature and density.
3.1.8 Photospheric Opacity
The opacity at a wavelength k in a particular atomic transition is the product of the

number of particles per unit volume capable of absorbing radiation, N, and the
transition cross section, rk
kk ¼ N r k :
Because in the following we deal with continuum processes that are smooth
functions of wavelength, we leave out the index k and imply a suitable mean
opacity in the visible part of the spectrum.
The photoionization of the negative hydrogen ion, H−, represents the dominant
absorption process in the conditions of the solar photosphere. Hydrogen is the most
abundant element of the solar gas. In the photosphere hydrogen can form a negative
ion that consists of one proton and two electrons. This ion has only one bound state
with a binding energy of 0.754 eV. The absorption cross section of negative
hydrogen is large and, even though only a small fraction of hydrogen exists in this
form, the opacity due in particular to the H− photoionizations is dominant in the
solar photosphere at visible wavelengths. For this process we can roughly write
k ¼ NH p rH2 ;
where NH is the number of negative hydrogen ions per unit volume and
rH = 52.91775 pm is the radius of the hydrogen in its ground state according to the
Bohr’s atomic model.
3.1.9 Hydrogen Ionization Equilibrium
The determination of the number of negative hydrogen ions requires an analysis of

the processes that make this ion form or, vice versa, disappear, that is, the analysis
of the ionization equilibrium of negative hydrogen. This equilibrium is represented
by the relation
H þ e $ H ;
where the double arrow indicates that processes of formation and dissociation occur
in both directions. We assume that the number of these processes is sufficiently high
that the thermodynamic equilibrium holds. In this case, the product between the
number densities of the reactants, neutral hydrogen, NH, and electrons, Ne, divided
by the density of the resulting negative hydrogen, NH , is a function of the tem-
perature and satisfies the Saha equation
NH Ne =NH ¼ u ðTÞ;
where the Saha function for the equilibrium of negative hydrogen u− is
u ¼ ðUH =UH Þ2ð2p me kB T=h2 Þ3=2 e EH =ðkB TÞ :
UH and UH are the partition functions of the negative and neutral hydrogen
respectively, and EH ¼ 0:754eV is the H− ionization potential. For T = 5778 K,
UH ¼ 1, and UH = 2, which leads to
u ¼ ð2=1Þ 2
½2 3:14 9:11 1028 g 1:38 1016 ergK1 5778K=ð6:63 1027 ergsÞ2 3=2
exp½0:754eV 1:60 1012 erg eV1 =ð1:38 1016 ergK1 5778KÞ
¼ 9:32 1020 cm3
The Saha equation for negative hydrogen relates the density of this ion to the
electron and neutral hydrogen densities. For a simple hydrogen gas consisting of
electrons, protons, neutral H atoms and negative H ions, the hydrogen ionization
equilibrium is completed by a second condition that refers to the formation of
neutral hydrogen from proton and electron and to the inverse process of dissocia-
tion, that is,
p þ e $ H:
The number densities of the species involved in the neutral hydrogen equilib-
rium satisfy another Saha equation
Np Ne =NH ¼ u þ ðTÞ;
where Np is the proton number density and the Saha function u+ is
u þ ¼ ðUp =UH Þ2ð2pme kB T=h2 Þ3=2 e EH =ðkB TÞ :

Here Up is the partition function for a proton and EH = 13.598 eV is the H

ionization potential. For T = 5778 K, it results in Up = 1, and, then,
u þ ¼ ð1=2Þ 2
½2 3:14 9:11 1028 g 1:38 1016 ergK1 5778K=ð6:63 1027 ergsÞ2 3=2
exp½13:6eV 1:601012 erg eV1 =ð1:38 1016 ergK1 5778KÞ
¼ 1:49 109 cm3
3.1.10 Electric Charge and Mass Conservation
The two Saha equations for hydrogen contain as unknown quantities the four
number densities of electrons, protons, neutral H atoms and negative H ions.
However, there are two more conditions that make the problem determined. The
first condition is a consequence of the conservation of the electric charge. If we
require that on the whole the gas is neutral, then the total amount of positive and
negative charges must be equal to 0. Thus, denoting by e the modulus of the
electron charge, we get
eNp eNe eNH ¼ 0
and, hence,
Np ¼ Ne þ NH :
The second condition derives from the conservation of the total mass, so the sum
of the four unknown mass densities must give the total gas density, that is,
Np mp þ NH mH þ NH mH þ Ne me ¼ q:
In order to simplify the Saha and conservation equations, we evaluate the rel-
ative importance of the four number densities Np, NH, NH– and Ne. These densities
are determined by photoionizations and collisional ionizations. The wavelength, k,
is related to the ionization potential of negative hydrogen by
EH ¼ hm ¼ hc=k;
which leads to
k ¼ hc=EH
¼ 6:6262 1027 erg s 2:9979 1010 cm s1 =ð0:754 eV 1:6022 1012 erg=eVÞ
¼ 1:6443 104 cm ¼ 1:6443 lm:
Because the photons that can ionize negative hydrogen have energies greater or equal
to EH−, this is their longest wavelength and occurs in the near infrared. In the photo-
sphere, at temperatures of about 6000 K, most of photons have wavelengths in the visible
range and, hence, have sufficient energy to ionize negative hydrogen. On the contrary,
there are far fewer photons that are able to ionize neutral hydrogen, whose ionization
potential, EH = 13.598 eV, is a factor 20 higher than the H− ionization potential.
Moreover, the most probable kinetic energy computed at the effective temperature is
kB T ¼ 1:38 1016 erg K1 5778 K ¼ 7:97 1013 erg=1:60 1012 erg eV1
¼ 0:498 eV;
Consequently, many gas particles have energies of the order of the H− ionization
potential so that collisional ionizations are more frequent for negative hydrogen
than for neutral hydrogen. As a result, the hydrogen in the photosphere is only
weakly ionized and negative ions are an even smaller amount than the neutral
hydrogen number. In summary, there is the following hierarchy between the
number densities of negative hydrogen, electrons and neutral hydrogen
NH Ne NH ;
where the symbol means much less than.

Now we are in the position to simplify the equations of ionization equilibrium
and charge and mass conservation. Neglecting the contribution of negative
hydrogen to charge conservation, we obtain
Np ¼ Ne :
Moreover, neglecting the contributions of negative hydrogen, protons and

electrons to mass conservation, we get
NH ¼ q=mH :
Thus, in our approximated photosphere made only of hydrogen, the proton

number equals the electron number and the gas density is due to neutral hydrogen.
Inserting these approximations into the Saha equation for the ionization of neutral
hydrogen, we estimate the electron number density in terms of the gas density. In
fact,

u þ ðTÞ ¼ Np Ne NH ¼ Ne Ne =ðq=mH Þ ¼ mH Ne2 q;
from which

Ne2 ¼ u þ q mH ;
and
1=2
Ne ¼ u þ q=mH :
Finally, to evaluate the negative hydrogen number density, we insert the electron
and neutral hydrogen number densities into the Saha equation of negative hydro-
gen. Thus, we obtain
1=2 .
u ðTÞ ¼ NH Ne =NH ¼ q u þ q=mH ðNH mH Þ;
from which
1=2 .
NH ¼ q u þ q=mH ðmH u Þ
and
.
1=2 3=2
NH ¼ u þ q3=2 mH u :
3.1.11 Evaluating Density and Pressure
Now we turn back to the expression of the optical depth equal to 2/3
2=3 ¼ k H ¼ NH prH2 H:

–
Replacing the density of H in the opacity, we get
1=2
2=3 ¼ u þ q3=2 prH2 H=ðm3=2 u Þ:
This equation contains only the gas density and temperature, so it allows us to
infer the density at s = 2/3 as a function of the effective temperature. In detail, we
find that
.
3=2 1=2
q3=2 ¼ 2mH u 3 u þ p rH2 H ;
from which, raising both sides of the equality to 2/3, we get

1=3 4=3
q ¼ ð2 u Þ2=3 mH =ð32=3 u þ p2=3 rH H2=3 Þ
.h 1=3
¼ ð2 9:32 1020 cm3 Þ2=3 1:67 1024 g 32=3 1:49 109 cm3
i
3:142=3 ð5:29 109 cmÞ4=3 ð1:37 107 cmÞ2=3 Þ
¼ 0:941 107 g cm3 :
This estimate of the density at s = 2/3 is a factor 2 smaller than the value
q ¼ 2:06 107 g cm3 ;
that an accurate model of the solar photosphere provides for the density at the level
with temperature T = 5778 K. The accurate model in question is the model VALC
constructed for the quiet photosphere by Vernazza, Avrett, and Loeser (1981).
Knowing the temperature and density we can derive the pressure at s = 2/3 by
means of the equation of state. We find

P ¼ kB q T=mH ¼ 1:38 1016 egr K 0:941 107 g cm3 5778K 1:67 024 g
¼ 4:49 104 dyn cm2 ¼ 4:49 104 Ba;
a value that is slightly lower than the value
P ¼ 6:07 104 Ba;
at T = 5778 K in the model VALC. Note that Ba is the symbol for Barye, that is,
the pressure unit in the CGS system; hence,
1 Ba ¼ 1 dyn cm2 :
The density of the photosphere is only about a tenth of a millionth of the average
density of the entire Sun. This is a result of the hydrostatic equilibrium. To balance
the weights of its internal layers, pressure gradients are required that are stronger for
deeper layers. Thus, passing from inner to outer layers, the density steadily
decreases. Using our approximated surface values for the pressure and scale height,
the hydrostatic equilibrium provides an average pressure gradient in the photo-
sphere of
dP=dz ¼ P=H ¼ 1:65 104 Ba=137 km ¼ 120 Ba per km:
Because the pressure, density and also the temperature as we will see soon,
rapidly increase inside the Sun, one surface value for each physical parameter is not
enough information; we need the full variations of the thermodynamic parameters
over the solar radius. This enlarged information constitutes a model of the solar
structure. In order to construct such a model we exploit the current equilibrium
conditions for the Sun, like the hydrostatic equilibrium that has already been dis-
cussed. Because the equilibrium conditions determine the variations of the struc-
tural parameters but not their absolute values, in the following section we start the
model from the photosphere, where we have already estimated the local physical
parameters.
3.2 A Model of the Photosphere
3.2.1 Radiation Intensity and Flux
The vertical variation of the photospheric temperature can be determined from the
observed solar intensity and flux in the visible and infrared wavelength range. Let
us then start from the definitions of radiation intensity and flux.
In Chap. 1 we have defined the amount of radiation energy dEk flowing in the
direction s through the surface dS, orthogonal to s, in the time interval dt, in the
solid angle dx around s and in the wavelength band dk, as the product
dEk ¼ Ik dS dt dk dx:
The radiation intensity Ik represents, then, the energy radiated from a point in a
single direction, per unit normal area, time, wavelength and solid angle.
If we are interested in the energy that crosses an area dA in all directions, we
need to consider that dA may be inclined with respect to the radiation direction
s (Fig. 3.6).
Fig. 3.6 The radiation energy dEk that flow in the direction s and crosses the area dA, is
proportional to the solid angle dx ¼ sin h du dh (panel a, where the z axis is normal to dA, h is the
inclination of s with respect to z, and u is the azimuth of s in the plane to where dA belongs). dEk is
also proportional to the area, dS = cos h dA, normal to s (panel b, where the area dA and dS are
seen in profile)
3.2 A Model of the Photosphere 91
Denoting with z the normal to dA and with h the inclination angle of s with
respect to z, the normal area dS is related to dA by (Fig. 3.6 panel b)
dS ¼ cos h dA:
The elementary amount of energy that is radiated across dA in all directions is

the integral of dEk over the solid angle, i.e.
Z Z Z Z
dEk ¼ Ik dS dt dk dx ¼ I cos h dA dt dk dx ¼ dA dt dk Ik cos h dx:
x x x x
Figure 3.6, panel a shows that the elementary solid angle is
dx ¼ sin h du dh;
where u is the azimuth of the line of sight s.

Thus, the integral over dx is a double integral over u and h
Z Z2p Zp
Ik cos h dx ¼ du Ik cos h sin h dh;
x 0 0
where 2p is the range covered by u, and h ranges from 0, for s = z, to p, for s = −z.
Moreover, we let l = cos h, so that
dS ¼ cos h dA ¼ l dA
and
dl ¼ d ðcos hÞ ¼ sin h dh:
Thus
Z Z2p Zp Z2p Z1

Ik cos h dx ¼ du Ik cos h sin h dh ¼ du Ik l ðdlÞ
x o 0 o 1
Z2p Z1
¼ du Ik l dl;
o 1
where l ranges from 1, for s = z, to −1, for s = −z.

Therefore, the elementary amount of energy that is radiated across dA in all
directions is
Z Z2p Z1
dEk ¼ dA dt dk du Ik l dl:
x o 1
The integral
Z2p Z1
Fk ¼ du Ik l dl;
o 1
is the radiation flux across dA. Thus Fk is the radiant energy per unit area, time and
wavelength. It is obtained by adding the intensities along all lines-of-sight each one
weighted by the cosine of its inclination with respect to the normal to the crossed
area. Notice that in general, the radiation flux is a vector whose components rep-
resent the energy rate per unit area normal to each of the three orthogonal axes
respectively. In particular, Fk is the z component of the radiation flux.
When the photosphere is modeled by parallel and homogeneous planes that are
orthogonal to the direction of gravity with the z axis in the opposite direction of
gravity, the radiation intensity along a direction s depends on both z and h but does
not depend on the angle u. Thus, the integral of the intensity over u gives simply a
multiplication by 2p and the flux is
Z1
Fk ¼ 2p Ik l dl:
1
Exercise 3.6. Assume that the photosphere emits like a black body and show that
the monochromatic emission flux, Fk is p times the Planck function, i.e. Fk = p Bk
Let us now consider the observed solar flux.
3.2.2 The Spectrum
Comparing the solar flux observed in the visible and infrared wavelength ranges
with a black body spectrum, we realize that the photosphere is only approximately
represented by a black body at the effective temperature Te = 5778 K (Fig. 3.7).
The observed spectrum is rather close to a black body for wavelengths
k > 0.7 lm while it appears clearly brighter almost everywhere at smaller wave-
lengths. Moreover, the Planck function is throughout a smooth function of wave-
length, but the photospheric spectrum displays many regions of suddenly reduced
flux, which have a variety of widths and depths (Fig. 3.8).
Fig. 3.7 The spectrum of solar flux from 0.3985 to 1 lm inferred from the measurements of Labs
and Neckel (1962) and Neckel and Labs (1984). The solar spectrum follows approximately the
curve of the black body radiation at 5778 K (in orange) and is closer to it in the infrared
These spectral features are called absorptions lines. This name derived from
their appearance on a photographic plate as a series of dark vertical bands, like a bar
code of the photosphere. They are also called Fraunhofer lines, because they were
discovered by the German scientist Joseph von Fraunhofer (1787–1826). The
spectral ranges free of lines are designated as continua.
The differences of the solar spectrum with the black body spectrum dramatically
increase the diagnostic potential of the solar light. Thus, an effort at their inter-
pretation is necessary. Because the formation of the solar spectrum is a complex
problem, we limit ourselves in the following to a number of basic concepts that are
useful to construct an approximate model of the photosphere.
3.2.3 The Origin of Continua and Lines
Spectral continua are produced in free-bound, bound-free and free-free atomic

transitions. All these transitions involve a free electron. In a free-bound transition
the free electron recombines to an atom or ion; in a bound-free transition a bound
electron becomes free, and in a free-free transition the kinetic energy of a free
electron, which is close to an atom or ion, changes because of interaction with a
third particle or a photon. An example of bound-free process is represented by the
photoionization of negative hydrogen that, as we know, is the most frequent
absorption process occurring in the photosphere. The inverse processes of H−
recombination are free-bound processes that give the dominant contribution to the
solar emission in the visible continuum.
Fig. 3.8 The solar flux in the range of the spectral line Ha at k = 0.6563 lm with high spectral
resolution. Each spectral point is integrated over a wavelength band of the order of 0.1 pm. The
same line is also visible in the central part of the spectrum of Fig. 3.7 but with a much lower
resolution, each point being integrated over a band as large as 2 nm
Spectral absorption lines are produced when atoms or ions make bound-bound
transitions. In these transitions the energy gained (lost) by an electron jumping
between two atomic bound levels equals the energy of an absorbed (emitted)
photon, or the energy that a colliding particle loses (gains). For instance, the Ha line
at k = 0.6563 lm, plotted in Fig. 3.8, corresponds to the transition between the first
and the second excited level of the hydrogen atom (Fig. 3.9).
Worked Exercise 3.7 The Ha line center wavelength.
The wavelength, k, associated with the Ha line center, can be computed using
the level energies of the Bohr’s model for the hydrogen atom
EH =22 EH =32 ¼ hm ¼ hc=k;
where EH = 13.60 eV is the ionization potential of the neutral hydrogen. Thus, we

get
k ¼ hc=½EH ð1=22 1=32 Þ

¼ 6:626 1027 erg s 2:998 1010 cm s1 13:60 eV 1:602 1012 erg=eV 5=36
¼ 6:565 105 cm ¼ 0:6565 lm:
Exercise end.
If the identification of a chemical element responsible for a series of spectral
lines may be easy, it is more difficult to compute the emergent intensity and flux at
Fig. 3.9 The energy levels of

the hydrogen atom. The
principal quantum number
n is given on the left hand side
of each level and the binding
energy in eV is given on the
right. The red arrow between
the levels n = 2 and n = 3
shows the energy jump of the
electron that originates the
absorption line Ha at
k = 0.6563 lm in the
spectrum of the solar
photosphere
each wavelength. In fact, once we abandon the simple assumption that the photo-
sphere is in thermodynamic equilibrium at the effective temperature, the radiation
intensity is no more given by the Planck function, but it becomes an unknown
quantity that has to be determined. This is accomplished using the equation of
radiation transfer.
3.2.4 Equation of Radiation Transfer
In absence of thermodynamic equilibrium there is no longer a detailed balance of

the collisional and radiative processes that redistribute particles and photons over
their energies. In particular, over a distance ds along a line of sight s, the number of
emitted photons of a given energy differs from the number of photons with the same
energy which are absorbed. Thus, the radiation intensity Ik varies from point to
point and all we can say is that its variation, dIk, over ds must match the energy
difference per unit normal area, time, wavelength and solid angle due to the
unbalance between the emission, jkds, and the absorption, kk Ikds, (Fig. 3.10), i.e.
dIk ¼ ðjk kk Ik Þ ds:
This is a condition of energy conservation for radiation which is called the

equation of radiation transfer. Note that the distance increment ds is physically
small and mathematically infinitesimal, as is the increment dz in the equation of
hydrostatic equilibrium.
Fig. 3.10 Radiation transfer along a line of sight. While crossing a photospheric layer of length ds
along a line of sight s, a light beam of intensity Ik varies by the quantity dIk = (jk–kkIk)ds that is the
difference between energy emission and absorption along ds per unit normal area, time,
wavelength and solid angle
In order to introduce the optical depth into the radiation transfer equation, we
generalize the definition of optical depth to the case of an opacity that varies with
the distance s along the line of sight. The opacity is constant over an infinitesimal
distance increment ds, thus the optical depth increment, dsk, corresponding to ds, is
simply the product
dsk ¼ kk ds:
Note that the sign minus means that for an outward line of sight the distance
increases outwards while the optical depth increases inwards. In particular the
optical depth is zero at infinity distance from the photosphere where the observer is
located, i.e.
sk ðs ¼ 1Þ ¼ 0:
The optical depth sk(s) along the line of sight of a point P at a distance s from a
reference point inside the photosphere, is the integral of the optical depth increment
between optical depths 0 and sk and, hence, results from the opacity contributions
of all the layers between P and the observer. Then
Zsk Zs
sk ð s Þ ¼ dsk ¼ kk ds
o 1
and
Z1
sk ð s Þ ¼ kk ds:
s
Turning back to the equation of radiation transfer, we factorize the product of kk

and ds
dIk ¼ ðjk kk Ik Þ ds ¼ ðjk kk =kk kk Ik Þ ds ¼ ðjk =kk Ik Þkk ds:
Because dsk = −kk ds, we obtain
dIk ¼ ðIk jk =kk Þ dsk :
The ratio of emission to opacity, jk/kk, is called the source function and is
denoted by Sk. Then, the equation of radiation transfer becomes
dIk ¼ ðIk Sk Þ dsk ;
or
dIk =dsk ¼ Ik Sk ;
where dIk/dsk is the intensity gradient with the optical depth along the line of sight
and
Sk ¼ jk =kk:
3.2.5 Solution of the Transfer Equation
In absence of emission Sk = 0 and the equation of radiation transfer reduces to
dIk =dsk ¼ Ik;
from which
dIk =Ik ¼ dsk :
This equation for Ik has the same form of the equation of hydrostatic equilibrium
for pressure but with the scale height replaced by the unity optical depth sk = 1.
Then, given two points P0 and P along s at optical depths sk0 and sk, respectively,
the solution is
Ik ðsk Þ ¼ Ik ðsk0 Þeðsk s0 Þ :
In absence of emission, the radiation intensity is attenuated by the exponential

factor eðsk sk0 Þ while moving from P0(sk0 ) to P(sk) along s.
Exercise 3.8 Show that the radiation intensity crossing normally a slab with width
L and constant opacity k is attenuated by e−kL.
On the other extreme, for vanishing opacity the transfer equation is
dIk ¼ jk ds:
Given two points P0 and P along s at distances s0 and s, respectively, the

solution is
Zs
Ik ðsÞ ¼ Ik ðs0 Þ þ jk ðs0 Þds0 ;
s0
where s′ is the distance of a generic point P′ of s located between P0 and P. Thus,

the intensity at P is the intensity coming from a deeper point P0 plus the contri-
butions due to the emission from all the interposed layers.
With both non-zero emission and absorption, the general solution of the radia-
tion transfer equation is (Exercise 3.9)
Zs0
0
ðsk0 sk Þ
Ik ðsk Þ ¼ Ik ðsk0 Þe þ Sk ðs0k Þeðsk sk Þ d s0k :
sk
were sk, sk′ and sk0 are the optical depths along s of the points P, P′ and P0,
respectively, and P′ is located between P0 and P.
Exercise 3.9 Verify that the solution given in text effectively satisfies the transfer
equation.
The intensity at P(sk) is the sum of two terms. The first term is the contribution
of the intensity Ik(sk0 ) that comes from a deeper point P0(sk0 ) along s, which is
exponentially attenuated by the optical thickness separating P from P0. The second
term is due to the contribution of the source function Sk(sk′) from all the points P′
between P0, and P, each one exponentially attenuated by the optical thickness
separating P from P′.
Let us now discuss an approximate solution of the equation of radiation transfer
that expresses the radiation intensity in terms of the source function computed at
unity optical depth along the line of sight. This formula is named after the English
scientist sir Arthur Eddington (1882–1944) and the French astronomer Daniel
Barbier (1907–1965).
3.2.6 The Eddington-Barbier Approximation
We begin by approximating the differentials that appear in the radiation transfer

equation with the finite differences between two points at optical depths sk1 and sk2
along the line of sight. Thus, the equation
dIk ¼ ðIk Sk Þ dsk
takes the form
Ik ðsk2 Þ Ik ðsk1 Þ ¼ ½Ik ðsk1 Þ Sk ðsk1 Þ ðsk2 sk1 Þ:
Then, we let sk2 = sk to denote any optical depth, and take sk1 = sk + 1, that is
the level one photon mean free path deeper than the level sk. Thus, the equation
becomes
Ik ðsk Þ Ik ðsk þ 1Þ ¼ ½Ik ðsk þ 1Þ Sk ðsk þ 1Þ ð1Þ
and
Ik ðsk Þ Ik ðsk þ 1Þ ¼ Sk ðsk þ 1Þ Ik ðsk þ 1Þ:
Cancelling out the term Ik(sk + 1), which is common to both sides, finally we get
Ik ðsk Þ ¼ Sk ðsk þ 1Þ:
Notice that the choice sk1 = sk + 1 is the only one such that the knowledge of a
further value of intensity is not required. The intensity at a fixed optical depth along
a line of sight equals the source function computed at the same optical depth
increased by 1. This result is called the Eddington-Barbier approximation. The
Eddington-Barbier approximation confirms that the observed solar intensity origi-
nates at unit optical depth along the line of sight.
Exercise 3.10 Show that the Eddington-Barbier approximation can be obtained
from the solution of the radiation transfer equation with a linear Sk(sk) for an
optically thick atmosphere where sk0 approaches ∞.
3.2.7 LTE
The solution of the radiation transfer equation is only a formal solution till we
specify the source function as a function of the optical depth. In the absence of TE
the source function has to be determined from all the emission and absorption
processes that are relevant at the wavelength under examination and, therefore,
depends on the populations of the atomic energy levels. For example, in the case of
a bound-bound transition the line source function has the general form
Sk ¼ jk =kk ¼ ð2hc2 =k5 Þ=½N1 g2 =ðN2 g1 Þ 1:
where N2 and N1 are the upper and lower level populations (Exercise 3.11).
Exercise 3.11 Verify the form of the line source function given in the text.
The atomic level populations are determined by the condition that they remain
constant in time. This condition is named statistical equilibrium (SE) and implies
that the sum of all the collisional and radiative transitions that populate a given level
in a unit time interval equals the sum of all the collisional and radiative transitions
depopulating it. Note that this is a global balance and not the detailed balance that
occurs in TE.
Thus, the determination of the radiation intensity, which appears in the transfer
equation, is coupled to that of the atomic level populations, which are the unknown
quantities of the SE equations. We can solve this problem if we model the pho-
tosphere under the assumption of local thermodynamic equilibrium (LTE), which is
the simplest way to expand on TE.
In thermodynamic equilibrium there is one uniform temperature that fixes the
particle velocity distribution, which is given by the Maxwellian distribution, the
atomic level populations, which follow the Boltzmann and Saha equations, and the
source function and the radiation intensity, which are both equal to the Planck
function. In LTE, with a solar temperature that varies layer by layer, we assume that
the TE energy distribution functions for the gas particles and the source function for
photons remain those of TE but must be evaluated at the local temperature of each
solar layer. Note that the assumption on the photon energy distribution is not
expanded to include Ik = Bk. This is because the departure from TE is just due to
the energy loss corresponding to the solar luminosity.
Under the LTE assumption, with the temperature varying with optical depth, T
(sk), the source function equals the Planck function calculated at the local tem-
perature, i.e.
Sk ¼ Bk ½Tðsk Þ ðLTEÞ:
Thus, we can write the transfer equation in the form
dIk ¼ ðIk Bk Þ dsk
and its solution is
Zsk0
0
Ik ðsk Þ ¼ Bk ðsk0 Þeðsk0 sk Þ þ Bk ðs0k Þeðsk sk Þ d s0k :
sk
As a consequence of LTE, the transfer equation directly contains the temperature

through the Planck function and its solution does not require the detailed calculation
of the microscopic processes of emission and absorption. This certainly is an
important simplification, but the effective validity of LTE is warranted only a
posteriori, that is, after a detailed calculation is made. We expect that the approx-
imation of local thermodynamic equilibrium improves for the internal layers that
are denser and more opaque. In these inner layers, the photons that are emitted and
not reabsorbed, and therefore form the net outward flux, become an increasingly
smaller fraction of the total number of existing photons. Moreover, the frequent
collisions, in particular with the fast free electrons, contribute in a determinant
manner to redistribute atoms and ions over their energy states, and, also, to couple
the source function to the local gas temperature.
We are now in the position to interpret the observed solar intensity and flux in
the visible and infrared wavelength range.
3.2.8 Center to Limb Intensity
The solar disk in the visible continuum appears brighter at center than at limb: this
effect is known as the limb darkening (Fig. 3.11).
The limb darkening indicates that the observed intensity decreases with
decreasing l. The intensity variation, which is smooth close to center and sharp
toward limb, is well fitted by a fifth order polynomial in l with the coefficients
depending on wavelength (Pierce and Slaughter 1977). However, we limit
Fig. 3.11 The limb

darkening clearly shows up in
this image of the solar disk in
the visible continuum
obtained at Big Bear Solar
Observatory, USA
ourselves to the approximation that the emergent continuum intensity is a linear

function of l, independently of the wavelength in the visible range. The linear
center-to-limb of the continuum intensity Ic has the form
Ic ð0; lÞ ¼ Ic ð0; 1Þ ð2 þ 3lÞ=5;
where Ic(0,1) is the emergent intensity at the disk center and Ic(0,l) is emitted at a
distance RS(1 − l2)0.5 from the disk center and, hence, with an inclination
h = arcos l with respect to the vertical (Exercise 3.12).
Exercise 3.12 Show that the intensity emitted at a distance RS(1 − l2)0.5 from the
disk center has an inclination h = arcos l with respect to the vertical.
Then, the ratio Ic (0,l)/ Ic (0,1), called the limb darkening function, is
Ic ð0; lÞ=Ic ð0; 1Þ ¼ ð2 þ 3lÞ=5:
The linear approximation to center-to-limb is satisfactory at disk center but

rather crude at the extreme limb.
Exercise 3.13 The coefficients of the fifth order polynomial in l that Pierce and
Slaughter give for the limb darkening at k = 0.5038 lm are in order of increasing l
power: 0.23546, 1.13724, −0.42727, −0.56643, 1.19616, and −0.57516. Compare
this polynomial with the simple linear limb darkening given in the text.
3.2.9 The Temperature Variation
The temperature variation with optical depth can be determined from the observed
limb darkening. We replace the optical depth sk along the line of sight, which
depends on the inclination, with the vertical optical depth skz, which is defined at
the z level as
Z1
skz ¼ kk dz:
z
Because z is the orthogonal projection of s onto the vertical (Fig. 3.12),

it is
s ¼ z=l
and
ds ¼ dz=l:
Thus, sk and skz are simply related by

Fig. 3.12 A point P along

the line of sight s inclined by
h with respect to the vertical
z axis of a plane parallel
photosphere has coordinate
s along s and z along the
vertical. Because z is the
orthogonal projection of
s onto the vertical, then
s ¼ z=cos h ¼ z=l:
Z1 Z1
sk ¼ kk ds ¼ kk dz=l ¼ skz =l
s z
and Ik(sk) is a function of both skz and l
Ik ðsk Þ ¼ Ik ðskz; ; lÞ:
Under the assumption of LTE, the Eddington-Barbier approximation in terms of

the vertical optical depth is
Ik ðskz; ; lÞ ¼ Ik ðsk Þ ¼ Bk ðsk þ 1Þ ¼ Bk ½ðsk þ 1Þl ¼ Bk ðskz þ lÞ:
The intensity along a line of sight inclined by an angle h = arcos(l) with respect
to the vertical and at any vertical optical depth skz, is approximated by the Planck
function computed at skz + l. Thus, while the Planck function is isotropic, the
radiation intensity depends on the inclination of the line of sight through the term l.
Hereafter we will deal only with the vertical optical depth, hence, we denote it
simply as sk, omitting the subscript z.
In particular, at infinite height, corresponding to the observer’s location, sk = 0
and the intensity emerging from the photosphere is
Ik ð0; ; lÞ ¼ Bk ðsk ¼ lÞ:
where sk denotes the vertical optical depth. The emergent intensity along a line of
sight inclined by h is approximated by the Planck function computed at the vertical
optical depth equal to l = cos h. In particular, the intensity at the disk center, where
l = 1, differs from the intensity at the limb, where l tends to 0.
Exercice 3.14 Find the assumptions under which the formal solution of the transfer
equation allow to write the radiation intensity Ik(0,l) emerging from the photo-
sphere as
Z1
Ik ð0; lÞ ¼ Sk ðsk Þ e sk =l d sk =l;
0
where sk denotes the vertical optical depth.

We are now in the position to determine the temperature variation from the limb
darkening. In fact, LTE and the Eddington-Barbier approximation allow us to
describe the visible continuum as
Bc ðsc ¼ lÞ ¼ Ic ð0; lÞ
and therefore, using the observed linear limb darkening,
Bc ðsc ¼ lÞ ¼ Ic ð0; 1Þ ð2 þ 3lÞ=5:
Thus,
Bc ðsc ¼ 1Þ ¼ Ic ð0; 1Þ
and
Bc ðsc Þ ¼ Bc ðsc ¼ 1Þð2 þ 3sc Þ=5:
This equation is an implicit relation between temperature and optical depth.

Since the Planck function integrated over all wavelengths is explicitly related to
temperature by the Stefan-Boltzmann law
Z1
B¼ Bk dk ¼ rT 4 =p;
0
we integrate the relation over wavelength. In order to do that, we assume that the
opacity and, hence, optical thickness, do not depend on wavelength, This is known
as the grey approximation and, together with LTE, implies that the source function
and, hence, radiation intensity, have the same wavelength dependence as that of a
black body, Such a behavior is actually close to the observed behavior of intensity
in the visible and infrared continua. Thus, we get
BðsÞ ¼ Bðs ¼ 1Þ ð2 þ 3 sÞ=5;
where s denotes the vertical optical depth independent of wavelength. In practice, s

is computed from the opacity at k = 0.5 lm, or from a suitable average opacity
which is called the Rosseland opacity after the Norwegian astrophysicist Svein
Rosseland (1894–1985).
Worked Exercise 3.15 The Rosseland opacity.

The Rosseland mean opacity kR comes from the average of 1/kk with the tem-
perature derivative of the Plank function as weighting function, i.e.
Z1 Z1
1 1 dBk dBk
¼ dk dk:
kR kk dT dT
0 0
Note that this kind of mean, which is called harmonic, weights the lowest values
of opacity more than the highest values. Moreover, because the temperature
derivative does not depend on the wavelength integral, it is
0 1 1
Z1 Z
dBk
dk ¼ d @ Bk dkA=dT ¼ dðrT 4 =pÞ=dT ¼ 4rT 3 =p:
dT
0 0
The Rosseland mean for kk is introduced into the relation between the temper-
ature gradient and total radiation flux discussed in Chap. 4.
Exercise end.
By virtue of the Stefan-Boltzmann law, we get
rT 4 ðsÞ=p ¼ ½rT 4 ð1Þ=p ð2 þ 3 sÞ=5;
and
T 4 ðsÞ ¼ T 4 ð1Þ ð2 þ 3 sÞ=5:
T(1), which is the temperature at unit optical depth, is related to the effective
temperature by (Exercise 3.16)
T 4 ð1Þ ¼ 5Te4 =4
Exercise 3.16 Verify that in the grey photosphere the temperature at unit optical
depth is about 330° hotter than the effective temperature.
Inserting the effective temperature in the trend of the temperature with optical
depth we obtain
T 4 ðsÞ ¼ T 4 ð1Þ ð2 þ 3sÞ=5 ¼ Te4 ð5=4Þ ð2 þ 3sÞ=5 ¼ Te4 ð2 þ 3sÞ=4

¼ Te4 ð1=2 þ 3s=4Þ;
and
TðsÞ ¼ Te ð1=2 þ 3s=4Þ1=4 :

Fig. 3.13 Temperature as a function of the decimal logarithm of optical depth. Black solid line
refers to the approximated grey photosphere discussed in the text. Vertical thin solid lines define
the interval where the approximation is valid. Red solid line refers to the VALC accurate model,
which is in scale of the optical depth at k = 0.5 lm
The temperature structure that we have obtained corresponds to the Eddington

approximation to the grey photosphere. Here, the adjective grey emphasizes the
assumption that opacity is independent of wavelength. Our grey photosphere is
shown in Fig. 3.13.
Because of the limited resolution in l due to instrumental and possible atmo-
spheric effects, the limb darkening can be obtained with sufficient accuracy only for
l > 0.1, thus, our T(s) is extrapolated outside the optical depth interval
−1.0 < log s < 0. For comparison, the figure shows also the photospheric tempera-
ture variation from the accurate VALC model, which is discussed in more detail later.
The grey approximation is rather close to the VALC for −1.5 < log s < 0.5. Outside
this range the accurate model departs from the grey temperature gradient. Actually, in
these layers, new physical phenomena that we have neglected play a major role.
Small optical depths, log s < −1.5 correspond to outer layers where low densities
make the LTE inadequate. Moreover, in the inner layers, for log s > 0.5, (s > 3.16),
the energy transfer by radiation becomes secondary and the temperature structure is
primarily determined by convection, as we will see in Chap. 4.
3.2.10 Temperature as Function of Height
Limb darkening inversion gives the temperature versus optical depth. The equation
of hydrostatic equilibrium allows to obtain the temperature as function of height.
Optical depth and height scales are connected through opacity
ds ¼ kdz
Visible opacity is largely produced by the H− photoionization. In Sect. 3.1 we

found that this opacity is given by

1=2 3=2
k ¼ u þ q3=2 p rH2 = mH u ;
where p, rH and mH are constant, q is the gas density, and the Saha functions u+ and
u− depend only on T. Actually, the continuum opacity at visible wavelengths can
be approximated by a power law in both q and T such as
k ¼ kqa T b ;
with k in cm−1, k ¼ 1:3 1021 , a = 1.74 and b = 6.76 (Fig. 3.14).

Thus
ds ¼ k dz ¼ kqa T b dz
and
Z1 Z1
sðzÞ ¼ k dz ¼ k qa T b dz;
z z
where z denotes the integration height.
Fig. 3.14 Photospheric opacity versus optical depth compared with the approximate power law
given in the text (black curve). The accurate values (red curve) are from the VALC model
We assume T = constant and keep the exponential solution of the hydrostatic

equilibrium
qðzÞ ¼ qðz0 Þeðzz0 Þ=H :
Then, the opacity is
Z1 Z1 a
0
0
sðzÞ ¼ k q T dz ¼ k
a b
qðz0 Þeðz z0 Þ=H Tðz0 Þb dz0
z z
Z1 Z1
0 0
¼ kqa0 T0b eaðz z0 Þ=H dz0 ¼ k0 eaðz z0 Þ=H dz0 ;
z z
where z0 > z and
k0 ¼ k qa0 T0b ¼ k qðz0 Þa Tðz0 Þb
is the opacity at z0. Performing the integration gives
Z1
eaðz z0 Þ=H d½aðz0 z0 Þ=H ¼ ðk0 H=aÞ½eaðzz0 Þ=HÞ 1
0
sðzÞ ¼ ðk0 H=aÞ z
z
¼ ðk0 H=aÞeaðzz0 Þ=H :
At s(z) = 1 we have
1 ¼ ðk0 H=aÞeaðzz0 Þ=H :
Assuming z = z0 = 0 at s(z) = 1, we then get
sðzÞ ¼ ðk0 H=aÞeaz=H
and
sð0Þ ¼ 1 ¼ k0 H=a:
Note that this relation, which is a consequence of the power law scaling of
opacity with density, confirms our previous assumption that photon mean free path
and scale height are of the same order of magnitude.
Thus,
sðzÞ ¼ eaz=H ;
and
z ¼ ðH=aÞln ðsÞ:
The optical depth is usually given in scale of the decimal logarithm, log, then
z ¼ ðH=aÞ 2:30 log ðsÞ;
where ln (10) = 2.3025851.

Exercise 3.17 Find the relation between natural and decimal logarithms.
Exercise 3.18 Under the same approximations used in text, find the pressure as
function of the average optical depth.
3.2.11 Complete Model Photosphere
In LTE the particle energy distributions are those of thermodynamic equilibrium at

the local temperature, thus two thermodynamic parameters are sufficient to deter-
mine all particle physical parameters (Exercises 3.19 and 3.20).
Exercise 3.19 Under the same approximations used in text, compute an approxi-
mate density qa(z) and compare it with the VALC density (use Table 3.1 between
z = −75 and 515 km].
Exercise 3.20 Verify that under LTE two thermodynamic parameters are sufficient
to determine the number density of the hydrogen atoms in the first excited state. For
sake of simplicity, assume a pure hydrogen gas.
This implies that the temperature and pressure variations with height represent
already a complete model photosphere. T(z) and P(z) that we obtained in this
Section are plotted in Fig. 3.15.
The height interval where the approximate temperature is valid has been
determined with the function z(s) given in the text. The approximate pressure is
closer to the pressure of the accurate model in the deeper layers than in the higher
layers, where the scale height of the accurate model is smaller.
Worked Exercise 3.21 The Eddington-Barbier approximation for the emergent
flux.
Since photons entering the photosphere from outside are negligible, we restrict
the flux integral to the outward directions, i.e.
Z1 Z1
Fk ð0Þ ¼ 2p Ik ð0; lÞl dl ¼ 2p Ik ð0; lÞl dl
1 0
We exploit LTE and the Eddington-Barbier approximation for intensity in order

to express the emergent flux as
Table 3.1 The VALC model
110
z, km s0.5 T, K NH, cm−3 Ne, cm−3 P, Ba q, g cm−3

2543 0.000E+000 447,000 1.005E+009 1.205E+009 1.425E−001 2.349E−015
2298 3.712E−008 141,000 3.205E+009 3.839E+009 1.434E−001 7.494E−015
2290 3.969E−008 89,100 5.041E+009 5.961E+009 1.416E−001 1.179E−014
2280 4.491E−008 50,000 9.038E+009 9.993E+009 1.376E−001 2.113E−014
2274 4.952E−008 37,000 1.201E+010 1.318E+010 1.348E−001 2.808E−014
2271 5.234E−008 32,000 1.378E+010 1.498E+010 1.331E−001 3.222E−014
2267 5.657E−008 28,000 1.567E+010 1.677E+010 1.315E−001 3.665E−014
2263 6.124E−008 25,500 1.718E+010 1.812E+010 1.303E−001 4.017E−014
2255 7.110E−008 24,500 1.797E+010 1.881E+010 1.305E−001 4.203E−014
2230 1.030E−007 24,200 1.862E+010 1.943E+010 1.334E−001 4.355E−014
2200 1.426E−007 24,000 1.932E+010 2.009E+010 1.354E−001 4.517E−014
2160 1.977E−007 23,500 2.051E+010 2.120E+010 1.419E−001 4.795E−014
2129 2.427E−007 23,000 2.163E+010 2.219E+010 1.460E−001 5.058E−014
2120 2.562E−007 22,500 2.231E+010 2.276E+010 1.470E−001 5.216E−014
2115 2.640E−007 21,000 2.403E+010 2.402E+010 1.462E−001 5.619E−014
2113 2.674E−007 18,500 2.732E+010 2.620E+010 1.436E−001 6.390E−014
2109 2.754E−007 12,300 4.092E+010 3.306E+010 1.326E−001 9.569E−014
2107 2.801E−007 10,700 4.673E+010 3.535E+010 1.282E−001 1.093E−013
2104 2.877E−007 9500 5.239E+010 3.705E+010 1.241E−001 1.225E−013
2090 3.243E−007 8440 6.127E+010 3.799E+010 1.228E−001 1.433E−013
2080 3.507E−007 8180 6.541E+010 3.780E+010 1.239E−001 1.530E−013
2070 3.770E−007 7940 6.960E+010 3.783E+010 1.254E−001 1.628E−013
2050 4.299E−007 7660 7.705E+010 3.792E+010 1.297E−001 1.802E−013
2016 5.203E−007 7360 9.075E+010 3.811E+010 1.401E−001 2.122E−013
(continued)
3 The Photosphere
1990 5.903E−007 7160 1.033E+011 3.858E+010 1.505E−001 2.417E−013
1925 7.717E−007 6940 1.380E+011 4.028E+010 1.840E−001 3.227E−013
1785 1.212E−006 6630 2.601E+011 4.771E+010 3.055E−001 6.082E−013
1605 1.958E−006 6440 6.386E+011 6.005E+010 6.778E−001 1.493E−012
1515 2.420E−006 6370 1.048E+012 6.456E+010 1.070E+000 2.450E−012
1380 3.286E−006 6280 2.273E+012 7.600E+010 2.233E+000 5.315E−012
1280 4.084E−006 6220 4.200E+012 7.486E+010 4.031E+000 9.822E−012
3.2 A Model of the Photosphere
1180 5.075E−006 6150 7.865E+012 8.108E+010 7.413E+000 1.839E−011

1065 6.861E−006 6040 1.711E+013 9.349E+010 1.577E+001 4.000E−011
980 9.148E−006 5925 3.147E+013 1.041E+011 2.840E+001 7.359E−011
905 1.239E−005 5755 5.546E+013 1.049E+011 4.855E+001 1.297E−010
855 1.553E−005 5650 8.135E+013 1.064E+011 6.987E+001 1.902E−010
755 2.537E−005 5280 1.864E+014 8.838E+010 1.495E+002 4.358E−010
705. 3.288E−005 5030 2.935E+014 7.664E+010 2.242E+002 6.864E−010
655 4.452E−005 4730 4.794E+014 8.085E+010 3.443E+002 1.121E−009
605 7.022E−005 4420 8.119E+014 1.112E+011 5.450E+002 1.899E−009
555 1.456E−004 4230 1.382E+015 1.733E+011 8.879E+002 3.232E−009
515 3.014E−004 4170 2.096E+015 2.495E+011 1.327E+003 4.902E−009
450 1.017E−003 4220 3.989E+015 4.516E+011 2.556E+003 9.327E−009
350 5.626E−003 4465 9.979E+015 1.110E+012 6.767E+003 2.334E−008
250 2.670E−002 4780 2.315E+016 2.674E+012 1.680E+004 5.413E−008
150 1.117E−001 5180 4.917E+016 6.476E+012 3.869E+004 1.150E−007
100 2.201E−001 5455 6.866E+016 1.066E+013 5.689E+004 1.606E−007
50 4.395E−001 5840 9.203E+016 2.122E+013 8.065E+004 2.152E−007
(continued)
111
112

0. 9.953E−001 6420. 1.166E+017 6.433E+013 1.137E+005 2.727E−007
−25 1.683E+000 6910 1.261E+017 1.547E+014 1.325E+005 2.949E−007
−50 3.338E+000 7610 1.317E+017 4.645E+014 1.527E+005 3.080E−007
−75 7.445E+000 8320 1.365E+017 1.204E+015 1.740E+005 3.192E−007
In the order the columns report the height z, optical depth at 0.5 lm s0.5, temperature T, H number density NH, electron number density Ne, gas pressure P, and
density q
3 The Photosphere
Fig. 3.15 Temperature (upper panel) and pressure (lower panel) as function of height in the
photosphere. Red and black curves refer to the VALC and the approximate model constructed in
the text, respectively. The level z = 0 corresponds to s = 1 in the approximate model and s0.5 = 1
in the VALC. Vertical thin solid lines define the interval where the approximate temperature is
valid
Z1 Z1
Fk ð0Þ ¼ 2p Ik ð0; lÞl dl ¼ 2p Bk ðsk ¼ lÞl dl:
0 0
At each wavelength the emergent flux is the integral over l of the Planck
function at sk = l, multiplied by l. The range of l corresponds to the variation
from disk center to limb. To evaluate this integral we use the grey approximation
and the limb—darkening that is linear with l and independent of wavelength. Thus,
the emergent flux integrated over all wavelengths is
Z1 Z1
Fð0Þ ¼ 2p Bðs ¼ lÞl dl ¼ 2p ½Bðs ¼ 1Þð2 þ 3lÞ=5l dl
0 0
Taking all quantities independent of l out of the integral and replacing the
integral of a sum with the sum of integrals, we obtain
Z1 Z1
Fð0Þ ¼ 2pBðs ¼ 1Þð2=5Þ l dl þ 2p Bðs ¼ 1Þð3=5Þ l2 dl
0 0
Since it is
Z1 Z1
l dl ¼ ½l 2
=210 ¼ 1=2 and l2 dl ¼ ½l3 =310 ¼ 1=3;
0 0
we get

2 1 3 1
Fð0Þ ¼ 2p Bðs ¼ 1Þ þ ¼ p Bðs ¼ 1Þð4=5Þ:
5 2 5 3
Finally, by means of the relation between the values of the Planck function at
optical depths 1 and 2/3, we find
Fð0Þ ¼ p Bðs ¼ 1Þð4=5Þ ¼ pð5=4ÞBðs ¼ 2=3Þð4=5Þ ¼ p Bðs ¼ 2=3Þ:
Since
Z1
B¼ Bk dk ¼ rT 4 =p;
0
we get
Fð0Þ ¼ pBðs ¼ 2=3Þ ¼ p rT 4 ðs ¼ 2=3Þ=p ¼ rT 4 ðs ¼ 2=3Þ ¼ rTe4 :
Therefore, s = 2/3 is the vertical optical depth from where the total flux emerges.
Exercise end.
In summary, based on the grey, LTE, and Eddington-Barbier approximations,
we have constructed an approximate photospheric model that has been compared to
an accurate model. Let us now briefly discuss how accurate models are constructed.
3.2.12 Accurate Models of Photosphere
Two types of models exist: semi-empirical and theoretical. Our approximate model
is semi-empirical because T(z) has been inferred from observation, i.e. the linear
limb darkening in the visible continuum. Accurate semi-empirical models deter-
mine T(sk) from a set of observations and, in order to cover as much of the
photosphere as possible, they exploit the possibility to diagnose layers at different
depths either by changing position on the disk or by selecting different wavelengths.
In fact, using the Eddington-Barbier approximation for the sake of simplicity, we
write
Ik ð0; lÞ ¼ Sk ðsk ¼ lÞ
and the observed average depth, d, is such that
sk ¼ l kk d;
where the symbol means roughly equal to; hence
d l=kk :
Let us keep l fixed and consider different wavelengths. At those wavelengths

where the opacity and, hence, optical thickness is smaller, the photosphere is more
transparent and the emergent radiation corresponds to the source function at deeper
levels, where the temperatures, and hence the Planck function, are higher.
Consequently, if LTE is assumed, the emergent intensity is greater at these
wavelengths. Vice versa, at the wavelengths where the opacity is greater, the
emergent radiation corresponds to a shallower source function and, if LTE holds, to
a smaller emergent intensity. As the emergent spectrum suggests, opacity variations
with wavelength occur smoothly in the continuum and sharply in the absorption
lines. The continuum absorption coefficient k(k) at s = 1 due to the H− bound-free
and free-free processes is shown in Fig. 3.16.
It has a characteristic wave like shape with a bump around 0.9 lm, close to the
red extreme of the visible spectrum, and a trough at 1.6 lm in the infrared. As we
already know, the bump is due to the H− bound-free absorption that has a threshold
Fig. 3.16 Opacity at s = 1 due to the H− bound-free and free-free as function of wavelength
at k = 1.64 lm corresponding to the H− ionization energy. The increase at

k > 1.64 lm is due to radiation absorption by means of free electrons in the
presence of neutral hydrogen atoms (H− free-free). The maximum excursion in the
continuum opacity in the figure is about a factor 3. A much larger opacity excursion
can be achieved in the absorption lines, where the opacity due to a bound-bound
transition adds to the continuum opacity. In fact, many so-called weak lines are just
slight perturbations of continuum, whereas a number of strong lines can be a factor
102 more opaque and consequently stop most of outward radiation flux. Typical line
widths range from 10−3 nm for weak lines to 1 nm for the strongest lines.
Worked Exercise 3.22 Line broadening.
The line broadening is described by the line profile w(k) introduced in Chap. 1
when discussing the basic emission and absorption processes. Returning to the two
level atom of Fig. 1.14, if each level has one well-defined energy value, the line
profile may be represented by a function which is infinitely thin and peaked at the
line center wavelength k0 = hc/DE, where DE is the level energy difference. This
function is called a d function (d is a Greek letter, name delta) and is such that for
any function f(k) it is
Z1 Z1
f ðkÞdðk k0 Þ dk ¼ f ðk0 Þ dðk k0 Þ dk ¼ f ðk0 Þ:
1 1
There are three physical processes that make an energy level span a non-zero
energy range. They are: radiative or natural broadening, collisional or pressure
broadening and thermal or Doppler broadening.
Natural broadening is due to the finite lifetime that an electron has when is
sitting in an atomic excited level. Because of the Heisenberg uncertainty principle, a
time uncertainty corresponds to an energy uncertainty (see Chap. 6). As a conse-
quence, a bound electron can usually access not only one, but a range of energies
with a probability which is given as function of wavelength by a Lorentzian line
profile

wðkÞ ¼ c= pðDk2 þ c2 Þ ;
where the parameter c is inversely proportional to the level lifetime and Dk =

− k0 is the wavelength distance from the line center k0.
Pressure broadening is due to weak, but continuous, interactions of the radiating
atom with the surrounding gas particles, which shift up and down the energy levels.
The resulting line profile may be approximated by a Lorentzian function where the
width c is the frequency at which perturbing collisions occur.
The wavelength of the radiation that an atom emits or absorbs depends on the
atom’s velocity according to the Doppler formula (Christian Doppler was an
Austrian physicist, 1803–1853). The wavelength shift of the rest wavelength k due
to the atom’s velocity va is equal to vak/c, where c is the light velocity. Because the
one-dimensional velocity distribution along a line-of-sight is a Gaussian function,
the line profile for Doppler broadening is
pffiffiffi 2
wðkÞ ¼ ½1=ð pDkD ÞeðDk=DkD Þ :
The so-called Doppler width DkD is the most probable Doppler shift, i.e.
DkD ¼ v0 k0 =c;
where
v0 ¼ ð2 kB T=mÞ1=2
is the most probable thermal velocity.

All the three broadening mechanisms usually occur together, thus the general
line profile is the convolution of a Gaussian of width DkD and a Lorentzian function
of width c, and has the form
Z1
pffiffiffi pffiffiffi
ey =½a2 ðv yÞ2 dy;
2
wðkÞ ¼ ð p=DkD ÞHða; vÞ ¼ ½a=ð p DkD Þ
1
with v = (k − k0)/DkD and a = c/DkD. This general line profile reduces to a

Gaussian function for a 1 while has a Gaussian core with large Lorentzian wings
for a 1, as Fig. 3.17 shows for the Ha line.
Fig. 3.17 Ha line absorption coefficient at s = 0.22, where T = 5455 K and NH = 6.87 1016
cm−3 in the VALC model. The plotted absorption coefficient has the form kðkÞ ¼ N2 r0 Hk: N2 is
the number of hydrogen atoms in the first excited level. r0 is the cross section at line center and has
pffiffiffi
the form r0 ¼ p k20 re f =DkD: where k0 = 0.6566 lm is the wavelength of the Ha line center,
re ¼ e=me c ¼ 2:82 103 pm is the classical radius of an electron. f is the oscillator strength. f is
2
a quantistic correction to the classical cross section and represents the number of classical
oscillators that are required to give the right absorption. For Ha,f = 0.6408. DkD = v0k0/c is the line
Doppler width, with the thermal velocity v0 given by v0 = (2kBT/m)1/2. Finally, Hk is the line
profile normalized to be 1 at line center. It is Hk ¼ Hða; vÞ ¼ ða=pÞR 1 1 e
y2
=½a2 ðv yÞ2 dy;
with v = (k − k0)/DkD. The function H is the convolution of a Gaussian of width DkD and a
Lorentzian function of width c. For this figure the parameter a = c/DkD was fixed to a = 10−1 and
the Doppler width to DkD = 20 pm
Exercise end.
Figure 3.17 shows as an example the Ha line absorption coefficient versus
wavelength at s = 0.22.
The VALC model is an example of an accurate semi-empirical model. Its
temperature structure was obtained by the best fit to observations in continua and
spectral lines, which were performed without the simplifying assumption of LTE.
This approach has allowed us to estimate the solar temperature in the photosphere
and beyond into the chromosphere and transition region, as Fig. 3.18 and Table 3.1
show. The VALC temperature begins to increase with height for z > 500 km.
The level z = 500 km is called the temperature minimum and marks the top of
the photosphere and the base of the chromosphere.
It is possible to construct fully theoretical models of the photosphere where the
temperature structure is inferred from equations. One example of a theoretical
model is the solar model in the stellar model grid called MARCS, whose T(z) is also
plotted in Fig. 3.18 (Gustafsson et al. 2008). MARCS models assume that radiation
Fig. 3.18 Temperature versus height for the VALC (black curve) and C7 (dashed line)
semiempirical models, and the MARCS solar theoretical model (red curve)
and convection are the only ways to transfer energy, which is different to the VALC
model. Consequently, the MARCS solar model has no chromospheric temperature
rise. We have also plotted the temperature variation of the model C7, which is an
improved version of the earlier model VALC (Avrett and Loeser 2008). The dif-
ferent temperatures give an idea of the uncertainties that exist also between accurate
models. Table 3.1 reports a number of physical parameters from the VALC model.
A further aspect in which accurate models are by far superior to our approximate
analytical approach is represented by the treatment of the equation of state and
opacities. We limited our considerations to hydrogen and helium to evaluate the
mean molecular weight, and H− photoionizations to estimate the opacity. Actually,
many transitions of different atoms and ions have to be treated in details in order to
get accurate opacities in the continua and spectral lines. Moreover, the metals, even
though relatively few in number, affect in a significant manner the electron pressure.
Thus, a basic ingredient of any accurate model is the chemical composition, that is,
the amount of the various elements forming the gas, which we describe in the
following.
3.2.13 The Chemical Composition
The black body spectrum does not depend on the material the body is made of but
depends only on its temperature. On the contrary, the radiation emitted by the Sun
and stars carry in the spectral lines the signatures of the chemical elements forming
the gas. The absorption and emission spectra of each chemical element are char-
acterized by the same and unique sequence of lines, because they correspond to the
allowed transitions between its energy levels. Therefore, the identification in a
stellar spectrum of a number of those lines, in the same wavelength sequence as
observed in laboratory, ensures that particular element is present on the star. As a
matter of example, the positions in the visible spectrum of some absorption lines of
the Balmer series, which correspond to all the possible light absorptions that start
from the first excited level of hydrogen, are plotted In Fig. 3.19.
The Balmer line wavelengths, kn, are related the H level energies by
hc=kn ¼ EH =22 EH =n2 ;
Fig. 3.19 The positions of

the hydrogen lines of the
Balmer series in the visible
spectrum. Note that the line
intensities in figure are
arbitrary and do not
correspond to the observed
absorptions. The Greek letters
used as indexes in the
symbols of the Balmer lines
are a, b, c, d and e (names:
alpha, beta, gamma, delta and
epsilon)
where EH is the ionization potential of neutral hydrogen and n is the principal

quantum number which takes the values 3, 4, 5 and so on.
Exercise 3.23 Show that according to the Bohr’s model the first five lines of the
Balmer series, denoted with the symbols Ha, Hb, Hc, Hd and He, occur at the
wavelengths 0.6565, 0.4863, 0.4342, 0.4103 and 0.3971 lm, respectively.
The absorption line strength depends on the amount of the particular element
producing the line. In a plot of intensity as function of wavelength, the line strength
is the area that the line takes away from the continuum, which is called the line
equivalent width (Fig. 3.20).
The equivalent width W is the integral over k of the fraction of continuum
intensity removed by the line, that is
Z
W ¼ ½ðIc Ik Þ=Ic dk;
where Ic indicates the continuum level and the integral extends over the whole
range of wavelengths spanned by the line. The line equivalent width depends on the
number density of atoms Na capable of absorbing radiation in the line. For instance,
in the case of the Balmer series, it depends on the number of hydrogen atoms in the
first excited level. The relation W(Na) has different forms according to the line
strengths. In particular, for weak lines W is directly proportional to Na, as shown in
Exercise 3.24.
Worked Exercise 3.24 Equivalent widths of weak lines.
The relation between line equivalent widths of an element with its abundance is
easily understood in the case of weak lines. If we sketch the line with a rectangular
Fig. 3.20 The solid line

sketches an absorption line
(compare with Fig. 3.7). The
horizontal dashed line
indicates the level of the
continuum Ic. The equivalent
width is defined as the
integral of Ic − Ik (in green)
divided by Ic over the
wavelength range spanned by
the line
hole having intensity Il, which is less than the continuum Ic, and wavelength width
Dkl, the equivalent width W is simply the product
W ¼ Dkl ðIc Il Þ=Ic :
Using the Eddington-Barbier approximation and LTE to express the emergent

intensity and denoting with sc the continuum optical depth and with sl the addi-
tional optical thickness due to the line opacity, we get
W ¼ Dkl ½Bðsc ¼ 1Þ Bðsc ¼ 1 sl Þ=Bðsc ¼ 1Þ;
where we have neglected the smooth variation of the Planck function over the line
width. For a weak line sl 1, hence we can approximate the difference between
the values of the Planck function at two depths by means of the differential, that is,
dB
Bðsc ¼ 1Þ Bðsc ¼ 1 sl Þ ¼ sl :
ds
Moreover, it is
sl ¼ Na rl ;
where Na is the number of atoms capable of absorbing and rl is the transition cross
section. Inserting the last two results in W, we find
dB dB
W ¼ Dkl ½Bðsc ¼ 1Þ Bðsc ¼ 1 sl Þ=Bðsc ¼ 1Þ ¼ Dkl B sl ¼ Dkl B Na rl :
ds ds
Exercise end.
The number of absorbers is in turn related to the abundance of the element
producing the line through the excitation and ionization equilibria. Thus, the
equivalent width depends on the physical structure of the photosphere in addition to
the element abundance and the transition cross section, even under the LTE
assumption which allows the use of the Boltzmann and Saha equations. Therefore,
determining the element abundance requires the analysis of several lines at the same
time, and, especially for strong lines, we are not able to use the LTE assumption,
which makes the problem much more complicated. Despite these difficulties, the
analysis of the solar spectrum allows us to state with confidence that the plasma
forming the solar photosphere consists of a mixture of chemical elements among
which hydrogen is by far the most abundant with more than 90% of the total
number of atoms. Helium represents just under 10% of the total number of atoms,
while the abundance of all the other elements together, the metals, is of the order of
0.1%.
Table 3.2 Relative element abundances in the solar photosphere and in meteorites (CI
chondrites). All values, which are in unit of the total number of hydrogen atoms per cubic cm
in the photosphere, have been multiplied by the factor 1012 and transformed in decimal logarithms;
thus, the photospheric abundance of hydrogen is 12. Values in squared brackets are indirect solar
estimates (Asplund et al. 2005)
Element Photosphere Meteorites Element Photosphere Meteorites
H 12.00 8.25 ± 0.05 Ru 1.84 ± 0.07 1.77 ± 0.08
He [10.93 ± 0.01] 1.29 Rh 1.12 ± 0.12 1.07 ± 0.02
Li 1.05 ± 0.10 3.25 ± 0.06 Pd 1.69 ± 0.04 1.67 ± 0.02
Be 1.38 ± 0.09 1.38 ± 0.08 Ag 0.94 ± 0.24 1.20 ± 0.06
B 2.70 ± 0.20 2.75 ± 0.04 Cd 1.77 ± 0.11 1.71 ± 0.03
C 8.39 ± 0.05 7.40 ± 0.06 In 1.60 ± 0.20 0.80 ± 0.03
N 7.78 ± 0.06 6.25 ± 0.07 Sn 2.00 ± 0.30 2.08 ± 0.04
O 8.66 ± 0.05 8.39 ± 0.02 Sb 1.00 ± 0.30 1.03 ± 0.07
F 4.56 ± 0.30 4.43 ± 0.06 Te 2.19 ± 0.04
Ne [7.84 ± 0.06] −1.06 I 1.51 ± 0.12
Na 6.17 ± 0.04 6.27 ± 0.03 Xe [2.27 ± 0.02] −1.97
Mg 7.33 ± 0.09 7.53 ± 0.03 Cs 1.07 ± 0.03
Al 6.37 ± 0.06 6.43 ± 0.02 Ba 2.17 ± 0.07 2.16 ± 0.03
Si 7.51 ± 0.04 7.51 ± 0.02 La 1.13 ± 0.05 1.15 ± 0.06
P 5.36 ± 0.04 5.40 ± 0.04 Ce 1.58 ± 0.09 1.58 ± 0.02
S 7.14 ± 0.05 7.16 ± 0.04 Pr 0.71 ± 0.08 0.75 ± 0.03
Cl 5.50 ± 0.30 5.23 ± 0.06 Nd 1.45 ± 0.05 1.43 ± 0.03
Ar [6.18 ± 0.08] −0.45 Sm 1.01 ± 0.06 0.92 ± 0.04
K 5.08 ± 0.07 5.06 ± 0.05 Eu 0.52 ± 0.06 0.49 ± 0.04
Ca 6.31 ± 0.04 6.29 ± 0.03 Gd 1.12 ± 0.04 1.03 ± 0.02
Sc 3.05 ± 0.08 3.04 ± 0.04 Tb 0.28 ± 0.30 0.28 ± 0.03
Ti 4.90 ± 0.06 4.89 ± 0.03 Dy 1.14 ± 0.08 1.10 ± 0.04
V 4.00 ± 0.02 3.97 ± 0.03 Ho 0.51 ± 0.10 0.46 ± 0.02
Cr 5.64 ± 0.10 5.63 ± 0.05 Er 0.93 ± 0.06 0.92 ± 0.03
Mn 5.39 ± 0.03 5.47 ± 0.03 Tm 0.00 ± 0.15 0.08 ± 0.06
Fe 7.45 ± 0.05 7.45 ± 0.03 Yb 1.08 ± 0.15 0.91 ± 0.03
Co 4.92 ± 0.08 4.86 ± 0.03 Lu 0.06 ± 0.10 0.06 ± 0.06
Ni 6.23 ± 0.04 6.19 ± 0.03 Hf 0.88 ± 0.08 0.74 ± 0.04
Cu 4.21 ± 0.04 4.23 ± 0.06 Ta −0.17 ± 0.03
Zn 4.60 ± 0.03 4.61 ± 0.04 W 1.11 ± 0.15 0.62 ± 0.03
Ga 2.88 ± 0.10 3.07 ± 0.06 Re 0.23 ± 0.04
Ge 3.38 ± 0.05 3.39 ± 0.05 Os 1.45 ± 0.10 1.34 ± 0.03
As 2.29 ± 0.05 Ir 1.38 ± 0.05 1.32 ± 0.03
Se 3.33 ± 0.04 Pt 1.64 ± 0.03
Br 2.56 ± 0.09 Au 1.01 ± 0.15 0.80 ± 0.06
Kr [3.28 ± 0.08] −2.27 Hg 1.13 ± 0.18
Rb 2.60 ± 0.15 2.33 ± 0.06 Tl 0.90 ± 0.20 0.78 ± 0.04
(continued)

Element Photosphere Meteorites Element Photosphere Meteorites
Sr 2.92 ± 0.05 2.88 ± 0.04 Pb 2.00 ± 0.06 2.02 ± 0.04
Y 2.21 ± 0.02 2.17 ± 0.04 Bi 0.65 ± 0.03
Zr 2.59 ± 0.04 2.57 ± 0.02 Th 0.06 ± 0.04
Nb 1.42 ± 0.06 1.39 ± 0.03 U <−0.47 −0.52 ± 0.04
Mo 1.92 ± 0.05 1.96 ± 0.04
Accurate values of the element abundances in the solar photosphere are listed in
Table 3.2 together with those measured in the meteorites that have fallen on Earth
keeping the original chemical composition since the time of their formation.
Many solar abundances appear similar to the meteoric ones. This result is not
accidental and is explained by the origin of our solar system from one early cloud of
gas and dust. The differences between solar and meteoric abundances that exist for
some elements are due to the evolution of the Sun during the about 4.5 billion years
of its life, which is the topic of Chap. 7.
3.2.14 Differences Between Photosphere and the Earth’s

Atmosphere
To summarize the characteristics of the solar photosphere and make them familiar,
let us compare the photosphere with the Earth’s troposphere, which is the part of the
Earth’s atmosphere closer to surface. At the ground level the physical conditions are
described by a temperature of 15 °C equal to 273 + 15 = 288 K, a density of
1.23 10−3 g cm−3 and a pressure of 1 atmosphere, which is equal to
1.013250 106 Ba. The number of particles is about 2.5 1019 per cubic cm and
their mean free path is about 10−5 cm. Neglecting weather perturbations, the
hydrostatic equilibrium holds in the troposphere and pressure decreases exponen-
tially with height with a scale height of 8.4 km, a little bit less than the altitude of
Mount Everest, which reaches 8848 m. As the chemical composition is concerned,
the terrestrial atmosphere is a mixture consisting mainly of nitrogen (diatomic
molecules, N2, with atomic weight 28 and percentage of 78%) and oxygen (O2,
atomic weight 32 and percentage 21%), so that the mean molecular weight is about
29. Finally, the acceleration of gravity is 9.8 m s−2.
Exercise 3.25 Verify that 29 is the value of the mean molecular weight for the
Earth atmosphere.
We define the photospheric level at s = 2/3 to correspond to the Earth’s ground
level. There, the temperature Te = 5778 K is a factor 20 greater than the Earth’s
temperature. The chemical composition of the Sun shows significant differences
with that of the Earth. In the photosphere hydrogen (mostly neutral atoms, H, with
atomic weight 1) constitutes a bit more than 90% of the particles, followed by
helium (neutral atoms, He, atomic weight 4) with a bit less than 10%. Then, the
mean molecular weight is 1.27, a factor 23 less than that of the Earth. Density is
2.06 10−7 g cm−3, almost factor 6000 less than it is in the troposphere. Pressure
is 6.07 104 Ba, which is about 1/17 of one atmosphere. The scale height is
137 km, which is a factor 16 greater than that of the Earth; this is mainly due to the
high temperature since the reduced mean molecular weight roughly balances the
greater surface gravity, about 28 times that of the Earth. The number of particles is
about 7.6 1016 per cubic cm, that is a factor 385 less than the terrestrial one,
because the solar gas is less dense than the air. The mean free path for the solar
particles is greater than for the terrestrial particles because in the solar case the
particles per unit volume are in average fewer and with smaller cross sections.
Therefore, in synthesis, the gas at the base of the photosphere is much hotter and
thinner than it is the gas of the Earth’s atmosphere at sea level.
We complete the Sun–Earth comparison noting that the particles are not the only
constituents of both atmospheres since a further basic component is light. For Earth
the primary source of photons is just the Sun so that the light flux is mainly entering
the terrestrial atmosphere. There photons are largely transmitted, even if the solar
light is also partly scattered, which is the process that makes the sky blue, and partly
absorbed. The absorbed fraction is variable. It corresponds to about a 15% of the
total amount when the Sun is at zenith and the air is dry, it increases both as the
solar inclination increases and in conditions of great humidity or dust, and it
strongly depends on the wavelength of the light. Finally, the solar light is absorbed
by the ground, which reemits it at large (infrared) wavelengths that correspond to a
mean temperature of the Earth’s crust of 15 °C. In the case of the photosphere the
prevailing photon flux is outwards and the inward flux from outside space can be
neglected. Because opacity depends on density, the photon mean free path is of the
order of the photospheric scale height.
The last consideration refers to the way all this information has been gathered. In
the case of the Earth we perform the necessary measurements with suitable
instruments directly at the sea level, or, as physicists say, in situ (Latin). In the solar
case we have defined as a reference the mean level from where photons emerge and
derived the values of physical parameters not in situ but at the respectable distance
of one astronomical unit. To do that, we have relied on the observation of the solar
light as well as the theoretical assumptions of the hydrostatic and thermodynamic
equilibria, and about the nature of the solar opacity.
Exercise 3.26 Verify that the number of H atoms capable of absorbing radiation in
the Ha line at s = 0.22 in the VALC model is N2 = 1.04 108, when LTE is
assumed. Use the information given in the caption of Fig. 3.17.
Exercise 3.27 In stellar physics the abundances of hydrogen, helium and metals are
usually given in unit of mass and denoted with X, Y and Z, respectively. On the base
of Table 3.2, which gives solar abundances per particle, and neglecting the less
abundant elements, estimate the solar values of X, Y and Z.
Exercise 3.28 Estimate the pressure scale height of the Earth’s troposphere.
Exercise 3.29 A container is filled with water. What is the weight of the displaced
water when 1 kg of iron, whose density is 7.874 g cm−3, is put in the water? What
happens when the iron is replaced with 1 kg of ice, whose density is
0.9167 g cm−3?
Exercise 3.30 Estimate the mean free path of the photons capable of being
absorbed at the Ha line center (use the information given in Fig. 3.17).
Exercise 3.31 Show that, under the approximations used in the text when dis-
cussing the H− ionization equilibrium, the electron number density scale height is
half of the gas density scale height.
Exercise 3.32 Using the approximation Ne = (u + q/m)1/2 discussed in the text and
the VALC model listed in Table 3.1, estimate the electron number density as
function of height up to z = 500 km. Note that compared with the accurate model
the approximate electron number is slightly higher below z = 0 km, because we
considered only H, hence m = mH instead of 1.27 mu, and is increasingly lower
above z = 0 km, because we neglected the ionization of the metals with low ion-
ization potentials.
Exercise 3.33 Show that the horizontal components of the radiation flux are zero at
any level of a plane parallel homogeneous photosphere.
ExerciseR 3.34 The mass column density is defined as dm = −qdz, thus
1
mðzÞ ¼ z qðzÞ dz, amd it is the mass per unit area. Show that the hydrostatic
equilibrium implies that the pressure is simply P = gSm.
Exercise 3.35 Show that assuming only H and He with number density
NHe = 0.1NH and neglecting ionization, the mass column density is
m = 1.1NHkBT/gS (m is defined in the previous exercise).
Exercise 3.36 Using the ionization potential ECa = 6.11 eV and the partition
functions U1 = 1.32 and U2 = 2.30, verify that under photospheric conditions most
of Calcium atoms are in the first ionization stage, denoted by Ca II.
Exercise 3.37 Using an excitation energy E2 = 3.12 eV for the first stable Ca II
energy level and the statistical weights g1 = 2 and g2 = 2 for the ground and the
excited level, respectively, estimate that under photospheric conditions most of
Ca II ions are in the ground state.
Exercise 3.38 Using the data given in the text, estimate the total gas mass which is
in the photosphere and that in the troposphere.
Chapter 4
The Convection Zone
Abstract What does the structure of the Sun look like below the optical depth
s = 2/3? Photospheric models show that both temperature and density rapidly grow
with depth at the bottom. The opacity, which depends on a positive power of the
temperature and density, increases even more rapidly and the optical depth is much
higher than 1 in a few tens of kilometers. This means that the layers beneath the
photosphere are completely invisible to our telescopes. Fortunately, the physical
conditions in the interior of the Sun can be theoretically determined. To construct a
simplified analytical model of the internal structure of the Sun, we use the same
procedure on which sophisticated numerical models are based. We start with the
known physical parameters of the photosphere, these are the upper boundary con-
ditions, and the appropriate equilibrium conditions, which constrain the variation of
the physical parameters in gradually inner layers, allow us to determine the structure.
In the first section of this chapter we will introduce the unusual conditions of the
solar interior, estimating the central pressure and temperature, and describing the
variation of the opacity, which is a relevant parameter for the structure. The rest of
the chapter describes the interior region that lies just below the photosphere and is
called the convection zone. Convective motions control the energy transfer in a
spherical shell which covers about 1/3 of the solar radius. This is a drastic change
compared with the photosphere, where the radiative energy transfer dominates. The
temperature structure depends on the way the energy flows. Thus, in the second and
third sections we derive the forms that the temperature gradient takes when either
radiation or convection controls the energy flux. In the last section we introduce a
criterion to choose between the convective and radiative energy transport and, then,
construct an approximate model of the convection zone, which is compared to an
accurate model, as we have done for the photosphere. Finally, we estimate the
extension of the convection zone and discuss the validity of the hydrostatic equi-
librium when convective motions occur.

128 4 The Convection Zone
4.1 Glimpses of the Solar Interior
Moving from the photosphere to the Sun’s interior, we stop using the approxi-
mation of plane and parallel layers and use, instead, spherical layers. Due to the
spherical symmetry, all physical properties are uniform on any internal spherical
layer. Thus, the properties that depend on the position vary only with the radius,
that is, the distance from the center. The radiation intensity at a wavelength k, Ik,
which is a function of the position and line of the sight, depends on the radial
distance r and the inclination of the line of sight h with respect to the radius, i.e.
Ik = Ik (r, h). The equilibrium conditions, which determine the increments of
physical properties, apply to layers which are spherical coronas of radius r,
thickness equal to the radius increment, dr, and volume dV = 4pr2 dr (Fig. 4.1).
The layers must be sufficiently thin enough that the increments of solar prop-
erties through them are infinitesimal. In practice, the layers have thicknesses which
are small fractions of the pressure scale height. Scale height increases with the
temperature in the interior, so the layer width ranges from few kilometers at surface
up to several hundred kilometers at center, but still remains thousands of times
smaller than the solar radius.
Before discussing the structure of the layers that lie just below the photosphere,
we must first have an idea of the peculiar conditions of the solar interior, estimating
the pressure and temperature at the solar center.
Fig. 4.1 A section of the gas sphere of center O and radius RS that represents the Sun in an
interior model. All physical properties depending on position vary only with the radial distance
r from O. Thus, they are constant over layers which are spherical coronas having radius r, surface
S = 4 p r2 and width equal to the radius increment dr. The elementary corona volume is
dV = Sdr = 4 p r2 dr. Note that the top extreme of the solar radius RS is the level s = 2/3 in the
photosphere, whose thickness is neglected in this context
4.1 Glimpses of the Solar Interior 129
4.1.1 Pressure
In order to estimate the central pressure we extend the hydrostatic equilibrium to the
interior of our present Sun (a check is given at the end of this chapter) and rewrite
the equation in the form
dP ¼ q g dr:
The gravity g on the spherical surface at radial distance r is given by
g ¼ G M=r 2
and varies with r and the mass M that is contained within the sphere of radius r and
is responsible for the gravitational force on the sphere’s surface. G = 6.67 10−8
dyn cm2 g−2 is the gravitational constant. The equation of the hydrostatic equi-
librium gives us an estimate of the pressure at the Sun’s center. We take qS, RS/2
and MS/2 as constant mean values for density, radius and mass, respectively; thus,
an average gravity is
g ¼ GðMS =2Þ=ðRS =2Þ2 ¼ 2gS :
Then, we integrate between the surface, where the pressure is P0 and the radius
is RS, and the center, where the pressure is Pc and the radius is 0,
ZPc Z0
dP ¼ qS 2gS dr:
P0 RS
Thus,
½PPPc0 ¼ 2qS gS ½r 0RS ;
Pc P0 ¼ 2qS gS RS
and
Pc ¼ P0 þ 2qS gS RS ¼ 6:07 104 Ba þ 2 1:4 g cm3 2:7 104 cm2 s1 7:0 1010 cm
¼ 5:3 1015 Ba:
This value of the central pressure is roughly 100 billion the surface pressure. At
r/RS = 5.10 10−4, or 355 km from center, the accurate model of Senesi et al. has
P = 2.33 1017 Ba, which is a factor 50 larger (Senesi, 2012, private communi-
cation. For a description of the adopted evolutionary code, see: Dell’Omodarme
et al., 2012; Tognelli et al., 2011).
4.1.2 Temperature
In order to estimate the central temperature we use the equation of state
P ¼ RqT=l
and simply assume l = 1, P equal to the central pressure and q to the mean density.
Thus,
Tc ¼ lPc =ðRqS Þ ¼ 5:3 1015 Ba=ð8:31 107 erg K1 g1 1:41 g cm3 Þ
¼ 4:5 107 K:
At r = 355 km, the Senesi et al. model has T = 1.55 107 K, which is about a
factor 3 smaller. This central temperature is almost 3 thousand of the effective
temperature.
The very high values of temperature and pressure at the Sun’s center suggest that
the solar interior is a place very different from the photosphere. Thus, we expect
that new physical processes may play a role in determining the structure. Let us start
considering the absorption processes that dominate the opacity at high
temperatures.
4.1.3 Opacity
The opacity of solar gas is a function of the temperature and density because the
number of radiation absorbers depends on temperature and density through the
Boltzmann and Saha equations, and, moreover, cross sections and absorption
profiles may also depend on the temperature and density. Figure 4.2 shows the
opacity as a function of temperature in the solar interior.
The photospheric opacity is primarily due to H− photoionizations. This kind of
opacity rapidly grows with the increasing temperatures at the bottom of the pho-
tosphere and immediately below, because the partial hydrogen ionization ensures
high numbers of both neutral H atoms and electrons, to which the number of H−
ions is proportional (Sect. 3.1). At larger depths and temperatures, the negative
hydrogen completely ionizes and disappears, but the opacity continues to rise due to
the photoionizations from the excited levels and ground state of neutral H atoms. At
log T = 4.7, the opacity due to H photoionizations has a peak with a maximum
value of about kR = 30 cm−1. At log T = 5, the opacity reaches a maximum slightly
lower than that of H photoionizations, which corresponds to photoionizations from
the excited and ground levels of the neutral helium.
Exercise 4.1 Estimate the temperatures where the H and He opacity peaks occur,
using the ionization potentials EH = 13.598 eV and EHe = 24.587 eV. Compare the
results with Fig. 4.2 and comment on it.
4.1 Glimpses of the Solar Interior 131
Fig. 4.2 The logarithm of opacity as a function of the logarithm of the temperature in the interior
of the Sun. The red curve represents the Rosseland mean opacity of the accurate model (Senesi
et al.). H− bound-free (b-f) and free-free (f-f) absorptions prevail at low temperatures, log T < 4.4.
The two bumps that correspond to H and He photoionizations, occur at about log T = 4.7 and log
T = 5, respectively. At larger temperature, logT > 5.2, b-f absorptions from metals dominate.
Close to the solar center, at log T > 7, in addition to b-f absorptions from metals, f-f transitions and
Thomson scattering contribute about 50% of the Rosseland mean opacity. The black curves
represent the approximate opacity power laws for H−, H and He peaks and metals
For even larger temperatures, log T > 5.3, H and He atoms are almost com-
pletely ionized and, hence, their photoionizations become rare. In these layers the
dominant contribution to opacity comes from the photoionizations of the many
excited levels of the elements heavier than helium, called metals. This type of
opacity increases with increasing density while decreases with increasing temper-
ature, because the number of metals decreases in higher and higher ionization
states. Thus, moving toward the center, metal opacity first decreases and then
increases according to the dominant and opposite effects of T and q.
Finally, close to the center, at log T > 7, free electrons absorb, on average,
roughly as much radiation as metals do by photoionizations. The relevant processes
are free-free transitions, with an electron accelerated and a photon iabsorbed in the
presence of an atom or ion, and the Thomson scattering, with a free electron
absorbing a photon and reemitting it in a different direction (Joseph John Thomson
was an English physicist, 1856–1940).
In summary, the opacity has two wide maxima in the inner layers with tem-
peratures between 2 104 and 2 105 degrees. Outside this region, the opacity
decreases both towards the surface, where its variation is very steep, and towards
the center, where it smoothly reaches a minimum before rising again, because of the
sharp density increase for temperatures higher than 2 106 degrees.
In practice, we approximate the mean opacity with power laws in terms of T and
q, which have different exponents according to the temperature and dominant
absorption process (as we will discuss in Worked Exercise 5.2). Thus, we assume
kR ¼ 1:3 1021 q1:74 T 6:76 cm1 ;
for T\2:5 104 K; ðlog T\4:4Þ, and H– photoionizations;
kR 30 cm1 ;
for 2.5 104 K < T < 1.6 105 K, (4.4 < log T < 5.2), and H and He pho-
toionizations, and, finally,
kR ¼ 1:5 1024 q2 T 7=2 cm1 ;
for T > 1.6 105 K (log T > 5.2), and metal photoionizations, free-free transitions
and Thomson scattering.
After having described the opacity variation in the solar interior, we introduce
the relation between the opacity and the temperature gradient that holds when the
energy flux is transported by radiation.
4.2 The Radiative Temperature Gradient
4.2.1 The Diffusion Approximation
In the photosphere one photon mean free path lk is of the order of 100 km and is
roughly equal to the mean distance from the solar surface. The sharp increase of the
opacity below the photosphere implies that lk = 1/kk rapidly decreases inwards and
for T > 2 104 K, lk 1 cm. Thus, in the interior, lk is very small compared to
the solar radius, which is now the scale of the distance from the surface. Due to the
short mean free path the radiation transfer in the interior is a diffusion process in
which the photon flux in a direction is proportional to the space variation of the
photon density in that direction. Because of the spherical symmetry of the Sun, only
the radial component Fk of the radiation flux does not vanish and the radiation
density Dk varies only with r. Thus, the diffusion equation for radiation has the form
Fk ¼ Cd dDk =dr;

4.2 The Radiative Temperature Gradient 133
where Cd is the diffusion coefficient and the sign minus is needed to get a positive
outward flux from a negative density gradient. Worked Exercise 4.2 shows that Cd
is given by
1
Cd ¼ clk ;
3
where c is the light velocity.

Worked Exercise 4.2 Relation of the mean free path to the diffusion coefficient.
In 1905 Einstein applied the diffusion equation to the microscopic random
motion of small particles pushed on by molecules in a liquid. The prototype of this
kind of motion, named the Brownian motion, was that of the plant pollen immersed
in water observed many years before by the British botanist Robert Brown (1773–
1858). According to Einstein, the one-dimensional probability P(x,Dt) of finding a
particle at a distance x from its origin after a given time Dt is a Gaussian distribution
N
Pðx; DtÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi ex =ð4Cd DtÞ ;
2
2 pCd Dt
where N is the particle number density and Cd the diffusion coefficient. Thus, Cd is
related to the Gaussian width, which represents the mean square deviation of a
particle in random motion, 〈x2 〉 , i.e.
hx2 i ¼ 2Cd Dt:
The numerical factor, which is 2 for the one dimensional case, becomes 6 in
three dimensions. Here, we substitute 〈x2〉 with the square of the mean free path
l and 6 with 3 as proposed by Smoluchowski (Marian Ritter von Smoluchowski
was a Polish physicist, 1872–1917). Thus, the relation becomes
l2 ¼ 3Cd Dt
and the diffusion coefficient is

Cd ¼ l2 =ð3DtÞ:
When applying the diffusion coefficient to radiation, l becomes the photon mean
free path lk and Dt the average time between an absorption and the next. Dt can be
calculated from the light velocity c and the mean free path as
Dt ¼ lk =c:
Substituting this relation into the diffusion coefficient gives

1
Cd ¼ c lk
3
Exercise end.
Because the radiation field in the solar interior is very close to that of a black body,
the space variation of Dk can be computed from the value that Dk has in TE, which
is (Worked Exercise 4.3)
Dk ¼ 4pBk =c:
Worked Exercise 4.3 The energy density.

Radiation energy density, Dk, is the radiation energy per unit volume. Thus, the
amount of energy, dEk, that flows in the time interval dt through a unit area dS
perpendicular to a direction s, per unit wavelength and solid angle, is
dEk ¼ Dk dV=ð4pÞ ¼ cDk dtdS=ð4pÞ;
where dV = cdtdS is the volume of the parallelepiped of face dS and width cdt, and
c is the light velocity (Fig. 4.3). Note that Dk/4p is the fraction of energy along a
single direction, because of isotropy.
On the other hand, under TE the radiation intensity is the Planck function Bk,
and the energy flowing in dt through dS, per unit wavelength and solid angle, is
dEk ¼ Bk dtdS:
Therefore, it must be
cDk dtdS=ð4pÞ ¼ Bk dtdS;
cDk =ð4pÞ ¼ Bk
and the radiation density is
Dk ¼ 4pBk =c:
Exercise end.
Fig. 4.3 The amount of

radiation energy that flows in
the time interval dt trough a
unit area dS perpendicular to a
direction s, is contained inside
the parallelepiped of width
cdt, face dS and volume
dV = cdtdS
4.2 The Radiative Temperature Gradient 135
Thus, the radiative diffusion equation becomes
1
Fk ¼ c lk d ð4pBk =cÞ=dr;
3
Because we are interested in the temperature gradient and the Planck function
depends on r through T, we write
Fk ¼ ð4p=3Þlk ðdBk =dT ÞdT=dr
Then, we integrate over wavelengths to get the total flux
Z1 Z1
F¼ Fk dk ¼ ð4p =3Þ lk ðdBk =dT Þdk dT=dr:
0 0
The integral in squared brackets is connected with the Rosseland mean opacity,
which has been defined in the Exercise 3.14 as
Z1
1 dBk 1
¼ dk ð4rT 3 =pÞ:
kR dT kk
0
Thus, we can write
Z1 Z1
dBk 1
lk ðdBk =dT Þdk ¼ dk ¼ 4rT 3 pkR
dT kk
0 0
and the radiative flux becomes
Z1
F ¼ ð4p =3Þ lk ðdBk =dT Þdk dT=dr ¼ ð4p=3Þ 4rT =ðpkR Þ dT=dr 3
and finally
F ¼ 16rT 3 ðdT=dr Þ=ð3kR Þ:
Therefore, the radiation flux is directly proportional to the temperature gradient

and inversely proportional to the mean Rosseland opacity.
The radiative temperature gradient is
dT=dr ¼ 3kR F=ð16 rT 3 Þ:

Denoting by L the luminosity at distance r, that is, the total energy flowing
across the surface 4p r2 in one second, which is constant with r in absence of
radiation energy sources and sinks, we get
L ¼ 4p r 2 F
and
dT=dr ¼ 3kR L=ð64pr 2 rT 3 Þ:
Exercise 4.4 Apply the radiative temperature gradient equation to the photosphere
and find the grey temperature stratification. Discuss the approximations which are
required to get the result.
Exercise 4.5 Estimate that at the base of the photosphere the radiative temperature
gradient is about 12 K per km [Hint: Take kR = Ds0.5/Dz = (0.9953–0.4395)/
50 km−1 from the VALC model in Table 3.1].
In the next section we determine an approximate expression for the radial gra-
dient of the temperature when the energy flux is transported by convection.
4.3 The Convective Temperature Gradient
4.3.1 Convection
The rapid growth of the opacity below the photosphere, which makes the inner
structure completely opaque, requires an increasing radiative temperature gradient
to keep the energy flux constant, as the diffusion equation for radiation shows. In
such conditions, the radiative energy transport is inefficient and convection starts on
in order to effectively smooth temperature differences. We first describe the general
character of the convective energy transport and then the physical conditions that
mark the onset of convective motions.
Let us consider the ordinary experience of heating water in a pot on a cooker.
Initially, the heat that the cooker supplies is transported to higher water layers by
conduction, without apparent fluid motion. In this situation at the microscopic level
energy is exchanged by collisions between water molecules, which lift up the
excess thermal agitation in the water layers closer to the flame. However, con-
duction cannot dissipate large amounts of heat in a short time. Thus, when due to
the continuous heating of the stove, the temperature of the bottom water becomes
sufficiently high compared to that of the surface water, the most efficient way to
redistribute energy inside is the ascending motion of hot water and descending
motion of cold water. Water layers in the pot are not uniform but there are warmer
and colder areas at the same level. At the bottom the overheated water expands,
becomes less dense, compared with the relatively colder water surrounding it, and
4.3 The Convective Temperature Gradient 137
moves upwards due to the buoyancy force. Vice versa, at the surface, the cold water
moves downwards being denser and heavier than the surrounding warm water. This
way of energy transport by motions of fluid elements with different temperatures is
called convection. The convective motion takes place in an ordered manner,
through the so called convective cells, where in a continuous circular movement,
the hot water ascends, cools at the surface, and descends warming up again at the
bottom. The convective cells remain until there is a sufficient temperature difference
between the bottom and surface of the water.
Similarly, in the solar convection zone where, due to the high opacity, the energy
flux would require a very large radiative temperature gradient to be fed, convective
motions are triggered and the sinking cold gas reduces the energy of the inner levels,
while the hot gas moving outwards increases the energy of the higher levels. The
existence of solar convection is confirmed by the appearance of the photospheric
granulation, which is observed on the solar disk at high spatial resolution (Fig. 4.4).
Granulation is a rather regular cellular pattern of brightness which originates by
the penetration into the photosphere of the underlying convection.
Due to its peculiar physical conditions, the solar convection has important dif-
ferences from that of the heated water in a pot. The most important difference is
stratification. In the case of the solar convection zone the density difference between
the surface and bottom is enormous, whilst in a pot of heated water this difference is
negligible. As we will see in Sect. 4.3, the convection zone extends from the base
Fig. 4.4 The photospheric granulation as observed by G. Scharmer with the Swedish Vacuum
Solar Telescope on 10 July 1997. To the left, the visible solar disk, with the Jupiter and Earth sizes
reported on the same scale, for comparison, at the low right corner. To the right, a high resolution
image of a 30,000 km 40,000 km disk area clearly shows the granulation structure. Bright
granules have sizes of about 1000 km (which is the distance unit of the horizontal axis and
approximately the length of Italy) and last on average about 10 min. Each granule is surrounded by
a dark background, which is called intergranule. Granulation is the overshooting into the
photosphere of the internal convective motions: bright granules are hot gas masses moving
outwards while dark intergranules are cold and sinking material
of the photosphere, were density is about 2 10−7 g cm−3, for almost

2.5 105 km, down to about r = 2/3 RS, a depth to which we roughly attribute the
average density of 1.4 g cm−3. Thus, the density decreases more than a million
times from the bottom to the top of the convection zone. Due to the high density the
convective energy flow can be transported at the bottom of the convection zone
with very small temperature differences between the upward and downward moving
gas. Actually, the temperature contrast between hot and cold elements is estimated
to a few degrees with respect to an average gas temperature of the order of a million
degrees K. On the contrary, at the surface the temperature contrast is very strong,
because of the low density; and, moreover, because cold elements are more
transparent and lose energy by radiation. Thus, at the surface, temperature differ-
ences of about 2000 degrees between hot and cold gas over a 6000 K average
temperature can occur. As a consequence, surface cold elements sink with high
speed while the bottom hot gas moves upwards smoothly and nearly uniformly.
Actually, the convective gas velocities can be larger than 1 km s−1 at the surface
and only a few m s−1at the bottom.
In conclusion, convection in a pot of water on a cooker is controlled by the water
heated at the bottom and moving up, whereas solar convection is primarily fed by
the gas that cools at the surface and sinks. Our understanding of the solar con-
vection has made a lot of progress after the advent of supercomputers and the
consequent diffusion of three-dimensional numerical simulations of the upper
convective layers. The description given in this paragraph is partly based on their
results (Spruit, 1997).
Now we’ll look at the relation between the radial gradient of temperature and the
convective energy flux.
4.3.2 The Convective Temperature Gradient
Solar and stellar structure models commonly use the approximate theory of the
mixing length in order to treat the convective energy transport and determine the
temperature structure in the convection zone. The mixing length theory treats the
convective moving elements like gas molecules. The mean free path of the hot and
cool elements, before they mix with the surrounding gas, is called the mixing length
and is usually assumed to be of the order of 1.5 times the pressure scale height.
The energy flux Fc transported by a gas element moving with velocity v in the radial
direction and with an excess temperature DT over its surroundings is
Fc ¼ vCP qDT
where q is the density and CP the specific heat at constant pressure (see Exercise
4.6).
Worked Exercise 4.6 Compute the gas specific heats.

The heat amount dQ required to increase the temperature of a mass element m by
dT in a reversible way can be written as
dQ ¼ CmdT;
where C is an element property called specific heat. Note that Q is not a state
function and the symbol d denotes an infinitesimal variation but not an exact
differential. For a gas two specific heats are usually defined, one for constant
volume, CV, and one for constant pressure, CP. Thus, considering hereafter the heat
for unit mass
CV ¼ ðdQ=dT ÞV
and
CP ¼ ðdQ=dT ÞP ;
where the indexes T and P mean that the transformations occur at constant volume
and pressure, respectively. CV has a simple expression in terms of the internal
energy U. The first law of thermodynamics relates the internal energy, heat and
work variations as
dU ¼ dQ PdV;
Note that an expanding gas makes a positive work, PdV, and U decreases, while
a compressed gas made a negative work and U increased. Moreover, U, Q and V are
per unit mass. Thus, at constant V,
dU ¼ dQ ¼ CV dT
and
CV ¼ dU=dT:
Cp has a simple expression in terms of the enthalpy H, which is a state function

defined by
H ¼ U þ PV:
Using the first law of thermodynamics we get for an enthalpy variation
dH ¼ dU þ d ðPV Þ ¼ dQ PdV þ PdV þ VdP ¼ dQ þ VdP

and, at constant P,
dH ¼ dQ ¼ CP dT:
Thus,
CP ¼ dH=dT:
For a perfect gas there is a simple relation between the specific heats and the gas
constant R. The enthalpy variation is
dH ¼ dU þ d ðPV Þ:
Using the state equation in the form
PV ¼ RT=l;
where V = 1/q is the specific volume (i.e. per unit mass), we get
d ðPV Þ ¼ RdT=l
By inserting this result and the specific heats, dH becomes
CP dT ¼ CV dT þ RdT=l:
Thus,
Cp ¼ CV þ R=l
and
R=l ¼ CP CV :
The internal energy per unit volume of a perfect gas is the product of the number
density of particles N and the mean energy of a particle 3kBT/2, i.e.
U ¼ 3NkB T=2 ¼ 3RqT=ð2lÞ:
Then, the specific energy per unit mass is
U ¼ 3RT=ð2lÞ
and the specific heat at constant volume becomes
CV ¼ dU=dT ¼ 3R=ð2lÞ:
Moreover, the specific heat at constant pressure is
CP ¼ CV þ R=l ¼ 3R=ð2lÞ þ R=l ¼ 5R=ð2lÞ:
Exercise end.
This expression is used to find a relation between the convective energy flux and
the actual temperature gradient of the convection zone, after that appropriate
averages of v and DT over convective elements have been defined. Actually, we
emphasize that the calculation of the temperature gradient in presence of convection
is not straightforward and, moreover, as we will see later, while in the inner layers
the gradient can be related to the adiabatic gradient, in the outer layers the gradient
is in between the adiabatic and radiative gradients and cannot be precisely com-
puted by the mixing length theory. Here we use a simpler approach than that of the
mixing length theory.
Exercise 4.7 Estimate the convective energy flux Fc = vCPqDT for q = 10−7 g cm−3,
v = 1 km s−1 and DT = 500 K and for q = 1 g cm−3, v = 1 m s−1 and
DT = 5 K. Compare the results with the solar flux at r = RS. [Hint. Take CP = 5R/2
as for a monoatomic perfect gas with l = 1].
We assume that solar convective motions occur on a shorter time scale than the
energy losses by radiation. Thus, moving gas elements do not exchange energy with
the surroundings and convective motions are adiabatic (a Greek word which means
not permeable). Actually, the adiabatic assumption holds when the gas is very
opaque and, hence, is a good approximation deep in the convection zone but fails
close to the surface. The adiabatic assumption allows to state simple relations
between the thermodynamic parameters of a convective element, i.e.
P ¼ c1 qc ;
P ¼ c2 T c=ðc1Þ ;
P ¼ c3 T 1=ðc1Þ ;
where c1, c2, c3 are constant and c is the adiabatic exponent, which is discussed in
Exercise 4.9. c = 5/3 for a perfect gas with monoatomic molecules.
Exercise 4.8 Use the gas state equation to derive the second adiabatic relation from
the first relation.
Worked Exercise 4.9 The adiabatic processes.
We consider an infinitesimal, reversible and adiabatic compression (or expansion),
which occurs between two gas states characterized by the values T1, q1 and T, q,
respectively. The first law of thermodynamics sets the internal energy variation as
dU ¼ dQ PdV:
For a perfect gas, it is
dU ¼ CV dT:
Because the process is adiabatic the exchanged heat is
dQ ¼ 0:
From the state equation
P ¼ RqT=l ¼ RT=ðlV Þ;
the exchanged work is
PdV ¼ RTdV=ðlV Þ ¼ RTdq=ðlqÞ:
Thus, the first law becomes
CV dT ¼ RTdq=ðlqÞ:
and
dT=T ¼ ðR=CV Þdq=ðlqÞ:
The gas constant is also the difference of the specific heats at constant pressure
and volume (Exercise 4.6)
R=l ¼ CP CV
and, by definition,
c ¼ CP =CV ;
hence,
R=ðl CV Þ ¼ ðCP CV Þ=CV ¼ CP =CV CV =CV ¼ c 1:
Thus,
dT=T ¼ ðc 1Þdq=q:
This equation can immediately be integrated between two states characterized by

temperatures and densities T1, q1 and T, q, respectively. Thus
d lnT ¼ ðc 1Þd lnq;
d lnT ¼ d lnqc1 ;
q
½lnT TT1 ¼ lnqc 1 q ;
1
lnT lnT1 ¼ lnqc1 lnqc1

1 ;
lnðT=T1 Þ ¼ lnðqc1 =qc1

1 Þ;
T=T1 ¼ ðq=q1 Þc1
and, eventually,
q=q1 ¼ ðT=T1 Þ1=ðc1Þ :
Note that because for a perfect gas it is
CV ¼ 3R=ð2lÞ
and
CP ¼ 5R=ð2lÞ;
it is also
c ¼ CP =CV ¼ 5=3:
Exercise end.
In order to extend the adiabatic relations from the single convective elements to
the whole surrounding medium, we also assume that at each level the elements are
in pressure equilibrium with the surrounding gas and convection is so efficient that
the mean gas temperature rapidly approaches the adiabatic temperature. We restate
that the temperature is actually close to the adiabatic temperature only in the deeper
layers of the convection zone. Under these conditions, the convective temperature
gradient can be determined straightforwardly from the radial derivative of the
second adiabatic relation between P and T
dP=dr ¼ d½c2 T c=ðc1Þ =dr ¼ c2 ½c=ðc 1Þ T ½c=ðc1Þ1 dT=dr
Multiplying and dividing by T the right hand side, we get

dP=dr ¼ c2 ½c=ðc 1Þ½ T ½c=ðc1Þ =TdT=dr
and, hence, since P ¼ c2 T c=ðc1Þ ;
dP=dr ¼ ½c=ðc 1ÞðP=T ÞdT=dr:
Thus, the adiabatic temperature gradient, which we use as a rough approxima-

tion to the temperature gradient in presence of convection, is
ðdT=dr Þa ¼ ½ðc 1Þ=cðT=PÞdP=dr:
Note that both the temperature and pressure gradients are negative because T and
P decrease with increasing r. The quantity (c − 1)/c is the ratio of the natural
logarithms of the adiabatic temperature and pressure gradients
ðc 1Þ=c ¼ ½dT=ðTdr Þa =½dP=ðPdr Þa ¼ ðd lnT=dr Þa =ðd lnP=dr Þa
and takes the value
ðc 1Þ=c ¼ ð5=3 1Þ=ð5=3Þ ¼ 0:4
in the case of a monoatomic perfect gas. With this value for (c − 1)/c and the
known values of the effective temperature and pressure scale height, the absolute
value of the adiabatic gradient at s = 2/3 is
ðdT=dr Þ
¼ 0:4ðT=PÞðP=H Þ ¼ 0:4Te =H ¼ 0:4 5778 K=137 km ¼ 17 K km1 :
a
At the same level the radiative temperature gradient is about 12 K km−1

(Exercise 4.5). Thus, at the base of the photosphere the adiabatic temperature
gradient is close to the radiative temperature gradient. In the next section we see
that the comparison between the radiative and adiabatic temperature gradients sets
the criterion for the occurrence of convection.
4.4 The Convection Zone Model
4.4.1 Convective Stability Criterion
To state the stability of the solar gas against convection, we require that a con-
vective motion, which starts as a local perturbation, does not grow but fades
instead. We consider a mass element which is in equilibrium with its surroundings
4.4 The Convection Zone Model 145
Fig. 4.5 The dynamical stability criterion. A mass element, which was in equilibrium with the
surroundings at the level r, is perturbed and lifted radially up to the level r + dr. The motion is
adiabatic (DQ = 0) and with pressure equilibrium between element and surroundings (DP = 0).
Then, if the element density at r + dr is larger than that of the surroundings (Dq > 0), the
buoyancy force −gDq is directed inwards and level r is stable
at the level r and is then perturbed and lifted up radially to the level r + dr
(Fig. 4.5).
We can assume that the element motion is slow enough that pressure differences
with the surroundings are immediately smoothed by sound waves, but also it is
sufficiently fast so that energy exchanges by radiation are too slow to occur. Thus,
the element is lifted up adiabatically and in pressure equilibrium with the
surroundings.
Let us denote by Df the difference in the generic physical parameter f between
the mass element and the surrounding gas at the same level. Then, during the
motion, it is DP = DQ = 0. Initially, at r, it is also DT(r) = 0 and Dq(r) = 0.
Finally, at r + dr, it will be
DT(r + dr) 6¼ 0 and Dq(r + dr) 6¼ 0. In particular, if Dq(r + dr) > 0, the ele-
ment is heavier than the surrounding gas and will be drawn back to its starting
position.
Therefore, the condition for dynamical stability is in terms of the weight per unit
volume
gqe ðr þ dr Þ [ gqðr þ dr Þ
and
qe ðr þ dr Þ [ qðr þ dr Þ;
where qe and q denote the element and surrounding densities, respectively. For dr
sufficiently small, by the definition of differential, the relation becomes
qe ðr Þ þ ðdqe =dr Þdr [ qðr Þ þ ðdq=dr Þdr;
and because Dq(r) = qe(r) − q(r) = 0, we get
ðdqe =dr Þdr [ ðdq=dr Þdr
and
dqe =dr [ dq=dr:
It is worthwhile to express the disequality in terms of the temperature gradient,

which is directly related to the energy flux. Then, as Worked Exercise 4.10 shows,
the condition for stability becomes
dTe =dr\dT=dr:
Because the element motion is adiabatic, it is dTe/dr = (dT/dr)a. Moreover,

because the surrounding medium is dynamically stable, it is dT/dr = (dT/dr)r. Thus,
we can write
ðdT=dr Þa \ðdT=dr Þr
and, because the temperature gradients are negative

ðdT=dr Þ
\
ðdT=dr Þ
:
r a
Therefore, to determine whether or not the solar gas is stable against dynamical
perturbations at a given level it is necessary to compare the radiative and adiabatic
temperature gradients.
Worked Exercise 4.10 Derive the Schwarzschild criterion for dynamical stability.
To express the dynamical stability condition
dqe =dr [ dq=dr
in terms of the temperature gradient, we take the radial derivative of the gas state
equation
q ¼ Pl=ðRT Þ
taking into account that P, T and l vary with r. Thus, we get for the gas element and
the ambient medium
dqe =dr ¼ ðdPe =dr Þl=ðRT Þ þ ðPl=RÞd ð1=Te Þ=dr þ ðdle =dr ÞP=ðRT Þ
and

dq=dr ¼ ðdP=dr Þl= Rg T þ ðPl=RÞd ð1=T Þ=dr þ ðdl=dr ÞP=ðRT Þ
Note that the element is in complete equilibrium with the surroundings when it
starts to move. Moreover, dPe/dr = dP/dr, because the motion is in pressure
equilibrium with the ambient. Thus, the stability condition becomes
ðPl=RÞd ð1=Te Þ=dr þ ðdle =dr ÞP=ðRT Þ [ ðPl=RÞd ð1=T Þ=dr þ ðdl=dr ÞP=ðRT Þ;
which, multiplyng by R/P and performing the derivative of 1/T, we arrange as
ld ð1=Te Þ=dr þ ðdle =dr Þ=T [ ld ð1=T Þ=dr þ ðdl=dr Þ=T;
lð1=T Þ2 dTe =dr þ ðdle =dr Þ=T [ lð1=T Þ2 dT=dr þ ðdl=dr Þ=T;
Multiplying by −T2/ l
dTe =dr ðT=lÞðdle =dr Þ\dT=dr ðT=lÞðdl=dr Þ;
dTe =dr dT=dr\ðT=lÞðdle =dr dl=dr Þ:
l is a function of P and T, which in the convection zone is determined by the

ionization equilibrium. Assuming that l(P,T) is the same function for both the
element and ambient, we get
dle =dr ¼ ð@l=@PÞT dPe =dr þ ð@l=@T ÞP dTe =dr
and
dl=dr ¼ ð@l=@PÞT dP=dr þ ð@l=@T ÞP dT=dr;
where @l=@PT and @l=@TP denote the P derivative at constant T and the
T derivative at constant P, respectively. Because dPe/dr = dP/dr,
dle =dr dl=dr ¼ ð@l=@T ÞP ðdTe =dr dT=dr Þ
and the stability condition becomes
dTe =dr dT=dr\ðT= lÞð@l=@T ÞP ðdTe =dr dT=dr Þ
and

1 ðT= lÞð@l=@T ÞP ðdTe =dr dT=dr Þ\0:
Therefore, because 1 − (T/l)(∂l/∂T)P > 0 (Exercise 4.11), we obtain that the

stability condition in terms of the temperature gradients is simply
dTe =dr\dT=dr:
Exercise end.
Exercise 4.11 Show that 1 − (T/l)(∂l/∂T)P > 0 [Hint. Use the state equation for a
perfect gas].
If the radiative temperature gradient is lower in modulus, that is, less steep than
the adiabatic temperature gradient

ðdT=dr Þ
\
ðdT=dr Þ
;
r a
the gas is stable and the energy transport is only radiative. This disequality between
radiative and adiabatic gradients constitutes the Schwarzschild criterion for
dynamical stability (Karl Schwarzschild was a German astronomer, 1873–1916).
Vice versa, when the radiative gradient is greater in modulus, i.e. steeper than the
adiabatic gradient

ðdT=dr Þ
[
ðdT=dr Þ
;
r a
then the layer is unstable and convection occurs.

Since
jdT=dr jr ¼ 3kR L=ð64pr 2 rT 3 Þ;
we expect the onset of convection in layers with high opacity (kR increases) or with
nuclear reactions concentrated near the center (L/r2 increases). Actually, many stars
have convective zones in their external layers, due to an increase of the b-b and b-f
absorption, as it occurs to the Sun, or convective cores, related to the onset of
nuclear reactions in high temperature layers close to the stellar center.
Worked Exercise 4.12 shows that the Schwarzschild criterion in the photosphere
and convection zone can be approximated with
Worked Exercise 4.12 An approximate version of the Schwarzschild criterion.
The Schwarzschild criterion for convective stability is

ðdT=dr Þ
\
ðdT=dr Þ
r a
with
jdT=drjr ¼ 3kR L=ð64pr 2 rT 3 Þ
and
jdT=dr ja ¼ ½ðc 1Þ=cðT=PÞjdP=dr j
To facilitate the comparison we make the pressure gradient appear in the

expression of the radiative temperature gradient. To this purpose, we take the
inverse of the squared radius from the hydrostatic equilibrium
dP=dr ¼ qGM=r 2 :
Thus,
1=r 2 ¼ jdP=dr j=ðqGMÞ
and the radiative gradient is
jdT=dr jr ¼ 3kR L=ð64pr 2 rT 3 Þ ¼ 3 LkR jdP=dr j=ð64prG MqT 3 Þ:
Taking the density from the state equation
q ¼ lP=ðRT Þ;
we get
jdT=dr jr ¼ 3RLkR ðT=PÞjdP=dr j=ð64p r G MlT 3 Þ:
We can take L = LS and approximately M = MS. The latter approximation is

suggested by the strong gravitational stratification of density and is good at a few
percent at the base of the solar convection zone. Thus, the constant part of the
radiative gradient, denoted by S, is
S ¼ 3RLS =ð64p r G MS Þ
and we can write the radiative gradient as
jdT=dr jr ¼ SðkR =lÞðT=PÞjdP=dr j=T 3 :
At this point the condition for convective stability
jdT=dr jr \jdT=dr ja
becomes
SðkR =lÞðT=PÞjdP=dr j=T 3 \½ðc 1Þ=cðT=PÞjdP=dr j;
that is,
SkR =ðlT 3 Þ\ðc 1Þ=c:
Exercise end.
SkR =ðlT 3 Þ\ðc 1Þ=c;

where the constant S is
S ¼ 3RLS =ð64p r G MS Þ:
In the photosphere, at s = 2/3, we have estimated that the radiative and con-
vective gradients are comparable, then, the left hand side and the factor (c − 1)/c
have roughly the same value at this level. In the higher layers for s < 2/3, the gas is
mostly made of neutral H and He and can be approximated by a monoatomic ideal
gas with (c − 1)/c = 0.4 and l = 1.27. In the left hand side kR decreases faster than
T3 with decreasing s, while l and (c − 1)/c in the right hand side are roughly
constant. Thus, we expect that in most of the photosphere the radiative gradient is
smaller in modulus than the adiabatic gradient and the photosphere is dynamically
stable (radiative).
This situation changes as we move to the convective zone where the temperature
is high enough to trigger the ionization of hydrogen and helium. The H and He
ionizations increase the opacity kR (Sect. 4.1), decrease (c − 1)/c by about a factor
4 (Exercise 4.13) and l by about a factor 2 (Exercise 4.14).
Worked Exercise 4.13 (c − 1)/c variation.
In the convection zone, where physical conditions allow for the H and He
ionization, the work done by the external pressure on a cold gas mass, which
adiabatically moves down toward higher pressure layers, serves partly to increase
the temperature and partly to ionize the hydrogen and helium it contains. Thus, the
compression of the solar gas produces a reduced temperature increment with respect
to that of an equal mass of a perfect gas. In other words, CP increases and, con-
sequently, the ratio (c − 1)/c = (CP/CV − 1)/(CP/CV) = (CP − CV)/ CP = R/(lCP)
reduces from the perfect gas value of 0.4 down to a minimum value of about 0.1.
Once that the gas is mostly ionized, (c − 1)/c turns to rise and reaches again the
value 0.4 at complete ionization.
Exercise end.
Worked Exercise 4.14 l variation.
Ionization also affects the average molecular weight. For a neutral mixture with
10 H atoms for each He atom it is l = 1.27. If all hydrogen is ionized, 10 H atoms
contribute 20 particles (10 protons and 10 electrons); moreover, if all helium is
twice ionized, each He atom contributes 3 particles (1 He nucleus, or a particle, and
2 electrons). Thus, when H and He ionizations are complete, the average molecular
weight is l = (10 + 4)/(20 + 3) = 0.61.
Exercise end.
The kR and l variations increase the left hand side of the stability criterion while
(c − 1)/c variation decreases the right hand side. As a consequence, the radiative
gradient is steeper than the adiabatic gradient and the convective zone is dynami-
cally unstable.
Because the three effects that set the temperature gradients are mainly related to
the hydrogen ionization, the convective zone is often called the hydrogen con-
vection zone.
An accurate calculation confirms that in the present Sun, convection is the
dominant mechanism of energy transport from the base of the photosphere down to
2/3 of the solar radius. Conversely, energy transport is secured by radiation in the
remaining layers, where the radiative temperature gradient is less steep than the
adiabatic temperature gradient. The extension of the convection zone is quantita-
tively estimated later in this section.
We are now ready for constructing a simplified model of the convective zone.
4.4.2 A Model of Convection Zone
To find the temperature in the convection zone we start with the expression of the
adiabatic temperature gradient, which approximates the actual temperature gradient
ðdT=dr Þa ¼ ½ðc 1Þ=cðT=PÞdP=dr:
The hydrostatic equilibrium equation gives the pressure gradient
dP=dr ¼ qG M=r 2 ;
that we replace into the expression of the adiabatic gradient

ðdT=dr Þa ¼ ½ðc 1Þ=cðTq=PÞGM=r 2 :
By the state equation
P ¼ RTq=l
in the form
Tq=P ¼ l=R;
we get

ðdT=dr Þa ¼ ½ðc 1Þ=cG M l= Rr 2 :
Let us take M = MS all over the convection zone. We consider also (c − 1)/c and
l constant, as it will be clarified later. Then, we denote the constant term in the right
hand side by the product W RS so that the constant W has the dimension of a
temperature and is defined by
W ¼ ½ðc 1Þ=cGMS l=ð RRS Þ
In this way, we get
dT=dr ¼ WRS =r 2
and, because d(1/r) = −1/r2,
dT ¼ ðWRS =r 2 Þdr ¼ WRS d ð1=r Þ:
At this point, the integration is immediate. If we denote by r0 and r two radial

distances and by T0 and T the corresponding temperatures, we get
ZT Zr
dT ¼ WRS d ð1= r Þ:
T0 r0
Thus,
½T TT0 ¼ WRS ½1=r rr0 ;
TT0 ¼ WRS ð1=r 1=r0 Þ
and finally
T ¼ T0 þ W ðRS =r RS =r0 Þ:
This equation gives the radial variation of the temperature in the convection zone
starting from a known temperature T0 at r0.
To construct an approximate model, we divide the convection zone into two
parts: the top, between r/RS = 1 and 0.95, where physical properties may depend on
the presence of both the radiation and convection, and hydrogen and helium may be
partly ionized, and the deeper zone, between r/RS = 0.95 and 0.70, where the H and
He ionizations are complete and the entire energy flow is transported by convection
except for the very bottom layers. We expect that a model, which is based on the
adiabatic temperature gradient and constant values of c and l, is more accurate in
the deep zone than in the top region. Figure 4.6 shows the model for the upper part
of the convection zone.
How simplified our model is may be seen from the first panel, which shows the
ratio (c − 1)/c. In our model it is constant at the intermediate value 0.26, corre-
sponding to c = 1.35, while in the accurate model it drops from 0.4 at r = RS, down
to 0.13 at about r = 0.999 RS. Below this level, (c − 1)/c gradually rises again to
0.4 once the H and the two He ionizations are complete. Ionization also affects the
average molecular weight. As the second panel shows, we have simply taken
the constant, intermediate value l = 0.9, but the average molecular weight of the
accurate model rapidly drops from the value of a neutral mixture (1.27) at the
surface to that of a fully ionized mixture (0.61) in depth.
In the third panel our simplified temperature, calculated with the initial point
r0 = RS and T0 = Te in the photosphere, is compared with the accurate model. Here,
the constant W is
W ¼ 5:35 106 K:
Exercise 4.15 Determine the value of the constant W in the top convection zone,
using c = 1.35 and l = 0.9.
The agreement between the approximate and accurate T variations is discreet. In
particular, the approximate model reproduces the steep T rise just below the pho-
tosphere with values larger than 105 K already at 0.99 solar radii.
In the fourth and fifth panel the approximate pressure and density variations have
been obtained from the adiabatic relations with T, which in terms of the initial
conditions at s = 2/3 are
P ¼ P0 ðT=T0 Þc=ðc1Þ
and
q ¼ q0 ðT=T0 Þ1=ðc1Þ :
Fig. 4.6 Model of the upper part of the convection zone. Our approximate model (solid lines) is
compared with the accurate model of Senesi et al. (red lines). In the order the six panels represent
respectively the adiabatic gradient (c − 1)/c, average molecular weight, temperature, pressure,
density, and mean Rosseland opacity. Note that all the model parameters are plotted as functions of
the fraction of solar radius r/RS, and temperature, pressure, density and opacity are decimal
logarithms
Fig. 4.6 (continued)
As for T, also for P and q the agreement between the simplified and accurate
models is in principle but not in detail.
Finally, the sixth panel shows the mean Rosseland opacity. As discussed in
Sect. 4.1, the approximate opacity is provided by the H− power law close to sur-
face, a constant plateau corresponding to the H and He peaks, and the Kramers
power law for metals in depth. Temperature and density necessary to compute the
power laws were taken from the Senesi et al. model. It is apparent that these
approximations schematically represent the accurate opacity variation.
Figure 4.7 shows the model for the deep part of the convection zone. Our
simplified model has been calculated starting from the values the physical param-
eters take in the accurate model at r0 = 0.95 RS. The ratio (c − 1)/c and the
Fig. 4.7 Model of the inner part of the convection zone. Our approximate model (solid lines) is
compared with the accurate model of Senesi et al. (red lines). From the upper left to lower down
the panels represent respectively the adiabatic gradient (c − 1)/c, average molecular weight,
temperature, pressure, density, and mean Rosseland opacity. Note that all the model parameters are
plotted as functions of the fraction of solar radius r/RS, and temperature, pressure, density and
opacity are decimal logarithms. The calculation of (c − 1)/c in the accurate model stops below r/
RS = 0.72
average molecular weight remain substantially constant down to the base of the
convection zone respectively to the values 0.4, characteristic of a perfect gas, and
0.61, characteristic of the complete H and He ionizations. With these values of
(c − 1)/c and l, the constant W entering in the approximate temperature variation is
W = 5.50 106 K.
Exetrcise 4.16 Determine the value of the constant W in the deep convection zone,
using c = 1.67 and l = 0.61.
As expected, here the agreement with the accurate model is better than close to
the surface. In particular, we note that near the bottom, at r = 0.71 RS, the tem-
perature exceeds 2 106 K, the pressure 5 1013 Ba and the density 0.1 g cm−3.
Table 4.1 shows the model of Senesi et al. for the convection zone.
Table 4.1 The convection zone model of Senesi et al.
160
r/RS M/MS T, K P, Ba q, g cm−3 kR, cm−1 |dT/dr|r/|dT/dr|a

0.9999995 1.000E+000 5780 6.514E+004 1.685E−007 5.547E−008 3.768E−001
0.9998981 1.000E+000 6592 1.047E+005 2.374E−007 2.998E−007 1.448E+000
0.9998011 1.000E+000 8452 1.512E+005 2.647E−007 4.160E−006 1.533E+001
0.9996354 1.000E+000 10057 2.477E+005 3.526E−007 3.397E−005 1.228E+002
0.9994398 1.000E+000 11149 4.059E+005 5.020E−007 1.307E−004 4.245E+002
0.9992179 1.000E+000 12073 6.651E+005 7.323E−007 3.824E−004 1.075E+003
0.9989705 1.000E+000 12944 1.090E+006 1.082E−006 9.883E−004 2.366E+003
0.9986972 1.000E+000 13813 1.786E+006 1.608E−006 2.384E−003 4.809E+003
0.9983968 1.000E+000 14712 2.926E+006 2.400E−006 5.526E−003 9.314E+003
0.9980674 1.000E+000 15666 4.795E+006 3.587E−006 1.250E−002 1.744E+004
0.9977065 1.000E+000 16696 7.857E+006 5.357E−006 2.787E−002 3.183E+004
0.9973105 1.000E+000 17827 1.288E+007 7.991E−006 6.162E−002 5.688E+004
0.9968755 1.000E+000 19083 2.110E+007 1.189E−005 1.354E−001 9.947E+004
0.9963961 1.000E+000 20499 3.457E+007 1.764E−005 2.951E−001 1.690E+005
0.9958658 1.000E+000 22120 5.665E+007 2.604E−005 6.359E−001 2.790E+005
0.9952763 1.000E+000 23995 9.283E+007 3.825E−005 1.379E+000 4.525E+005
0.9946173 1.000E+000 26198 1.521E+008 5.580E−005 2.900E+000 6.940E+005
0.9938759 1.000E+000 28817 2.492E+008 8.087E−005 5.739E+000 9.754E+005
0.9930355 1.000E+000 31958 4.084E+008 1.163E−004 1.083E+001 1.275E+006
0.9920757 1.000E+000 35763 6.692E+008 1.659E−004 2.006E+001 1.586E+006
0.9909700 1.000E+000 40452 1.097E+009 2.343E−004 3.088E+001 1.571E+006
0.9896831 1.000E+000 46361 1.797E+009 3.271E−004 3.795E+001 1.183E+006
0.9881675 1.000E+000 53916 2.944E+009 4.513E−004 3.721E+001 6.844E+005
0.9863637 1.000E+000 63498 4.825E+009 6.168E−004 3.290E+001 3.543E+005
(continued)
4 The Convection Zone
0.9842045 1.000E+000 75293 7.906E+009 8.397E−004 3.045E+001 1.959E+005
0.9816196 1.000E+000 89291 1.296E+010 1.145E−003 3.095E+001 1.224E+005
0.9785358 1.000E+000 105697 2.123E+010 1.565E−003 3.198E+001 7.746E+004
0.9748627 1.000E+000 125337 3.479E+010 2.136E−003 2.971E+001 4.267E+004
0.9704793 9.999E−001 149449 5.700E+010 2.902E−003 2.285E+001 1.889E+004
0.9652331 9.999E−001 179309 9.340E+010 3.927E−003 1.551E+001 7.262E+003
0.9589430 9.998E−001 216166 1.530E+011 5.296E−003 1.044E+001 2.755E+003
4.4 The Convection Zone Model
0.9514030 9.998E−001 261404 2.508E+011 7.133E−003 7.409E+000 1.098E+003

9.501E−001 9.997E−001 2.696E+005 2.716E+011 7.482E−003 7.056E+000 9.524E+002
9.474E−001 9.997E−001 2.861E+005 3.165E+011 8.202E−003 6.458E+000 7.289E+002
9.448E−001 9.997E−001 3.018E+005 3.630E+011 8.906E−003 5.998E+000 5.766E+002
9.424E−001 9.996E−001 3.164E+005 4.100E+011 9.585E−003 5.651E+000 4.710E+002
9.402E−001 9.996E−001 3.303E+005 4.575E+011 1.024E−002 5.381E+000 3.943E+002
9.369E−001 9.995E−001 3.506E+005 5.331E+011 1.122E−002 5.053E+000 3.093E+002
9.332E−001 9.994E−001 3.745E+005 6.307E+011 1.241E−002 4.761E+000 2.392E+002
9.297E−001 9.993E−001 3.965E+005 7.297E+011 1.354E−002 4.548E+000 1.925E+002
9.265E−001 9.993E−001 4.171E+005 8.302E+011 1.464E−002 4.394E+000 1.597E+002
9.235E−001 9.992E−001 4.365E+005 9.320E+011 1.569E−002 4.276E+000 1.356E+002
9.204E−001 9.991E−001 4.563E+005 1.043E+012 1.678E−002 4.186E+000 1.163E+002
9.152E−001 9.989E−001 4.905E+005 1.253E+012 1.873E−002 4.076E+000 9.120E+001
9.104E−001 9.988E−001 5.220E+005 1.468E+012 2.060E−002 4.002E+000 7.427E+001
9.060E−001 9.986E−001 5.514E+005 1.687E+012 2.239E−002 3.944E+000 6.212E+001
9.019E−001 9.984E−001 5.791E+005 1.910E+012 2.412E−002 3.900E+000 5.306E+001
8.980E−001 9.983E−001 6.054E+005 2.137E+012 2.580E−002 3.865E+000 4.605E+001
(continued)
161
162

8.936E−001 9.981E−001 6.359E+005 2.420E+012 2.780E−002 3.829E+000 3.937E+001
8.868E−001 9.977E−001 6.825E+005 2.894E+012 3.095E−002 3.770E+000 3.137E+001
8.807E−001 9.974E−001 7.259E+005 3.382E+012 3.398E−002 3.721E+000 2.575E+001
8.749E−001 9.971E−001 7.666E+005 3.883E+012 3.692E−002 3.678E+000 2.162E+001
8.696E−001 9.967E−001 8.053E+005 4.397E+012 3.978E−002 3.642E+000 1.849E+001
8.645E−001 9.964E−001 8.421E+005 4.923E+012 4.257E−002 3.613E+000 1.605E+001
8.594E−001 9.960E−001 8.800E+005 5.502E+012 4.550E−002 3.590E+000 1.398E+001
8.505E−001 9.954E−001 9.465E+005 6.615E+012 5.082E−002 3.578E+000 1.121E+001
8.423E−001 9.947E−001 1.009E+006 7.772E+012 5.598E−002 3.582E+000 9.279E+000
8.347E−001 9.940E−001 1.068E+006 8.972E+012 6.103E−002 3.586E+000 7.840E+000
8.276E−001 9.934E−001 1.124E+006 1.021E+013 6.597E−002 3.587E+000 6.729E+000
8.209E−001 9.927E−001 1.178E+006 1.150E+013 7.083E−002 3.584E+000 5.848E+000
8.145E−001 9.920E−001 1.230E+006 1.282E+013 7.562E−002 3.578E+000 5.134E+000
8.062E−001 9.911E−001 1.299E+006 1.471E+013 8.212E−002 3.565E+000 4.351E+000
7.949E−001 9.898E−001 1.394E+006 1.759E+013 9.143E−002 3.540E+000 3.501E+000
7.833E−001 9.883E−001 1.496E+006 2.100E+013 1.017E−001 3.506E+000 2.815E+000
7.712E−001 9.867E−001 1.604E+006 2.504E+013 1.130E−001 3.462E+000 2.260E+000
7.600E−001 9.850E−001 1.707E+006 2.932E+013 1.243E−001 3.418E+000 1.853E+000
7.494E−001 9.833E−001 1.807E+006 3.385E+013 1.355E−001 3.369E+000 1.543E+000
7.394E−001 9.817E−001 1.904E+006 3.862E+013 1.467E−001 3.317E+000 1.301E+000
7.299E−001 9.800E−001 1.999E+006 4.365E+013 1.578E−001 3.261E+000 1.108E+000
7.208E−001 9.783E−001 2.090E+006 4.893E+013 1.694E−001 3.231E+000 9.601E−001
7.121E−001 9.767E−001 2.173E+006 5.446E+013 1.819E−001 3.293E+000 8.698E−001
In the order the columns report the solar radius fraction r/RS, mass fraction M/MS, temperature T, pressure P, density q, mean opacity kR and the ratio of the
radiative and adiabatic temperature gradients in modulus |dT/dr|r/ |dT/dr|a
4 The Convection Zone
4.4.3 The Extension of the Convection Zone
We have acquired enough tools to verify quantitatively that the convective zone
extends from the solar surface to about 2/3 of the solar radius. To this aim we start
from the criterion for dynamic stability in the form obtained in Exercise 4.12, that is
S kR =ðlT 3 Þ\ðc 1Þ=c;
where
S ¼ 3RLS =ð64p rGMS Þ:
Inserting the opacity power law in terms of the temperature and density
kR ¼ k0 qa T b ;
where k0, a and b differ between the outer and deeper parts of the convection zone,
we get
S k0 qa T b =ðlT 3 Þ\ðc 1Þ=c:
Expressing the density in terms of the temperature by the adiabatic relationship
q ¼ q0 ðT=T0 Þ1=ðc1Þ ;
we get (because the exponent of the power of a power is the product of the
exponents)
a=ðc1Þ
S k0 ðqa0 T a=ðc1Þ =T0 ÞT b =ðlT 3 Þ\ðc 1Þ=c
and
ÞT a=ðc1Þ þ b3 l1 \ðc 1Þ=c

a=ðc1Þ
ðS k0 qa0 =T0
or, in terms of T/T0,
ðS k0 qa0 T0b3 ÞðT=T0 Þa=ðc1Þ þ b3 l1 \ðc 1Þ=c:
Exercise 4.17 From the approximate model of the photosphere in Chap. 3, derive
the relation q(T). Then, compare this “radiative relation” with the adiabatic relation
with c = 5/3 and verify that the adiabatic relation is closer to the q(T) relation of the
VALC model (Table 3.1) than the radiative relation.
Just above the top of the convection zone, it is T0 = 5780 K, P0 = 65,144 Ba

and q0 = 1.69 10−7g cm−3, from the Senesi et al. model. Moreover, with the
opacity dominated by the H− photoionizations, we have
k0 ¼ 1:3 1021 in cgs; a ¼ 1:74 and b ¼ 6:76:
Thus, with c− = 5/3 as for a monoatomic perfect gas,
a=ðc 1Þ þ b 3 ¼ 1:74=ð5=3 1Þ þ 6:76 3 ¼ 6:37
and
S k0 q0 a T0b3 ¼ 0:146:
Taking (c − 1)/c = 0.4 and l = 1.27 for a neutral H and He mixture, we get for
the condition for dynamical stability
0:146ðT=5780Þ6:37 =1:27\0:4:
The left hand side rapidly grows with increasing temperature, and overcomes the
right hand side for temperatures (Fig. 4.8)
T [ 6750 K:
Fig. 4.8 Convective stability criterion near the top of the convection zone. Solid line refers to the
ratio (c − 1)/c, which is proportional the adiabatic temperature gradient, and dotted line to the term
that is proportional to the radiative gradient and is described in the text. The radiative gradient is
greater in modulus than the convective gradient for T > 6750 K, which represents the upper
boundary of the convection zone
Therefore, for temperatures above 6750 K, the radiative gradient in modulus

overcomes the convective gradient and convection starts. From Table 4.1 we can
see that in the Senesi et al. model the onset of convection is located slightly higher
at T = 6252 K.
With increasing temperatures below the base of the photosphere, both (c − 1)/c
and l decrease and their reduction strengthens the convective instability. In the
deeper layers of the convection zone, where metal photoionizations dominate the
opacity, it is
k0 ¼ 1:5 1024 in cgs; a ¼ 2 and b ¼ 3:5:
Thus, with c = 5/3,
a=ðc 1Þ þ b 3 ¼ 2=ð5=3 1Þ 3:5 3 ¼ 3:5:
Moreover, at r0 = 0.95 RS, the accurate model gives T0 = 269,595 K,

P0 = 2.72 1011 Ba and q0 = 7.32 10−3g cm−3, hence
S k0 qa0 T0b3 ¼ 412:
With (c − 1)/c = 0.4 and l = 0.61 for a fully ionized H and He gas, the stability
condition becomes
412 ðT=269; 595Þ3:5 =0:61\0:4:
Now, the left hand side decreases with increasing temperature; so that moving
inwards, it becomes smaller than the convective term on the right hand side. In this
simplified model, the temperature for which this occurs is (Fig. 4.9).
T [ 1:96 106 K:
In the Senesi et al. model the bottom of the convection zone is located slightly
deeper, at about r/RS = 0.724 and T = 2.06 106 K (see Table 4.1). Therefore, for
temperatures greater than 2 106 K, the radiative temperature gradient becomes
smaller than the convective temperature gradient in modulus, convection turns off
and the radiative part of the solar interior starts.
4.4.4 Checking the Hydrostatic Equilibrium
Convective motions have their maximum velocities at the top of the convection
zone, where velocities can locally be of the order of 10 km s−1. However, in order
to check the global stability of the hydrodynamic equilibrium we consider the space
averaged speed of such motions over many granules and intergranules, which does
not exceed 1 km s−1. In order to estimate the effect, we compute the dynamic
Fig. 4.9 Convective stability criterion near the bottom of the convection zone. Solid line refers to the
ratio (c − 1)/c, which is proportional the adiabatic temperature gradient, and dotted line to the term
proportional to the radiative gradient. The radiative gradient is smaller in modulus than the convective
gradient for T > 1.96 106 K, which represents the lower boundary of the convection zone
pressure and compare it with the gas pressure. The dynamic pressure is defined as
half the product of the motion squared velocity and the gas density (Fig. 4.10)
Pd ¼ 1=2 qv2
The gas pressure and density at the top of the convection zone are (Table 4.1)
P0 ¼ 6:51 104 Ba
and
q0 ¼ 1:69 107 g cm3 :
For velocities of the order of
v ¼ 1 km s1 ¼ 105 cm s1 ;
we get
Pd ¼ 1=2 qv2 ¼ 1=2 1:69 107 g cm3 ð105 cm s1 Þ2

¼ 0:85 107 g cm3 1010 cm2 s2
¼ 8:5 102 Ba:
Fig. 4.10 The dynamical pressure in a fluid with density q moving with velocity v is due to the
momentum that in average crosses per unit time any fluid section dA in the normal direction n (panel
a). Let dm be the mass that in the time t crosses dA, and h the velocity inclination with respect to n,
then, dm occupies the volume dV = vt cos h dA and the momentum along n is v cos h dm ¼
v cos h qdV ¼ q v cos h vt cos h dA ¼ q v2 cos2 h t dA: The force along n, F, is the momentum per
unit time, that is F ¼ qv2 cos2 h dA and the force per unit area, that is the associated pressure P, is
P ¼ qv2 cos2 h: Because dA has an arbitrary R p orientation, we consider an averaged fluid inclination
by replacing cos2h with the mean ð1=pÞ 0 cos2 h dh: This integral can be computed by applying
the Pythagorean theorem to the right triangle with unity hypotenuse and an angle equal to h, whose
R pcos h2 and sin 2h (panel b);
sides are R p thus,it isR cos
2
h þ sin2 hR ¼ 1: By integrating over h between 0
p p
and p, 0 ðcos h þ sin hÞ dh ¼ 0 dh and 0 cos2 h dh þ 0 sin2 h dh ¼ ½hp0 ¼ p: Panel c shows
Rp Rp
that the two integrals (corresponding to the dashed areas) are equal, 0 cos2 h dh ¼ 0 sin2 h dh:
Rp R p
Then, we can write 2 0 cos2 h dh ¼ p and ð1=pÞ 0 cos2 h dh ¼ 1=2. Therefore, the dynamical
pressure is Pd ¼ 1=2 q v : 2
The dynamic pressure at the top of the convection zone is only a few percent of
the gas pressure and, therefore, convective motions represent only small pertur-
bations that do not substantially affect the global mechanical equilibrium of the
solar gas. Moreover, if the mechanical balance between pressure gradient and
gravity broke, the Sun would be dramatically affected and a large variation of the
solar radius would occur in a short time interval, named the dynamic time scale. As
shown in Chap. 7, the dynamic time scale can be computed from
td ¼ 1=ðGqÞ1=2 ;
where q is the gas density and G = 6.670 10−8 dyn cm2 g−2 is the gravitational
constant. To estimate td for the present Sun we use the average density, hence
td ¼ 1=ðGqÞ1=2 ¼ 1=ð6:67 108 dyn cm2 g2 1:4 g cm3 Þ1=2 ¼ 1=ð9:4 108 s2 Þ1=2
¼ 1=3:1 104 s1 ¼ 3:3 103 s 1 h:
Our daily experience shows that the solar radius does not show any significant
variation over such a short time scale. In conclusion, we are very confident that the
present Sun satisfies with accuracy the condition of the hydrostatic equilibrium.
Exercise 4.18 Starting from the equation of radiation transfer along the line of
sight,
dIk =ds ¼ Sk kk Ik ;
show that for a spherical symmetric Sun the radiation transfer equation in spherical
coordinates takes the form

ldIk =dr þ 1 l2 =r dIk =dl ¼ Sk kk Ik ;
with Ik = Ik(r,h) and l = cos h.

Exercise 4.19 Show that because of the high opacity in the solar interior, the term
[(1 − l2)/r] dIk/dl can be neglected in the radiation transfer equation for a spherical
symmetric Sun, which then reduces to that for plane parallel layers with the radial
distance r replacing the height z.
Exercise 4.20 A layer where all the energy is transported by radiation is said to be
in radiative equilibrium. Under this condition the total emission must equal the total
absorption. Show that in a plane parallel slab in radiative equilibrium the radiation
flux is constant. [Hint. Integrate the radiation transfer equation over all directions
and wavelengths.]
Exercise 4.21 Estimate the temperature where the H− photoionizations peak.

Exercise 4.22 Estimate the temperature where the photoionizations of He II (he-
lium ion with one electron) peak, using the He II ionization potential of 54.42 eV.
Exercise 4.23 Assuming that H and He are completely ionized at the Sun’s center
and neglecting metals, determine the electron number density.
Exercise 4.24 Using Table 4.1, estimate the radiative diffusion coefficient at the
top and bottom of the convection zone.
Exercise 4.25 Estimate the total radiation energy density and photon number
density at the Sun’s center.
Exercise 4.26 The granulation intensity contrast is defined by

Ck ¼ 2 Ikg Iki = Ikg þ Iki
where Igk and Iik are the maximum and minimum intensities in the granule and
intergranule, respectively. Use the Wien approximation of the Planck function to
show that it is

Ck hc= kB kT 2 Tg Ti ;
where Tg − Ti is the temperature difference between granule and intergranule and

T is the mean temperature.
Exercise 4.27 Assuming that the granulation contrast is Ck = 0.4 at k = 0.5 lm
and T = 6000 K, (see Exercise 4.26), determine the temperature difference between
granule and intergranule.
Exercise 4.28 Using Table 4.1, estimate the radial distances where the opacity due
to the H and He photoionizations peak. Compare these distances with those where
the ratio (c − 1)/c has two minima (Fig. 4.6, first panel).
Chapter 5
The Radiation Zone
Abstract As discussed in the previous chapter, below 0.7 solar radii and at tem-
peratures higher than 2 106 K, convection switches off and radiation comes back
to transport the energy outwards. As we will discuss in detail in the next chapter,
the solar energy is produced by nuclear fusion reactions that take place in the core,
which is the hottest, central part of our star and extends up to about ¼ of the solar
radius. Thus, in the layers that lie above the core and below the convection zone,
between about 0.25 Rs and 0.7 Rs, nuclear reactions vanish and the energy transfer
is controlled by radiation as in the photosphere. This region is called the radiation
zone.
Actually, unlike the photosphere, with temperatures of about 6000 K and the
maximum of the radiation intensity at k 0.5 lm, in the visible wavelength range,
the radiation zone is characterized by temperatures about thousand times higher,
with radiation concentrated at wavelengths thousand times smaller ðk 0:5 nmÞ
and typical of “X rays.”
In this chapter we first describe the physical conditions that characterize the
radiation zone and, then, construct and evaluate an approximate analytical model of
its structure.
5.1 The Physical Conditions
Let us collect the equations that determine the physical conditions of the radiation
zone.
Exercise 5.1 Verify that, according to Wien’s law, the Planck function at
T = 2 106 K peaks in the X rays wavelength range.

172 5 The Radiation Zone
5.1.1 Temperature
As for the photosphere we assume that in the radiation zone there is no energy
source or well, but only radiation energy transfer with a net flux of photons in the
outward direction. Then, the temperature gradient is determined by the radiative
diffusion equation.
dT=dr ¼ 3LkR =ð64pr 2 rT 3 Þ:
Here the radiation flux F, which is the energy per second, per unit area, and decreases
with the inverse of the squared radius, has been expressed in terms of the luminosity,
L = 4pr2F, that is, the total energy per second crossing each spherical layer of radius
r. The luminosity in the radiation zone is constant and equal to the surface value.
L ¼ LS ¼ 4pR2S rTe4 ;
As anticipated in Sect. 4.1, the Rosseland mean opacity kR can be approximated

by power laws in density and temperature
kR ¼ k0 qa T b ;
where k0, a and b are constant. Let us consider the opacity power laws in some
more detail.
5.1.2 The Opacity Power Laws
In a Thomson scattering a free electron absorbs a photon that is then reemitted in a

different direction. This process can be treated according to classical mechanics.
Thus, the incident radiation is modeled by a monochromatic plane wave, which sets
the electron in oscillatory motion, and the emitted radiation at large distances from
the oscillating free electron is assumed to be a spherical wave. With this approach,
it is possible to show that the absorption coefficient of the Thomson scattering is
kkT ¼ rT Ne ;
where Ne is the electron number density,
rT ¼ 8pre2 =3
is the scattering cross section, and

re ¼ e= me c2 ¼ 2:82 103 pm
is the classical radius of an electron. In LTE Ne depends on the ionization equilibria

of all the elements that can contribute electrons, according to the Saha equations,
5.1 The Physical Conditions 173
but when H and He are fully ionized, as it occurs in the radiation zone, we can use
the approximated formula
Ne ¼ qð1 þ XÞ=ð2mu Þ;
where X is the H mass fraction (see Exercise 5.2). Because rT is independent of

wavelength, the Rosseland mean opacity for the Thomson scattering equals the
monochromatic opacity, thus
kRT ¼ kkT ¼ rT Ne ¼ 4pre2 qð1 þ XÞ=ð3mu Þ

¼ 0:200 qð1 þ XÞ cm1 :
The opacity power laws for b-f and f-f transitions were first derived by the Dutch
physicist Hendrick Kramers (1894–1952). In a f-f absorption an electron passes
close to an atom or ion and is accelerated absorbing a photon. Thus, if the electron
velocity is v1 before the absorption, and v2 after, the absorbed photon energy is

hm ¼ me v22 v21 =2:
In a b-f transition, an atom or ion is photoionized by a photon having an energy
hm ¼ Ei þ me v21 =2
where Ei is the atomic ionization potential and me v21 =2 is the kinetic energy of a
bound electron that becomes free; hence, m Ei/h.
In order to understand the temperature and density dependences that appear in
the Kramers’ opacity laws, we consider a f-f absorption. Neglecting metals, which
amounts to 0.02 in mass, free electrons interact with H and He ions, hence the f-f
monochromatic absorption coefficient is
kkff ¼ rHff Ne NH þ rHeff Ne NHe ;
where rHff and rHeff are the cross sections per unit particle for an electron interacting
with a H and a He ion, respectively, and Ne, NH and NHe are the electron, H and He
ion number densities, respectively. Denoting, as usually, with X the H mass fraction
and with Y that of He, the corresponding number densities are
NH ¼ qX=mu
and
NHe ¼ qY=ð4mu Þ;
where q is the gas density and mu the atomic mass unit. Exercise 5.2 shows that
Ne ¼ qð1 þ XÞ=ð2mu Þ:
Kramers has shown that the cross sections are proportional to rff Z 2 k3 , where Z
is the charge number of the ion, k the photon wavelength and rff is independent of
Z and k. Thus, because Z = 1 for H and Z = 2 for He,
rHeff ¼ 4rHff ¼ 4rff k3
and
kkff ¼ rHff Ne NH þ rHeff Ne NHe
¼ rff k3 Ne NH þ 4rff k3 Ne NHe
¼ rff k3 Ne ðNH þ 4NHe Þ
¼ rff k3 ½qð1 þ XÞ=ð2mu Þ½qX=mu þ 4qY=ð4mu Þ;
and, finally,
kkff ¼ rff k3 q2 ð1 þ XÞðX þ YÞ=ð2m2u Þ:
The interaction probability is proportional to the time during which the electron
and ion are sufficiently close, which in turn is inversely proportional to the electron
velocity. Thus, using the most probable thermal velocity
v ¼ ð2kB T=me Þ1=2 ;
we expect that thee cross section scales as
rff 1=v T 1=2 :
Therefore, we get
kkff k3 q2 T 1=2 :
When performing the Rosseland mean the factor k3 contained in kkff gives a
factor T−3 in the mean. Thus,
kR T 3 q2 T 1=2 q2 T 7=2 :
Eventually, inserting the dependence on the chemical composition, we get
kRff ¼ k0 ð1 þ XÞðX þ YÞq2 T 7=2 :

with
k0 = 3.68 1022 in cgs (Bisnovatyi-Kogan, 2000).
A similar formula also holds for b-f absorptions
kRbf ¼ k0 ð1 þ XÞZq2 T 7=2 ;
where
k0 ¼ 4:34 1025 in cgs:
Note that in the b-f case, the factor Z replaces (X + Y), because, when H and He
are fully ionized, the relevant contribution is due to metals. In both power laws
sometimes it is added a factor close to unity, named the Gaunt factor (after the
English physicist John Arthur Gaunt, 1904–1944), in order to account for quantum
mechanics effects in the cross sections.
In the radiation zone the opacity is dominated by photoionizations of metals,
with free-free absorptions by electrons and Thomson scattering playing only a
minor role. Thus, in the radiative diffusion equation we get
kR ¼ k0 ð1 þ XÞZq2 T 7=2 :
However, it is worthwhile to emphasize that the opacity laws represent an useful

analytical approximation, but in order to construct accurate solar and stellar models,
opacity tables and numerical codes that take many opacity sources into a detailed
account are required (e.g. http://opalopacity.llnl.gov/opal.html).
Worked Exercise 5.2 Determine the electron number density in the radiation zone.
The number density of H nuclei is
NH ¼ qX=mu
and the number density of He nuclei is
NHe ¼ qY=ð4mu Þ;
where q is the gas density, mu the atomic mass unit and X and Y are the mass
factions of hydrogen and helium, respectively. Analogously, the number density of
the nuclei having mass fraction Zi and atomic weight li is
Ni ¼ qZi =ðli mu Þ:
In the radiation zone, H and He are completely ionized and, then, give a number
of free electrons equal to qX/(mu) and
2qY=ð4mu Þ ¼ qY=ð2mu Þ;
respectively. For the metals, which are partly ionized, a reasonable approximation is
that each element gives a number of free electrons equal to half its atomic weight,
that is,
ðli =2ÞqZi =ðli mu Þ ¼ qZi =ð2mu Þ:
P the total electron contribution from metals is estimated to qZ/(2mu), where

Thus,
Z = iZi is the heavier elements mass fraction.
Therefore, the electron number density is
Ne ¼ qX=mu þ qY=ð2 mu Þ þ qZ=ð2mu Þ

¼ ðq=mu ÞðX þ Y=2 þ Z=2Þ:
Because X + Y + Z = 1, we get
X þ Y=2 þ Z=2 ¼ X þ ð1 XÞ=2 ¼ ð1 þ XÞ=2;
and the electron number density finally is
Ne ¼ qð1 þ XÞ=ð2mu Þ:
Exercise end.
Exercise 5.3 Using the approach of Exercise 5.2, show that in the radiation zone
the mean molecular weight can be approximated by
l ¼ 1=ð2X þ 3Y=4 þ Z=2Þ:
5.1.3 Pressure
The pressure gradient is fixed by the hydrostatic balance
dP=dr ¼ qGM=r 2 :
On the right hand side gravity is
g ¼ GM=r 2
where M is the mass contained in a sphere of radius r and responsible for the
gravitational attraction on the overlying layers. As in the convection zone gravity
varies with radius; however in the radiation zone the situation is more complex,
because M cannot be taken as almost constant but appreciably decreases with

decreasing radius.
5.1.4 State Equation
Pressure, density and temperature are related by the state equation
P ¼ RqT=l:
In the radiation zone the average molecular weight is close to l = 0.61, the value
that we have computed for the inner convection zone considering a fully ionized
mixture of 90% hydrogen and 10% helium and neglecting metals.
The equations for the temperature and pressure gradients, the opacity law and the
state equation are 4 constrains for the 5 quantities T, P, kR, q and M; we need yet
another condition if we want to determine all the parameters as a function of r.
5.1.5 Mass Conservation
The definition of density as the ratio between the mass and its volume provides the
necessary additional condition, which in differential terms is
q ¼ dM=dV:
The mass increment dM is contained within a spherical corona of thickness

dr and radius r, which has volume dV = 4pr2dr (Fig. 4.1). Thus, the density is
q ¼ dM=ð4p r 2 drÞ
and the mass increment is
dM ¼ 4p r 2 q dr:
This condition is a mass balance involving all the mass inside the elementary
corona of radius r: this mass is conserved and no source, well or flow of matter is
present.
Actually, the Sun has a mass loss due to the solar wind, which will be discussed
in Chap. 7, and to the radiation emitted into space, which is equivalent to a mass
because of the mass-energy relation. In total, the Sun loses about 5 109 kg per
second, which is a completely negligible amount compared to its total mass of
almost 2 1030 kg. Moreover, for as long as the Sun will last in its present stage,
which corresponds to a lifetime of the order of 1010 year as we shall see in the next
chapter, the total mass loss is less than one thousandth of the solar mass. Then,
taking the present solar mass as a constant is a very good approximation.
Exercise 5.4 Verify that during the Sun’s lifetime the mass loss is less than 10−3
M S.
Exercise 5.5 Estimate the mass within a spherical corona that has width Dr =
500 km and is located just below the photosphere.
Exercise 5.6 Find the mean solar density by integrating the mass conservation
equation.
In summary, in the radiation zone the 3 unknown T, P and M as functions of
r are constrained by the 3 equations of energy transport, hydrostatic equilibrium and
mass conservation
dT=dr ¼ 3LS kR =ð64pr 2 rT 3 Þ;

dP=dr ¼ qGM=r 2 ;
dM ¼ 4pr 2 q dr:
Moreover, we have related the two more parameters q and kR to T and P by

means of the state equation
P ¼ RqT=l
and
a Kramers’ opacity law
kR ¼ k0 q2 T 3:5 :
5.2 The Radiation Zone Model
To find our approximate model we start considering the diffusion equation
dT=dr ¼ 3LS kR =ð64pr 2 rT 3 Þ:
The right hand side of the equation is a ratio of temperature to length. Then,
multiplying numerator and denominator by T
dT=dr ¼ 3TLS kR =ð64pr 2 rT 4 Þ;
we can define a scale height as

5.2 The Radiation Zone Model 179
HT ¼ 64pr 2 rT 4 =ð3LS kR Þ
and get
dT=dr ¼ T=HT :
The temperature scale height HT varies with r as HT * r2T4/kR. We expect that

the increase of T4 with decreasing r roughly compensates the r2 and k−1R decreases.
In the accurate model HT varies by almost a factor 2 in the radiation zone, with a
mean value of 1.77 HT0 , where HT0 is the value of HT at the base of the convection
zone. In our simplified model we take HT constant and equal to the mean value.
Note that we assume the base of the convection zone as the initial radial distance r0
and take the boundary values of physical parameters from the accurate model. Thus,
we get
HT ¼ 2:50 105 km:
Exercise 5.7 Using kR0 ¼ 3:22 cm1 , T0 ¼ 2:06 106 K, from the Senesi et al.
model at r0 = 0.7234460 RS, and Ls = 3.84 1033 erg s−1, show that
HT = 2.50 105 km.
With a constant scale height the temperature grows exponentially with
decreasing radius
T ¼ T0 eðr0 rÞ=HT :
Let us now consider the equation of hydrostatic equilibrium
dP=dr ¼ qGM=r 2
and, in particular, the pressure scale height. Substituting the density from the state
equation
q ¼ lP=ðRTÞ;
the hydrostatic equilibrium becomes

dP=dr ¼ lPGM= RTr 2
and
dP=dr ¼ P=HP ;
with the pressure scale height given by

HP ¼ RTr2 =ðlGMÞ:
We assume that the P scale height is constant. Actually, the inward decrease of
r2 is roughly balanced by the simultaneous growth of the T/M ratio, in which the
temperature increases and the mass decreases. The accurate model of Senesi et al.
shows that the actual scale height variation is less than 20%, and
HP ¼ 0:98HP0 ¼ 5:29 104 km:
is a good mean value. With a constant scale height, the hydrostatic pressure grows
exponentially with decreasing radius starting from the value P0 at the reference
radial distance r0.
P ¼ P0 eðr0 rÞ=HP :
Exercise 5.8 Using the values P0 = 4.73 1013 Ba, T0 = 2.06 106 K, and
M0 = 0.979 MS = 1.95 1033 g, taken from the accurate model at
r0 = 0.7234460 RS, verify that HP = 0.98 HP0 = 5.29 104 km.
From the state equation we can find that the density also grows exponentially
with decreasing radius
q ¼ q0 eðr0 rÞ=Hq ;
with a density scale height given by
Hq ¼ 1=ð1=HP 1=HT Þ:
Exercise 5.9 Derive the density scale height from the state equation inserting the
exponential temperature and pressure variations given in the text.
Using the values given above for HT and HP, we get
Hq ¼ 6:71 104 km:
Exercise 5.10 Verify the value for Hq given in the text.

Thus, in the approximate model, T, P and q exponentially increase toward the
Sun’s center, with the scale heights such that
HP \Hq \HT ;
which means that P has the fastest growth, followed by q and T.

In order to complete the model we have to determine the mass as a function of
radius by solving the mass conservation equation
dM ¼ 4pr 2 q dr:
Substituting the density solution
dM ¼ 4pr 2 q0 eðr0 rÞ=Hq dr;
and integrating between r0 and r, which is any level in the radiation zone, we get
ZM Zr
0
dM ¼ 4pr 2 q0 eðr0 r Þ=Hq dr 0 :
M0 r0
On the left hand side integration is immediate. On the right hand side we take the
constant terms out of the integral
Zr
0
r0 =Hq
M M0 ¼ 4pq0 e r 2 er =Hq dr 0 :
r0
In the integral, we make the substitution
x ¼ r 0 =Hq
Thus,
0
r 0 ¼ Hq x; r 02 ¼ Hq2 x2 ; er =Hq ¼ ex ; dr 0 ¼ Hq dx
and
Zr Zx
02 r 0 =Hq 0
r e dr ¼ Hq2 x2 ex Hq dx
r0 x0
Zx
¼ Hq3 x2 ex dx;
x0
where x0 = r0/Hq. Since it is

Z
x2 ex dx ¼ ðx2 þ 2x þ 2Þ ex ;
R
Exercise 5.11 Show that x2 ex dx ¼ ðx2 þ 2x þ 2Þ ex .
The integral becomes
Zr
0
r 02 er =Hq dr 0 ¼ Hq3 ½ðx2 þ 2x þ 2Þ ex xx0
r0
¼ Hq3 ½ðx20 þ 2x0 þ 2Þ ex0 ðx2 þ 2x þ 2Þ ex
Expressing x in terms of r and substituting the result into the mass conservation
equation, we get
M Mr0 ¼ 4pq0 er0 =Hq Hq3 ½ðr02 =Hq2 þ 2 r0 =Hq þ 2Þ er0 =Hq
ðr 2 =Hq2 þ 2r=Hq þ 2Þ er=Hq
and
M ¼ M0 þ 4pq0 Hq3 ½ðr02 =Hq2 þ 2r0 =Hq þ 2Þ eðr0 r0 Þ=Hq

ðr 2 =Hq2 þ 2r=Hq þ 2Þ eðr0 rÞ=Hq :
Finally, because er0 r0 =Hq ¼ e0 ¼ 1 and q0 eðr0 rÞ=Hq ¼ q, the mass variation
becomes
M ¼ M0 þ 4pHq3 ½ðr02 =Hq2 þ 2 r0 =Hq þ 2Þ q0 ðr 2 =Hq2 þ 2 r=Hq þ 2Þ q:
Starting from the value M0 = 0.979 MS at the base of the convection zone, mass
decreases inwards according to the negative term on the right hand side, which is
proportional to the exponentially decreasing q.
Our approximate model of the radiation zone is compared to the accurate model in
Fig. 5.1. The approximate model, based on exponential variations and constant scale
heights for temperature, pressure and density, appears to be rather satisfactory for T,
P, q and M, as the first four panels show. The fifth panel compares the accurate
Rosseland mean opacity with a Kramers’ opacity law kR ¼ kR0 q2 T 3:5 , where q and
T are taken from the accurate model and the constant kR0 has been fitted to the
accurate opacity at r0. The approximate opacity increases inwards faster than the
accurate opacity to become a factor 10 larger at the bottom of the radiation zone.
Finally, the sixth panel shows the variations of the pressure and temperature
scale heights in the accurate model, which have been discussed above in order to
get the mean scale heights for the approximate model.
The radiation zone model of Senesi et al. is also listed in Table 5.1. The average
molecular weight, not listed in the table, is very close to the value 0.61 of a mixture
with 90% hydrogen and 10% helium, both completely ionized.
Note that the luminosity starts to slightly differ from the surface value just at the
bottom of the radiation zone. There, at 0.25 solar radii, the temperature exceeds
Fig. 5.1 The radiation zone model in scale of the solar radius fraction. In the order from top to
bottom the first five panels report the decimal logarithms of the temperature, pressure and density,
the mass fraction and the decimal logarithm of the Rosseland mean opacity. Red lines refer to the
model of Senesi et al. and solid lines to our approximate model. Last panel shows the temperature
(solid line) and pressure (dotted line) scale heights in the accurate model; both the heights are
normalized to their respective values at the base of the convection zone


Table 5.1 The radiation zone model of Senesi et al.
186
r/RS M/MS T, K P, Ba q, g cm−3 kR, cm−1 L/LS

7.248E−001 9.791E−001 2.050E+006 4.652E+013 1.640E−001 3.226E+000 1.000E+000
7.159E−001 9.774E−001 2.137E+006 5.194E+013 1.762E−001 3.261E+000 1.000E+000
7.075E−001 9.758E−001 2.215E+006 5.762E+013 1.891E−001 3.340E+000 1.000E+000
6.923E−001 9.726E−001 2.346E+006 6.929E+013 2.157E−001 3.556E+000 1.000E+000
6.778E−001 9.692E−001 2.464E+006 8.246E+013 2.450E−001 3.817E+000 1.000E+000
6.644E−001 9.659E−001 2.570E+006 9.670E+013 2.760E−001 4.004E+000 1.000E+000
6.521E−001 9.626E−001 2.665E+006 1.120E+014 3.085E−001 4.217E+000 1.000E+000
6.406E−001 9.593E−001 2.752E+006 1.285E+014 3.426E−001 4.450E+000 1.000E+000
6.298E−001 9.560E−001 2.834E+006 1.460E+014 3.783E−001 4.698E+000 1.000E+000
6.197E−001 9.526E−001 2.910E+006 1.647E+014 4.155E−001 4.953E+000 1.000E+000
6.102E−001 9.493E−001 2.983E+006 1.846E+014 4.543E−001 5.216E+000 1.000E+000
6.012E−001 9.460E−001 3.052E+006 2.056E+014 4.946E−001 5.481E+000 1.000E+000
5.921E−001 9.424E−001 3.123E+006 2.295E+014 5.397E−001 5.769E+000 1.000E+000
5.763E−001 9.358E−001 3.249E+006 2.777E+014 6.280E−001 6.318E+000 1.000E+000
5.620E−001 9.291E−001 3.366E+006 3.308E+014 7.223E−001 6.873E+000 1.000E+000
5.488E−001 9.225E−001 3.476E+006 3.889E+014 8.223E−001 7.434E+000 1.000E+000
5.367E−001 9.158E−001 3.582E+006 4.521E+014 9.283E−001 8.000E+000 1.000E+000
5.254E−001 9.092E−001 3.682E+006 5.206E+014 1.040E+000 8.571E+000 1.000E+000
5.149E−001 9.025E−001 3.779E+006 5.945E+014 1.158E+000 9.147E+000 1.000E+000
5.050E−001 8.959E−001 3.872E+006 6.739E+014 1.281E+000 9.728E+000 1.000E+000
4.957E−001 8.892E−001 3.962E+006 7.589E+014 1.410E+000 1.031E+001 1.000E+000
4.869E−001 8.826E−001 4.049E+006 8.497E+014 1.545E+000 1.091E+001 1.000E+000
4.786E−001 8.759E−001 4.134E+006 9.463E+014 1.686E+000 1.150E+001 1.000E+000
4.707E−001 8.693E−001 4.216E+006 1.049E+015 1.832E+000 1.211E+001 1.000E+000
(continued)
5 The Radiation Zone
r/RS M/MS T, K P, Ba q, g cm−3 kR, cm−1 L/LS
4.566E−001 8.565E−001 4.370E+006 1.264E+015 2.130E+000 1.328E+001 1.000E+000
4.431E−001 8.432E−001 4.523E+006 1.512E+015 2.463E+000 1.452E+001 1.000E+000
4.306E−001 8.299E−001 4.671E+006 1.786E+015 2.819E+000 1.578E+001 1.000E+000
4.191E−001 8.166E−001 4.814E+006 2.088E+015 3.198E+000 1.705E+001 1.000E+000
4.083E−001 8.033E−001 4.953E+006 2.419E+015 3.600E+000 1.834E+001 1.000E+000
3.982E−001 7.900E−001 5.088E+006 2.778E+015 4.026E+000 1.965E+001 1.000E+000
3.887E−001 7.767E−001 5.221E+006 3.169E+015 4.476E+000 2.097E+001 1.000E+000
5.2 The Radiation Zone Model
3.797E−001 7.634E−001 5.350E+006 3.591E+015 4.950E+000 2.231E+001 1.000E+000

3.712E−001 7.501E−001 5.478E+006 4.046E+015 5.448E+000 2.366E+001 1.000E+000
3.630E−001 7.368E−001 5.604E+006 4.535E+015 5.971E+000 2.503E+001 9.999E−001
3.552E−001 7.235E−001 5.727E+006 5.059E+015 6.518E+000 2.642E+001 9.999E−001
3.478E−001 7.103E−001 5.850E+006 5.619E+015 7.089E+000 2.783E+001 9.998E−001
3.387E−001 6.933E−001 6.004E+006 6.389E+015 7.855E+000 2.964E+001 9.997E−001
3.253E−001 6.667E−001 6.243E+006 7.727E+015 9.139E+000 3.256E+001 9.995E−001
3.128E−001 6.401E−001 6.479E+006 9.233E+015 1.053E+001 3.556E+001 9.991E−001
3.010E−001 6.135E−001 6.714E+006 1.092E+016 1.202E+001 3.864E+001 9.986E−001
2.898E−001 5.869E−001 6.949E+006 1.281E+016 1.362E+001 4.180E+001 9.978E−001
2.795E−001 5.612E−001 7.177E+006 1.483E+016 1.528E+001 4.494E+001 9.966E−001
2.717E−001 5.412E−001 7.355E+006 1.654E+016 1.664E+001 4.740E+001 9.951E−001
2.629E−−001 5.178E−001 7.565E+006 1.872E+016 1.833E+001 5.034E+001 9.926E−001
2.549E−001 4.960E−001 7.763E+006 2.093E+016 1.998E+001 5.313E+001 9.897E−001
2.426E−001 4.616E−001 8.080E+006 2.480E+016 2.276E+001 5.773E+001 9.836E−001
In the order the columns report the solar radius fraction r/RS, mass fraction M/MS, temperature T, pressure P, density q, Rosseland mean opacity kR and
luminosity fraction L/LS
187
8 106 K, the density 20 g cm−3 and the pressure 2 1016 Ba. In these condi-
tions the probability that two protons collide and merge into a heavier nucleus,
overcoming the Coulomb repulsion due to their positive electric charge, is no
longer negligible. This process is the first of a sequence that starting from 4 protons
eventually creates a helium nucleus and simultaneously transforms part of the
proton masses in energy. The seat of these nuclear fusion reactions is the central
part, or solar core, which is the subject of the following chapter.
Exercise 5.12 Verify the numerical value of the constant in the formula for the
Thomson scattering
kRT ¼ 4pre2 q ð1 þ XÞ=ð3mu Þ ¼ 0:200 qð1 þ XÞ cm1 :
Exercise 5.13 Using the approximate formulas given in the text and data from
Table 5.1, estimate the opacities due to b-f., f-f absorptions and Thomson scattering
at the top and bottom of the radiation zone.
Exercise 5.14 Verify that the Rosseland mean opacity of a sum of two
monochromatic opacities from different sources is not equal to the sum of the two
individual Rossend mean opacities.
Exercise 5.15 Using Table 5.1, verify that the mean molecular weight in the
radiation zone is close to l = 0.61.
Exercise 5.16 Apply the Schwarzschild criterion given in Chap. 4 in order to
verify that the bottom of the radiation zone is stable.
Exercise 5.17 Verify that the mass loss due to the solar luminosity amounts to
about 4 109 kgs−1.
Chapter 6
The Core
Abstract Our feeling that the Sun always shines in the same way is confirmed by
observations and theory: the solar luminosity now is constant over time at less than
a percent, and will retain unchanged value, within a few tenths, for billions of years.
The luminosity is the energy per second that the Sun loses into space; thus, its
constancy over time suggests that inside the Sun there is a source that produces
exactly the same amount of energy lost per second. The energy source is provided
by the fusion reactions of hydrogen into helium, primarily through the
proton-proton (pp) cycle. The hydrogen burning occurs in the hot central part of the
Sun, which is called the core and is the subject of this chapter.
In the following we turn back to the nuclear reactions of the first branch of the pp
cycle, which have been introduced in Chap. 1. We treat in detail the cross section of
the chain first step, in which two protons fuse into a deuteron. This reaction is the
slowest in the cycle and involves both a collision of two protons and a b+ decay of a
proton into a neutron. We also show how to take properly into account the mean
over the Maxwellian proton energy distribution, in order to determine the nuclear
energy production rate and, hence, the luminosity. Finally, we construct an
approximate model of the solar core and compares it with the accurate model of
Senesi et al.
6.1 Cross Sections
6.1.1 Solar Nuclear Reactions
The fusion of hydrogen into helium is the mechanism that generates the solar
energy. The most frequent reactions realizing this process in the Sun constitute the
first branch of the proton–proton chain (ppI). An introduction to the ppI chain has
been given in Chap. 1, together with a summary of the basic nuclear physics. In the

190 6 The Core
first step two protons merge into a nucleus of deuterium, with the emission of a
positron and an electronic neutrino
p þ p ! 2 H þ e þ þ ve þ 1:442 MeV:
The energy release of the reaction, given on the right end side in MeV, includes
the average energies of the resulting particles and emitted photons. The positron
quickly annihilates with an electron, releasing two c rays
e þ þ e ! 2c:
In the second step, the deuteron interacts with a third proton to form tritium with
a photon emission
2
H þ p ! 3 He þ c þ 5:494 MeV:
The first two steps must occur twice to produce the two 3He nuclei that the
subsequent step requires. In this last step, two 3He nuclei combine to form an a
particle and release two protons
3
He þ 3 He ! 4 He þ 2p þ 12:86 MeV:
Thus, in a cycle 4 protons and 2 electrons transform themselves into a helium

nucleus and a total amount of 26.732 MeV is released as energies of the produced
particles and emitted c rays. According to the mass—energy relation, the total
energy release equals the squared light velocity multiplied by 0.028mp, that is, the
mass defect of an a particle with respect to the sum of four proton and two electron
masses.
Worked Exercise 6.1. The pp chains.
The ppI, ppII and ppIII chains are three different reaction sequences that allow
the formation of helium from hydrogen. The ppI chain is the most frequent chain
occurring in the Sun and has been described in the text. The ppI occurrence fre-
quency is about 85%, but this value depends on the solar model and, in particular,
on the temperature structure. After that 3He and 4He are formed in a ppI chain, the
following reactions can occur giving rise to the ppII chain, with a frequency of
about 15%,
3
He þ 4 He ! 7 Be þ c þ 1:586 MeV:
7
Be þ e ! 7 Li þ ve þ c þ 0:864 MeV:
7
Li þ p ! 2 4 He þ 17:346 MeV:
The ppIII chain (0.02% frequency) occurs with the following reactions, after that
Be7 is formed in a ppII chain,
6.1 Cross Sections 191
7
Be þ p ! 8 B þ c þ 0:137 MeV:

8
B ! 8 Be þ e þ þ ve þ 15:078 MeV:
8
Be ! 24 He þ 2:995 MeV,
where the symbol *denotes an unstable nucleus.

Exercise end.
Worked Exercise 6.2 The CNO cycle.
The CNO cycle is an alternative to the pp cycle that also realizes the H fusion in
He. According to current models the CNO cycle provides only 1.2% of the total
solar energy.
The nuclear reactions of the CNO cycle are:
12
C þ p ! 13 N þ c þ 1:944 MeV:
13
N ! 13 C þ e þ þ ve þ 2:220 MeV
13
C þ p ! 14 N þ c þ 7:551 MeV:
14
N þ p ! 15 O þ c þ 7:297 MeV
15
O ! 15 N þ e þ þ ve þ 2:754 MeV:
15
N þ p ! 4 He þ 12 C þ 4:966 MeV:
We note that C, N and O isotopes act as catalysts and their number densities are
unaffected by the cycle reactions.
The total energy release to produce an a particle by a CNO cycle is again
26.732 MeV, as for the pp cycle.
Exercise end.
Exercise 6.3 Verify that the total energy release per formation of a a particle is
26.732 MeV in the CNO cycle and the ppII and ppIII chains.
In addition to the amount of energy released in a single fusion process, we need
to know the probability that the process occurs, hence, its cross section. Moreover,
since the number of reactions per second is a function of the number densities of the
interacting particles, we have also to account for the Maxwellian distribution of
their kinetic energies. All these factors together allow determining the total energy
produced per second and per unit volume, that is, the nuclear energy production
rate.
Let us first consider the cross sections and, in particular, that of the first reaction
of the ppI chain, in which two protons collide and merge into a deuterium nucleus.
192 6 The Core
6.1.2 The First Reaction of the ppI Chain
The result of a proton-proton collision depends on the proton kinetic energies and
interacting forces. As long as the two protons are sufficiently far, they repel each
other, because they are both positive charges, thus subject to the Coulombiam
electrostatic repulsion force, which has a long range of action. For simplicity, let us
assume that a proton approaches the other proton that is at rest. Thus, the moving
proton has a constant total energy
Etot ¼ mp v2 =2 þ V;
which is the sum of the kinetic energy, mpv2/2, and the positive potential energy
V ¼ e2 =r;
where r is the distance between the two protons and
e ¼ 4:80325 1010 statC
is the proton charge in statCoulomb, the electrostatic unit (esu) described in

Exercise 6.4.
Worked Exercise 6.4 The esu unit of electric charge.
The esu unit of electric charge is derived from the Coulomb law for the elec-
trostatic force in the form
F ¼ q1 q2 =r 2 ;
where q1 and q2 are the interacting electric charges, r their distance, and the sign
minus implies that the repulsive forces acting between charges which are both
positive or both negative, are negative. Thus, the esu charge unit, or statCoulomb
(symbol statC), is fixed by those of force and distance, that is
0:5
1 statC ¼ 1 dyn cm2 ¼ 1ðerg cmÞ0:5 :
In particular, for the proton charge we get
e ¼ 4:80325 1010 statC:
The electromagnetic unit (emu) of the electric charge is named Coulomb

(symbol C) and is derived from the electric current density, J, which is measured in
Ampère (symbol A, after André-Marie Ampère, French physicist, 1775–1836). 1 C
is the amount of charge that a current of 1 A transports in one second, that is,
1 C ¼ 1 A 1 s:
In particular, for the proton charge we get
e ¼ 1:60217662 1019 C:
Thus,
1 statC ¼ 3:3356 1010 C:
Exercise end.
The kinetic energy decreases with decreasing r while the potential energy increases.
However, the potential energy reaches a maximum, Vm, at a minimum distance, rm,
where the two protons are subject to the strong interaction, which is attractive and at
short range of action. If the moving proton goes beyond Vm, the two protons merge
in a nucleus. However, only two protons cannot coexist in a nucleus for a long time.
In order to get a stable nucleus of deuterium, one of them has to decay into a
neutron. The proton decay is a rare event under control by the weak interaction. If it
occurs, the resulting deuterium nucleus represents a potential well for the initially
moving proton, with a negative potential energy corresponding to the binding
energy between the proton and neutron. The variation of the proton potential energy
with distance is plotted in Fig. 6.1.
We expect that the edge of the deuteron potential well occurs at a distance, rm,
larger than 2 Fm = 2 10−13 cm, which roughly is the deuteron size. To this
amount it must be added the distance over which the strong interaction acts, which
Fig. 6.1 Schematic variation of the potential energy V of two protons with their distance
r. Denoted by rm the distance corresponding to the maximum potential energy, Vm, the two protons
are free for r > rm and bound into a nucleus for r < rm
194 6 The Core
is about 1.5 Fm. Thus, we approximately take rm = 3.5 Fm. At this distance the
maximum repulsive potential energy is
2
Vm ¼ e2 =rm ¼ 4:80 1010 erg0:5 cm0:5 =3:5 1013 cm
¼ 2:31 1019 =3:5 1013 erg ¼ 0:66 106 erg:
Since the number of ergs corresponding to 1 eV is

EeV ¼ 1:6022 1012 erg/eV ¼ 1:6022 1012 erg/ 106 MeV
¼ 1:6022 106 erg/MeV,
we get

Vm ¼ 0:66 106 erg ¼ 0:66 106 erg/ 1:60 106 erg=MeV ¼ 0:41 MeV:
In order to allow a proton to fall into the nuclear potential well, and the fusion to
occur, it is necessary that the proton overcomes the repulsive potential barrier Vm.
With an example derived from the macroscopic mechanics, the proton might be
compared with a billiard ball that has to climb a slope, which is a gravitational
potential barrier, before falling into a hole, which is a gravitational potential well
(Fig. 6.2).
Fig. 6.2 The gravitational potential energy as function of the distance from the center of a hole is
a gravitational analogue of the potential energy of a proton pair. To get the same variation as in
Fig. 6.1, we take a slope height, h, such that the product hr = constant, that is, a hyperbolic
profile. A sphere with mass m and initial kinetic energy E oversteps the slope with top height hm
and goes into the hole only if E = ½mv2 > mghm
While approaching the hole the potential energy of the ball increases and the
kinetic energy decreases, but the total energy, which is the sum of the kinetic and
potential energies, remains constant. Thus, according to classical mechanics, the
ball enters the hole only if its initial kinetic energy is higher than the gravitational
potential energy maximum, that is,
E ¼ mv2 =2 [ mghm ;
where m and v are the mass and initial velocity of the ball, respectively, g the
Earth’s gravity acceleration and hm the maximum height of the slope. If E < mghm,
the ball reaches a height h0 such that E = mgh0, and from there comes back.
Exercise 6.4. Show that for m = 100 g and hm = 1 m, a ball must have a kinetic
energy equal at least to E = 0.98 107 erg = 0.98 J to overstep the slope and fall
in the hole.
In the solar core at temperatures of the order of 107 K, the proton most probable
kinetic energy is
E ¼ kB T ¼ 8:62 105 eV K1 107 K ¼ 862 eV ¼ 0:862 keV,
where the Boltzmann constant in unit of eV per K degree is
kB ¼ 1:38062 1016 erg/K

¼ 1:38062 1016 erg K1 = 1:6022 1012 erg eV1
¼ 8:6171 105 eV K1 :
Thus, the proton kinetic energy is about 500 times smaller than the potential barrier
maximum Vm = 0.41 MeV. Basing on classical mechanics, we expect that a proton
reaches a distance r0 from the target proton such that e2/r0 = E, hence, r0 = e2/E, and is
rejected. Only a proton with an energy that is 500 times the most probable energy should
overcome the barrier. But the number of the protons that classically overstep the barrier,
is completely negligible, since, according to the Maxwell distribution, it is proportional
to ev =v0 e500 10217 . At the nuclear level, however, classical mechanics is no
2 2
more sufficient and quantum mechanics must be taken into account. In particular,
due to the so-called tunnel effect, which is unusual for our common sense, protons
have a significant probability of moving through the repulsive potential barrier even
with a kinetic energy lower than the barrier maximum. Before discussing the tunnel
effect, in the following we introduce some relevant concepts of quantum mechanics.
6.1.3 De Broglie Wavelength
Quantum particles can behave like waves. This characteristic allows protons to
move through the electrostatic potential barrier with an energy which would be
196 6 The Core
insufficient for a classical particle. Following the approach proposed by De Broglie,

the motion of any particle is governed by the propagation of a wave with a
wavelength k related to the particle momentum p by
k ¼ h=p;
where h is the Planck constant. Moreover, the wave frequency m is related to the
particle energy by the same formula that holds for photons, i.e.
m ¼ E=h;
where E is the total energy of the particle, which in the general case includes the
kinetic and potential energies and the energy equivalent to the rest mass of the
particle according to the Einstein mass-energy relation. Since the Planck constant
has a very small value
h ¼ 6:6262 1027 erg s,
the wavelength associated with a particle is significant only if the momentum is

sufficiently small. Assuming T = 107 K as a typical temperature in the solar core,
the proton speed is
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
v¼ 2kB T=mp ¼ 2 138 1016 erg K1 107 K=1:67 1024 g
¼ 4:06 107 cm s1 ;
which is about one thousandth of the light speed. The corresponding proton
momentum is
p ¼ mp v ¼ 1:67 1024 g 4:06 107 cm s1 ¼ 6:78 1017 g cm s1 :
Thus, the wavelength associated to the proton is
k ¼ h=p ¼ 6:63 1027 erg s1 =6:78 1017 g cm s1 ¼ 0:963 1010 cm:
The proton wavelength in the solar core is of the order of 1000 Fm and, hence, it
is much greater than the proton radius (about 1 Fm) and comparable to the average
minimum distance at which protons can approach by thermal agitation. The situ-
ation is different for a billiard ball, which has a negligible wavelength compared to
its size.
Exercise 6.5 Show that a ball with a mass equal to 100 g and a speed of 3 m s−1,
has a De Broglie wavelength equal to k = 2.21 10−31 cm.
Thus, the motion of a billiard ball is properly treated by classical mechanics,
while the motion of a proton in the solar core follows the laws of quantum
mechanics. Then, let us consider more in detail the wave nature of quantum
particles.
6.1.4 Wave Function
The waves that are associated to the motion of quantum particles are different from
classical waves. In fact, they are neither the oscillation of material particles, as for
sound waves, nor the oscillation of force fields, as for electromagnetic waves.
Instead, the wave associated to a particle is related to the probability of finding that
particle in a certain position in space and at a certain instant in time. Strictly
speaking and for the one-dimensional case, to each particle we associate a complex
wave function, f(x, t), such that the probability of finding the particle in the space
interval [x, x + dx] and in the time interval [t, t + dt] is
Pðx; tÞ dx dt ¼ f ðx; tÞf ðx; tÞ dx dt;
where f* denotes the complex conjugate of f. Thus, the probability density, P(x, t),
is the squared modulus of the wave function f. When we represent complex
numbers as vectors in the plane C, the squared modulus of a number equals the sum
of the squares of the real and imaginary parts (Fig. 6.3).
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffi
jfj ¼ f f ¼ ðfr ifi Þðfr þ ifi Þ ¼ fr2 þ ifr fi ifr fi þ fi2 ¼ fr2 þ fi2 :
As examples of wave function we consider two simple cases representing free

particles. If we know that the particle has exactly energy E and momentum p, the
wave function is
f ðx; tÞ ¼ cos½2pðmt x=kÞ þ i sin½2pðmt x=kÞ;
Fig. 6.3 Geometric

representation of complex
numbers. The number
f = fr + ifi is represented by
the vector whose components
are the real part, fr, along the
x axis and the imaginary part,
fi, along the y axis. The
length, or modulus, of f is
198 6 The Core
where the real and imaginary parts are monochromatic waves (described in
Fig. 1.12). The frequency m and wavelength k are related to E and p, respectively,
by the De Broglie postulates, thus m = E/h, k = h/p and we can write
f ðx; tÞ ¼ cos½2pðmt x=kÞ þ i sin½2pðmt x=kÞ

¼ cos½2pðEt=h px=hÞ þ isin½2pðEt=h px=hÞ
¼ cos½ðEt pxÞ=h þ isin½ðEt pxÞ=
h;
with
h ¼ h=ð2pÞ ðpronounced hbarÞ:
From this wave function we get that the probability density of finding the particle
in the space interval [x, x + dx] at a certain time t0 is
Pðx; t0 Þ ¼ f ðx; t0 Þf ðx; t0 Þ ¼ cos2 ½ðEt0 pxÞ=

h þ sin2 ½ðEt0 pxÞ=h ¼ 1:
This means that the particle of which we know exactly the momentum has equal
probability to stay anywhere in the space, that is, the uncertainty in its position is
infinite. Actually this is a limit case, because usually we can measure the
momentum and position of a particle with certain finite degrees of uncertainty, let’s
say Dp and Dx, respectively. In this case, the wave function is a wave packet, that is,
a superposition of an infinitely large number of monochromatic waves, each with
infinitesimally differing energies and momenta within the finite ranges DE and Dp,
respectively. The real part of a wave packet at a certain time t0 is plotted in Fig. 6.4.
Fig. 6.4 The real part of a wave packet f(x, t0) as a function of the position x at a certain time t0
The monochromatic waves composing the packet are all in phase at x = 0, the
center of the packet, and progressively get out of phase with increasing and
decreasing x, because of the differing momenta. Beyond a length Dx the amplitude
of the wave packet vanishes because there are equal numbers of waves having the
same amplitude and opposite phase. As a consequence P(x, t0) = 0 outside Dx,
which then represents the space interval where the particle can be located. The
uncertainty in the wave momentum Dp is related to that in the position by
ðDx DpÞ=h 1:
The exact value of the constant on the right side depends on the relative
amplitudes of the component waves of the packet. This relation is a result of the
Fourier analysis, which can be intuitively understood because the larger is the range
of waves with different momenta entering into the superposition, the smaller is the
position range outside which they become completely out of phase.
The relation between the uncertainties in position and momentum of a particle,
which we have discussed in connection with its wave-like nature, is one of the
fundamental limits that quantum mechanics sets on the precision with which an
observer can determine the physical parameters of a dynamical system. These limits
are summarized by the uncertainty principle.
6.1.5 The Uncertainty Principle
As a consequence of the wave nature of quantum particles we can no longer

simultaneously determine with absolute precision all the quantities that characterize
their motion. In particular, denoting by Dx, Dy and Dz the uncertainties in the x,
y and z components of the particle’s position, and by Dpx, Dpy and Dpz the
uncertainties in the corresponding momentum components, the products of the
uncertainties in each direction cannot be lower than the Planck constant divided by
2p, that is.
Dx Dpx h;
Dy Dpy h;
and
Dz Dpz h:
Moreover, denoting by Dt the uncertainty in the time at which the particle passes
through a given point and by DE the simultaneous uncertainty in its energy, the
product of these uncertainties cannot be lower than the Planck constant, that is,
200 6 The Core
Dt DE h:
Let us turn back to the tunnel effect and the cross section of the proton-proton
collisions that occur in the solar core.
6.1.6 Tunnel Effect
The motion of two colliding protons is governed by a wave function which is the
solution of an equation, called the Schrödinger equation, where the potential barrier
in Fig. 6.1 appears as a known term (Erwin Schrödinger was an Austrian physicist,
1867–1961). The Schrödinger equation for the wave function f(x, t) of a
non-relativistic particle with mass m, moving under the influence of a
time-dependent, one-dimensional potential V(x, t) is
h2 @ 2 f @f
h ;
:þV f ¼ i
2m @ x2 @t
where ∂2f/∂x2 is the second partial space derivative of f at constant t, ∂f/∂t is the first
partial time derivative of f at constant x, and i = √−1 is the imaginary unit number.
Here we limit to describe the wave function that satisfies the Schrödinger
equation in the case of a Coulombian potential barrier. The real part of this wave
function at a certain time t0 is plotted in Fig. 6.5.
Fig. 6.5 Real part at fixed time of the wave function of a proton colliding with another proton
(red curve). The increasing wavelength of the incident wave is due to the momentum lost when the
kinetic energy is spent against the repulsive potential of the Colombian barrier (solid black curve).
The horizontal straight line represents the proton initial kinetic energy, E = 0.2 MeV, which is
lower than the potential energy top, Vm = 0.41 MeV. The proton wave function decreases
exponentially inside the barrier and tunnels into the potential well
At large distances from the maximum of the repulsive potential barrier, r >> rm,
the wave function of the incident proton is that of a free particle with energy E,
whose real part is a harmonic oscillation at the De Broglie wavelength, k = h/p.
Approaching another proton, the wavelength of the incident proton increases,
because the momentum is gradually lost while the kinetic energy is spent to
overcome the repulsive potential. At r = r0, the potential energy equals the initial
kinetic energy, V = e2/r0 = E. Inside the barrier, the wave function is an evanescent
wave, that is, its amplitude exponentially decreases with decreasing r. Thus, the
probability of finding the proton within the barrier is different from zero even if
E < Vm. and, classically, the proton should stop and reverse the motion. The pos-
sibility that quantum particles penetrate a repulsive potential barrier with a kinetic
energy lower than the top of the barrier is called tunnel effect.
Now we are in the position to determine the cross section of the first reaction of
the ppI chain.
6.1.7 Cross Section
The cross section rpp for the proton-proton fusion into deuteron is written as a
product of three terms in order to point out the quantities that depend on energy,
that is
rpp ¼ Tr S=E:
The factor 1/E derives from the proton cross section estimated by taking the De
Broglie wavelength as radius, that is,
pffiffiffiffiffiffiffiffiffiffiffi 2
pk2 ¼ pðh=pÞ2 ¼ pðh= 2mp E Þ ¼ ph2 =ð2mp =EÞ:
The term Tr is related to the tunnel effect and represents the proton wave
transmission through the potential barrier. Tr depends on the ratio between the
proton probability densities inside and outside the Coulombian barrier, respectively.
The calculation by wave mechanics shows that Tr has the form
pffiffiffi
Tr ¼ e31:28= E
ðE in keVÞ:
Finally, the factor S is the residual probability that one of the two colliding
protons decay in a neutron. This b+ decay allows the formation of a stable nucleus
of deuterium in place of an unstable nucleus with two protons. As described in the
first chapter, in a proton, which has electric charge +1 (in unit of e), there are two up
quarks, having charge +2/3, and a down quark, with charge −1/3. A neutron, which
is electrically neutral, is made of an up quark and two down quarks. Thereby, when
a proton decays in a neutron through a weak interaction, a quark in the instable two
proton nucleus changes flavor from up to down, emitting a positron and an electron
202 6 The Core
neutrino. Since the neutron mass is larger than the proton mass, the b+ decay
requires energy from the environment to occur and, hence, does not take place for
an isolated proton, which is stable. Vice versa, the mass of two protons is larger
than that of a deuteron, thus, their fusion releases energy. On the basis of accurate
calculations, we can take S approximately constant
S ¼ 3:8 1046 keV cm2 :
Putting together the three factors as function of energy, the cross section of the
first reaction of the ppI chain is
rpp ¼ Tr S=E
pffiffiffi
¼ 3:8 1046 e31:28= E
=E cm2 ðE in keVÞ:
6.2 Reaction Rates
The knowledge of a reaction cross section allows us to calculate the number of

reactions that occur per unit of time and volume, or reaction rate.
Let us assume that there are Na particles per unit volume that can interact with
other Nb particles and denote by dNa and dNb the numbers of particles per unit
volume with velocities in the ranges [va, va + dva] and [vb, vb + dvb], respectively.
Referring to Fig. 3.2 and the relative discussion, the number of collisions that occur
to a particle a over a distance L is equal to the number of particles b contained in the
cylinder having L as length and the cross section rab as base, that is, rabLdNb. If
v = va−vb is the relative velocity between the particles a and b, and t the time
necessary to travel L, then L = vt and the number of interactions per unit time of a
particle a is
rab LdNb =t ¼ rab vtdNb =t ¼ rab vdNb :
The reaction rate, drab, between all particles dNa and dNb is then obtained
multiplying by dNa, that is,
drab ¼ rab vdNa dNb :
We can also write
drab ¼ Na Nb rab vðdNa =Na ÞðdNb =Nb Þ;
where dNa/Na and dNb/Nb are the probabilities of finding particles a and b in given
velocity intervals and are provided by the Maxwell velocity distribution, introduced
in Chap. 1.
6.2 Reaction Rates 203
The cross section depends only on the relative velocity of the particles, but not
on the absolute velocity of their center of gravity. Thus, we replace the probability
of finding a particle pair with velocities va and vb with that of finding a pair with
relative velocity v, which is still a Maxwellian, i.e.
h i 2 2
drab ¼ Na Nb v rab =ðpv0 Þ3=2 ev =v0 4pv2 dv:
In order to get the total reaction rate, rab, we have to sum over all relative
velocities. Thus,
Z1 Z1 h i 2 2
rab ¼ drab ¼ Na Nb vrab =ðpv0 Þ3=2 ev =v0 4pv2 dv;
0 0
where
v0 ¼ ð2 kB T=mÞ1=2 ;
is the most probable speed and
m ¼ ma mb =ðma þ mb Þ
is the reduced mass of two particles with masses ma and mb, respectively. Denoting
with the brackets, < > , the mean over the Maxwellian
Z1 h i 2 2
hv rab i ¼ v rab =ðpv0 Þ3=2 ev =v0 4pv2 dv;
0
the reaction rate is
rab ¼ Na Nb hv rab i:
At this point, to obtain the energy produced per second and per unit volume in a
nuclear reaction that we label with the subscript k, we have just to multiply the
reaction rate rk by the released energy Ek. Thus, the energy, e, that is produced per
second and unit volume in all the nuclear reactions is
X X
e¼ Ek rk ¼ Ek Nak Nbk hvk rk i;
k k
P
where the symbol k indicates the summation over all the reactions.
The energy production rate, e, depends on the thermodynamic parameters
through the number densities of the particles involved in the reactions. Moreover, it
depends on the temperature through the average of the cross sections over the
Maxwell distribution, as we see now in detail.
204 6 The Core
6.2.1 The Average over the Maxwell Distribution
We turn back to the first reaction of the pp cycle and perform the average over the
Maxwell distribution substituting the cross section as function of energy found in
Sect. 6.1
Z1 h i 2 2

vrpp ¼ v rpp =ðpv0 Þ3=2 ev =v0 4pv2 dv
0
Z1 n pffiffiffi .h io
v 3:8 1046 e31:28= E Eðpv0 Þ3=2 ev =v0 4pv2 dv:
2 2
¼
0
Taking the term that is independent of velocity
3:8 1046 4p=ðpv0 Þ3=2 ¼ 1:52 1045 =ðp1=2 v0 Þ

3=2
out of the integral, and recalling that the exponent of a product of two exponentials
is the sum of exponents, we get
h i Z1 h pffiffiffi i
45
eð31:28= E þ v =v0 Þ v2 v=E dv:
3=2 2 2
vrpp ¼ 1:52 10 =ðp1=2 v0 Þ
0
It is worthwhile to transform the integral over velocity modulus into an integral

over energy. Thus, as Exercise 6.5 shows, we get
h i Z1 n pffiffiffi o
vrpp 45 2 1=2 3=2
¼ 3:04 10 =ðm p v0 Þ e½31:28= E þ E=ðkB TÞ dE cm3 s1 ðE in keVÞ:
0

Worked Exercise 6.6 Transforming vrpp into an integral over energy.
In order to transform the integral over velocity into an integral over energy in the
mean of rpp over the Maxwellian
Z1 h pffiffiffi i
3=2
vrpp ¼ 1:52 1045 =ðp1=2 v0 Þ eð31:28= E þ v =v0 Þ v2 v=E dv;
2 2
we make the substitution

E ¼ mv2 =2:
Thus, it is
v2 ¼ 2E=m
and
dE ¼ 2mv dv=2 ¼ mv dv;
hence
vdv ¼ dE=m:
Then,
h i Z1 n pffiffiffi o
45
e½31:28= Eþ2E=ðmv0 Þ 2E=ðmEÞ dE=m:
3=2 2
hvrpp i ¼ 1:52 10 =ðp1=2 v0 Þ
0
The most probable energy is
E0 ¼ mv20 =2 ¼ kB T:
Thus,
v0 ¼ ð2kB T=mÞ1=2 ;
3=2
v0 ¼ ð2kB T=mÞ3=4
and
3=2
m2 v0 ¼ m2 ð2kB T=mÞ3=4 ¼ m23=4 ð2kB TÞ3=4
¼ m5=4 ð2kB TÞ3=4 :
Then
Z1 n pffiffiffi o
45
vrpp ¼ f3:04 10 =½p 1=2
m 5=4
ð2kB TÞ 3=4
g e½31:28=E þ E=ðkB TÞ
dE cm3 s1 ðE in keVÞ:
0
Exercise end.
To estimate the integral, one might think of using the most probable energy. But,
just at the energies where the proton number is larger, the cross section is very small.
However, because the tunneling probability increases with energy as e−31.28/√E,
206 6 The Core

the largest contribution to the average vrpp comes from the protons that have
energies which are about one order of magnitude larger than E0 and,hence, speeds
three times greater than v0. The maximum of the integrand of vrpp is named the
Gamow peak, after George Gamow (Russian physicist, 1904–1968).
6.2.2 The Gamow Peak
The Maxwell distribution exponentially drops with energy as
eE=kB T ;
while the tunneling cross section grows as

pffiffiffi
e31:28= E
:
Thus, their product

pffiffiffi
e½E=ðkB T Þ þ 31:28= E
;

which appears in the integrand of the mean vrpp , has a peak, as shown in
Fig. 6.6.
Fig. 6.6 The Gamow peak. Black solid lines refer to the energy dependences in the proton fusion
cross section, which are due to the Maxwellian energy distribution (label eE=kB T ), and to the
tunneling of the Colombian repulsive barrier, (label e−b/√E). The product of the two factors has a
pronounced maximum known as the Gamow peak (red curve). Note that the vertical axis scale is
logarithmic to highlight small values
The maximum of an exponential with negative exponent corresponds to the

minimum of the absolute value of its exponent; this minimum is obtained by
vanishing the energy derivative, that is,
h pffiffiffiffii
d E=ðkB T Þ þ 31:28= E =dE ¼ 0
and, hence
pffiffiffiffi
1=ðkB T Þ 31:28= EG 2 E G ¼ 0;
where EG denotes the energy of the Gamow peak. Thus, to explicit EG we get

3=2
1=ðkB T Þ ¼ 31:28= 2EG ;
3=2
EG ¼ 15:64 kB T
and
EG ¼ ð15:64 kB TÞ2=3 ; with kB T in keV:
Exercise 6.7 Show that EG = 5.66 keV for T = 107 K.

For T = 107 K, we get kBT = 0.862 keV, and EG = 5.66 keV, then the Gamow
peak occurs at an energy which is about 7 times the most probable energy.
By approximating
the Gamow peak with an appropriate Gaussian function, the
average vrpp can be written in terms of Gamow energy as

vrpp ¼ 4 1040 s2 es cm3 s1 ;
with
s ¼ 3EG =kB T ¼ 3ð15:64 kB TÞÞ2=3 =kB T

¼ 18:76ðkB T Þ1=3 ðkB T in keVÞ:
Exercise
6.8 Show that at T = 107 K, we get s = 19.7
−46 3 −1
and vrpp = 4 10 cm s .

At typical temperatures of the solar core vrpp is of the order of 10−46 cm3 s−1

(Exercise 6.7). This value of vrpp may appear very small, however the reaction
rate depends also on the square of the proton number density. The number of
protons per unit volume is related to the density by the relation
208 6 The Core
Np ¼ Xq=mp ;
where X is the hydrogen mass fraction and mp the proton mass.

The mass density at T = 107 K is q = 44 g cm−3, and X = 0.64, as we will see
later. Thus
Np ¼ Xq=mp ¼ 0:64 44 g cm3 =1:67 1024 g ¼ 1:7 1025 protons cm3 :
Dividing by 2 the number of target protons in order to avoid a double account in

our case of collisions between identical particles, in the end the rate of proton
fusions into deuteron, rpp, is

rpp ¼ Np2 vrpp =2 ¼ ð1:7 1025 cm3 Þ2 4 1046 cm3 s1 =2
¼ 6 104 reactions s1 cm3
In conclusion, despite the low probability that two protons merge into a deuteron,
due mainly to the small probability of the proton b + decay, roughly sixty thousand
proton-proton fusions per cubic centimeter take place in the core every second.
6.2.3 Lifetime
The rate

2rpp =Np ¼ Np vrpp
represents the fraction of protons that disappear by fusion per second, where the
factor 2 enters since a deuteron requires two protons to be formed. The inverse
number

Np = 2rpp ¼ 1=ðNp vrpp Þ
is the number of seconds between two subsequent reactions and, hence, represents
the average time that protons survive to fusion. For example, if the fraction of
protons which fuses per second was 1, all the protons would disappear in 1 s and
the average life of protons with respect to fusion would be only 1 s. At the other
extreme, if the fusing proton fraction was 0, the average life time would be infinity.
As Exercise 6.9 shows, we estimate that the average life spp of protons for fusion
into deuteron at a level with T = 107 K is
spp ¼ 4:6 1012 years:

Worked Exercise 6.9 Fusion lifetime.

At a level in the core with T = 107 K, we know that vrpp = 4 10−46
cm3 s−1 and Np = 1.7 1025 cm−3. Thus, at this level the proton lifetime spp for
fusion into deuteron is

spp ¼ 1=ðNp vrpp Þ ¼ 1=ð1:7 1025 cm3 4 1046 cm3 s1 Þ ¼ 1:5 1020 s
¼ 1:5 1020 s=ð3:16 107 s y1 Þ ¼ 4:6 1012 y:
Exercise end.
The value of spp decreases close to the solar center, where the temperature and
density are lager, and approaches spp = 1010 years, that is, the time required to burn
all the hydrogen contained in the core. With such a long lifetime the first step of the
ppI chain occurs much more slowly than all the other steps. Actually, it is the
proton fusion into deuterium to determine the time that the Sun spends in the
current state, which is maintained by the pp cycle.
Worked Exercise 6.10 Solar luminosity.
We can estimate the solar luminosity from the rate of the first reaction of the ppI
chain. We neglect ppII and ppIII because they are less frequent than ppI. Moreover,
because the first step of the ppI chain occurs much more slowly than the other steps,
we assume that the total energy production rate is simply the product of the energy
amount released in a cycle and the proton fusion rate rpp.
Assuming that fusion releases the energy E = 26.7 MeV at a constant rate
rpp = 7 105 cm−3s−1, inside a volume V = 4pr3/3 with radius r = 0.25 ∙ RS, we
get the following estimate for the luminosity
L ¼ E rpp V ¼ E rpp 4 3:14 ð0:25RS Þ3 =3

¼ 26:7 106 eV 1:60 1012 erg eV1 6
104 cm3 s1 4 3:14ð0:25 6:96 1010 cmÞ3 =3
¼ 0:17 1033 erg s1 :
This number is approximately a factor 20 less than the measured luminosity

LS = 3.844 1033 erg s−1.
Exercise end.
Despite of the approximate calculations, Exercises 6.9 and 6.10 show that once
we have identified the hydrogen fusion into helium as the solar energy source, we
can determine two basic solar parameters like lifetime and luminosity.
6.2.4 Energy Production Rate
The energy per second and unit volume produced by the nuclear burning, or nu-
clear energy production rate, enters in the energy conservation equation, which is
required for modelling the core, as we discuss in the next section. The energy rate of
210 6 The Core
the entire pp cycle is the sum over all reactions of the energy that each reaction
releases multiplied by the corresponding number density of reactions per second
X
e¼ E k rk ;
k
where Ek and rk are the released energy and reaction rate, respectively, for the
k reaction.
We limit to consider only the pp cycle that is the dominant energy source for the
present Sun. Since the first branch of the cycle is the most frequent branch, we
neglect the ppII and ppIII reactions in the sum. Moreover, because the first step of
the ppI chain occurs much more slowly than the other steps, we approximate the
total energy produced per second with the product of the energy released in the
entire cycle, Eppc, and the reaction rate of the proton fusion into deuteron, that is,
X
e¼ Ek rk ¼ Eppc rpp ¼ Eppc Np2 vrpp =2:
k
The number of protons per unit volume is
Np ¼ Xq=mp ;
where X is the hydrogen mass fraction, q the density and mp the proton mass.
Thus,

e ¼ Eppc ðXq=mp Þ2 vrpp =2:
We have estimated that

vrpp ¼ 4 1040 s2 es cm3 s1
with
s ¼ 18:76ðkB TÞ1=3 ðkB T in keVÞ:
Moreover, we get
Eppc ¼ 26:7 MeV ¼ 29:7 MeV 1:60 106 erg=MeV

¼ 4:00 105 erg:
Thus, the energy production rate per unit volume is

e ¼ Eppc ðXq=mp Þ2 vrpp =2
¼ ½4 105 erg=ð1:67 1024 gÞ2 ð4 1040 cm3 s1 =2 Þ ðXqÞ2 s2 es
¼ 3:06 103 X 2 q2 s2 es erg cm3 s1 ðq in g cm3 Þ:
and in terms of T
1=3
e ¼ 3:06 103 q2 X 2 ½18:76ðkB TÞ1=3 2 e18:76ðkB TÞ erg g1 s1 ðkB T in KeVÞ:
In the following we use this formula for the energy production rate but with
slightly more accurate coefficients, that is
2=3 3:37=T 1=3

e ¼ 2:9 104 q2 X 2 T9 e 9 erg g1 s1 ;
where T9 is the temperature in unit of 109 K.

The expression of e allows us to complete the set of equations necessary to
calculate the solar core model.
6.3 The Core Model
6.3.1 Luminosity
The luminosity is created by the nuclear fusion reactions that occur in each volume
element of the core. In the Son’s center the luminosity is zero and increases
onwards with the increase of the contributing volume, up to reach the surface value
at the bottom of the radiation zone. Thus, the solar luminosity may strongly differ in
different layers of the core whereas is equal to the surface value in the photosphere,
convection and radiation zones. To determine the luminosity variation we need a
new equation, which is provided by the condition of energy conservation.
6.3.2 Energy Conservation
Inside the core, a spherical corona with radius r, thickness dr and volume dV,
provides an increment of luminosity dL equal to the energy production rate per unit
volume, e, multiplied by the volume dV, that is
dL ¼ e dV:
212 6 The Core
In order to get the mass M as unknown quantity instead of V, the volume

increment is usually replaced with the mass increment dM = qdV, where q is the
density. Then, denoting with
em ¼ e=q
the energy production rate per unit mass, we get
dL ¼ e dV ¼ edM=q ¼ em dM:
This equation expresses the energy conservation because all the energy that is
produced by nuclear reactions contributes to the outward radiation flux. Thus, there
is no heating or cooling and the Sun’s structure remains unchanged over time. In the
photosphere, convection and radiation zones, where the temperature is too low for
the onset of nuclear reactions, we can write
e ¼ dL=dV ¼ 0
and immediately find
L ¼ constant:
We are now ready to construct a model of the solar core.
6.3.3 Model
In summary, we can write for the solar core the following four equilibrium con-
ditions, which represent in the order the energy transport, hydrostatic equilibrium,
mass and energy conservation
dT=dr ¼ 3LkR =ð64pr 2 rT 3 Þ

dP=dr ¼ qGM=r 2
dM ¼ 4 p r 2 qdr
dL ¼ em dM:
These equations are called the structure equations and allow us to determine the
values of the four parameters, temperature, pressure, mass and luminosity, as
function of the radius, when the other three parameters, density, opacity, and
nuclear energy production rate, are known. Actually, in LTE q, kR and em are related
to two thermodynamic parameters by the so called constitutive equations, which
represent the true physical ingredients of a solar and, more generally, a stellar
model. Here we adopt the analytical formulas
6.3 The Core Model 213
P ¼ qRT=l
kR ¼ kR0 q2 T 7=2
2=3 3:37=T 1=3
em ¼ 2:9 104 qX 2 T9 e 9 ;
which have been discussed in the previous text. They allow us to get analytical
solutions of the structure equations but with a rather high degree of approximation
compared with the accurate models, which are based on comprehensive theories
and use sophisticated numerical techniques.
We point out that the opacity, energy production rate and mean molecular
weight depend upon the chemical composition. In particular, because nuclear fusion
burns H to create He at different rates in different layers, the mass fractions X and
Y are function of the radial distance.
6.3.4 Boundary Conditions
To solve the structure equations, we need to know the solar parameters at an initial
level. As we have previously done, we can use the physical conditions at the bottom
of the preceding layer, in this case the radiation zone, at r = 0.25RS. However,
because the volumes contributing mass and luminosity vanish approaching the
center, it must also be
M ¼ 0 at r ¼ 0
and
L ¼ 0 at r ¼ 0:
These two boundary conditions in the center are valid for any model of solar
structure.
We start with determining T and P and use the diffusion and hydrostatic
equations
dT=dr ¼ 3LkR =ð64pr 2 rT 3 Þ
and
dP=dr ¼ qGM=r 2 :
We solve these equations with the same approximations we made for the radi-
ation zone. Thus, we put them in the form
214 6 The Core
dT=dr ¼ T=HT
and
dP=dr ¼ P=HP ;
and assume that the scale heights
HT ¼ 64pr 2 rT 4 =ð3LkR Þ
and
HP ¼ Rg Tr 2 =ðlGMÞ
are constant at the values they have at r = 0.25RS (Table 5.1). Then, the solutions
are
T ¼ T0 eðr0 rÞ =HT
and
P ¼ P0 eðr0 rÞ =HP ;
with
HT ¼ 2:11 105 km
and
HP ¼ 5:05 104 km:
Exercise 6.11 Verify the values of HT and HP given in the text.

From the state equation
q ¼ lP=ðRTÞ;
we find the density
q ¼ q0 eðr0 rÞ =Hq ;
where
Hq ¼ 1=ð1=HP 1=HT Þ
and, hence

Hq ¼ 1= 1=5:05 104 km 1=2:11 105 km ¼ 6:64 104 km:
We expect that these solutions for T, P and q fail near the solar center, because
for r approaching 0, M and L approach 0 as the volume, i.e. as r3; hence, HP, HT and
Hq scale as r−1 and approach ∞.
For M and L we have boundary conditions both at r = 0 and r = 0.25RS, but we
use those at the center which are exact.
If we assume that q is constant at its central value qc, the mass increases in
proportion to the volume
M ¼ 4pqc r 3 =3
(Exercise 6.12).
Exercise 6.12 Taking q = qc = constant in the mass conservation equation, find
that the mass increases in proportion to the volume as M = 4pqcr3/3.
To improve the approximation we take the density variation into account and get
the equation
dM ¼ ¼ 4pr 2 qc er =Hq dr;
with
qc ¼ q0 er0 =Hq ¼ 21:1 g cm3 eð0:256:9610 ¼ 290 g cm3 :

5
km=6:64104 kmÞ
This equation is analogous to the mass equation discussed for the radiation zone
and can be solved using the same approach to get
h
i
M ¼ 4 p qc Hq3 2 r 2 =Hq2 þ r=Hq þ 2 er=Hq :
Exercise 6.13 Find the mass variation given in the text for an exponential q.
Thus, the mass increases with radius less steeply than at constant density.
Exercise 6.13 Taking e = ec = constant in the energy conservation equation, find
that the luminosity increases in proportion to the volume as L = 4pecr3/3.
In order to solve the energy conservation
dL ¼ em dM;
we take advantage from the fact that em as function of M fits rather well an
exponential function
216 6 The Core
em ¼ emc eM =Mh
with
emc ¼ 16:4 erg s1 g1
and a scale height
Mh ¼ 0:124 M0
(Fig. 6.7).
Thus, we integrate the energy conservation between M = 0 and M, where the
luminosity is 0 and L, respectively
ZL ZM ZM
dL ¼ em dM ¼ emc eM =Mh dM:
0 0 0
It is straightforward to continue with
ZM
L ¼ emc eM =Mh dM
0
ZM
¼ emc Mh eM =Mh dM= Mh
0
h iM
¼ emc Mh eM =Mh
0
and finally get
L ¼ emc Mh ð1 eM =Mh Þ:
In order to determine L(r), it is sufficient to Insert M(r) that we have previously

found.
The approximate and accurate models are shown in Figs. 6.8 and 6.9. In par-
ticular, Fig. 6.8 shows the variations of T, P, q, M and L.
The overall agreement is rather satisfactory for the mass and luminosity frac-
tions, while the temperature and, more strongly, pressure and density depart from an
exponential function approaching the center (see also Table 6.1 for a comparison of
accurate and approximated model parameters close to center).
We leave as exercise to find the variations of T, P and q in the neighborhood of
the center, which are also plotted in Fig. 6.8.
Exercise 6.14 Verify that in the neighborhood of the center T and P scale as
Fig. 6.7 The decimal logarithm of the energy production rate per unit mass as a function of the
mass fraction. The exponential function given in the text (solid line) is compared with the accurate
values of Senesi et al. (red line)
T 4 ¼ Tc4 ec kRc r 2 =ð8rÞ

and
P ¼ Pc 2pGq2c r 2 =3:
The corresponding variation of q can be determined from the state equation.

In Fig. 6.9 we have plotted the molecular weight, Rosseland mean opacity and
energy production rate as functions of the radius fraction.
The agreement for l is poor. Actually, the approximated molecular weight is
simply the constant value l = 0.61, which corresponds to a mixture of 90%
hydrogen and 10% helium, both totally ionized. On the contrary, in the accurate
model, l grows from 0.61 at r = 0.25 RS up to 0.82 at the center. The growth is due
to the hydrogen fusion into helium, which has reduced the number of particles
keeping substantially constant the mass. If we start with an initial mixture of 40 H
nuclei and 4 He nuclei and assume that 20 protons have already been transformed
into 5 a particles, the mass of 20 protons and 9 a particles is 20 + 9 4 = 56 in
atomic mass units. The number of particles includes 29 nuclei, 20 electrons from H
and 9 2 = 18 electrons from He, i.e. 29 + 20 + 18 = 67 particles on total. Thus
l = 56/67 = 0.83. Such a value of l, which is very close to l = 0.82 of the
accurate model, indicates that in the past life the Sun has burned about half of its
hydrogen reservoir in the central part.
The Rosseland mean opacity of the Senesi et al. model takes accurately into
account the contributions from metal photoionizations, free-free transitions and
218 6 The Core
Fig. 6.8 The core model. Black solid curves refer to our approximate model and red curves to the
accurate model of Senesi et al. In the order from top to bottom the panels represent the decimal
logarithms of the temperature, pressure and density, the mass and luminosity fractions (relative to
the surface values). The variations of T, P and q in the neighborhood of the center are also plotted
according to the formulas given in Exercise 6.14 (dashed lines). All quantities are plotted as
functions of the radius fraction r/RS
Thomson scattering, whereas the approximate opacity is based on the Kramers’

formula for b-f and f-f absorptions
kR ¼ kR0 q2 T 7=2 cm1 :
Note that in the Kramers’ formula q(r) and T(r) were taken from the Senesi et al.
model and the constant kR0 was determined by the condition that kR(r0) equals the
220 6 The Core
Fig. 6.9 The core model. Black solid curves refer to our approximate model and red curves to the
accurate model of Senesi et al. In the order from top to bottom the panels represent the average
molecular weight and the decimal logarithms of the Rosseland mean opacity and nuclear energy
production rate. All quantities are plotted as functions of the radius fraction r/RS
accurate value at r = 0.25RS. This constrain yields a better approximate opacity

than the opacity computed with kR0 ¼ 1:5 1024 and plotted in Fig. 4.2.
The approximate energy production rate per unit volume has been computed
with q(r) and T(r) taken from the Senesi et al. model.
The core model of Senesi et al. is listed in Table 6.2.
222 6 The Core
Table 6.1 Physical conditions at r = 0.25RS and r = 0

Parameter r = 0.2496915 RS r = 5.101102 10−4 RS r=0
T, K 7.89 10 6
15.5 106 18.0 106
P, Ba 2.25 10 16
2.33 1017
6.98 1017
−3
q, g cm 21.1 150 290
kR, cm−1 55.0 190 210
e, erg s−1 cm−3 7.04 2.47 103 7.34 103
M/MS 0.482 1.40 10−8 0
L/LS 0.987 9.14 10−41 0
The values in columns 2 and 3 are derived from the Senesi et al. model and those in the last
column from the approximate model
Exercise 6.15 Using the data listed in Table 6.2 verify that the energy production
rate per unit mass em is close to the exponential mass function plotted in Fig. 6.7.
Exercise 6.16 In the text the potential energy barrier encountered in a two proton
collision has been compared with a slope surrounding a hole that a billiard ball has
to climb. How the comparison holds despite that the protons repel each other
whereas the Earth attracts the ball?
Exercise 6.17 Verify that the wave function of a free particle with energy E and
momentum p given in the text is a solution of the Schrödinger equation in the
special case of a potential V(x, t) = constant.
Exercise 6.18 Let us define the wavenumber k as the reciprocal of the wavelength
k, i.e. k = 1/k and k is the number of wavelengths per unit distance. Show that the
superposition of two monochromatic waves with very close frequencies and
wavenumbers
f 1 ¼ sin½2pðmt kxÞ
and
f 2 ¼ sin½2pðm þ dmÞt ðk þ dk ÞxÞ

Table 6.2 The core model of Senesi et al.
r/RS M/MS T, K P, Ba q, g cm−3 kR, cm−1 L/LS e, erg s−1cm−3
2.504E−001 4.836E−001 7.877E+006 2.228E+016 2.096E+001 5.477E+001 9.877E−001 6.866E+000
2.419E−001 4.597E−001 8.098E+006 2.504E+016 2.293E+001 5.800E+001 9.832E−001 9.267E+000
2.336E−001 4.357E−001 8.324E+006 2.805E+016 2.502E+001 6.133E+001 9.777E−001 1.245E+001
2.254E−001 4.118E−001 8.555E+006 3.133E+016 2.723E+001 6.474E+001 9.708E−001 1.664E+001
6.3 The Core Model
2.173E−001 3.879E−001 8.791E+006 3.490E+016 2.958E+001 6.825E+001 9.625E−001 2.214E+001

2.093E−001 3.639E−001 9.034E+006 3.880E+016 3.208E+001 7.182E+001 9.522E−001 2.931E+001
2.013E−001 3.400E−001 9.284E+006 4.306E+016 3.473E+001 7.550E+001 9.398E−001 3.862E+001
1.936E−001 3.171E−001 9.532E+006 4.751E+016 3.744E+001 7.910E+001 9.253E−001 5.013E+001
1.883E−001 3.011E−001 9.709E+006 5.084E+016 3.943E+001 8.169E+001 9.135E−001 6.001E+001
1.829E−001 2.852E−001 9.891E+006 5.437E+016 4.152E+001 8.432E+001 9.001E−001 7.170E+001
1.775E−001 2.692E−001 1.008E+007 5.814E+016 4.371E+001 8.708E+001 8.849E−001 8.555E+001
1.694E−001 2.457E−001 1.036E+007 6.415E+016 4.717E+001 9.131E+001 8.588E−001 1.108E+002
1.637E−001 2.293E−001 1.057E+007 6.868E+016 4.974E+001 9.436E+001 8.377E−001 1.324E+002
1.580E−001 2.134E−001 1.078E+007 7.342E+016 5.241E+001 9.748E+001 8.144E−001 1.573E+002
1.488E−001 1.882E−001 1.112E+007 8.161E+016 5.697E+001 1.026E+002 7.715E−001 2.060E+002
1.390E−001 1.627E−001 1.149E+007 9.097E+016 6.217E+001 1.083E+002 7.191E−001 2.710E+002
1.288E−001 1.372E−001 1.189E+007 1.017E+017 6.809E+001 1.145E+002 6.559E−001 3.582E+002
1.177E−001 1.117E−001 1.233E+007 1.140E+017 7.497E+001 1.215E+002 5.796E−001 4.757E+002
1.089E−001 9.306E−002 1.267E+007 1.243E+017 8.081E+001 1.273E+002 5.142E−001 5.878E+002
1.025E−001 8.030E−002 1.293E+007 1.323E+017 8.535E+001 1.316E+002 4.639E−001 6.823E+002
9.504E−002 6.674E−002 1.322E+007 1.417E+017 9.082E+001 1.368E+002 4.050E−001 8.033E+002
8.650E−002 5.265E−002 1.354E+007 1.528E+017 9.739E+001 1.430E+002 3.370E−001 9.588E+002
7.782E−002 4.008E−002 1.386E+007 1.642E+017 1.043E+002 1.493E+002 2.699E−001 1.133E+003
6.836E−002 2.845E−002 1.420E+007 1.766E+017 1.121E+002 1.564E+002 2.015E−001 1.340E+003
(continued)
223
224
r/RS M/MS T, K P, Ba q, g cm−3 kR, cm−1 L/LS e, erg s−1cm−3

5.839E−002 1.855E−002 1.453E+007 1.894E+017 1.203E+002 1.638E+002 1.377E−001 1.568E+003
4.653E−002 9.838E−003 1.487E+007 2.035E+017 1.298E+002 1.723E+002 7.662E−002 1.842E+003
3.436E−002 4.121E−003 1.517E+007 2.161E+017 1.386E+002 1.801E+002 3.336E−002 2.102E+003
2.449E−002 1.529E−003 1.535E+007 2.242E+017 1.443E+002 1.851E+002 1.267E−002 2.274E+003
1.685E−002 5.052E−004 1.545E+007 2.288E+017 1.476E+002 1.881E+002 4.238E−003 2.374E+003
1.161E−002 1.662E−004 1.550E+007 2.310E+017 1.493E+002 1.896E+002 1.402E−003 2.424E+003
6.497E−003 2.925E−005 1.553E+007 2.324E+017 1.503E+002 1.904E+002 2.477E−004 2.456E+003
2.344E−003 1.436E−006 1.555E+007 2.330E+017 1.507E+002 1.906E+002 1.218E−005 2.469E+003
5.101E−004 4.254E−008 1.555E+007 2.331E+017 1.508E+002 1.900E+002 3.611E−007 2.471E+003
From left to right the columns list the solar radius fraction r/RS, mass fraction M/MS, temperature T, pressure P, density q, Rosseland mean opacity kR,
luminosity fraction L/LS and energy production rate per unit volume e
6 The Core
has the approximate form
f 1 þ f 2 ¼ 2cos½2pðtdm=2xdk=2Þ sin½2pðmt kxÞ:
Interpret this result.

Chapter 7
Evolution
Abstract The fusion of hydrogen into helium in the solar core implies that the H
mass fraction decreases with time while the He mass fraction increases. This pro-
gressive change of the chemical composition affects the mean molecular weight,
opacity and nuclear energy production rate. Thus, the solar structure changes with
time, but the Sun evolves on a very long time scale. The oldest fossils found on Earth
date back to about 109 years ago. Since life is extremely sensitive to the amount of
solar energy that the Earth receives, the thermodynamic structure of the Sun
109 years ago was not too different from that of today. Actually, 109 y = 1 Gy is an
appropriate unit for the evolutionary times of the Sun and many stars. Turning back
in time, the hydrogen and helium number densities in the core approach the ratio of
10 to 1, which is presently measured in the atmosphere. The period of the solar life
during which hydrogen is gradually transformed into helium is called the main
sequence phase. In this section we describe the whole evolution of the Sun including
the pre and post main sequence phases. We will start with the solar formation, which
started as the result of a gravitational collapse of interstellar material.
7.1 Pre-main Sequence Evolution
7.1.1 Onset of Gravitational Collapse
Like any other star the Sun was formed from the interstellar medium by a gravita-
tional instability of the medium itself. The interstellar material is a mixture of gas,
dust and radiation that fills the space between stars in a galaxy. The thermodynamic
conditions of the interstellar medium, which can have a large range of values, are
schematically represented by a model with three phases corresponding to tempera-
tures T < 300 K, T * 104 K and 106 K, respectively. Densities are always very
low, with numbers ranging between 10−4 and 106 particles per cm3. Dust amounts
only to about 1% by mass and the gas, which is 99%, has about X = 0.7, Y = 0.28
and Z equal to a few percent (for the Sun, Z = 0.02). If a sufficiently large amount of
interstellar material is compressed into a relatively smaller volume, self-gravitation

228 7 Evolution
Fig. 7.1 Formation of a shock wave from a body moving in a medium with a velocity v equal to
or greater than the sound speed cs. When v = cs (panel a), the wave fronts (represented by circles
of radius increasing with the time lag) cannot escape the source. The resulting pile of waves forms
a large amplitude front (represented in profile by the yellow line). At this speed the shock wave is a
high pressure front perpendicular to the velocity of the body and moving with it. When v > cs
(panel b), the wave fronts lag behind the source in a cone-shaped region with the source at the
vertex. The edge of the cone forms a supersonic shock wave front with unusually large amplitude
(yellow lines)
may become large enough to overcome the internal pressure, make continuous the
shrinking process and, thus, trigger what is called a gravitational collapse. We do
not know exactly the cause for the gravitational collapse that gave rise to our Sun,
but we can imagine two possible scenarios. In both cases a shock wave is responsible
for the initial compression. This particular type of wave can be generated by a body
that moves in a material medium with a speed equal to or greater than the sound
speed in the medium. As seen in Fig. 7.1, the compression waves that the body
creates add together in the direction of motion and laterally to it and amplify.
A shock wave is characterized by a very steep peak of the pressure, temperature and
density. In particular, the pressure peak can start a gravitational collapse in the
interstellar medium. The spiral arms of our Galaxy are a possible source of shock
waves. In fact, they are rapidly rotating with a speed that reaches 240 km s−1 in the
position of our Sun, at about 26,100 light years from the galactic center (a light year
is the distance that light travels in vacuum in 365.25 days and is equal to about 9.46
1017 cm). Such a speed is widely supersonic because the speed of sound in a cool
interstellar material is of the order of 1 km s−1 (Exercise 7.1).
Exercise 7.1 Assuming T = 50 K and l = 1, show that the interstellar medium has
a speed of sound cs = [ckBT/(lmu)]1/2 = 0.830 km s−1.
One of these arms, made up of stars and interstellar gas with T * 50 K and
q * 10−23 g cm−3, would have produced the shock wave that compressed the
surrounding material and, thus, began the formation process of the Sun and many
other stars.
Alternatively, we may associate the parent shock wave to the explosion of a
supernova.
7.1 Pre-main Sequence Evolution 229
7.1.2 The Contribution of Core-Collapse Supernovae
Different kinds of stars end their lives with catastrophic and spectacular explosions.
The sudden appearance in the sky of a new star led the ancient astronomers to label
these objects as novae (new in Latin). In the 1930s Fritz Zwicky (Swiss astronomer,
1898–1974) introduced the concept of supernova to point out the existence of
exceptional cases of novae due to the large amount of energy, up to 1053 erg, released
into space in a few days/weeks by these gigantic explosions. In particular, stars with
8 MS or more quickly consume their nuclear fuel. We know that fusion reactions can
occur only till to the formation of iron (Fig. 1.16 and relative discussion). Thus, a
massive star close to the end of its evolution has a stratified chemical composition, with
a core of iron and concentric shells of lighter elements out to helium and hydrogen. Iron
nuclei are the most tightly bound of all atomic nuclei, therefore iron fusion cannot
proceed and there is not release of energy. As a consequence the core of the star is no
longer able to generate an outward radiation pressure capable to balance gravity and the
star violently implodes. Due to high matter density the electron capture by proton,
p + e− ! n + me, converts the core of the star to a degenerate neutron gas charac-
terized by q * 1014 g cm−3 and Ne and Np < 0.01 Nn (for more details on the
properties of a degenerate gas see Sect. 7.3). In a fraction of a second, the iron core with
a mass comparable to that of our Sun shrinks to the size of the Earth and collapses into a
neutron star, whose size is only a few kilometers across. If the mass left in the core is
larger than *3 MS the core eventually collapses by gravity into a black hole, an object
with such a strong gravity that not even light can escape from inside it. The gravita-
tional collapse releases an enormous amount of energy, more than 100 times what the
Sun will radiate over its entire lifetime. Due to the large energy absorption and radiation
pressure, most of the stellar material outside the core is superheated and projected into
space at the speed of 104 km s−1 in a gas with T > 106 K (in the case of the supernovae
associated with c ray bursts, this speed can be 3–4 times larger). These stellar explo-
sions are responsible for the creation of approximately half of the neutron-rich atomic
nuclei heavier than iron. This nucleosynthesis is due to repeated neutron captures,
which occur before neutrons can decay to protons and during the photodisintegration of
the iron core. Moreover, the explosion creates a forward shock wave at the outermost
edge of the ejected material, corresponding to the abrupt transition between the inner
material with high pressure and temperature and the outer interstellar medium.
Therefore, supernovae in the Milky Way could produce the shock waves that
triggered a gravitational collapse of the interstellar cloud and eventually led to the
formation of our Solar System five billions years ago. But there is actually more: by
enriching the initial interstellar material of heavy elements, such as C, O, N, Ca, P,
S and others, supernovae have provided the necessary ingredients for the emer-
gence of life on the Earth.
7.1.3 Jeans Criterion
The contraction of a cloud of interstellar matter continues if the hydrostatic equi-

librium is broken and gravity overcomes the opposite internal pressure gradients.
230 7 Evolution
In energy terms, this means that the gravitational potential energy of the cloud in
modulus, |Eg|, is greater than two times the kinetic energy, Ek, that is,

Eg [ 2Ek :
Expressing these energies in terms of cloud parameters allows us to discuss

some important aspects of the contraction process.
As Exercise 7.2 shows, the gravitational potential energy of a spherical star with
mass M and radius R, that is, the work of gravity to build the star, is

Eg ¼ GM 2 R:
Worked Exercise 7.2 The gravitational potential energy.

To calculate Eg, for simplicity’s sake we consider the gravitational contraction
process as the progressive approach of material from infinite distance onto a
spherical symmetric aggregate of mass m and radius r. Approaching a further mass
element dm by a distance −dr to the center of the mass sphere, gravity does a work

Gmdm r 2 dr;
which is positive because the force and displacement are in the same direction.
Thus, the increment of gravitational energy dEg when a new mass element dm is
moved from infinity to distance r is the integral
Zr Zr
2
dEg ¼ Gmdm r 2 dr ¼ Gmdm 1 r dr
1 1
¼ Gmdm=r;
since
Zr Zr
2
1 r dr ¼ d ð1=r Þ ¼ 1=r þ 0 ¼ 1=r:
1 1
Due to spherical symmetry, we assume that the mass element dm is a spherical

corona of radius r and thickness dr. Then, the total work to accrete a total mass M is
ZM ZM
Eg ¼ dEg ¼ ðGm=r Þ dm:
0 0
In the integral m is a function of r. In the simplified case of constant density it is

m ¼ qV ¼ q4p r 3 3
and
dm ¼ q4pr 2 dr:
Thus, we can transform the integral over mass into an integral over radius, with
r ranging between 0 and the final radius of the star, R,
ZM ZR

Eg ¼ ðGm=r Þ dm ¼ G q4p r 3 ð3r Þ q4p r 2 dr
0 0
ZR

¼ G q2 16p2 3 r 4 dr:
0
Since
ZR ZR

r dr ¼
4
d r 5 5 ¼ R5 5
0 0
and

q¼M 4pR3 3 ;
the gravitational potential energy is

Eg ¼ G q2 16p2 3 R5 5

¼ G M 2 16p2 R6 9 16p2 3 R5 5

¼ 3G M 2 ð5 RÞ:
Because the mass inside a star is not evenly distributed but concentrated towards
the center, the gravitational energy is actually greater than the value computed with
constant density. Thus, finally, we keep the formula but replacing the factor 3/5
with 1, that is,

Eg ¼ GM 2 R:
Exercise end.
Note that the gravitational potential energy is negative and Eg = 0 corresponds
to masses which are infinitely distant from each other. Thus, −Eg is the energy
necessary to smash the star bringing back all its mass components at infinite dis-
tance among them, like the binding energy inside a nucleus.
To estimate the average kinetic energy of the cloud, we multiply the mean
kinetic energy per unit mass of a single particle and the total mass of the cloud
232 7 Evolution
Ek ¼ ½ð3=2ÞkB T=ðlmu ÞM ¼ 3kB TM=ð2lmu Þ:
In the spirit of the calculation we also replace the factor 3/2 with 1, i.e.
Ek ¼ kB TM=ðlmu Þ;
and write the condition for the onset of contraction as

Eg [ Ek :
Thus, we get

G M 2 R [ kB TM=ðlmu Þ
and, finally,
GM=R [ kB T=ðlmu Þ:
This relation is called the Jeans criterion, and constrains the mass, radius, and
temperature of a cloud in order that a gravitational collapse starts.
A lower limit to the cloud mass is
M [ RkB T=ðGlmu Þ:
The minimum mass for a cloud to collapse is also called the Jeans mass, MJ, i.e.
MJ ¼ RkB T=ðGlmu Þ:
From the mean density

q ¼ MJ 4pR3 3 ;
the radius of the Jeans mass is
R ¼ ½3MJ =ð4pqÞ1=3 :
The radius can be used to get the Jeans mass in terms of temperature and density
MJ ¼ RkB T=ðGlmu Þ ¼ ½3MJ =ð4pqÞ1=3 kB T=ðGlmu Þ;

.
1=3
MJ MJ ¼ ½3=ð4pqÞ1=3 kB T=ðGlmu Þ;
2=3
MJ ¼ ½3=ð4pqÞ1=3 kB T=ðGlmu Þ
and, finally
MJ ¼ ½3=ð4pqÞ1=2 ½kB T=ðGlmu Þ3=2 :
Using this formula for an interstellar medium with T = 50 K, q = 1023 g cm−3

and l = 1 (pure H gas), we can find that the mass involved in a gravitational
collapse is at least a thousand times the solar mass.
Exercise 7.3 Show that using the values of T, q and l for the interstellar medium
quoted in the text, the Jeans mass is
MJ ¼ 1:20 103 MS :
7.1.4 Isothermal Collapse and Fragmentation
At the beginning of a gravitational collapse, the cloud is transparent and the work
done by gravity is spent to collisionally excite the transitions between the rotational
states of the cloud molecules, which are quantized, and does not significantly
increases the thermal agitation. The molecules, mostly H2, that are present at the
low temperatures of the interstellar medium, rapidly disperse the contraction energy
by getting de-excited and releasing radiation mainly at infrared and millimetric
wavelengths. Therefore, the first stage of a gravitational collapse can be approxi-
mated as isothermal. At a constant temperature, the Jeans mass decreases with
increasing density as
.
MJ 1 q1=2 :
Once the contraction is triggered, the cloud density grows and, hence, smaller
parts of the cloud can meet the Jeans criterion and continue the gravitational col-
lapse on their own. This process, called fragmentation, implies that the Sun and
stars usually were not born alone but in groups, or clusters, of many stars.
7.1.5 Adiabatic Collapse
With the contraction in progress, the cloud opacity grows and the collapse is no
longer isothermal. Now, the gravitational energy increases the internal energy
because molecules and neutral atoms, which are collisionally excited, are more
likely de-excited by another collision rather than by a photon emission. In this way
the temperature and pressure of the cloud increase. Neglecting radiation losses, we
234 7 Evolution
can assume that the collapse is adiabatic in this phase. Thus, temperature and
density are related by
T qc1 :
and the Jeans mass is proportional to

.
MJ T 3=2 q1=2 q3c=23=21=2 q3c=22 :
With c = 5/3, we get
MJ q1=2 :
Therefore, in the adiabatic phase of the collapse, the minimum allowed mass
grows with increasing density and this mechanism puts an end to the fragmentation
of the cloud. If we assume that fragmentation ends when the last fragment becomes
optically thick, i.e.
Rkm q 1;
where km is the opacity per gram, then the radius R of the smallest forming stars is
given by
R ¼ 1=ðkm qÞ:
Because

q¼M 4pR3 3 ;
we get

R ¼ 4pR3 3 ðkm M Þ ¼ 4pR3 ð3km M Þ:
Thus, the smallest radius is

R R3 ¼ 4p=ð3km M Þ;
R13 ¼ 4p=ð3km M Þ;
R2 ¼ 4p=ð3km M Þ;
R2 ¼ 3km M=ð4pÞ;
.
R ¼ ð3km M Þ1=2 ð4pÞ1=2 :
Let us use the expression of the Jeans mass in terms of the stellar radius, that is,
MJ ¼ RkB T=ðGlmu Þ:
Substituting the value of the smallest radius into the Jean mass gives an estimate
of the smallest stellar masses formed, which we denotes by Mm. Thus,
.
Mm ¼ ð3km Mm Þ1=2 ð4pÞ1=2 kB T=ðGlmu Þ
.
¼ ð3km M Þ1=2 kB T 2p1=2 Glmu ;
.
Mm Mm1=2 ¼ ð3km Þ1=2 kB T 2p1=2 Glmu ;
.
Mm1=2 ¼ ð3km Þ1=2 kB T 2p1=2 Glmu ;
. 2
Mm ¼ 3km ðkB T Þ2 2p1=2 Glmu :
With temperature and opacity values typical of a cool interstellar medium, this
formula gives masses of the last fragment of about 0.1 Ms. This is the order of
magnitude of the smallest and most common stellar masses.
Exercise 7.4 Show that with a temperature T = 100 K and a low opacity value of
the order of km = 0.1 g−1 cm2, the mass of the last cloud fragment is Mm = 3.67
1032 g = 0.184 Ms.
7.1.6 ProtoSun
Once the fragmentation of the cloud stops, the collapse of a single fragment, or
protostar, goes on using the gravitational energy to dissociate the molecules and
ionize neutral atoms (proto is a Greek word meaning first).
7.1.7 Sizes
The initial radius, Ri of the protoSun can be estimated from the formula of the Jeans
mass solved with respect to R,
Ri ¼ MJ Glmu =ðkB T Þ:
Thus, with T = 100 K and MJ = 1 Ms, we can estimate that the size of the
primordial solar cloud was more than 2 105 times the present radius.
Exercise 7.5 Show that in the beginning the protoSun had R = 2.31 105 RS.
After the contraction has dissociated molecules and ionized hydrogen, the radius
of the protoSun has been reduced to a value Rf, such that the difference of the
236 7 Evolution
gravitational potential energies before and after contraction equals the energy
absorbed in the dissociation and ionization processes, i.e.

GM 2 Ri þ GM 2 Rf ¼ MEd =ð2mu Þ þ M Ei =mu ;
where Ed = 4.5 eV is the H2 dissociation energy, Ei = 13.6 eV the H ionization

energy, and 2mu and mu are the masses of molecular and atomic hydrogen,
respectively. This allows us to infer that Rf is
Rf ¼ 1=½1=Ri þ ðEd =2 þ Ei Þ=ðmu GM Þ;
which in the case of the Sun gives
Rf ¼ 8:74 1012 cm ¼ 126 RS :
Exercise 7.6 Verify that
Rf ¼ 1=½1=Ri þ ð Ed =2 þ Ei Þ=ðmu GM Þ
and, for the Sun, Rf = 8.74 1012 cm = 126 RS.

Thus, at the end of this stage the protoSun extended almost to half the present
Sun-Earth distance.
7.1.8 Temperature
An upper limit to the temperature that the protoSun has reached after the dissoci-
ation and ionization phase can be obtained by assuming that all the energy that the
Sun has absorbed is finally converted into kinetic energy, that is,
M ðEd =2 þ Ei Þ=mu ¼ Ek ¼ 3kB TM=ð2mu Þ:
Thus,
3kB T=2 ¼ Ed =2 þ Ei ;
T ¼ ðEd þ 2Ei Þ=ð3kB Þ

¼ ð4:5 eV þ 2 13:6 eVÞ 1:60 1012 erg eV1 3 1:38 1016 erg K1
¼ 1:23 105 K:
No nuclear reaction could occur in the interior of the protoSun with such rela-
tively low temperatures.
7.1.9 Contraction Times
We can estimate how long the solar primordial cloud took to contract. During
contraction the Newton law of motion replaces hydrostatic equilibrium. Thus, for a
spherical corona with mass
dm ¼ qdV ¼ qSdr ¼ 4pr 2 qdr;
which moves with an acceleration a from a radial distance r toward the center of the
cloud with mass m, we can write

adm ¼ Gmdm r 2 þ Pðr ÞdS Pðr þ dr Þ dS;
where P(r) and P(r + dr) are the pressures on the bottom and top of the spherical
corona, respectively.
Replacing dm

a qSdr ¼ GmqSdr r 2 þ Pðr Þ S Pðr þ dr ÞS;
we cancel S

aqdr ¼ Gmqdr r 2 þ Pðr Þ Pðr þ dr Þ:
Replacing the pressure increment with the differential dP = (dP/dr)dr

aqdr ¼ Gmqdr r 2 ðdP=dr Þ dr;
we divide by qdr and finally get

a ¼ Gm r 2 dP=ðqdr Þ:
The acceleration is provided by the unbalance between gravity and pressure

gradient. Obviously if a = 0, then

dP=ðqdr Þ ¼ Gm r 2
and the protoSun is in hydrostatic equilibrium. During the collapse, when gravity
dominates, we can neglect the pressure gradient and approximate the acceleration
simply with

a ¼ Gm r 2 :
238 7 Evolution
The free fall time, tff, is defined as the average time for a spherical cloud with
radius R and mass M to completely collapse. Then, the mean speed during collapse
is of the order of

v ¼ R tff
and the mean acceleration, in absolute value, is given by

.
jaj ¼ v tff ¼ R tff2 :
Substituting this estimate into the Newton law gives

.
jaj ¼ R tff2 ¼ GM R2 :
where M is the total mass of the cloud.

Thus,
.
tff2 R ¼ R2 ðGM Þ;

tff2 ¼ R3 ðGM Þ;
and, finally,
.
tff ¼ R3=2 ðGM Þ1=2 :
In terms of density

q¼M 4pR3 3 ¼ M 4R3 ;
where in the spirit of the calculation, we have cancelled 3 and p, the free fall time is
1=2
tff ¼ R3 GM ¼ ½1=ð4GqÞ1=2
and, finally,
.h i
tff ¼ 1 2ðGqÞ1=2 :
Using the density of the interstellar medium, q = 10−23 g cm−3, into the free fall
time, we can estimate the duration of the first phase of the gravitational contraction,
when the cloud was transparent
.h i .h 1=2 i
tff ¼ 1 2ðGqÞ1=2 ¼ 1 2 6:67 108 dyn cm2 g2 1023 g cm3

¼ 6:12 1014 s 3:16 107 sy1
¼ 1:94 107 y
Inserting into the free fall time the solar mass MS and the initial radius of the
protoSun Ri, we estimate the time required for the second phase of the gravitational
collapse, when the cloud is opaque and H2 is dissociated and H ionized,
1=2 h 3 . i1=2
tff ¼ R3i ðGMs Þ ¼ 1:61 1016 cm 6:67 108 dyn cm2 g2 1:99 1033 g

¼ 1:77 1011 s 3:16 107 sy1 ¼ 5:61 103 y
These estimates indicate that the first transparent phase of the contraction is
much slower than the subsequent opaque phase.
In conclusion, we have identified some basic physical mechanisms that are at
work during the formation and first evolution of the Sun. However, the numerical
values we computed for some relevant parameters represent only order of magni-
tude estimates. Moreover, we have neglected the roles played by the magnetic fields
and rotation of the cloud while actually the physical processes involved in a stellar
accretion are more complex than our schematic description may suggest. Thus, as
observed in many young stars, also the primordial solar cloud probably went
through an accretion disk, with two opposite outward jets of material along the
rotation axis of the disk (Fig. 7.2). Then, a further fragmentation occurred, which
gave rise to the planets and minor bodies of the solar system.
7.1.10 The Hayashi Track
Now let us look in more detail at our Sun as it appeared a few hundred million years
after it had become a single object and mass accretion from the progenitor mass
cloud was ended. Then, the solar mass is almost equal to the present mass but the
radius and luminosity are several times larger. Once hydrogen is ionized inside, the
pressure rises and the contraction slows down: the protoSun is getting closer to
hydrostatic equilibrium. The surface has a temperature close to 3000 K. The core is
much warmer than the surface, but nuclear reactions are not yet active. The interior
is very optically thick and first reaches hydrostatic equilibrium. The high gas
opacity also implies a high radiative temperature gradient; hence, the excess thermal
energy is entirely transported by convection and radiated into space at the surface.
Once that even the outermost layers have reached hydrostatic equilibrium, the
radius has been reduced to about 4 times the present radius, while the luminosity is
several times the present luminosity.
240 7 Evolution
Fig. 7.2 The primordial solar cloud has been probably a rotating accretion disk with two opposite
outward jets of material along the rotation axis
During this phase, the Sun obeys a relationship between luminosity and effective
temperature that is named the Hayashi track after the Japanese astrophysicist
Chushiro Hayashi (1920–2010). The name track refers to the plot of this rela-
tionship into a Cartesian diagram where the x-axis represents effective temperatures
and the y-axis luminosities. This kind of diagrams is an important tool to understand
stellar evolution and is called the Hertzsprung and Russell (or HR) diagram, after
the two astronomers who introduced it (Ejnar Hertzsprung, Danish, 1873–1967 and
Henry Norris Russell, American, 1877–1957). The Hayashi tracks identify the loci
of totally convective stellar structures in the HR diagram and are the impassable
right limits for stars in that diagram.The Sun’s Hayashi track in the HR diagram is
plotted in Fig. 7.3.
Exercise 7.7 Show that in the HR diagram the curves at constant stellar radius are
straight lines inclined by 104° with respect to the x axis.
Worked Exercise 7.8 Spectral type, color index and luminosity class.
Spectral type and color index are two quantities related to the effective tem-
perature, Te, that are both used as alternative for the x axis of the HR diagram.
The spectral types classify stars on the base of the spectral lines occurring in
their observed spectra, as Table 7.1 shows. Each type is denoted by a letter (O, B,
A, F, G, K, M in the order of decreasing Te) and divided in 10 subtypes (from 0 to 9
in the order of decreasing Te). The Sun has the spectral type G2.
A color magnitude is a measure of the decimal logarithm of the stellar emission
flux in a restricted wavelength range obtained with a spectral filter. The color
magnitude defined by a filter with a spectral band C(k), is
Fig. 7.3 The Sun’s Hayashi track in the HR diagram (red piecewise straight line). Note that the
HR diagram is the log Te–log L plane, with the log Te axis in reversed order, that is, Te increases
toward the origin. When the evolution is controlled by a slow contraction in almost hydrostatic and
thermal equilibrium, the position of the protoSun in the HR diagram descends along the Hayashi
track with decreasing L and slightly increasing Te, as indicated by the arrow. The red disk at the
beginning of the track represents the Sun on the birth line, at an age of 1 My. The symbol ‘’, at
Te = 5578 K and L = LS, is the current position of the Sun on the MS, at an age of 4.57 Gy. The
dotted line represents the zero age main sequence (ZAMS), that is, the positions that stars with
different Te and L occupy at the onset of the H burning
Table 7.1 Spectral types

Spectral Origin of characteristic spectral lines Corresponding effective
type temperature, K
O HeII, HeI, OIII, NIII, SiIII, HI 41,000
B HeI, HI, OII, CII, NII, SiII 31,000
A HI (strongest lines), CaII, MgII, FeII 9500
F HI (weaker lines), CaII, ionized metals 7240
G HI (weaker lines), CaII, ionized and neutral metals 5920
K CaII (strongest lines), neutral metals (strong lines), 5300
HI (weak lines)
M Neutral atoms (strong lines), TiO 3850
In the order from left to right columns list the spectral type notation, the atoms or ions that
originate the most prominent spectral lines, and the effective temperature corresponding to the
hottest stars of each type
242 7 Evolution
2 1 3
Z
MC ¼ 2:5 log4 CðkÞFk dk5 þ constant:
0
The magnitude scale is such that a –2.5 magnitude difference corresponds to a

flux ratio of 10 and inspires to the human eye response to brightness.
A color index is the difference of two color magnitudes. For instance, the
B − V color index is defined by
B V ¼ MB MV :
The blue, MB, and visible, MV, magnitudes are defined by two filters, CB(k) and
CV(k), with bandwidths of about 100 nm and central wavelengths at 440 nm and
548 nm, respectively.
The constant in the B − V index is defined in such a way that B − V = 0 for the
spectral type A0. The Sun has B − V = 0.650 ± 0.005.
Stellar luminosities are also grouped in different intervals, or luminosity classes,
as listed in Table 7.2. Each class is denoted by a Roman number (from I to V in the
order of decreasing luminosity, with the most luminous stars distinguished in Ia and
Ib). Note that the luminosity range associated with a class varies with Te; for
instance, all the main sequence stars are dwarfs (Fig. 7.3). The present Sun is a
dwarf star, or luminosity class V.
Exercise end.
7.1.11 Virial Theorem
In a steady state configuration governed by hydrostatic equilibrium, we expect that

the gravitational potential energy in modulus, –Eg, and the kinetic energy, Ek, which
corresponds to the internal energy of the Sun, Eint, are of the same order of mag-
nitude. Actually, the overall energy balance is adjusted by the virial theorem, which
exactly requires
1
Ek ¼ Eg :
2
Table 7.2 Luminosity Luminosity class Description

classes
Ia Bright supergiants
Ib Supergiants
II Bright giants
III Giants
IV Subgiants
V Dwarfs
While the Sun moves along the Hayashi track, the only energy input is supplied
by the gravitational contraction, which occurs so slowly that can be represented by
a sequence of states in hydrostatic equilibrium, each slightly different from the
following state. According to the virial theorem, half of the energy supplied by the
gravitational contraction is transformed into internal energy of the Sun, thus sus-
taining the temperature and pressure and ensuring the hydrostatic equilibrium. The
remaining half of the gravitational energy is dispersed in the space through the
radiation emitted from the surface.
The total energy is the sum of the internal and gravitational potential energies
Etot ¼ Eint þ Eg ;
where we neglected the kinetic energy of convective motions, in agreement with the
assumption of hydrostatic equilibrium.
Because of the virial theorem
Eg ¼ 2Ek ¼ 2Eint ;
thus
Etot ¼ Eint 2Eint ¼ Eint
and also
1 1
Etot ¼ Eg þ Eg ¼ Eg :
2 2
The total energy is negative, which means that the Sun is gravitationally bound,
is equal to half the gravitational energy, and, in absolute value, equals the internal
energy.
During the slow contraction in quasi hydrostatic equilibrium, the total energy
does not remain constant over time but decreases with a rate that is equal to the
luminosity, that is,
dEtot =dt ¼ L:
Moreover,

dEg dt ¼ 2dEtot =dt ¼ 2L\0
and
dEint =dt ¼ dEtot =dt ¼ L [ 0:
Thus, the primordial Sun emitting radiation into space contracts and gets hotter
while a rigid body that loses energy cools. The slow contraction continues until the
244 7 Evolution
central temperature does not exceed several million degrees and, hence, nuclear
fusion reactions can start. We can estimate the time required by the slow contraction
on the Hayashi track, tsc, from the equation

d Eg dt ¼ 2L;
which requires for the average energy, time and luminosity

Eg tsc ¼ 2L:
We take the present values of the solar luminosity and mass, and 10 times the
present radius, Thus, we get

tsc ¼ Eg ð2LS Þ ¼ GMS2 ð20RS LS Þ
2 .
¼ 6:67 108 cm3 g1 s2 1:99 1033 g 20 6:96 1010 cm 3:844 1033 erg s1

¼ 4:93 1013 s 3:16 107 s y1 ¼ 1:56 106 y ¼ 1:56 My:
Therefore, the slow contraction of the Sun along the Hayashi track lasted for a
few million years. This particular lifetime is also called the Kelvin-Helmholtz time
(Hermann Helmholtz was a German scientist, 1821–1894).
7.2 Main Sequence Phase
When hydrogen fusion finally starts in the core, the nuclear energy source quickly
becomes capable of producing the amount of energy dispersed into space by
radiation. The total energy becomes constant and the Sun reaches the thermal
equilibrium. In this condition both the gravitational and internal energies are con-
stant, thus the Sun neither contracts nor is heated and the pressure and temperature
structure remains almost unchanged for a long time. This is the main sequence
(MS) phase, in which the Sun is at present and will remain for about additional
5 Gyr. The model constructed in the first part of the book corresponds to this
evolutionary phase.The name main sequence refers to the particular relationship
between effective temperature and luminosity obeyed by the stars that generate
energy through the nuclear fusion of hydrogen into helium (Fig. 7.3). Stars during
the MS phase are also called dwarfs due to their luminosity class (Worked Exercise
7.7). Thus, the present Sun is a yellow dwarf because of its color (Chap. 2).
7.2 Main Sequence Phase 245
7.2.1 Stability of the Solar Equilibrium
The main sequence phase is characterized by both the hydrostatic and thermal
equilibrium of the Sun. Neglecting the slow chemical evolution that nuclear fusion
causes in the core, the solar equilibrium is stable, that is, any small perturbation of
the structure is damped and the Sun immediately recovers its initial state. Suppose,
for example, that the temperature in the core has a slight increase, which occurs so
quickly to be considered an adiabatic perturbation. Then, the density increases as
q T 1=c1 ;
and the nuclear energy production rate per unit mass increases, roughly as
em qT 4 ;
and the luminosity also increases, let say from L to Ln. The total energy is no longer
constant with time and its variation in the time interval dt is equal to the luminosity
increase, that is,
dEtot =dt ¼ Ln L [ 0:
Thus

dEg dt ¼ 2ðLn LÞ [ 0
and
dEint =dt ¼ ðLn LÞ\0:
The gravitational potential energy increases and the internal energy decreases,
that is, the Sun expands and cools. In this way the reaction damps both the tem-
perature and density perturbations. The temperature and density reductions decrease
in turn the energy production rate bringing again the luminosity and total energy to
their equilibrium values. Conversely, a negative temperature fluctuation causes the
opposite reaction, i.e. the Sun shrinks and heats up, which equally restores the
initial equilibrium conditions.
Therefore, by virtue of the hydrostatic balance and nuclear fusion mechanism,
the Sun works as a huge thermostat. The thermostat was been perfect if hydrogen,
that is its nuclear fuel, would not gradually be exhausted.
246 7 Evolution
7.2.2 Equations of Evolution
During the main sequence phase solar luminosity is produced by the fusion of
hydrogen into helium. We can estimate the duration of this phase, tMS , using the
Kelvin-Helmholtz time that we write now as
tMS ¼ Enuc =LS ;
where LS is the present luminosity. Enuc is the total nuclear energy produced by the
H fusion, which we can estimate from the mass—energy relation with a mass equal
to the solar mass multiplied by the mass defect of an a particle in unit of 4 mp, that
is 0.028/4 = 0.007 (Chap. 1). Thus

tMS ¼ Enuc =LS ¼ 0:007MS c2 LS
2 .
¼ 0:007 1:99 1033 g 2:99 1010 cms1 3:84 1033 ergs1

¼ 3:26 1018 s 3:16 107 s y1 ¼ 1:03 1011 y:
This value is a factor 10 greater than the lifetime calculated with an accurate
approach that accounts for the hydrogen burning layer by layer.
Nuclear fusion reactions alter the chemical composition of the solar interior. The
equations that describe the time variations of the involved elements are called
equations of evolution. Here we limit to consider the change in the percentage of
hydrogen. Considering that hydrogen is completely ionized in the core, the
hydrogen mass fraction, X, is in terms of the gas density

X ¼ mp Np q;
where Np is the proton number density and mp the proton mass.

Denoting by em the energy production rate per unit mass and by Q the energy
produced in one pp cycle, the number of cycles that occur per second per unit mass
is em/Q. The H mass fraction dX lost in the time dt is given by the mass lost in a
cycle, 4mp, times the number of cycles occurring in dt, that is,

dX ¼ 4mp em dt Q
and the X loss rate is

dX=dt ¼ 4mp em Q:
The nuclear energy rate em has the approximate expression in terms of density
and temperature
2=3 1=3
em ¼ 2:9 104 q X 2 T9 e3:37=T9 erg g1 s ;
where T9 = 109 K. We define
2=3 3:37=T91=3
eX ¼ 2:9 104 q T9 e ;
thus,
em ¼ eX X 2
and the evolution equation becomes

dX=dt ¼ 4mp eX X 2 Q:
Since X is dimensionless, the right hand side has the dimension of an reciprocal
of a time and we write the hydrogen evolution equation as

dX=dt ¼ X 2 t
with

t¼Q 4mp eX :
Assuming that t remains constant, we can solve the equation of evolution by

separation of variables
dX=X 2 ¼ dt=t:
We integrate between the times t = 0 and t, in which the H mass fractions are X0
and X, respectively
ZX Zt
dX=X ¼
2
dt0 =t:
X0 0
R
Because dX X 2 ¼ 1=X, we get
1=X þ 1=X0 ¼ t=t;

1=X ¼ 1=X0 t=t;
1=X ¼ 1=X0 þ t=t;
1=X ¼ ð1 þ X0 t=tÞ=X0 :
Finally,
248 7 Evolution
Fig. 7.4 The time variation of the hydrogen mass fraction during the MS phase. The time t on the
x axis is in unit of the halftime t1/2, and the H mass fraction on the y axis in unit of the initial value
X0
X ¼ X0 =ð1 þ X0 t=tÞ:
The mass fraction X is halved at the time t1/2 such that
t1=2 ¼ t=X0
and approaches 0 only for infinite time. This simple chemical evolution is plotted in
Fig. 7.4.
Exercise 7.9 Assuming the values of the solar center for the temperature and
density and that of surface for the initial H mass fraction, show that the hydrogen
half-life is t1/2 = 3.03 Gy.
With the values of the solar center for temperature and density and that of
surface for the initial H mass fraction, we can evaluate that the hydrogen half-life is
t1=2 ¼ 3:03 Gy:
This value is slightly smaller than the current age of the Sun of 4.57 Gy.
Assuming that for the present Sun approximately half of the hydrogen at the center
has been burnt, we get NHe/Np = 0.5 and

X ¼ mp Np q ¼ mp Np mp Np þ 4NHe ¼ Np Np þ 4NHe ¼ 1 1 þ 4NHe =Np
¼ 1=ð 1 þ 4 0:5Þ ¼ 1=3 ¼ 0:33:
This is close to the value X = 0.355 of the Senesi et al. model.

The time variation of the He mass fraction is immediately obtained from the
condition
X þ Y þ Z ¼ 1:
Neglecting metals, we get
dX=dt þ dY=dt ¼ 0
and
dY=dt ¼ dX=dt:
Thus, of course, the He mass fraction decreases by the same amount that the
hydrogen mass fraction increases.
Note that the equations of evolution limited to hydrogen and with a constant
coefficient are a simplification in the spirit of constructing an analytical approximate
model of the Sun. Actually, an accurate approach to the evolution of the chemical
composition due to the H fusion requires that the proper radial and time variations
of all the chemical species involved in nuclear reactions are taken into account.
We can now complete the description of the procedure which is used to calculate
an accurate standard model of the present Sun.
7.2.3 Evolution Sequence
Because the present chemical composition is the evolution of the initial composi-
tion, in order to determine the present solar model we must construct a sequence of
models which are each the evolution of the previous model, with the contents of the
elements involved in nuclear reactions that are modified according to the equations
of evolution. The beginning of the sequence corresponds to the tip of Hayashi line
and is characterized by a uniform chemical composition. Actually, the initial
models may be unrealistic because dating to ages the Sun was still in the parent
cloud and not yet in thermal equilibrium. The first real model, or birth line model,
has an age of about 1 My and is in hydrostatic and thermal equilibrium.
Each model is the solution of the four structure equations and three constitutive
equations, that we have previously described and Table 7.3 summarizes in the form
adopted in the text.
At each step of the evolution sequence, the model integration starts from rea-
sonable guesses of the radius and luminosity, which are, then, adjusted until the
mass and luminosity are zero at the solar center. The age of the Sun, tS = 4.57 Gy,
is a rather well known quantity, based on the age of the meteorites that were formed
in the early solar system. On the contrary, the initial abundance of helium, Y0, and
250 7 Evolution
Table 7.3 Structure and constitutive equations in the form adopted in the text

Energy transport, radiation ðdT=dr Þr ¼ 3LkR ð64pr 2 rT 3 Þ
Energy transport, convection ðdT=dr Þa ¼ ½ðc 1Þ=c ðT=PÞ dP=dr

Hydrostatic equilibrium dP=dr ¼ qGM r 2
Mass conservation dM ¼ 4 p r 2 q dr
Energy conservation dL ¼ em dM
State equation P ¼ qRT=l
Opacity, Kramers‚ law kR ¼ kR0 qa T b
Energy production rate 2=3 3:37=T91=3
em ¼ 2:9 104 qX 2 T9 e
The Schwarzschild criterion allows us to choose between radiation or convection energy transport
the mixing length, lm, that enters into the convection energy flux, are two more
uncertain parameters. Thus, usually an iterative process fits Y0 and lm until the Sun
model corresponding to the time tS reaches the present radius and luminosity.
Figures 7.5 and 7.6 show some characteristics of the model in the evolution
sequence that represents the Sun on the birth line, at the beginning of the Hayashi
track (data from Senesi et al.). Then, the Sun was only 106 years old. Nuclear
reactions had not yet altered the original uniform chemical composition.
Temperatures, pressures and densities were everywhere smaller than those of the
present Sun. The effective temperature was 4515 K, that is, 0.78% the present
effective temperature, the radius was 2.19 times the present radius and the lumi-
nosity 1.79 times the present luminosity.
Let us now look at the physical consequences that the chemical evolution during
the main sequence phase has on the internal structure of the Sun.
7.2.4 Main Sequence Evolution
The continuous transformations of protons into a particles during the MS phase

progressively reduce the particle number density in the core and, hence, the pres-
sure. The pressure reduction causes a slow contraction of the core, which restores
the hydrodynamic equilibrium and also heats the core at expense of gravity.
Accordingly, nuclear fusion becomes more efficient and luminosity increases. The
increased luminosity also makes the Sun expand and the radius and surface tem-
perature gradually increase. As a result, after about 4.5 Gy spent in the MS phase
until today and about half of the initial hydrogen burnt in the core, the Sun’s
luminosity has increased by 43%, radius by 12% and effective temperature by 3.2%
(Schröder and Smith, 2008, which is our reference for the MS and post-MS evo-
lution data). At an age of 7.13 Gy the Sun reaches the maximum effective tem-
perature of 5820 K and a luminosity L = 1.26 LS. From that moment on, the solar
Fig. 7.5 The Sun’s structure on the birth line, at an age of 1 My (Senesi et al.). In the order from
top to bottom the panels show the decimal logarithms of the temperature, pressure and density as
functions of the radius fraction. Black line refers to the protoSun and red line to the present Sun
surface gradually become colder, but the luminosity continues to grow. At an age of
10 Gy the hydrogen fusion into helium in the core is turned off, the luminosity and
radius reach 1.84 LS and 1.37 RS, respectively, and the effective temperature drops
to 5751 K, that is, virtually equal to the present value.
252 7 Evolution
From the astronomical point of view the physical changes that accompany the
MS phase are small compared with those occurring in other evolution phases.
However, these changes are very drastic in terms of life on the Earth.
7.2.5 MS Evolution and Life on the Earth
Solar radiation is the first factor determining the temperature of the Earth’s surface.
The Earth absorbs only a fraction of the solar radiation incident on it, which we
denote by 1−A, while reflects the remaining fraction, A, backwards. The quantity
A is called the albedo, with a Latin word. A depends on the surface characteristics of
reflecting material; here we limit to consider only global mean values for A. When,
in the course of its evolution, the Sun has a luminosity L, the solar energy that the
Earth absorbs per second, Pa, is

Pa ¼ Lð1 AÞR2E 4D2E :
Here RE and DE are the Earth radius and Earth-Sun distance, respectively, and
the factor pR2E/(4pD2E) = R2E/(4D2E) represents the area fraction that the Earth covers
on the spherical surface of radius DE (Fig. 2.7 and relative discussion).
In order to determine the Earth surface temperature, TE, we neglect the green-
house effect in the atmosphere and radioactive heat sources in the interior, and
simply assume that the Earth reemits all the energy it absorbs, like a black body.
Thus, the total emitted power, Pe, is
Fig. 7.6 The Sun’s structure on the birth line, at an age of 1 My (Senesi et al.). In the order from
top to bottom the panels show the H and He mass fractions as functions of the radius fraction.
Black line refers to the protoSun and red line to the present Sun. Note that in the present Sun the
higher values of H and lower values of He for r/RS > 0.4 are produced by microscopic diffusion,
because, due mainly to gravity, lighter elements slowly diffuse toward the surface while heavier
elements tend to diffuse toward the center
254 7 Evolution
Pe ¼ 4pR2E rTE4 :
Because Pa = Pe, we get

4pR2E rTE4 ¼ Lð1 AÞR2E 4D2E ;

TE4 ¼ Lð1 AÞ 16prD2E ;
1=4
TE ¼ Lð1 AÞ 16prD2E :
This formula gives the Earth surface temperature as a function of the solar
luminosity, albedo and Earth-Sun distance. To calibrate the formula, we take the
present values of the solar luminosity and Earth-Sun distance, and a constant mean
albedo A = 0.3, which represents a mix of water and ground. Then, we get
(
.h
TE ¼ 3:84 1033 erg s1 1 0:3 16 3:14 5:67 105 erg s1 cm2 K4
)1=4
i
2 2
1:50 10 cm
13
¼ 255 K:
The present average temperature at the Earth surface is 15 °C =

15 + 273 K = 288 K, that is, 33 K more. We may speculate that this higher tem-
perature is mainly due to extra warming by the Earth atmosphere. However, we
simply add the estimated temperature defect and get the final formula
1=4
TE ¼ Lð1 AÞ 16prD2T þ 33 K:
This simplified approach clearly shows how sensitive the average temperature on
the Earth surface is to the variations of the solar luminosity. In particular, looking at
the past, a luminosity that at the beginning of the MS phase was a 30% lower than
that of the present Sun, with the other parameters unchanged, would have produced
an Earth surface temperature of −7 °C.
Exercise 7.10 Verify that L = 0.7 LS implies TE = 266 K.
More accurate calculations confirm such low temperatures and imply that the
entire Earth was covered by ice at that time. A complete glaciation is at odds with
the present physical conditions on the Earth. In fact, the ice albedo is higher than
the present Earth’s average value, about 0.6 instead of 0.3, hence, the subsequent
increase to the present luminosity would not have been able to melt the ice cover.
Actually, the Earth climatic variations depend on other factors besides solar radi-
ation, thus, we may reconcile solar evolution and Earth surface temperature
Fig. 7.7 Solar constant variation from 12 January 1976 to 1 February 2017. The day number on
the x axis is such that 1 corresponds to 1 January 1980, 5 103 to 8 September 1993 and 104 to 15
May 2007. The horizontal red line corresponds to the average value of 1360.9 Wm−2. The plot is
based on a composite dataset constructed with the approach described in Fröhlich (2006). The
dataset is version 42_65_1702 in file composite_42_65_1702.dat. It includes new data from the
VIRGO Experiment on the cooperative ESA/NASA Mission SoHO, and is available from ftp.
pmodwrc.ch/pub/data/irradiance/composite/. Courtesy of PMOD/WRC, Davos, Switzerland
assuming a different chemical composition of the Earth’s atmosphere in the past. In

particular, a larger amount of carbon dioxide, which is compatible with the lack of
the plants that transform it into oxygen through the chlorophyll photosynthesis,
could have strengthened the greenhouse effect and prevented a freezing of the Earth
surface in spite of the reduced solar radiation. In this way, just the greenhouse
effect, which is of great concern today, could have saved the Earth in the past from
complete glaciation.
Turning back to the present and to a shorter time scale, the solar constant varies
by a few per thousand during an eleven years long magnetic activity cycle
(Fig. 7.7).
Considering this variation for L in the approximate formula shows that the
magnetic activity cycle should not produce appreciable variations of the Earth’s
surface temperature.
Finally, projecting into the future, at the age of 10 Gy corresponding to the end
of the MS phase, the Sun will have L = 1.84 LS. According to our formula, this lead
to TE = 57 °C, a temperature that should make life extremely difficult on Earth.
Exercise 7.11 Verify that L = 1.84 LS implies TE = 57 °C.
Thus, the end of the main sequence phase should also mark the beginning of the
end for the Earth as a habitable planet. Nevertheless the evolution of the Sun will
continue after the end of the main sequence phase.
256 7 Evolution
7.3 Post-main Sequence Evolution
7.3.1 Hydrogen Shell Fusion and Red Giant Phase
During the MS phase the hydrogen burning in the core causes a slow growth of the
solar luminosity roughly at constant effective temperature. The progressive reduc-
tion of hydrogen requires a continuous increase of the temperature to keep the same
efficiency of nuclear reactions. The temperature rise makes fusion start in the layers
surrounding the core that are yet rich in hydrogen. Eventually, when the core
becomes made only of helium, hydrogen burning occurs only in a shell surrounding
the central region. Note that with increasing temperature the CNO cycle becomes
more efficient than the pp cycle, hence, when hydrogen is to exhaust it is the CNO
cycle that produces most of energy and drives the evolution.
The beginning of the hydrogen shell burning is marked by an expansion of the
outer layers at roughly constant luminosity. Then, the effective temperature,
T4e * L/R2, decreases and the increased temperature gradient makes convection
stronger in the layers overlying the shell.
Continuing to burn, also the hydrogen in the shell is progressively consumed and
converted into helium. Thus, the mass of the helium core continuously increases
while the hydrogen burning shell progressively moves towards the outer layers.
Due to the absence of energy sources, the helium core initially is almost isothermal,
but such a situation cannot persist. The helium accretion makes the weight increase,
hence, a contraction occurs that increases the density and restores the hydrostatic
balance. During this contraction, as long as the hydrogen shell remains the only
source of nuclear energy, the Sun moves in the HR diagram along a line that is
parallel to the Hayashi track, but now Te decreases and L increases with age, thus
the Sun climbs the track. The hydrogen shell fusion lasts about 1.75 Gy before
switching off. In most of this time, 1.5 Gy, the Sun expands, up to 3.3 RS, its
surface cools down to approximately 4300 K, and luminosity becomes almost
3 times greater. The Sun now is a sub red giant, a type of star whose name
emphasize the increased size (giant) and the low surface temperature (red color).
During the subsequent 0.25 Gy, when most of the hydrogen surrounding the
core is fused into helium, the luminosity growth, the expansion of the outer layers
become even stronger and the surface temperature has a further decrease. Climbing
the Hayashi track, the Sun passes from sub red giant to red giant (Fig. 7.8).
At the end of this phase the Sun has R = 256 RS; that is slightly more than 1 AU,
so that it has incorporated Mercury, Venus and Earth inside. The Sun is now 2730
times brighter than today, and its surface has an effective temperature of 2602 K.
During the red giant phase the solar wind from the outer layers becomes stronger
and, hence, the mass loss is high. At the end of the red giant phase the Sun’s mass is
reduced to 0.668 MS, according to Schröder and Smith (2008). However, it is
difficult to exactly determine how much of the solar mass is dispersed into space,
which makes somewhat uncertain the evolution of the Sun after the main sequence
phase.
7.3 Post-main Sequence Evolution 257
Fig. 7.8 The track in the HR diagram during the evolution of the Sun from yellow dwarf to red
giant (red piecewise straight line). The previous evolution is represented in black. (See Fig. 7.3
and related description)
7.3.2 Mass Loss
At present, the solar wind is a continuous mass flow from the solar corona into the
heliospheric space, which is modulated by the magnetic activity cycle. The wind is
mostly composed of electrons and protons, and a few percent of a particles. The
mass loss rate due to the wind is on average
dM=dt ¼ 2:4 1012 g s1 ¼ 3:8 1014 MS y1 :
At the end of the 1950s Parker first theoretically described in a consistent way
the solar wind on the basis of a stationary mass flow in an isothermal and spherical
symmetric corona (Eugene N. Parker, born 1927, is an American astrophysicist).
However, Parker’s theory did not include the acceleration mechanism. Moreover,
recent observations clearly demonstrate the absence of spherical symmetry. In fact,
the wind has a slow component that has velocities of about 400 km s−1 and is
observed over the equatorial areas, and a fast component that has velocities of about
750 k ms−1 and is observed at higher solar latitudes. The structure of the wind is
also related to that of the magnetic fields. Thus, the sources of the slow component
are associated to the closed magnetic loops interconnecting sunspots, while the fast
component originates from the coronal holes, which are magnetic regions with open
field lines spreading toward the heliosphere.
When the Sun becomes a red giant, it is plausible that the solar wind becomes
more massive and slower than the present wind. Actually, such type of wind is
258 7 Evolution
observed in many red giant and supergiant stars. The solar evolution model of
Schröder and Smith assumes the following expression for the mass loss rate

dM=dt ¼ 4 1013 g ðL=LS Þ ðR=RS Þ ðM=MS Þ MS y1 ;
where L, R and M are the luminosity, radius and mass, and η is a fudge factor,
which is used to adapt the formula to a selected group of stars. For instance, for the
mass loss associated with the present solar wind, η = 0.05. This expression is a
version of the empirical formula of Reimers (1975, 1977), which is based on the
observation of a set of giant and supergiant stars.
Thus, including mass loss in the evolution sequence requires that at each time
step both the chemical composition and total mass of the Sun are updated according
to the evolution equations and mass rate equation, respectively.
7.3.3 Degenerate Gas
The helium fusion begins when the Sun’s core reaches temperatures T * 108 K
and densities q * 106 g cm−3. In these conditions, the electron gas component is
no longer an ideal gas but, instead, a degenerate gas, whose pressure depends much
more on density than on temperature. The relation between the thermodynamic
variables of a degenerate gas is a consequence of the Pauli’s exclusion principle.
This principle of quantum mechanics holds for all the elementary particles called
fermions. Any elementary particle has a quantized intrinsic angular momentum, or
spin, which is
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
S ¼ ðh=2pÞ sðs þ 1Þ;
where the spin quantum number s can be any non-negative integer and half-integer
number. Fermions are characterized by half-integer values of the spin quantum
number, such as s = 1/2, 3/2, and so on. In particular, electrons, protons and
neutrons, all are fermions with s = 1/2. The exclusion principle prevents more than
one fermion from occupying the same quantum state, which is defined by a given
set of quantum numbers. The exclusion principle for fermions adds to the uncer-
tainty principle, which requires that any quantum particle, localized between the
points x and x + dx and with a momentum within p and p + dp, must have
dxdp [ h=ð2pÞ:
Quantum mechanics effects become important and electron degeneracy occurs

when the average distance between electrons de is smaller than the De Broglie
wavelength kDB, i.e.
de \kDB :
The interelectronic distance decreases when density increases while the De

Broglie wavelength decreases with increasing temperature. In the red giant core, the
density is sufficiently high and the temperature not too large that the electron gas is
degenerate. In this condition the electron energy distribution is different from the
Maxwell distribution and follows the Fermi-Dirac distribution (Paul A. M. Dirac
was an English physicist, 1902–1984). Because of the limits imposed by the
exclusion and uncertainty principles, in the Fermi-Dirac distribution fewer electrons
can access lower energy states than in the Maxwell distribution. The gas pressure is
a measure of the particle momentum, which is related to energy by
pffiffiffiffiffiffiffiffiffi
p¼ 2mE :
Thus, a degenerate gas, whose particles are forced to occupy energy states
relatively higher, exerts a higher pressure than an ideal gas does at the same
temperature and density. For T approaching 0 the degeneracy becomes complete,
that is, all the quantum states with lower energies are occupied and the only way to
make the pressure increase is to increase the number of particles and, hence, the
density. In these conditions, the pressure is no longer sensitive to temperature and
the degenerate state equation becomes
P q5=3 :
Exercise 7.12 Gas degeneracy occurs when the average distance between particles
d is smaller than the De Broglie wavelength kDB, i.e. d < kDB. Show that this
relation implies
N [N
where
.
N ¼ ð3mkB T Þ3=2 h3 ;
and N, m and T are the particle number density, mass and temperature, respectively.
Verify the condition for the electron degeneracy in the red giant Sun with
T = 108 K and q = 106 g cm−3.
7.3.4 Helium Fusion
At the end of the red giant phase, the degenerate helium core has acquired a critical
mass of about 0.5 MS and has approximately reached T = 108 K and q = 106 g cm−3
260 7 Evolution
at the center. These physical conditions light a new nuclear power source: helium
fusion. In the helium cycle three helium nuclei merge into a carbon nucleus passing
through an unstable beryllium nucleus, according to the scheme
4
He þ 4 He ! 8 Be
8
Be þ 4 He ! 12 C:
The average energy released in the formation of carbon by 3 a particles is

7.273 MeV. Almost the same energy amount is then produced when a carbon
nucleus interacts with a helium nucleus to form oxygen
12
C þ 4 He ! 16 O þ c:
Other heavier nuclei can also be formed by captures of a particles and protons at
the core boundary.
An important characteristic of the helium fusion is its strong temperature
dependence. In particular, in the vicinity of T = 108 K, the energy production rate
eHe is approximately given by
eHe ¼ 4:4 108 q2 Ya3 T840 erg g1 s1 ;
where Ya is the mass fraction of a particles and T8 = T/108 K. This implies, for
example, that a 10% greater temperature makes the energy production rate increase
by 1.140 = 45 times. The result is that the hot, dense gas at the solar center is highly
explosive, that is, a small temperature change can release a large amount of energy.
Moreover, in the end of the red giant phase, a physical phenomenon occurs in the
core that breaks the thermostatic mechanism that had secured stability since the
arrival on the main sequence. This phenomenon is the electron degeneracy.
7.3.5 Thermal Instability and Helium Flash
During the helium burning the pressure in the Sun’s core is dominated by electrons,
which are twice as many as a particles. Degeneracy does not occurs for a particles
which have spin s = 0, and a larger mass than that of electrons and, hence, a smaller
momentum for the same energy. On the contrary, electrons are a strongly degen-
erate gas, to which we can apply with good approximation the equation for a
completely degenerate gas. This has dramatic consequences for the thermal
stability.
The onset of the helium burning produces energy that raises temperature. In a
perfect gas any temperature increase results in a pressure increase that expands and
cools the core. Such a reaction reduces the nuclear energy production rate and
restores the thermal equilibrium. But the solar core behaves in a quite different way
when it is controlled by a strongly degenerate electron gas. In this case the pressure
is almost independent of temperature, thus, the core does not expand and not cool in
response to the temperature increase produced by helium fusion. Because of its
strong temperature sensitivity, the energy released by He fusion rapidly increases
making the temperature grow even more and departing from thermal balance. In a
few hours the core temperature increases by almost 80% and the luminosity, which
is approximately proportional to the nuclear energy production rate, increases by
1.840, that is, more than 109, a value comparable to that of an entire galaxy. This
quick and explosive phase is called helium flash. The helium fusion proceeds by
self-stimulation until temperatures are reached high enough to remove the electron
degeneracy and restore the usual thermostatic mechanism that stabilizes the thermal
equilibrium. Detailed calculations show that the electron degeneracy can be
removed only in the layers that absorb most of the released energy and overly the
layer where the flash was triggered. The process then proceeds through a series of
secondary flashes, separated in time and progressively closer to the solar center,
until the degeneracy is completely removed in the core and helium begins to burn in
a not-explosive way.
7.3.6 Horizontal Branch and Asymptotic Giant Branch
At an age slightly more than 12 Gy (Schröder and Smith, 2008, Table 1), helium
starts to fuse in the solar core in quiescent mode. At the same time, hydrogen fusion
also occurs in the outer layers that are still sufficiently hot and hydrogen-rich. This
phase is named the horizontal branch (HB), after the sequence at almost constant
luminosity that low mass central helium burning stars, with the same core mass but
varying envelope mass, produce in the HR diagram. The Sun abruptly reaches the
HB locus after the helium flash at the end of the red giant branch (Fig. 7.9).
Now the Sun has L = 33.7 LS, R = 11.2 RS, and Te = 4667 K. In addition,
because of the mass loss during the red giant phase, the Sun keeps only 66.8% of its
initial mass. However solar luminosity cannot remain constant too long. Powered
by the fusions of both hydrogen and helium, the luminosity starts to grow until the
Sun becomes very bright. The effective temperature decreases, radius grows and
mass loss turns to be strong. Because the track in the HR diagram is similar to that
of a red giant, this phase is called the asymptotic giant branch (Fig. 7.9). At the end
of this phase, which lasts 0.13 Gy, at the tip of the asymptotic branch, the Sun has
L = 2090 LS, R = 149 RS, Te = 3200 K and M = 0.546 MS. During the phase of
asymptotic giant branch the Sun goes through a series of instabilities, or thermal
pulses, which are characterized by abrupt adjustments of the core pressure and
temperature. In the evolution model we are following, the last and strongest of these
pulses coincides with the end of the asymptotic branch and sets L = 4170 LS,
R = 179 RS, Te = 3467 K and M = 0.544 MS.
262 7 Evolution
Fig. 7.9 The Sun’s horizontal branch and asymptotic giant branch in the HR diagram (red
piecewise straight line). The previous evolution is represented in black
7.3.7 Planetary Nebula
At the end of the asymptotic giant branch the strong outward radiation flux exerts
additional pressure on the outer layers of the Sun, which are only weakly bound to
the core by gravity. As a consequence, the outer layers are expelled from the Sun by
means of a massive wind. Thus, the envelope disperses into space and leaves
behind an electron degenerate core made of C and O. The core is surrounded by a
thin layer where both H and He burning occurs but gradually turns off.
The expulsion of the outer layers continues until the entire outer envelope is
ejected by stripping away the core. The result is called a planetary nebula and is
today observed in other stars as one of the most spectacular views in the sky
(Fig. 7.10).
According to the evolution model of Schröder and Smith, the mass loss during
the last 3 104 y of the asymptotic branch phase is less than 0.01 MS and con-
tributes to form the circumsolar envelope of the planetary nebula. Thus, the solar
nebula probably looks like IC 2149, a nebula which was formed around a low-mass
star and is observed in the Auriga constellation (Fig. 7.11).
7.3.8 White Dwarf and Evolution Last Phases
When the envelope of the planetary nebula is entirely dissipated, only the core
remains and appears as a hot star with about half the present solar mass. This
evolved Sun is very unusual compared with its present structure. Because of the
Fig. 7.10 Two observed planetary nebulae. Left panel: NGC 6751, which is located in the
constellation Aquila Image credit: James Long & the ESA/ESO/NASA Photoshop FITS Liberator.
Right panel: the Ring Nebula, or M57, in the Lyra constellation. Image Credit: NASA, ESA, and
the Hubble Heritage (STScI/AURA)—ESA/Hubble Collaboration
Fig. 7.11 The planetary nebula IC 2149 in the constellation of Auriga. (Image credit: Patrick A.
Young, http://mmtao.org)
very strong pressure, the material is greatly compressed and in a state of degen-
eracy. Even atomic nuclei are squeezed more strongly than in the ordinary matter.
Such a star is called a white dwarf. In this phase, the Sun is white, because its
surface temperature is of the order of 104, 105 K. A white dwarf has a radius similar
264 7 Evolution
Fig. 7.12 The Sun’s tracks in the HR diagram as a planetary nebula and a white dwarf (red
piecewise straight line). The previous evolution is represented in black
to the Earth’s radius, that is, about 0.01 RS and a luminosity comparable with the
present luminosity (Fig. 7.12).
The age is about 15 Gy, but the exact time necessary to the Sun to become a
white dwarf is uncertain. The most striking number of a white dwarf is the density
value. The fact that about half solar mass is crushed in a volume equal to that of the
Earth implies that the order of magnitude of density is 106 g cm−3 for the average
density and 108 g cm−3 at the center.
At this stage the Sun still evolves for several Giga years; but its fate is sealed.
Cooling by radiation loss and without nuclear and gravitational energy sources, the
Sun becomes first a brown dwarf and finally dies as a black dwarf.
Exercise 7.13 Using the condition
N [ N;
where
.
N ¼ ð3mkB T Þ3=2 h3 ;
and N, m and T are the particle number density, mass and temperature, respectively,
find the values of N and N for both protons and electrons under the conditions of the
present Sun’s center. The results suggest that protons are far from degeneracy
whereas partial electron degeneracy occurs.
Exercise 7.14 At 1 AU the solar wind typically has a proton number density
Np = 10 protons cm−3 and a velocity v = 500 km s−1. Determine the mass loss in
g s−1 and MS y−1.
Exercise 7.15 Estimate the total nuclear, gravitational and thermal energies at
disposal to the Sun.
Exercise 7.16 Using the results of the previous exercise, show that only the nuclear
energy due to the hydrogen fusion can supply the total energy emitted during the
solar life.
List of Exercises and Answers
Chapter 1 Exercise subject Number Answer

Energy balance for atomic 1.1 (1) m(v21− v22)/ 2 = 3EH /4
transitions (2) 5EH /36 = Ep
(3) Ep −EH = mv2/ 2
Change of units 1.2 1 Å = 10−7 mm = 102 pm
1 Fm = 10−12 mm = 10−3pm
Change of units 1.3 kB = 3.84 10−30 kWh K−1
Change of units 1.4 Negative values for kBT
Pythagorean theorem 1.5 In the text
Maxwell distribution 1.6 Worked exercise
Derivatives 1.7 Worked exercise
Maxwell distribution 1.8 Worked exercise
Partition function 1.9 In the text
Statistical weights 1.10 In the text
H ionization fractions 1.11 Worked exercise
Free electron degeneracy 1.12 Worked exercise
Saha distribution 1.13 Worked exercise
Integrals 1.14 Worked exercise
Planck function 1.15 Worked exercise
Wien displacement law 1.16 In the text
Approximations to Bk 1.17 In the text
Planck function 1.18 Worked exercise
ElectronVolt 1.19 Worked exercise
Change of units 1.20 In the text
Mass-energy relation 1.21 m = 2.423 10−32 g
pp chain energy 1.22 In the text
Bohr atomic model 1.23 r = 0.47 10−8 cm
Molecular size 1.24 lm = 3.1 10−8 cm
Avogadro number 1.25 NA = 6.022 1023
Mean molecular weight 1.26 la = 29
(continued)

Notes in Physics, DOI 10.1007/978-3-319-64961-0
268 List of Exercises and Answers
(continued)
Gas characteristics 1.27 kBT = 4 10−14 erg
v = 0.4 km/s
Gas characteristics 1.28 P = 1.05 106 dyn/cm2 = 1.05 106 Ba
Spherical corona 1.29 In the text
Saha equation 1.30 In the text
Electromagnetic spectrum 1.31 10−5, 10−8, 10−11, 10−12,
10−13, 10−16, 10−18, 10−20,
10−22 erg
Maas-energy relation 1.32 In the text
Chapter 2 Number Answer
Exercise subject
Earth’s orbit 2.1 a = 1.016 AU, p = 0.984 AU
AU measurement 2.2 Worked exercise
AU measurement 2.3 The Earth and Venus orbits are not perfectly circular
AU measurement 2.4 The Sun is not a good mirror for radio waves
Angular diameter 2.5 Worked exercise
Lorentzian 2.6 LMax = 1/(p c)
function
Convolution 2.7 I(r) = (I0/p) {arctan[n(1−r/Rs)/2]− arctan[n(−1−r/Rs)/2]}
Convolution 2.8 As the distance from the maximum point increases, a
Gaussian function vanishes much faster than the Lorentzian
function with the same half width. Thus, the neighborhood
of any point r where the point spread function is e.g. greater
than 1% of its maximum value, is much broader for the
Lorentzian function than for the Gaussian function
Centripetal 2.9 In the text
acceleration
Centripetal 2.10 r = 26.4 Gm
acceleration
Solar constant 2.11 Worked exercise
Solar flux 2.12 In the text
Centripetal 2.13 r = 42 Mm
acceleration
Gravitational 2.14 E ¼ G MAU
S mE
¼ 5:29 1040 erg
potential energy
Earth orbital 2.15 v ¼ 30 km=s
velocity
Escape velocity 2.16 v ¼ 42:1 km=s
Escape velocity 2.17 v ¼ 11:2 km=s
Kepler’s laws 2.18 P = 0.615, 1.881, 11.87, 29.65 years
Gravitational 2.19 In the text
potential energy
Energy 2.20 Et = 8.00 1015 erg
conservation Np = 2.62 1027
(continued)
List of Exercises and Answers 269
(continued)
Exercise subject
Photons and light 2.21 T = 4462 K
E = 2.50 1013 erg
Kepler’s law 2.22 aM = 383 Mm
Chapter 3 Exercise Number Answer
subject
Density stratification 3.1 Worked exercise
Density stratification 3.2 In the text
Pressure stratification 3.3 In the text
Boltzmann distribution 3.4 Worked exercise
Flux formation depth 3.5 Worked exercise
TE emission flux 3.6 In the text
Ha wavelength 3.7 Worked exercise
Radiation transfer 3.8 In the text
Center to limb 3.12 In the text
Radiation transfer 3.14 Optically infinite and plane parallel photosphere
Rosseland opacity 3.15 Worked exercise
Grey photosphere 3.16 In the text
Logarithms 3.17 ln(x) = ln(10) ∙ log(x)
P(s) 3.18 P(s) = P0 ∙ s1/a
q(z) 3.19 q(z) = q(z = 0)e−z/H
LTE 3.20 In the text
Eddington-Barbier 3.21 Worked exercise
approximation
Line broadening 3.22 Worked exercise
Balmer series 3.23 In the text
Equivalent width 3.24 Worked exercise
Mean molecular 3.25 In the text
weight
H level populations 3.26 In the text
Abundances 3.27 X = 0.75, Y = 0.25 neglecting metals. Z = 0.011
considering metals with abundance Nm/NH > 10−5
Pressure scale height 3.28 HE = 8.42 km
Archimedes’ principle 3.29 127 g. A fraction of ice emerges
Photon mean free path 3.30 140 m
Electron number 3.31 In the text
density
(continued)
(continued)
Chapter 3 Exercise Number Answer
subject
Electron number 3.32 In the text
density
Radiation flux 3.33 In the text
Mass column density 3.34 In the text
Mass column density 3.35 In the text
Calcium ionization 3.36 In the text
Calcium excitation 3.37 In the text
Photospheric and 3.38 Mpho = 1.79 1023 g, Mtro = 5.28 1021 g
tropospheric masses
Exercise subject
Opacity 4.1 H, logT = 5.2. He, logT = 5.5. The estimate neglects the ion
excitation that effectively decreases the ionization potential
Diffusion 4.2 Worked exercise
coefficient
Energy density 4.3 Worked exercise
Radiative 4.4 In the text
temperature
gradient
temperature
gradient
Specific heats 4.6 Worked exercise
Convective energy 4.7 Top: 1.04 109 erg cm2 s−1,
flux bottom: 1.04 1011 erg cm2 s−1
Adiabatic relations 4.8 In the text
Adiabatic process 4.9 Worked exercise
Schwarzschild 4.10 Worked exercise
criterion
Mean molecular 4.11 In the text
weight
Schwarzschild 4.12 Worked exercise
criterion
(c−1)/c 4.13 Worked exercise
l 4.14 Worked exercise
Temperature 4.15 In the text
structure
Temperature 4.16 In the text
structure
q(T) 4.17 In the text
(continued)
(continued)
Exercise subject
equation
equation
equilibrium
H− opacity 4.21 T = 8704 K
He opacity 4.22 logT = 5.8 K. The estimate neglects the He II excitation
Electron number 4.23 Ne = 1.2P/(2.3 kBT), hence, Ne = 5.7 1025
density electrons cm−3, using the Senesi et al. model at r = 355 km.
Diffusion 4.24 Top, Cd = 5.06 1016 cm2s−1. Bottom, Cd = 3.09 109
coefficient cm2s−1
Energy density 4.25 D = 4prT4c /c = 2.42 1019erg cm−3. Nph D/kBTc = 1.13
1028 photons cm−3
Granulation 4.26 In the text
contrast
Granulation 4.27 Tg − Ti = 500 K
contrast
H and He 4.28 r/RS = 0.9896831 and 0.9785358
photoionizations
Wien displacement law 5.1 In the text
Electron number density 5.2 Worked exercise
Mean molecular weight 5.3 In the text
Mass conservation 5.4 In the text
Mass conservation 5.5 DM 4pR2s q0eDr/HDr 10−8 MS
Mass conservation 5.6 In the text
T scale height 5.7 In the text
P scale height 5.8 In the text
q scale height 5.9 In the text
q scale height 5.10 In the text
Integrals 5.11 In the text
Thomson scattering 5.12 In the text
Opacities 5.13 Top:
kRT = 5.63 10−2
cm−1,
kRff = 0.133 cm−1,
kRbf =3.21 cm−1.
Bottom:
kRT = 7.13 cm−1,
kRff = 19.6 cm−1,
kRbf = 473 cm−1
(continued)
(continued)
Opacities 5.14 In the text
Mean molecular weight 5.15 Use the state equation
Schwarzschild criterion 5.16 In the text
Mass loss 5.17 In the text
Exercise subject
pp chains 6.1 Worked exercise
CNO cycle 6.2 Worked exercise
Energy per cycle 6.3 In the text
Esu units 6.4 In the text
De Broglie 6.5 In the text
wavelength

vrpp 6.6 Worked exercise
Gamow energy 6.7 In the text

vrpp 6.8 In the text
Fusion lifetime 6.9 Worked exercise
Luminosity 6.10 Worked exercise
Scale heights 6.11 In the text
Mass variation 6.12 In the text
Luminosity 6.13 In the text
variation
T and 6.14 In the text
P variations
em 6.15 In the text
Proton-proton 6.16 The two protons approach each other, whereas the ball turns
collision away from the Earth center
Schrödinger 6.17 In the text
equation
Wave 6.18 The resulting wave is the wave f1 modulated by the amplitude
superposition 2cos[2p(tdm/2 – xdk/2)], which is a wave with smaller
frequency and wavenumber
Interstellar medium 7.1 In the text
Gravitational energy 7.2 Worked exercise
Jean mass 7.3 In the text
Last cloud fragment 7.4 In the text
ProtoSun radius 7.5 In the text
ProtoSun radius 7.6 In the text
HR diagram 7.7 In the text
(continued)
(continued)
Spectral type, color index 7.8 Worked exercise
and luminosity class
Hydrogen half-life 7.9 In the text
TE vs L 7.10 In the text
TE vs L 7.11 In the text
Gas degeneracy 7.12 Ne = 2.99 1029 cm−3
Ne = 7.97 1026 cm−3
Gas degeneracy 7.13 Np = 3.19 1025 cm−3
Np = 3.82 1030 cm−3
Ne = 6.09 1025 cm−3
Ne = 4.86 1025 cm−3
Solar wind 7.14 dM/dt = 2.36 1012 g s1 = 3.76
10−14 MS y−1
Solar energies 7.15 En = (Eppc/4)MS/mp = 1.3 1052 erg, Eg = 3.8
1048 erg, Ek = 2.6 1048 erg
Solar energies 7.16 Ee = 1.2 1051 erg
References
Asplund, M., Grevesse, N., & Sauval, A. J. (2005). ASP Conference Series, 336, 25.
Avrett, E. H., & Loeser, R. (2008). Astrophysics Journal Supplementum, 175, 229.
Battaner, E. (1996). Astrophysical fluid dynamics, Cambridge: Cambridge University Press
Bisnovatyi-Kogan. (2000). Stellar physics (Vol. 1). Berlin: Springer
Castellani, V. (1985). Astrofisica stellare, I fondamenti fisici dell’interpretazione evolutiva di stelle
e galassie. Bologna: Zanichelli.
Clayton, D. D. (1968). Principles of stellar evolution and nucleosynthesis, NY: McGraw-Hill, Inc.
Dell’Omodarme, M., Valle, G., Degl’Innocenti, S., & Prada Moroni, P. G. (2012). Astronomy &
Astrophysics, 540, 26.
Eisberg, R. M. (1967). Fundamentals of modern physics. NY: Wiley
Fröhlich, C. (2006). Solar irradiance variability since 1978: Revision of the {PMOD} composite
during solar cycle 21. Space Science Reviews 125, 53–65. ftp://ftp.pmodwrc.ch/pub/Claus/
ISSI_WS2005/ISSI2005a_CF.pdf.
Gibson, E. G. (1973). The Quiet Sun. NASA ESP-303
Gustafsson, B., Edvardsson, B., Eriksson, K., Jørgensen, U. G., Nordlund, Å., & Plez, B. (2008).
Astronomy & Astrophysics, 486, 951.
Kippenhahn, R. & Weigert, A. (1990). Stellar structure and evolution, Springer
Labs, B., & Neckel, H. (1962). Z. Astrophys., 55, 269.
Landau, L. D., & Ktaigorodskij, A. I. (1972). La fisica per tutti, Editori Riuniti
Landi Degl’Innocenti, E. (2008). Fisica solare, Springer
Marion, J. B. (1975). Physics and the Physical Universe, NY: Wiley.
Neckel, H., & Labs, B. (1984). Solar Physics, 90, 205.
Noyes, R. W. (1982). The Sun. Our Star: Harvard University Press.
Pierce, A. K., & Slaughter, C. D. (1977). Solar Physics, 51, 25.
Reimers, D. (1975). Memoires of the Société Royale des Sciences de Liège, 8, 369.
Reimers, D. (1977). Astron. Astrophys., 61, 217.
Schröder, K.-P., & Smith, R. C. (2008). MNRAS, 386, 155.
Senesi, M. (2012). Private communication.
Spruit, H. C. (1997). Memorie SAIt, 68–2, 397.
Stix, M. (2002). The sun. An Introduction. Berlin: Springer.
Tognelli, E., Prada Moroni, P. G., & Degl’Innocenti, S. (2011). Astronomy Astrophysics, 533, 109.
Vernazza, J. E., Avrett, E. H., & Loeser, R. (1981). Astrophysics Journal Supplement, 45, 635.

Notes in Physics, DOI 10.1007/978-3-319-64961-0

(Undergraduate Lecture Notes in Physics) Giuseppe Severino (Auth.) - The Structure and Evolution of The Sun-Springer International Publishing (2017)

Uploaded by

Document Information

Original Title

Copyright

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

(Undergraduate Lecture Notes in Physics) Giuseppe Severino (Auth.) - The Structure and Evolution of The Sun-Springer International Publishing (2017)

Uploaded by

Copyright:

Undergraduate Lecture Notes in Physics

ULNP titles must provide at least one of the following:

More information about this series at http://www.springer.com/series/8917

The Structure and Evolution

ISSN 2192-4791 ISSN 2192-4805 (electronic)

© Springer International Publishing AG 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

referring to a preceding formula, I decided to rewrite the formula so as not to

Napoli, Italy Giuseppe Severino

1 Some Basic Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2.1.2 The Steady Sun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

4.1.1 Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

6.3.4 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

© Springer International Publishing AG 2017 1

1.1 Basic Structure of Matter

1.1.1 Elementary Particles and Fundamental Forces

1.1.3 Units of Measurement and Exponential Notation

1 AU ¼ 150 billion meters ¼ 150,000,000,000 m ¼ 15 1010 m:

1 m over 10 billion ¼ 1=10,000,000,000 m ¼ 1=1010 m ¼ 1010 m;

and atomic nuclei have sizes of the order of

1 m over 1 million billion ¼ 1=1,000,000,000,000,000 m ¼ 1=1015 m ¼ 1015 m:

103 m ¼ 1 millimeter ¼ 1 mm:

1 AU ¼ 15 1010 m ¼ 150 109 m ¼ 150 Gm:

Table 1.2 The known stable elements.

Table 1.2 (continued)

kB ¼ 1:38062 1016 erg K1 :

In terms of the density q the equation of state can be written

P ¼ NkB T ¼ mNkB T=m ¼ qkB T=m:

P ¼ q kB T=m ¼ qkB T=ðlmu Þ

R ¼ kB =mu ¼ 1:38062 1016 erg K1 =1:660538921 1024 g

is called the universal gas constant.

1.2 Thermodynamic Equilibrium

1.2.1 The Energy Distribution of Gas Particles

The macroscopic properties of a gas in thermodynamic equilibrium (whose

while its kinetic energy is

where v is the modulus of v (Fig. 1.4).

1.2.2 Maxwell Distribution

Consider a gas in thermodynamic equilibrium, characterized by a temperature

v2 ¼ v2x þ v2y þ v2z

dVv ¼ dv v dh v senh du ¼ v2 senh dv dh du:

The elementary volume, where the velocity modulus is approximately constant

The Maxwell distribution is represented in Fig. 1.8. It contains a squared

Expressing the derivative of a product, we get

Moreover, developing the derivative of a function of a function, we get

sufﬁcient to search for solutions where

f 0 ð xÞ ¼ df ð xÞ=dx ¼ lim ½f ðx þ DxÞf ð xÞ=Dx ¼ lim Df ð xÞ=Dx

Table 1.4 Derivatives of Name f(x) df(x)/dx

df ðx; yÞ ¼ ½@f ðx; yÞ=@xdx þ ½@f ðx; yÞ=@ydy

where ∂f(x,y)/∂x is the x derivative of f at y = constant and ∂f(x,y)/∂y is the

dN=N ¼ 2p1=2 Ep3=2 eE=Ep E 1=2 dE:

Worked Exercise 1.8 The Maxwell distribution in terms of energy.

¼ ð1=2Þð2E=mÞ1=21 2dE=m ¼ ð2mE Þ1=2 dE:

Ep ¼ mv2p =2 ¼ mð2kB T=mÞ=2 ¼ kB T:

\E [ ¼ 3 kB T=2 ¼ 3Ep =2:

dN=N ¼ 2ð3=2 þ 31=2Þ pð3=2 þ 1Þ Ep3=2 mð3=21 1=2Þ eðE=Ep Þ E ð11=2Þ dE

and, ﬁnally, the distribution of the kinetic energy at equilibrium is

dN=N ¼ 2p1=2 Ep3=2 eE = Ep E 1=2 dE:

1.2.3 Boltzmann Distribution

Ni =N ¼ gi eEi =ðkB TÞ =U;

and the ratio Ni/ Nj is

Ni =Nj ¼ gi eðEi Ej Þ=ðkB TÞ =gj ;

weights, g1/(g0 + g1) = 0.8, as T approaches ∞. Therefore, at high temperatures,

1.2.4 Saha Distribution

In TE the ionization processes of a chemical species are in equilibrium with the

where the Saha function u(T) is

f 0 ð xÞ ¼ df ð xÞ=dx ¼ lim ½f ðx þ DxÞf ð xÞ=Dx ¼ lim Df ð xÞ=Dx

df ðx; yÞ ¼ ½@f ðx; yÞ=@xdx þ ½@f ðx; yÞ=@ydy

dx dy dz dpx dpy dpz h3 :

x ¼ ½me =ð2kB T Þ1=2 v;

dx ¼ ½me =ð2kB T Þ1=2 dv;

jstk ¼ N2 B21 ðhc=kÞ½1=ð4pÞwk Bk :