Lecture Notes in Physics 907

Valerio Faraoni

Cosmological
and Black
Hole Apparent
Horizons
Lecture Notes in Physics

Volume 907

Founding Editors
W. Beiglböck
J. Ehlers
K. Hepp
H. Weidenmüller

Editorial Board
M. Bartelmann, Heidelberg, Germany
B.-G. Englert, Singapore, Singapore
P. Hänggi, Augsburg, Germany
M. Hjorth-Jensen, Oslo, Norway
R.A.L. Jones, Sheffield, United Kingdom
M. Lewenstein, Barcelona, Spain
H. von Löhneysen, Karlsruhe, Germany
J.-M. Raimond, Paris, France
A. Rubio, Donostia, San Sebastian, Spain
S. Theisen, Potsdam, Germany
D. Vollhardt, Augsburg, Germany
J.D. Wells, Ann Arbor, USA
G.P. Zank, Huntsville, USA
The Lecture Notes in Physics
The series Lecture Notes in Physics (LNP), founded in 1969, reports new
developments in physics research and teaching-quickly and informally, but with
a high quality and the explicit aim to summarize and communicate current
knowledge in an accessible way. Books published in this series are conceived as
bridging material between advanced graduate textbooks and the forefront of
research and to serve three purposes:
• to be a compact and modern up-to-date source of reference on a well-defined
topic
• to serve as an accessible introduction to the field to postgraduate students and
nonspecialist researchers from related areas
• to be a source of advanced teaching material for specialized seminars, courses
and schools
Both monographs and multi-author volumes will be considered for publication.
Edited volumes should, however, consist of a very limited number of contributions
only. Proceedings will not be considered for LNP.
Volumes published in LNP are disseminated both in print and in electronic
formats, the electronic archive being available at springerlink.com. The series
content is indexed, abstracted and referenced by many abstracting and information
services, bibliographic networks, subscription agencies, library networks, and
consortia.
Proposals should be sent to a member of the Editorial Board, or directly to the
managing editor at Springer:

Christian Caron
Springer Heidelberg
Physics Editorial Department I
Tiergartenstrasse 17
69121 Heidelberg/Germany
[email protected]

More information about this series at http://www.springer.com/series/5304

Valerio Faraoni

Cosmological and Black

Hole Apparent Horizons

123
Valerio Faraoni
Physics Department
Bishop’s University
Sherbrooke, QC, Canada

ISSN 0075-8450 ISSN 1616-6361 (electronic)

Lecture Notes in Physics
ISBN 978-3-319-19239-0 ISBN 978-3-319-19240-6 (eBook)
DOI 10.1007/978-3-319-19240-6

Library of Congress Control Number: 2015943318

Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2015
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.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media (www.

springer.com)
To my sisters Adriana and Lilly
Preface

Currently, substantial research efforts are devoted to understanding the physics of

horizons. A horizon is a surface which separates a region of space-time which
is accessible to an observer from one which is not and from which this observer
cannot receive light or other physical signals. This feature gives rise to interesting
physics: with the emphasis given in modern times to the role played by information
in theoretical physics, it is easy to guess that a horizon will produce interesting
physical phenomena. If entropy is understood as information entropy, then a horizon
which hides information should be attributed some entropy. This is in fact what
the black hole thermodynamics developed in the 1970s found. The discovery of
Hawking radiation from black hole horizons made possible the development of
black hole thermodynamics, a remarkable and beautiful construct which shows
that, indeed, there is very interesting physics associated with horizons. Already
in special relativity without gravity, uniformly accelerated observers experience
acceleration horizons. When gravity is introduced, one encounters black hole and
cosmological horizons. Then, studying black holes, one meets inner, outer, Cauchy,
and extremal horizons, and in cosmology there are particle, event, de Sitter, and
apparent horizons.
The pioneers who developed black hole mechanics and thermodynamics in the
1970s discussed stationary black holes and event horizons. Dynamical situations
such as gravitational collapse, black hole evaporation by Hawking radiation, and
black holes interacting with nontrivial environments and exchanging mass-energy
require that the concept of event horizon be generalized. Conceivable dynamical
situations include black holes accreting surrounding fluids, black holes immersed
in a cosmological background, and, most significantly, black holes emitting (and
possibly absorbing) Hawking radiation, which becomes significant in the last
evolutionary stages. If a black hole is placed in a nontrivial environment, its mass-
energy should be also the internal energy which we need to account for in the first
law of thermodynamics. This mass-energy must be defined carefully, usually with
some quasi-local notion, which in turn is sometimes related to the notion of apparent
horizon.

vii
viii Preface

In dynamical situations, it is not clear what is meant by “black hole” because

the most salient feature of a black hole is precisely its horizon, and the event
horizon familiar from stationary black holes turns out to be essentially useless for
practical purposes in dynamical space-times. This major obstacle appears because
the definition of event horizon requires the knowledge of the entire future of space-
time, which is physically impossible to achieve in nonstationary situations. The
ambiguity in the notion of “horizon” therefore implies murkiness in the concept of
“black hole” itself. Simultaneously, thanks to the increase in the power of modern
supercomputers, great theoretical efforts are now made to predict in detail the
waveforms of gravitational waves emitted by black holes. These waveforms are
needed to build banks of templates to separate signals from noise in the laser
interferometric detectors of gravitational waves. The notion of event horizon is of
little use in the numerical study of fast astrophysical processes producing those
gravitational waves. Instead, “black holes” are routinely identified with outermost
marginally trapped surfaces and apparent horizons in numerical research. Hence,
a part of the research community is still focused on event horizons, while another
part dismisses it altogether and uses horizon surfaces, the role of which is not yet
understood clearly. This dichotomy needs to be addressed, and this work is intended
to give a contribution in this direction.
This book contains a series of graduate-level lectures introducing the main
problems in this area of theoretical physics. The first three chapters are pedagogical
in nature, while the remaining two report a series of “case studies” to which the
concepts of apparent and trapping horizon are applied. They consist of relatively
rare analytic solutions of Einstein’s theory and of scalar-tensor and f .R/ gravity
which appeared in the literature and contain, at least in certain space-time regions,
black hole and cosmological apparent horizons. The dynamics of apparent horizons
can be rather bizarre and reserves several surprises. The phenomenology of apparent
horizons known thus far is described and analyzed. While this field of research is
definitely not settled and the last word is not said on any of the issues examined,
these lectures aim at collecting and summarizing the existing results and providing
an introduction and a toolkit for researchers approaching this field, especially
graduate students. An extensive bibliography refers the reader to specific points
which cannot be discussed in a single volume. I hope that these lectures will be
stimulating and that some of my readers will soon find new directions for this area
of research.

Sherbrooke, Canada Valerio Faraoni

April 2015
Acknowledgments

I am indebted to many people for their contributions to the material discussed in this
book, and my thanks go first of all to my collaborators and to my students: Thomas
Sotiriou, Vincenzo Vitagliano, Stefano Liberati, Angus Prain, Alex Nielsen, Audrey
Jacques, Andres Zambrano, and Roshina Nandra.
For discussions or comments, or for the occasional e-mail pointing out references
that I missed, I am grateful to Hideki Maeda, Matt Visser, Daniel Guariento,
John Barrow, Eric Poisson, Ivan Booth, Luciano Vanzo, Sergio Zerbini, Rituparno
Goswami, Bibhas Majhi, Andrzej Krasiński, Kayll Lake, Viqar Husain, Timothy
Clifton, Sergei Odintsov, Emilio Elizalde, Sebastiano Sonego, Nemanja Kaloper,
Bahram Mashhoon, Salvatore Capozziello, Valeri Frolov, Sarah Shandera, Carlos
Herdeiro, José Senovilla, and Rong-Gen Cai. Many thanks to Angus Prain for
drawing the figures and to Aldo Rampioni and Kirsten Theunissen of Springer for
their support and friendly assistance during the writing of this book.
This research is supported by Bishop’s University and by the Natural Sciences
and Engineering Research Council of Canada.

ix
Contents

1 Stationary Black Holes in General Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Stationary Black Holes of General Relativity . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Schwarzschild Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.1 Schwarzschild Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.2 Isotropic Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.3 Kruskal-Szekeres Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.4 Eddington-Finkelstein Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.5 Painlevé-Gullstrand Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.6 Kerr-Schild Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.7 Novikov Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4 Reissner-Nordström Metric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.1 Schwarzschild Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.2 Kruskal-Szekeres Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5 Kerr Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.6 Kerr-Newman Metric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.7 Energy Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2 Horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Null Geodesic Congruences and Trapped Surfaces . . . . . . . . . . . . . . . . . . 25
2.3 Rindler Horizons for Accelerated Observers
in Minkowski Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.4 Event Horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.5 Killing Horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.6 Apparent Horizons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.7 Trapping Horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.8 Isolated and Dynamical Horizons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

xi
xii Contents

2.9 Kodama Vector and Surface Gravity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.10 Surface Gravities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.10.1 Killing Horizon Surface Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.10.2 Surface Gravity of Marginally Trapped Surfaces. . . . . . . . . . . 44
2.10.3 Fodor et al. Surface Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.10.4 Isolated Horizon Surface Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.10.5 Proposal for Slowly Evolving Horizons . . . . . . . . . . . . . . . . . . . . 46
2.10.6 Other Proposals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.11 Spherical Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.12 Rindler Horizons Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.13 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3 Cosmological Horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1.1 FLRW Cosmologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2 Hyperspherical Coordinates for FLRW Space . . . . . . . . . . . . . . . . . . . . . . . 61
3.3 Kruskal-Szekeres Coordinates for de Sitter Space . . . . . . . . . . . . . . . . . . . 61
3.4 Painlevé-Gullstrand and Schwarzschild-Like
Coordinates for k D 0 FLRW Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.5 Schwarzschild-Like Coordinates for General FLRW Spaces. . . . . . . . 63
3.6 Painlevé-Gullstrand Coordinates for General FLRW Spaces . . . . . . . . 67
3.7 Congruences of Radial Null Geodesics in FLRW Space . . . . . . . . . . . . 68
3.8 Horizons in FLRW Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.8.1 Particle Horizon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.8.2 Event Horizon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.8.3 Hubble Horizon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.8.4 Apparent Horizon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.8.5 Trapping Horizon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.8.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.9 Dynamics of Cosmological Horizons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.10 Another Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.11 de Sitter Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.12 Thermodynamics of Cosmological Horizons in General Relativity. 91
3.13 Thermodynamics of de Sitter Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.14 Thermodynamics of Apparent/Trapping Horizons
in FLRW Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.15 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4 Inhomogeneities in Cosmological “Backgrounds”
in Einstein Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.2 Schwarzschild-de Sitter-Kottler Spacetime. . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Contents xiii

4.3 McVittie Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

4.3.1 Apparent Horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.3.2 Phantom McVittie Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.3.3 Nolan Interior Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.4 Charged McVittie Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.5 An Application to the Quantization of Black Hole Areas . . . . . . . . . . . 125
4.6 Generalized McVittie Spacetimes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.6.1 A Single Perfect Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.6.2 Imperfect Fluid Without Radial Flow of Material . . . . . . . . . . 130
4.6.3 Imperfect Fluid with Radial Flow of Material . . . . . . . . . . . . . . 131
4.6.4 “Comoving Mass” McVittie Solution . . . . . . . . . . . . . . . . . . . . . . . 133
4.6.5 More General Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.6.6 “Comoving Mass” Solution as an Attractor . . . . . . . . . . . . . . . . 137
4.6.7 Recent Developments and Scalar Field Sources . . . . . . . . . . . . 138
4.7 Sultana-Dyer Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.8 Husain-Martinez-Nuñez Spacetime. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.9 Fonarev Solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.10 Other Analytic Cosmological Black Hole Solutions
of the Einstein Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
4.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
5 Cosmological Inhomogeneities in Alternative Gravity . . . . . . . . . . . . . . . . . . 167
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
5.2 Brans-Dicke Cosmological Black Holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
5.2.1 Clifton-Mota-Barrow Black Holes . . . . . . . . . . . . . . . . . . . . . . . . . . 169
5.2.2 Conformally Transformed
Husain-Martinez-Nuñez Spacetime . . . . . . . . . . . . . . . . . . . . . . . . . 176
5.3 f .R/ Cosmological Black Holes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
A.1 Painlevé-Gullstrand Coordinates for General Spherically
Symmetric Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
A.2 Kodama Vector in FLRW Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
A.2.1 Pseudo-Painlevé-Gullstrand Coordinates . . . . . . . . . . . . . . . . . . . 194
A.2.2 Comoving Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Symbols and Acronyms

gab metric tensor

gab inverse metric tensor
g metric determinant
ab Minkowski metric
ra covariant derivative operator
bca Christoffel symbols
Rab Ricci tensor
R D Ra a Ricci scalar
Gab  Rab Einstein tensor
 12 gab R
Tab matter energy-momentum tensor
 cosmological constant
lPl Planck length
ka ;  a Killing vectors
Ka Kodama vector
 surface gravity
 Kodama time
a uniform acceleration of Rindler observers
2
d˝.2/ D d 2 line element on the unit 2-sphere
C sin2  d' 2
R areal radius
rN isotropic radius
 conformal time
a geodesic deviation vector
l expansion of a null geodesic congruence with tangent field la
MMSH Misner-Sharp-Hernandez mass
kB Boltzmann constant
T (absolute) temperature
S entropy
A horizon area

xv
xvi Symbols and Acronyms

k curvature index of FLRW space

a scale factor of the FLRW metric
H  aP =a Hubble parameter of FLRW space
q  Raa=Pa2 deceleration parameter of FLRW space
fP derivative of f with respect to comoving time
I C  future null infinity
J I C causal past of future null infinity
 equal by definition
D P equality valid in General Relativity in FLRW space with a perfect
fluid as the matter source

FLRW Friedmann-Lemaître-Robertson-Walker
AH apparent horizon
TH trapping horizon
EH event horizon
PH particle horizon
FOTH future outer trapping horizon
FITH future inner trapping horizon
PITH past inner trapping horizon
MOTS marginally outer trapped surface
MTT marginally trapped tube
MOTT marginally outer trapped tube
iff if and only if
Chapter 1
Stationary Black Holes in General Relativity

The noblest pleasure is the joy of understanding.

—Leonardo da Vinci

1.1 Introduction

Beginning with the Rindler horizons which appear for accelerated observers in
Minkowski space without gravity, one comes quickly to acknowledge the presence
of various types of horizons when gravity is introduced in spacetime: first one
encounters black hole horizons and cosmological horizons. Then, studying classical
and semiclassical black holes one is faced with inner, outer, Cauchy, and extremal
horizons. The early literature on black holes and the works which developed black
hole thermodynamics in the seventies had plenty to do with discussing stationary
black holes and event horizons (e.g., [23, 74, 75]). Dynamical situations such
as gravitational collapse, black hole evaporation due to Hawking radiation, and
black holes interacting with non-trivial environments and exchanging mass-energy
require that the concept of event horizon be generalized to some other construct
with which it is possible to work. Conceivable dynamical situations include black
holes accreting or expelling gravitating (i.e., non-test) matter; examples are Vaidya
spacetimes, black holes immersed in a cosmological “background” other than de
Sitter space, black holes emitting (and possibly also absorbing) Hawking radiation
(which becomes significant in the last evolutionary stages with backreaction playing
an important role), or black holes with variable mass because of other physical
processes. If a black hole is placed in a non-trivial environment, its mass-energy
should be also the internal energy which we need to account for in the first law of
thermodynamics. This mass-energy must be defined carefully; usually it is identified
with some quasi-local energy construct which is related to the notion of horizon. In
these lectures we will use the Misner-Sharp-Hernandez mass in spherical symmetry
and its generalization, the Hawking-Hayward quasi-local energy in the absence of
spherical symmetry.
Intuitively, an horizon is “a frontier between things observable and things
unobservable” [64]. Inequivalent notions of black hole horizon abound in the
technical literature and the terminology used features event, Killing, inner, outer,
Cauchy, apparent, trapping, quasi-local, isolated, dynamical, and slowly evolving

© Springer International Publishing Switzerland 2015 1

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6_1
2 1 Stationary Black Holes in General Relativity

horizons (Refs. [4, 7, 51, 74] are reviews of an extensive literature). For stationary
black holes some of these constructs coincide, but they are, in general, very different
or unrelated for dynamical black holes with masses and other physical parameters
which change with time. It is not clear what it is meant by “black hole” in dynamical
situations because the most salient feature of a black hole is precisely its horizon,
which is universally taken to signal the presence of a black hole. The ambiguity
in the appropriate notion of “horizon” therefore implies a serious ambiguity in the
concept of “black hole”.
The definition of event horizon inspired by (and historically attached to) station-
ary black holes turns out to be essentially useless for practical purposes in dynamical
spacetimes. This major obstacle manifests itself because knowing the event horizon
requires the knowledge of future null infinity, which is physically impossible to
achieve [3].
Astronomy is undergoing remarkable progress and it points out more and more
the important roles of stellar mass and supermassive black holes in astrophysical
processes. Great theoretical efforts are made to predict in detail the waveforms of
gravitational waves emitted by black holes. This programme is made possible by
the increase in power of modern supercomputers but it remains a very ambitious
goal. The notion of event horizon is of little use in the numerical study of the fast
dynamical evolution occurring in the gravitational collapse of a cosmic body, or in
the close inspiralling and merger of a black hole with its companion in a binary
system. “Black holes” are routinely identified with outermost marginally trapped
surfaces and apparent horizons in numerical work [6, 13, 73].
What about cosmological horizons? These surfaces are probably the playground
in which one should take baby steps in understanding horizon physics. The
cosmology textbooks discuss particle and event horizons in relation with early
universe inflation [42, 46, 49]. Different cosmological horizons, the apparent and
trapping horizons, are also used more and more. It was not long after the discovery
of Hawking radiation [32, 34] and the completion of black hole thermodynamics
that Gibbons and Hawking pointed out [28] that the event horizon of de Sitter space
behaves as a thermodynamic system and should be endowed with a temperature
and an entropy. The region of de Sitter space below the de Sitter horizon is static
and the horizon itself does not change in time, so it can be regarded to a certain
extent as the cosmological analogue of the Schwarzschild event horizon. There is an
important difference, though: a de Sitter horizon depends on the observer while the
Schwarzschild horizon does not. If the analogy carries through, then the analogue
of time-dependent black hole horizons would necessarily be the apparent and
trapping horizons of Friedmann-Lemaître-Robertson-Walker (FLRW) spacetime,
which evolve with the cosmic time.
Similar to black hole thermodynamics for event horizons, there have been many
attempts to formulate a meaningful thermodynamics for other horizon constructs
(e.g., [4, 14, 35]). The thermodynamics of black hole apparent, trapping, isolated,
and dynamical horizons has been scrutinized often in recent years and thermody-
namical studies of FLRW apparent horizons have also appeared.
1.1 Introduction 3

At the same time, there has been a resurgence of interest in theories of gravity
alternative to General Relativity: they are motivated by various reasons. First, there
is the search for a quantum theory of gravity, which promotes interest in low-energy
effective actions, which invariably contain ingredients foreign to Einstein gravity,
such as scalar fields coupled non-minimally to the curvature (which give a scalar-
tensor nature to the theory), higher derivative terms, or perhaps non-local terms.
Other sources of interest in alternative gravity are the serious attempts [17, 68] to
explain the current acceleration of the universe discovered with type Ia supernovae
without invoking an ad hoc dark energy [1]. Although not as well motivated, there
are also attempts to remove the need for dark matter in galaxies and clusters by
modifying gravity, given that dark matter particles still elude direct detection.
Black holes in alternative theories of gravity evade the no-hair theorem of
General Relativity and one can have non-trivial scalar hair due to interactions
between scalar fields and astrophysical black holes [15, 63, 69, 70], which could
lead to detectable effects [5, 36]. Analytic solutions describing time-dependent black
holes in these theories allow one to get a glimpse of the phenomenology to expect.
More interest in alternative gravity comes from the thermodynamics of spacetime
idea, according to which Einstein theory corresponds somehow to a state of
thermodynamic equilibrium [38] and extended gravities to some sort of excitation
[19] in some “space of theories”. This idea, which seems to fit well in the wider
context of emergent gravity approaches to the problem of quantizing gravity, has its
foundations in the analysis of local Rindler horizons of observers with worldlines
threading spacetime, and in the prescription that the entropy is equal to one quarter
of the horizon area.
There are many instances in which cosmological and time-evolving horizons
play a role in theoretical research in gravity. In these lectures we review the main
properties of cosmological and black hole time-varying horizons in both General
Relativity and extended theories of gravity, and we attempt to provide a unified
view of their physics for applications to various areas of gravitational theory. We
begin with cosmological horizons, and then we discuss horizons associated with
time-dependent black holes. Since only a few exact solutions of General Relativity
and of other theories of gravity are known for which the horizons are explicitly time-
dependent, we concentrate on spacetimes which describe black holes embedded in
cosmological “backgrounds”, which have been studied in some detail.
We begin our study by reviewing basic material in the first two chapters. First
we recall the stationary black hole solutions of the Einstein equations: this first
chapter is meant to be only a refresher since there is no point in repeating the vast
and excellent literature on “standard” black holes. Chapter 2 reminds the reader
of the basic tools used in the analysis of black hole and other horizons, i.e., null
geodesic congruences, and introduces various definitions of horizons appearing
in the literature. The following chapter analyzes the various horizons of FLRW
space, including the thermodynamics proposed for these horizons. We also discuss
several coordinate systems for FLRW spaces which are useful in the study of the
dynamics and thermodynamics of cosmological horizons. The following chapters
4 1 Stationary Black Holes in General Relativity

discuss analytic solutions of the field equations of various theories of gravity which
exhibit time-varying horizons and, often, horizons which appear and/or disappear in
pairs.
In the following, a spacetime .M ; gab / is described by a 4-dimensional manifold
M on which a metric tensor field gab with Lorentzian signature  C CC is defined.
We follow the notations and conventions of Wald’s book [74]. In particular, we use
units in which the speed of light c and Newton’s constant G are unity. The Riemann
tensor is given by

Rabc d D ac;b
d
 bc;a
d
C ace ebd  bce ead (1.1)

in terms of the Christoffel symbols abc of the metric gab . The Ricci tensor is the
contraction

Rac D Rabc b ; (1.2)

and the Ricci scalar is R D gab Rab .

1.2 Stationary Black Holes of General Relativity

For introductions to the theory of General Relativity and basic properties of its
stationary black hole solutions, we refer the reader to well known textbooks
[11, 48, 58, 74]; for a useful list of references on “standard” general-relativistic black
holes see [16, 24, 58]. Here we review background material used in the following
chapters and the main asymptotically flat black hole solutions of Einstein theory.
According to the no-hair theorems [12, 33, 37, 50, 66, 67], the Schwarzschild and
Kerr spacetimes and their charged (Reissner-Nordström and Kerr-Newman) gen-
eralizations are the generic asymptotically flat electrovacuum black hole solutions
of this theory. In General Relativity a black hole formed by gravitational collapse
will settle down to a state determined by only three parameters: its mass M, angular
momentum J, and electric charge Q, irrespective of the initial configuration, the
nature of the collapsing matter, and the details of the collapse. Perturbation analyses
show that perturbations are radiated away quickly according to laws established by
Price (a field with spin s will radiate away a multipole l  s in such a manner
that, in the late stage of collapse, the field decays with a power-law tail scaling with
time as t2lCpC1 , where p D 1 for initially static multipoles and p D 2 otherwise
[30, 60, 61]). A black hole characterized only by the three parameters M; J, and Q
can correspond to a very large number of possible configurations unobservable by
an observer located outside the horizon and then, heuristically, to a large entropy.
Let us review the classic black hole spacetimes in various coordinate systems. It
is often convenient to introduce coordinates tied to particular families of timelike
observers or to use null coordinates based on outgoing or ingoing null geodesics. It
1.3 Schwarzschild Spacetime 5

is sometimes useful to see the latter as the null limit of the worldlines of timelike
observers. All the spacetimes which we review in this chapter are asymptotically
flat and stationary.

1.3 Schwarzschild Spacetime

The Schwarzschild metric was discovered soon after Einstein introduced the theory
of General Relativity and is the prototype of a black hole spacetime. Several
coordinate systems have been developed in order to study it.

1.3.1 Schwarzschild Coordinates

The Schwarzschild line element is

 
2M dR2
ds2 D  1  dt2 C C R2 d˝.2/
2
(1.3)
R 1  2M
R

in Schwarzschild
r (or “curvature”) coordinates in which R is the areal radius, i.e.,
A
R , with A being the area of 2-spheres of symmetry, and where
4
2
d˝.2/  d 2 C sin2  d' 2 (1.4)

is the line element on the unit 2-sphere, which will be used throughout these
lectures. The metric (1.3) is a vacuum solution of the Einstein equations and it is
static, spherically symmetric, and asymptotically flat; it exhibits the well-known
event horizon at R D 2M, which corresponds to a singularity of the Schwarzschild
coordinates at which g00 D 0, but not to a true spacetime singularity—the invariants
of the curvature tensor are finite there. There is a true spacetime singularity at R D 0,
where the invariants of the curvature tensor diverge. The Schwarzschild coordinate
patch only covers the region R > 2M exterior to the horizon.
The concept of event horizon will be discussed in detail in Sect. 2.4. For now, it
is sufficient to know that light and massive particles which start out in the region
R < 2M cannot escape from it. The entire region R < 2M “below the horizon” is
not seen by observers located at radii R > 2M (the region “outside the horizon”).
The Schwarzschild line element represents a one-parameter family of metrics
parametrized by the mass parameter M. Only non-negative values of this parameter
are physical, while the limit M ! 0 gives Minkowski space.1

1
One should be careful, however, in taking limits of the spacetime geometry based on coordinate
systems: the M ! 0 limit of Schwarzschild space, really, produces either Minkowski space or a
6 1 Stationary Black Holes in General Relativity

1.3.2 Isotropic Coordinates

It is common to see the Schwarzschild metric presented using the isotropic radius rN
defined by
 
M 2
R  rN 1 C : (1.5)
2Nr

Isotropic coordinates .t; rN ; ; '/ consist of the usual Schwarzschild time t and polar
coordinates .; '/, with the Schwarzschild areal radius R replaced by the isotropic
radius rN . Using rN , there are two copies of the spacetime region R > 2M outside the
horizon because Eq. (1.5) gives
r !
R M 2M
rN D 1 ˙ 1 : (1.6)
2 R R

Isotropic coordinates for the Schwarzschild geometry were studied in detail in

Refs. [8, 76].
The definition (1.5) gives the relation between differentials
 
M2
dR D 1  2 dNr (1.7)
4Nr

and, using the relation

!2
2M 1 M
2Nr
1 D ; (1.8)
R 1C M
2Nr

it is easy to obtain the Schwarzschild line element in isotropic coordinates from

Eq. (1.3)
   
M 2
1  2N M 4 2 
ds2 D   r
 2
dt 2
C 1 C dNr C rN 2 d˝.2/
2
: (1.9)
1 C 2N
M 2Nr
r

The Schwarzschild event horizon R D 2M corresponds to rN D M=2, where the two

values (1.6) of the coordinate rN coincide.

Kasner metric [27] and, strictly speaking, a coordinate-free approach [55–57] is needed to make
limits rigorous.
1.3 Schwarzschild Spacetime 7

1.3.3 Kruskal-Szekeres Coordinates

The Kruskal-Szekeres coordinates [44, 72] replace the Schwarzschild coordinates

.t; R/ leaving the polar angles .; '/ untouched and are based on ingoing and out-
going radial null geodesics. Introduce first the Regge-Wheeler tortoise coordinate
[62, 77]
ˇ ˇ
ˇ R ˇ
r  R C 2M ln ˇˇ  1ˇˇ (1.10)
2M

defined in the range 1 < r < C1 corresponding to 2M < R < C1. This
coordinate is chosen so that
 
dR2 2M
D 1 .dr /2 (1.11)
1  2M=R R

and the 2-metric of the .t; r / surface is explicitly conformally flat,

 
2M 
ds2.2/ D 1 dt2 C .dr /2 : (1.12)
R

The differential of r is related to that of the areal radius R by

dR
dr D : (1.13)
1  2M=R

Then the null coordinates

u  t  r ; v  t C r (1.14)

(“retarded time” and “advanced time”, respectively) turn the line element (1.3) into
   
2 2M  2  2 2 2 2M
ds D 1  dt C .dr / C R d˝.2/ D  1  dudv C R2 d˝.2/
2
;
R R
(1.15)
where R.u; v/ is an implicit function of u and v defined by r .R/ D .v  u/ =2. Now
introduce the new null coordinates

U  eu=4M ; V  ev=4M ; (1.16)

where the upper sign refers to the exterior region (there are two copies of the region
R > 2M in these coordinates). The function R.U; V/ is given implicitly by
 
R
eR=2M  1 D UV (1.17)
2M
8 1 Stationary Black Holes in General Relativity

and one obtains the Schwarzschild line element in Kruskal-Szekeres coordinates

32M 3 R=2M
ds2 D  e dUdV C R2 d˝.2/
2
(1.18)
R
which is clearly regular at R D 2M.
The surfaces R D constant correspond to UV D constant, which describes
hyperbolae with two branches asymptotic to the U and V axes. The event horizon
R D 2M corresponds to U D 0 and/or V D 0, while the singularity R D
0 corresponds to two branches of the corresponding hyperbola UV D 1. The
spacetime region R > 2M covered by the Schwarzschild coordinates corresponds to
V > 0 and U < 0 (“region I” in Fig. 1.1), but it is clear that the Schwarzschild
manifold includes also the region U > 0 and V > 0 (“region II”). There are
also two other regions (“region III” and “region IV”) describing a Schwarzschild
white hole, i.e., the time-reversal of a black hole, which constitutes the maximal
extension of the Schwarzschild manifold. However, these regions are not accessible
to timelike or null particles because the Kruskal diagram describes vacuum and
must be cut off at the timelike surface of collapsing matter. The R D 0 singularity of
the Schwarzschild metric is spacelike because the coordinate R turns timelike when
crossing from outside to inside the event horizon R D 2M and for 0 < R < 2M a
surface R D constant is spacelike (much like a surface t D constant in the region
R > 2M outside the event horizon).

1.3.4 Eddington-Finkelstein Coordinates

The Eddington-Finkelstein coordinates [18, 21] use either the retarded or

the advanced time (1.14). The ingoing Eddington-Finkelstein coordinates are
.v; R; ; '/, the outgoing Eddington-Finkelstein coordinates are .u; R; ; '/.
In ingoing Eddington-Finkelstein coordinates .v; R; ; '/, using Eq. (1.13) and
dt D dv  dr D dv  dR= .1  2M=R/, one obtains
 
2M
ds2 D  1  dv 2 C 2dvdR C R2 d˝.2/
2
: (1.19)
R

In outgoing Eddington-Finkelstein coordinates .u; R; ; '/, dt D du C dr D du C

dR
12M=R
gives
 
2 2M
ds D  1  du2  2dudR C R2 d˝.2/
2
: (1.20)
R

(Note the difference in the sign of the off-diagonal term in outgoing and ingoing
coordinates.) The coordinates .v; R/ cover regions I and II of the Kruskal-Szekeres
diagram, while the coordinates .u; R/ cover regions III and IV. Clearly, since u
1.3 Schwarzschild Spacetime 9

U V

0
r=

II st
con
t=
2M
r=

III I

tim
IV eli
ke
wo
rld
lin
e

Fig. 1.1 The Kruskal-Szekeres plane: the horizon R D 2M corresponds to the U and V axes,
the singularity R D 0 to the hyperbola UV D 1, and region I is covered by the Schwarzschild
coordinates .t; R/. The Schwarzschild black hole is described by regions I and II and the white
hole by regions III and IV. The timelike worldline of a particle crossing the event horizon and
falling onto the singularity is also shown

describes outward-propagating radial null rays and these cannot exit from the
Schwarzschild event horizon, the coordinates .u; R/ cannot describe regions I and II
but they are useful to describe the white hole regions. In an Eddington-Finkelstein
.v; R/ or .u; R/ diagram, outgoing radial null geodesics do not propagate at 45ı
angles while ingoing radial null geodesics do. In fact radial null geodesics, which
have ds2 D 0 and d D d' D 0, satisfy
 
2M
dv  1  dv C 2dR D 0 :
R
10 1 Stationary Black Holes in General Relativity

Curves with dv D 0 (v D constant) correspond to ingoing rays of constant

slope dv=dR D 0 at 45ı angles with the axes of the .t; R/ plane, while dv D
2dR= .1  2M=R/ corresponds to rays of variable slope dv=dR D 2= .1  2M=R/.

1.3.5 Painlevé-Gullstrand Coordinates

Painlevé-Gullstrand coordinates for the Schwarzschild geometry [29, 54] are dis-
cussed in various references [25, 26, 43, 65]; a pedagogical introduction is given in
Ref. [47], which we follow here.
The three spatial coordinates .R; ; '/ are the usual Schwarzschild coordinates,
while the Painlevé-Gullstrand time T is the proper time of radial freely falling
observers who start from rest at infinity. The components of the timelike 4-velocity
ua of such observers in Schwarzschild coordinates can be found by solving the
geodesic equation, but it is easier to resort to the staticity of the Schwarzschild
spacetime and to energy conservation along timelike geodesics. The normalization
uc uc D 1 gives
 
0 2 2M 2 2M
.u / 1  D1 C .u1 /2 (1.21)
R R

while, denoting by  a D .@=@t/a the timelike Killing vector, the energy per unit
E
mass EQ  D uc c is conserved along timelike geodesics travelled by particles
m
of mass m [74]. It is
   
2M 2M
EQ D uc c D u  ./ 1  ı 0 D 1 u0 (1.22)
R R

which, in conjunction with Eq. (1.21), yields

2M
EQ 2 D 1  C .u1 /2 : (1.23)
R

Since the observer is infalling, it is u1 < 0. At R ! C1, this relation yields

0 12
B v.1/ C 1
EQ 2 D 1 C @ q A D 2
 2
.1/ ;
2
1  v.1/ 1  v.1/

where v.1/ is the three-dimensional velocity at infinity and .1/ is the correspond-
ing Lorentz factor. The energy per unit mass of this observer can be expressed as
1
EQ D u0.1/ D q : (1.24)
2
1  v.1/
1.3 Schwarzschild Spacetime 11

The Painlevé-Gullstrand time T is obtained by settingr2 v.1/ D 0. With this choice

2M
one obtains EQ D 1, u0 D .1  2M=R/1 , and u1 D  . Therefore, one has
R
r !
1 2M
u D ; ; 0; 0 ; (1.25)
1  2M
R
R
p !
2M=R
u D 1;  ; 0; 0 : (1.26)
1  2M
R

Now, it is

u D r T D @ T (1.27)

for a function T; this property is crucial for the introduction of the Painlevé-
Gullstrand time T. In fact, by integrating it using Eq. (1.26) one obtains @t T D 1
and

T D t C f .R/ (1.28)
p
2M=R
for D 0 while, for D 1, one obtains @R T D and
1  2M=R
Z p
2M=R
TD dR C g.t/ : (1.29)
1  2M=R

By comparing Eqs. (1.28) and (1.29) one obtains

Z p
2M=R
T DtC dR ; (1.30)
1  2M=R

which is immediately integrated to

0 ˇq ˇ1
r ˇ ˇ
ˇ
1 ˇ 2M
R
 1 ˇ
B R ˇC
T D t C 4M @ C ln ˇ q ˇA : (1.31)
2M 2 ˇ R C 1ˇ
ˇ 2M ˇ

2
2
A more general family of coordinates parametrized by the parameter p  1  v.1/ with
0 < v.1/ < 1 are introduced in [45] and discussed in [47]; it includes as special cases the Painlevé-
Gullstrand coordinates for p ! 1, Eddington-Finkelstein coordinates in the lightlike limit p ! 0,
and it is related to another family of coordinate systems characterized by p > 1 and discussed in
Refs. [25, 26].
12 1 Stationary Black Holes in General Relativity

As a check, by differentiating the last expression it is easy to see that T satisfies

Eq. (1.27). Contrary to the Kruskal-Szekeres coordinates, the Painlevé-Gullstrand
time is given as an explicit function of t andpR.
Since, from Eq. (1.30), it is dt D dT  2M=R .1  2M=R/1 dR, one obtains
the Schwarzschild line element in Painlevé-Gullstrand coordinates
  r
2 2M 2 2M
ds D  1  dT C 2 dTdR C dR2 C R2 d˝.2/
2
(1.32)
R R

or, alternatively,
r !2
2M
ds2 D dT 2 C dR C dT C R2 d˝.2/
2
: (1.33)
R

The metric (1.32) is clearly regular at the event horizon R D 2M but singular at
R D 0; it is non-diagonal and the three-dimensional surfaces T D constant are flat,
as can be seen by setting T D constant which gives ds2.3/ D dR2 C R2 d˝.2/ 2
, the
Euclidean line element in three dimensions.
The Painlevé-Gullstrand coordinates do not cover the white hole portion of the
Kruskal-Szekeres plane but only regions I and II because the radial freely falling
observers cross the future, but not the past, event horizon (see Ref. [47] for a
discussion).
Following Ref. [52] (the authors of which actually consider the more complicated
situation of non-static and non-asymptotically flat metrics),
p we can reintroduce
the speed of light c and define the quantity v.R/  2M=R to rewrite the line
element (1.32) as

ds2 D  c2  v 2 .R/ dT 2 C 2v.R/dTdR C dR2 C R2 d˝.2/
2
: (1.34)

1.3.6 Kerr-Schild Coordinates

A Kerr-Schild metric is an algebraically special metric of the form [40, 41, 71]

gab D ab C ka kb ; (1.35)

where ab is the flat Minkowski metric, is a scalar function, and ka is a null
geodesic vector with respect to both ab and gab :

ab ka kb D 0 ; gab ka kb D 0 ; (1.36)

k @c k D k rc k D 0 ;
c a c a
(1.37)
1.3 Schwarzschild Spacetime 13

where ka  ab kb . The inverse of the Kerr-Schild metric (1.35) is

gab D ab  ka kb (1.38)

since gab gbc D ıac . Kerr-Schild coordinates are those in which a Kerr-Schild metric
assumes explicitly the form (1.35). The Einstein field equations for Kerr-Schild
metrics in Kerr-Schild coordinates are linear [31]. The Kerr-Schild class of metrics
includes the Reissner-Nordström, Kerr-Newman, Vaidya, and pp-wave spacetimes.
Kerr-Schild coordinates for the Schwarzschild metric are .Qt; R; ; '/, where the time
coordinate used is [48]

Qt D v  R D t C r  R : (1.39)

The null geodesic vector corresponds to the tangent of the ingoing radial null
congruence and the Schwarzschild metric in Kerr-Schild coordinates is [48]
   
2 2M 4M 2M
ds D  1  dQt 2 C Q
dt dR C 1 C dR2 C R2 d˝.2/
2
; (1.40)
R R R

where D 2M=R and k D .1; 1; 0; 0/ in coordinates .Qt; R; ; '/.

1.3.7 Novikov Coordinates

The Novikov coordinates [48, 53] employ the comoving time  of geodesic
observers and the comoving radius
r
Rmax
R   1; (1.41)
2M

where Rmax is the largest R-coordinate attained by a test particle ejected near the
singularity R D 0 [48, 53]. The Schwarzschild line element in Novikov coordinates
is
  2
1 C R2 @R
ds2 D d 2 C dR2 C R2 d˝.2/
2
; (1.42)
R2 @R

where R D R .; R / is given implicitly by

s s !
   R .R=2M/2  3=2 R=2M
D ˙ 1 C R2  2
C 1 C R2 cos1 :
2M 2M 1 C R 1 C R2
(1.43)
14 1 Stationary Black Holes in General Relativity

1.4 Reissner-Nordström Metric

The Reissner-Nordström spacetime describes the geometry of a static, spherically

symmetric, asymptotically flat, electrically charged black hole which solves the
Einstein-Maxwell equations (therefore, not the vacuum, but the “electrovacuum”
Einstein equations) with a purely electric radial field.

1.4.1 Schwarzschild Coordinates

The Reissner-Nordström line element in Schwarzschild coordinates is

   1
2 2m Q2 2 2m Q2
ds D  1  C 2 dt C 1  C 2 dR2 C R2 d˝.2/
2
; (1.44)
R R R R

while the only non-vanishing components of the Maxwell tensor are those of the
electric field
Q
F 01 D F 10 D : (1.45)
R2
This line element describes a two-parameter family of metrics parametrized by the
mass m (which is the Arnowitt-Deser-Misner
 mass [2]) 2and the electric charge Q.
2m Q
The inverse metric component g11 D 1  C 2 vanishes at
R R
p
R˙ D m ˙ m2  Q2 ; (1.46)

and therefore there are two horizons, commonly called inner and outer horizon.3
The metric (1.44) describes a black hole spacetime when jQj  m and a naked
singularity when jQj > m. The case jQj D m describes an extremal black hole for
which the inner and outer horizons coincide, RC D R D m.
Since g00 vanishes at RC , the Schwarzschild coordinates .t; R; ; '/ become
singular there and they only cover the region R > RC .

1.4.2 Kruskal-Szekeres Coordinates

A Kruskal-Szekeres coordinate patch can be introduced which covers the region

R < R < C1, but it does not penetrate the inner horizon R D R and another

3
In the terminology to be introduced later, the outer horizon R D RC is an event and an apparent
horizon, while the inner horizon R D R is an apparent, but not an event, horizon, and is also a
Cauchy horizon which is unstable [9, 59].
1.4 Reissner-Nordström Metric 15

coordinate patch is needed there (that is, Kruskal-Szekeres coordinates are specific
to a single horizon). We introduce u  t  r and v  t C r as usual, the function
2m Q2
f .R/  1  C 2 , and the quantity
R R
 
f 0 .RC / 1 Q2
C  D 2 m : (1.47)
2 RC RC

Then, for R < R < C1, we define the Kruskal-Szekeres coordinates

UC D eC u ; VC D eC v ; (1.48)

with the upper sign for the exterior region R > RC and the lower one for the
interior region R < RC . These coordinates are well behaved near the outer horizon
but become singular near the inner horizon where r ! C1 (see Ref. [58]
for a detailed discussion). Near the outer horizon the metric in Kruskal-Szekeres
coordinates is
2
ds2 '  2 dUC dVC C R2C d˝.2/ 2
: (1.49)
C

A second patch of Schwarzschild coordinates, distinct from the one used for R >
RC , can be used in the region R < R < RC . To extend the metric inside the
inner horizon, define again u  t  r and v  t C r using the inner Schwarzschild
radial and the tortoise coordinates, in addition to

f 0 .R /
  ; (1.50)
2
and then

U D e u ; V D e v ; (1.51)

with the upper sign for R > R and the lower one for R < R . Near the inner
horizon R D R the line element is [58]

2
ds2 '  dU dV C R2 d˝.2/
2
; (1.52)
2

which is regular at R D R . The coordinates .U ; V ; 0; 0/ are regular at the inner

horizon but singular at the outer one R D RC .
For R < R , it is g00 D f < 0, hence the singularity at R D 0 is timelike,
contrary to the Schwarzschild R D 0 singularity which is spacelike. This means
that the Reissner-Nordström singularity can be avoided by observers in the region
R < R , which can go around it in 3-space, while in the Schwarzschild spacetime all
observers who have crossed the horizon R D 2M must meet the R D 0 singularity
within a finite time.
16 1 Stationary Black Holes in General Relativity

1.5 Kerr Spacetime

The stationary and axially symmetric Kerr metric [39] is interpreted as describing a
spinning black hole. The Kerr line element in Boyer-Linquist coordinates .t; r; ; '/
is
 
2mr 4mar sin2  ˙ 2
ds2 D  1  2 dt2  2
dtd' C 2 sin2 d' 2 C dr2 C 2 d 2

(1.53)
or, equivalently,
2 2
 2 ˙
ds2 D  dt C 2 sin2  .d'  !dt/2 C dr2 C 2
d 2 ; (1.54)
˙ 

where
2
D r2 C a2 cos2  ; (1.55)
 D r2  2mr C a2 ; (1.56)
 2
˙ D r2 C a2  a2  sin2  ; (1.57)
g03 2mar
!D D : (1.58)
g33 ˙

This line element describes a two-parameter family of spacetimes parametrized by

the mass M and the angular momentum per unit mass a (the total angular momentum
at spatial infinity given by the Komar formula is J D Ma). When a ! 0 the Kerr
metric reduces to the Schwarzschild one.
The inverse metric has components

˙
g00 D  2
; (1.59)

2Mar
g03 D g30 D  2
; (1.60)


g11 D 2
; (1.61)

1
g22 D 2
; (1.62)

  a2 sin2 
g33 D 2  sin2 
: (1.63)

The metric is singular at  D 0 and D 0; the first surface corresponds to

a coordinate singularity while the second one corresponds to a true spacetime
1.5 Kerr Spacetime 17

singularity. In fact, the Kretschmann scalar

  
48m2 r2  a2 cos2  4
 16a2 r2 cos2 
Rabcd Rabcd
D 12
(1.64)

diverges there.
The static limit surface is defined by g00 D 0, which yields
p
rSL D m C m2  a2 cos2  : (1.65)

This static limit surface is also defined by considering static observers, i.e.,
observers whose 4-velocity is parallel to the timelike Killing vector  a D .@=@t/a
of components  D .1; 0; 0; 0/,

a
ua D  a  p (1.66)
gcd  c  d

or, in components,
ı0
u Dp : (1.67)
jg00 j

This equation becomes invalid, and static observers no longer exist, when r
approaches the static limit rSL . When r  rSL all observers are forced to co-rotate
with the spacetime, a phenomenon known as the “dragging of inertial frames”.
The static limit is not an event horizon; there is an event horizon located at
p
rC D m C m2  a2 ; (1.68)

where the quantity  vanishes. The static limit touches the event horizon at the poles
 D ˙ =2, where rSL D rC , and the region between the static limit and the event
horizon is called ergosphere. The event horizon exists only if a  m, equivalent
to J  m2 . If a D m (or J D m2 ), the black hole is extremal. For a > m, the
Kerr metric describes a naked singularity and the extremal black hole, therefore,
constitutes a threshold between black holes and naked singularities in parameter
space.
There are two roots of g11 D 0, the equation which, as we will see later,
locates the horizons. Using Eq. (1.61), this condition is seen to be equivalent to
 D r2  2mr C a2 D 0, which has as roots the radius of the event horizon r D rC
and
p
r D m  m2  a2 : (1.69)

This surface is analogous to the inner horizon of the Reissner-Nordström spacetime.

The two horizons at r D r˙ coincide for an extremal black hole with a D m.
18 1 Stationary Black Holes in General Relativity

1.6 Kerr-Newman Metric

The Kerr-Newman spacetime is interpreted as describing an electrically charged

spinning black hole. It is a three-parameter (mass M, angular momentum per unit
mass a, and electric charge Q) family of metrics. The Kerr-Newman line element in
Boyer-Linquist coordinates is
2 2
 2 ˙
ds2 D  dt C 2 sin2  .d'  !dt/2 C dr2 C 2
d 2 (1.70)
˙ 

where
2
D r2 C a2 cos2  ; (1.71)
2 2 2
 D r  2mr C a C Q ; (1.72)
 2
˙ D r2 C a2  a2  sin2  ; (1.73)
 
a r2 C a2  
!D : (1.74)
˙
The Kerr-Newman metric reduces to the Kerr metric in the limit Q ! 0 and to the
Reissner-Nordström one when a ! 0. The static limit is given by
p
rSL . / D m C m2  Q2  a2 cos2  I (1.75)

the horizon is located at

p
rC D m C m2  Q2  a2 ; (1.76)

with the ergosphere comprising the region rC < r < rSL .

1.7 Energy Conditions

Here we summarize the point-wise energy conditions of General Relativity. The

energy conditions satisfied by a certain form of matter are formulated in terms of its
energy-momentum tensor Tab . In order to visualize an energy condition, it is useful
to see the form that it assumes for a perfect fluid characterized by the stress-energy
tensor

Tab D .P C / ua ub C Pgab : (1.77)

1.7 Energy Conditions 19

• The weak energy condition (WEC) is

Tab ta tb  0 for all timelike vectors ta : (1.78)

For the fluid (1.77), this condition becomes

0 and C P  0: (1.79)

• The dominant energy condition (DEC) consists of the WEC and of the extra
requirement that T ab ta be a null or timelike vector (i.e., Tab T b c ta tc  0) for any
timelike vector ta . For the fluid (1.77) the DEC assumes the form

 jPj ; (1.80)

(i.e., the speed of the energy flow does not exceed the speed of light).
• The null energy condition (NEC) consists of

Tab la lb  0 for all null vectors la I (1.81)

for the fluid (1.77), this means that

C P  0: (1.82)

Violations of the NEC are studied in the context of macroscopic traversable

wormholes and occur in cosmology if the expansion of the universe is super-
P > 0, where H is the Hubble parameter); in this case energy is
accelerated (H
called phantom energy.
• The null dominant energy condition (NDEC) consists of

Tab la lb  0 and T ab lb is null or timelike for any null vector la : (1.83)

The NDEC resembles the DEC but here la is null instead of timelike. The NDEC
for the fluid (1.77) amounts to

 jPj or D P : (1.84)

• The strong energy condition (SEC) consists of

 
1
Tab  Tgab ta tb  0 for any timelike vector ta (1.85)
2

or, for the fluid (1.77),

CP0 and C 3P  0 : (1.86)

20 1 Stationary Black Holes in General Relativity

This condition ensures that gravity is always attractive and is violated by a

positive cosmological constant (which satisfies P./ D  ./ ), during inflation,
or in a dark energy-dominated cosmological era with P <  =3.
Quantum systems are expected to violate all of the energy conditions, including
positivity of the energy density, on short timescales (e.g., [22]). However, while a
negative energy density is permitted for a quantum system during a certain interval
of time, it appears that later on the system more than compensates with a positive
energy density (quantum interest), thus respecting some averaged energy conditions.
Even at the classical level, alternative theories of gravity and the theory of a scalar
field nonminimally coupled to the Ricci curvature contain fields which can violate
all of the energy conditions [10, 20].

1.8 Conclusions

The previous list of spacetimes commonly associated with black holes is not
exhaustive. For detailed discussions of these spacetimes, the timelike and null
geodesics in them, and the thermodynamics of their horizons, see Refs. [11, 23,
48, 58, 71, 74, 75]. In the following chapters we assume that the reader has some
familiarity with these basic solutions of Einstein’s theory and we will focus on
causal barriers which preclude the knowledge of spacetime regions to families of
(timelike) observers.

Problems

1.1. Prove that u  t  r and v  t C r are null coordinates. Show that the
coordinates uN  t  r and vN  t C r, instead, are not null.
1.2. Establish the causal character of the inner and outer horizons of the Reissner-
Nordström metric (1.44).

References

1. Amendola, L., Tsujikawa, S.: Dark Energy, Theory and Observations. Cambridge University
Press, Cambridge (2010)
2. Arnowitt, R.L., Deser, S., Misner, C.W.: The dynamics of general relativity. In: Witten, L.
(ed.) Gravitation: An Introduction to Current Research, pp. 227–265. Wiley, New York (1962).
[Reprinted in arXiv:gr-qc/0405109]
3. Ashtekar, A., Galloway, G.J.: Some uniqueness results for dynamical horizons. Adv. Theor.
Math. Phys. 9, 1 (2005)
References 21

4. Ashtekar, A., Krishnan, B.: Isolated and dynamical horizons and their applications. Living Rev.
Relat. 7, 10 (2004)
5. Babichev, E., Charmousis, C.: Dressing a black hole with a time-dependent Galileon. J. High
Energy Phys. 1408, 106 (2014)
6. Baumgarte, T.W., Shapiro, S.L.: Numerical relativity and compact binaries. Phys. Rept. 376,
41 (2003)
7. Booth, I.: Black hole boundaries. Can. J. Phys. 83, 1073 (2005)
8. Buchdahl, H.A.: Isotropic coordinates and Schwarzschild metric. Int. J. Theor. Phys. 24, 731
(1985)
9. Burko, L.M., Ori, A. (eds.): Internal Structure of Black Holes and Spacetime Singularities, An
International Research Workshop, Haifa (IOP, Bristol, 1997)
10. Capozziello, S., Faraoni, V.: Beyond Einstein Gravity, A Survey of Gravitational Theories for
Cosmology and Astrophysics (Springer, New York, 2010)
11. Carroll, S.M.: Spacetime and Geometry—An Introduction to General Relativity (Addison-
Wesley, San Francisco, 2004)
12. Carter, B.: Axisymmetric black hole has only two degrees of freedom. Phys. Rev. Lett. 26, 331
(1970)
13. Chu, T., Pfeiffer, H.P., Cohen, M.I.: Horizon dynamics of distorted rotating black holes. Phys.
Rev. D 83, 104018 (2011)
14. Collins, W.: Mechanics of apparent horizons. Phys. Rev. D 45, 495 (1992)
15. Davis, A.-C., Gregory, R., Jha, R., Muir, J.: Astrophysical black holes in screened modified
gravity. J. Cosmol. Astropart. Phys. 1408, 033 (2014)
16. Detweiler, S.: Resource letter BH-1: black holes. Am. J. Phys. 49, 394 (1981)
17. De Felice, A., Tsujikawa, S.: f .R/ theories. Living Rev. Relat. 13, 3 (2010)
18. Eddington, A.S.: A comparison of Whitehead’s and Einstein’s formulas. Nature 113, 192
(1924)
19. Eling, C., Guedens, R., Jacobson, T.: Nonequilibrium thermodynamics of spacetime. Phys.
Rev. Lett. 96, 121301 (2006)
20. Faraoni, V.: Cosmology in Scalar-Tensor Gravity. Kluwer Academic, Dordrecht (2004)
21. Finkelstein, D.: Past-future asymmetry of the gravitational field of a point particle. Phys. Rev.
D 110, 965 (1958)
22. Ford, L.H., Roman, T.A.: Classical scalar fields and violations of the second law. Phys. Rev. D
64, 024023 (2001)
23. Frolov, V.P., Novikov, I.D.: Black Hole Physics, Basic Concepts and New Developments.
Kluwer Academic, Dordrecht (1998)
24. Gallo, E., Marolf, D.: Resource letter BH-2: black holes. Am. J. Phys. 77, 294 (2009)
25. Gautreau, R.: Light cones inside the Schwarzschild radius. Am. J. Phys. 63, 431 (1995)
26. Gautreau, R., Hoffmann, B.: The Schwarzschild radial coordinate as a measure of proper
distance. Phys. Rev. D 17, 2552 (1978)
27. Geroch, R.: Limits of spacetimes. Commun. Math. Phys. 13, 180 (1969)
28. Gibbons, G.W., Hawking, S.W.: Cosmological event horizon, thermodynamics, and particle
creation. Phys. Rev. D 15, 2738 (1977)
29. Gullstrand, A.: Allgemeine lösung de statischen eink örper-problems in der Einsteinschen
gravitations theories. Ark. Mat. Astron. Fys. 16, 1 (1922)
30. Gundlach, C., Price, R.H., Pullin, J.: Late-time behavior of stellar collapse and explosions. I.
Linearized perturbations. Phys. Rev. D 49, 883 (1994)
31. Gürses, M., Gürsey, F.: Lorentz covariant treatment of the Kerr-Schild geometry. J. Math. Phys.
16, 2385 (1975)
32. Hawking, S.W.: Black hole explosions? Nature 248, 30 (1970)
33. Hawking, S.W.: Black holes in general relativity. Commun. Math. Phys. 25, 152 (1972)
34. Hawking, S.W.: Particle creation by black holes. Commun. Math. Phys. 43, 199 (1975);
Erratum 46, 206 (1976)
35. Hayward, S.A.: General laws of black hole dynamics. Phys. Rev. D 49, 6467 (1994)
22 1 Stationary Black Holes in General Relativity

36. Herdeiro, C.A.R., Radu, E.: Kerr black holes with scalar hair. Phys. Rev. Lett. 112, 221101
(2014)
37. Israel, W.: Event horizons in static vacuum space-times. Phys. Rev. 164, 1776 (1967)
38. Jacobson, T.: Thermodynamics of spacetime: the Einstein equation of state. Phys. Rev. Lett.
75, 1260 (1995)
39. Kerr, R.P.: Gravitational field of a spinning mass as an example of algebraically special metrics.
Phys. Rev. Lett. 11, 237 (1963)
40. Kerr, R.P., Schild, A.: A new class of vacuum solutions of the Einstein field equations. In:
Atti del Convegno sulla relativitá generale: problemi dell’energia e onde gravitazionali, p. 222.
Barbera, Firenze (1965)
41. Kerr, R.P., Schild, A.: Some algebraically degenerate solutions of Einstein’s gravitational field
equations. Proc. Symp. Appl. Math. 17, 199 (1965)
42. Kolb, E.W., Turner, M.S.: The Early Universe. Addison-Wesley, Reading (1990)
43. Kraus, P., Wilczek, F.: Some applications of a simple stationary line element for the Schwarz-
schild geometry. Mod. Phys. Lett. A 9, 3713 (1995)
44. Kruskal, M.D.: Maximal extension of Schwarzschild metric. Phys. Rev. 119, 1743 (1960)
45. Lake, K.: A class of quasi-stationary regular line elements for the Schwarzschild geometry.
Preprint, arXiv:gr-qc/9407005
46. Liddle, A.R., Lyth, D.H.: Cosmological Inflation and Large Scale Structure. Cambridge
University Press, Cambridge (2000)
47. Martel, K., Poisson, E.: Regular coordinate systems for Schwarzschild and other spherical
spacetimes. Am. J. Phys. 69, 476 (2001)
48. Misner, C.W., Thorne, K.S., Wheeler, J.A.: Gravitation. Freeman, New York (1973)
49. Mukhanov, V.: Physical Foundations of Cosmology. Cambridge University Press, Cambridge
(2005)
50. Muller zum Hagen, H., Robinson, D.C., Seifert, H.J.: Black holes in static vacuum space-times.
Gen. Rel. Gravit. 4, 53 (1973)
51. Nielsen, A.B.: Black holes and black hole thermodynamics without event horizons. Gen. Rel.
Gravit. 41, 1539 (2009)
52. Nielsen, A.B., Visser, M.: Production and decay of evolving horizons. Class. Quantum Grav.
23, 4637 (2006)
53. Novikov, I.D.: PhD thesis, Shternberg Astronomical Institute, Moscow (1963)
54. Painlevé, P.: La méchanique classique et la théorie de la relativité. Comp. Rend. Acad. Sci.
(Paris) 173, 677 (1921)
55. Paiva, F.M., Romero, C.: On the limits of Brans-Dicke space-times: a coordinate-free approach.
Gen. Rel. Gravit. 25, 1305 (1993)
56. Paiva, F.M., Reboucas, M., MacCallum, M.: On limits of space-times: a coordinate-free
approach. Class. Quantum Grav. 10, 1165 (1993)
57. Paiva, F.M., Reboucas, M., Hall, G.S., MacCallum, M.: Limits of the energy-momentum tensor
in general relativity. Class. Quantum Grav. 15, 1031 (1998)
58. Poisson, E.: A Relativist’s Toolkit: The Mathematics of Black-Hole Mechanics. Cambridge
University Press, Cambridge (2004)
59. Poisson, E., Israel, W.: The internal structure of black holes. Phys. Rev. D 41, 1796 (1990)
60. Price, R.H.: Nonspherical perturbations of relativistic gravitational collapse. I. Scalar and
gravitational perturbations. Phys. Rev. D 5, 2419 (1972)
61. Price, R.H.: Nonspherical perturbations of relativistic gravitational collapse. II. Integer-spin,
zero-rest-mass fields. Phys. Rev. D 5, 2439 (1972)
62. Regge, T., Wheeler, J.A.: Stability of a Schwarzschild singularity. Phys. Rev. 108, 1063 (1957)
63. Rinaldi, M.: Black holes with nonminimal derivative coupling. Phys. Rev. D 86, 084048 (2012)
64. Rindler, W.: Visual horizons in world-models. Mon. Not. R. Astr. Soc. 116, 663 (1956).
[Reprinted in Gen. Rel. Gravit. 34, 133 (2002)]
65. Robertson, H.P., Noonan, T.W.: Relativity and Cosmology. Saunders, Philadelphia (1968)
66. Robinson, D.C.: Classification of black holes with electromagnetic fields. Phys. Rev. D 10, 458
(1974)
References 23

67. Robinson, D.C.: Uniqueness of the Kerr black hole. Phys. Rev. Lett. 34, 905 (1975)
68. Sotiriou, T.P., Faraoni, V.: f .R/ theories of gravity. Rev. Mod. Phys. 82, 451 (2010)
69. Sotiriou, T.P., Zhou, S.-Y.: Black hole hair in generalized scalar-tensor gravity. Phys. Rev. Lett.
112, 251102 (2014)
70. Sotiriou, T.P., Zhou, S.-Y.: Black hole hair in generalized scalar-tensor gravity: an explicit
example. Phys. Rev. D 90, 124063 (2014)
71. Stephani, H., Kramer, D., MacCallum, M., Hoenselaers, C., Herlt, E.: Exact Solutions of
Einstein’s Field Equations, 2nd edn. Cambridge University Press, Cambridge (2003)
72. Szekeres, G.: On the singularities of a Riemannian manifold. Publ. Mat. Debr. 7, 285 (1960)
73. Thornburg, J.: Event and apparent horizon finders for 3 C 1 numerical relativity. Living Rev.
Relat. 10, 3 (2007)
74. Wald, R.M.: General Relativity. Chicago University Press, Chicago (1984)
75. Wald, R.M.: The thermodynamics of black holes. Living Rev. Relat. 4, 6 (2001)
76. Weyl, H.: Zur Gravitationstheorie. Ann. Phys. (Leipzig) 54, 117 (1917)
77. Wheeler, J.A.: Geons. Phys. Rev. 97, 511 (1955)
Chapter 2
Horizons

The greatest deception men suffer is from their own opinions.

—Leonardo da Vinci

2.1 Introduction

We will now review standard tools used in the analysis of horizons and black hole
physics, most notably the congruences of null geodesics crossing a horizon. Many
textbooks present this standard material but we will recall it here anyway to provide
a more self-contained discussion. Then we will present the laws describing how
these null geodesic congruences are affected by gravity. Certain formulae which
are used in the calculations of the following chapters will also be introduced here.
The case of spherical symmetry is particularly important because most of the known
analytic dynamical solutions of Einstein theory and of alternative theories of gravity,
which are used to gain physical insight into both gravitational physics and the
physics of horizons, are spherically symmetric.

2.2 Null Geodesic Congruences and Trapped Surfaces

A null geodesic is a curve on the spacetime manifold which has null tangent la (i.e.,
la la D 0) and satisfies the geodesic equation

lb rb la D ˛. / la ; (2.1)

where is a parameter along the curve. The geodesic equation expresses the fact
that the tangent is transported parallel to itself as one moves along the geodesic, or,
the fact that this curve is “as straight as possible” in the curved spacetime.
The parameter can be chosen so that the geodesic equation simplifies to

l b rb l a D 0 (2.2)

© Springer International Publishing Switzerland 2015 25

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6_2
26 2 Horizons

or, in components,

d2 x dx˛ dxˇ
C  ˛ˇ D 0; (2.3)
d 2 d d

where x . / are the coordinates of points along the curve. If the geodesic equation
reduces to the simple form (2.3), we say that it is affinely parametrized and that is
an affine parameter. Two affine parameters and 0 can differ at most by an affine
transformation,
0
Da Cb (2.4)

where a and b are real constants (e.g., [78]).

Let O be an open region of the spacetime manifold; a congruence of curves in
O is a family of curves such that through every point of O passes one and only one
curve of the family. The tangents to these curves define a vector field on O (and,
conversely, every continuous vector field in O generates a congruence of curves,
those to which the vector field is tangent). If the field of tangents is smooth, we say
that the congruence is smooth. In particular, we can consider a congruence of null
geodesics with tangents la in the open region O.
Consider a congruence of affinely parametrized null geodesics with tangent la
and affine parameter , which satisfy la la D 0 and lb rb la D 0. Let us consider now
another parameter s which labels the various geodesics of the congruence in O,
so that the family of curves parameterized by and s generates a 2-dimensional
surface with coordinates and s. Points along the geodesics of the congruence
in O have spacetime coordinates x D x . ; s/ and, in analogy with the tangent
l D @x =@ , we can introduce the deviation vector with components   @x =@s.
By construction, it is

a la D 0 (2.5)

but a can still have a component along the curves, that is, parallel to la because la is
null (this would not be the case if we were considering timelike geodesics instead).
However, we can restrict ourselves to deviation vectors which are considered to
be equivalent if they differ only by a component along la . More precisely: define
the equivalence relation a  a0 , a0  a D bla for some real number b. It is
straightforward to show that  is an equivalence relation and we can consider
equivalence classes of deviation vectors.
The tangent space composed of all vectors orthogonal in this sense to la
constitutes a 2-dimensional vector space and we can consider its dual and the space
of tensors built with them (see Refs. [78, 95] for details). It can be proved that the
geodesic deviation vector a satisfies the geodesic deviation equation

D2 a
D Ra bcd ub c ud ; (2.6)
D 2
2.2 Null Geodesic Congruences and Trapped Surfaces 27

which expresses the fact that neighbouring geodesics deviate from each other
because of the spacetime curvature.
Consider now a congruence of null geodesics with tangent la parametrized by
an affine parameter , with la la D 0 and lc rc la D 0. The metric hab in the 2-space
orthogonal to la is constructed as follows [78]: select another null vector field na
such that nc nc D 0. The null vectors lc and nc are defined up to rescalings; here we
normalize them according to lc nc D 1, but other choices are possible and will be
used later. Then, the 2-metric orthogonal to la is

hab  gab C la nb C lb na : (2.7)

It is easy to verify that hab is purely spatial and ha b is a projection operator onto the
2-space orthogonal to la :

hab la D hab lb D 0 ; (2.8)

h a
a D 2; (2.9)
ha c hc b D ha b : (2.10)

Only the null congruence with tangent la is fixed and the choice of na is not unique,
but geometric and physically relevant quantities do not depend on it.
Let a be the geodesic deviation vector and define the tensor field [78, 95]

Bab  rb la ; (2.11)

which satisfies lb rb a D Ba b b and is orthogonal to the null geodesics, Bab la D

Bab lb D 0. The transverse part of the deviation vector (“relative velocity” of two
neighbouring geodesics) is

Q a  ha b b D a C .nc c /la : (2.12)

The orthogonal component of lc rc a , denoted by a tilde, is [78]

B
.lc rc a / D ha b hc d Bb c Q d  BQa d Q d : (2.13)

The transverse tensor BQ ab is decomposed into its symmetric and antisymmetric parts,
and the symmetric part is further decomposed into its trace and trace-free parts as
[78, 95]
 

BQ ab D BQ .ab/ C BQ Œab  hab C ab C !ab ; (2.14)
2

where the trace

  gab BQ ab D gab Bab D rc lc (2.15)

28 2 Horizons

is the expansion of the affinely-parametrized congruence (this quantity does not

depend on the vector na );


ab  hab (2.16)
2
is the expansion tensor;


ab  BQ .ab/  hab (2.17)
2
is the shear tensor; and

!ab  BQ Œab (2.18)

is the vorticity tensor. The expansion, shear, and vorticity tensors are purely
transversal:

ab la D ab lb D 0 ; (2.19)

ab l D ab l D 0 ;
a b
(2.20)
!ab la D !ab lb D 0 ; (2.21)

and the shear and vorticity are trace-free,  a a D ! a a D 0. The shear scalar and
vorticity scalar are

 2  ab  ab ; ! 2  !ab ! ab (2.22)

and they are non-negative. The propagation of the expansion along a null geodesic
is ruled by the Raychaudhuri equation [78, 95]

d 2
D    2 C ! 2  Rab la lb (2.23)
d 2
(which is invariant under redefinitions of na ), and similar propagation equations hold
for ab and !ab [95]. This equation describes how null rays are focused (d=d < 0)
or defocused (d=d > 0) by expansion itself, shear, rotation, and matter (related
to Rab through the Einstein equations). Conventional forms of energy in General
Relativity satisfy the positive curvature condition Rab la lb  0 and the old adage
“gravity always focuses” follows from this condition. Therefore, a gravitational lens
will focus light rays passing nearby. Shear acts as gravity while rotation acts in the
opposite direction. This effect is familiar in Newtonian gravity, where the rotation
of a massive body is associated with a centrifugal acceleration which counteracts
the gravitational attraction (except at the poles).
Thus far, we have considered affinely-parametrized geodesics. If the congruence
of null geodesics with tangent la is not affinely-parametrized, the geodesic equation
assumes the form
2.2 Null Geodesic Congruences and Trapped Surfaces 29

l c rc l a D  l a (2.24)

instead of lc rc la D 0, where the quantity  which measures the failure of la to

be affinely-parametrized is sometimes used, on a horizon, as a possible defini-
tion of surface gravity [72, 77]. The expansion of a congruence of non-affinely
parametrized null geodesics is then

 D rc l c   (2.25)

and the Raychaudhuri equation becomes [78]

d 2
D    2 C ! 2  Rab la lb : (2.26)
d 2
A compact and orientable surface has two independent directions orthogonal to
it, corresponding to ingoing and outgoing null rays. In the presence of spherical
symmetry, one is naturally led to study congruences of radial ingoing and outgoing
null geodesics with tangent fields la and na , respectively, which are orthogonal to the
2-spheres of symmetry. In this case the role of the vector na is played by the tangent
to the ingoing null geodesics (hence the use of the same symbol). To compute the
expansion of the null vector la when the geodesic to which it is tangent is not
necessarily affinely-parametrized, the relation

la nb C na lb
l D hab ra lb D gab C ra lb (2.27)
.nc ld gcd /

is useful. Here ha b acts as a projection tensor onto the two-dimensional surface

to which both la and na are normal. If la is null everywhere then the expression
lb ra lb in the third term on the right hand side vanishes identically. Equation (2.27)
is independent of the field equations of the theory of gravity and can be applied
when lc and nc are not normalized to satisfy lc nc D 1 (this normalization is usually
imposed but it is not necessary and it sometimes conflicts with other requirements
that one may want to impose on lc and nc , such as in the various possible definitions
of surface gravity on a horizon [72] or in calculations of quasi-local energy [49]).
The following are basic definitions for closed 2-surfaces in regard to the
ingoing and outgoing null geodesic congruences,1 which have expansions n and
l , respectively [4, 21, 24, 68]:
• A normal surface corresponds to l > 0 and n < 0 (this is the case, e.g., of a
2-sphere in Minkowski space in the absence of gravity).

1
While these 2-surfaces are usually assumed to be spacelike [4, 19, 21] this condition is not
imposed here.
30 2 Horizons

• A trapped surface [76] corresponds to l < 0 and n < 0. In this case, the
outgoing, in addition to the usual ingoing, future-directed null rays are converg-
ing instead of diverging—light propagating outward is dragged back by strong
gravity.
• A marginally outer trapped (or marginal) surface (MOTS) corresponds to l D 0
(where la is the outgoing null normal to the surface) and n < 0.
• An untrapped surface is one with l n < 0.
• An anti-trapped surface corresponds to l > 0 and n > 0 (both outgoing and
ingoing future-directed null rays are diverging).
• A marginally outer trapped tube (MOTT) is a 3-dimensional surface which can
be foliated entirely by marginally outer trapped (2-dimensional) surfaces.
It was proved by Penrose that, in General Relativity, if a spacetime contains
a trapped surface, the null energy condition holds, and there is a non-compact
Cauchy surface for the spacetime, then the spacetime contains a singularity [76].
Trapped surfaces seem to be essential features in the black hole concept and
notions of “horizon” of practical utility will be identified with the boundaries of
spacetime regions which contain trapped surfaces. The mathematical conditions for
the existence and uniqueness of MOTSs are not totally clear at the moment. In
general, a marginally outer trapped tube can be distorted smoothly, which implies
that MOTTs are non-unique [2, 21, 39].
We will now examine various types of horizons which appear in the literature on
black holes, cosmology, quantum field theory in curved spaces, and the thermody-
namics usually associated with these horizons.

2.3 Rindler Horizons for Accelerated Observers

in Minkowski Spacetime

In Minkowski space, consider a particle of position x.t/ moving with 3-velocity

u D dx=dt in an inertial frame. Its 3-dimensional acceleration is a  du=dt. The
transformation law of the 3-acceleration between this frame and another inertial
frame moving with constant speed v with respect to it is [87]

1 ax
a0x D ; (2.28)
3
.1  vux /3
1 ax vuy
a0y D 2
ay C ; (2.29)
2 .1  vux / .1  vux /
1 ax vuz
a0z D 2
az C ; (2.30)
2 .1  vux / .1  vux /
 1=2
where D 1  v 2 is the Lorentz factor. Let us define now uniform accelera-
tion: for simplicity, assume that the particle moves along the x-axis, so u D .u; 0; 0/.
2.3 Rindler Horizons for Accelerated Observers in Minkowski Spacetime 31

Clearly, one cannot proceed as in Newtonian theory by requiring that du=dt D con-
stant  a for constant acceleration because then u.t/ D u0 C at exceeds the speed
of light when t is sufficiently large, which is not possible in Special Relativity.
Therefore, a different definition is required (e.g., Ref. [38]). A particle has uniform
acceleration if and only if its acceleration has the same value at each instant in any
inertial frame comoving with the particle (i.e., in any frame moving with the same
velocity as the particle, in which the particle is instantaneously at rest). At different
times there are different inertial frames which are comoving but it is required that
they all measure the same acceleration a of the particle. In other words, it is required
that the particle move along a straight line and has constant proper acceleration a
(where the proper acceleration is the particle acceleration in the frame in which the
particle is instantaneously at rest).
In any instantaneously comoving frame (characterized by v D u), the particle
has velocity and acceleration

u0 D 0 ; (2.31)
0
du
D a D const. (2.32)
dt0
Then, the transformation rule for the acceleration between inertial frames gives (we
invert Eq. (2.28) by exchanging primed and unprimed quantities and changing v into
v)

du 1 du0
D (2.33)
dt 3 .1 C vu0 /3 dt0

and, setting u0 D 0 and du0 =dt0 D a D constant,

 3=2 du
1  u2 D a: (2.34)
dt
Z
dz z
This equation is integrated by remembering that q D p ,
.˛ 2  z2 /3 ˛ ˛ 2  z2
2

yielding the algebraic equation for u

u
p D a .t  t0 / (2.35)
1  u2

where we assumed that an integration constant u0 vanishes. This algebraic equation

is solved as

dx a .t  t0 /
DuD q ; (2.36)
dt
1 C a2 .t  t0 /2
32 2 Horizons

and this ordinary differential equation can be easily integrated to yield

q
1 1
x.t/  x0 D 1 C a2 .t  t0 /2  : (2.37)
a a
By squaring and collecting terms, one obtains
 2
x  x0 C 1a .t  t0 /2
 1 2   2 D 1 : (2.38)
1
a a

For simplicity we take the initial condition x0 D 1=a at t0 D 0, reducing

Eq. (2.38) to
 2
2 1 2
x t D ; (2.39)
a

a family of hyperbolae parametrized by the constant acceleration a and called

hyperbolic motions [25] or worldlines of Rindler observers [38]. The worldline of
the uniformly accelerated observer can be parametrized by its proper time  and has
equation
   sinh.a / cosh.a/ 
x . / D t. /; x./; 0; 0 D ; ; 0; 0 (2.40)
a a

and tangent
 
u D cosh.a /; sinh.a/; 0; 0 (2.41)

with uc uc D 1. Then, it is

dx
vDuD D tanh.a/ (2.42)
dt
and the Lorentz factor is

D cosh.a / : (2.43)
r
1
As is evident by writing t D ˙ x2  , these hyperbolae have the branches of the
a2
light cone through the origin t D ˙x as asymptotes. The effect of choosing arbitrary
initial condition x0 and arbitrary initial time t0 is only to move the origin. Each
hyperbola is the worldline of a uniformly accelerated observer travelling in the x-
direction. It is clear from Fig. 2.1 that the lines t D ˙x separate the spacetime in
two regions: the region to the left of this null cone through the origin is forever
unaccessible to the uniformly accelerated observer because signals sent from it
would have to travel faster than light to cross the line t D x and reach the observer.
This observer has an horizon, called Rindler (or acceleration) horizon. An horizon is
2.3 Rindler Horizons for Accelerated Observers in Minkowski Spacetime 33

Fig. 2.1 The Rindler horizon of a uniformly accelerated observer is the boundary of the shaded
region, i.e., the line t D x. A light signal emitted from the shaded region would have to travel faster
than light to reach the accelerated observer to the right of it

a causal barrier, a surface which separates spacetime into two regions—the region
beyond the horizon cannot causally influence the region in which the observer is
located and no information sent by the region beyond the horizon can ever leak out
and reach the observer. Vice-versa, signals sent by the observer can cross the event
horizon and propagate through the other region. The location of the event horizon
depends on the uniformly accelerated observer: different accelerated observers will
determine different acceleration horizons.
A uniformly accelerated observer in Minkowski spacetime is subject to the
Unruh effect of quantum field theory [34, 44, 89] (see, e.g., [27] for a pedagogical
exposition). Although an observer at rest in this Minkowski spacetime “sees” a
quantum vacuum corresponding to zero particles of a quantum field (usually taken,
for simplicity, to be a massless minimally coupled scalar field), the uniformly
accelerated observer will detect a thermal bath of particles in equilibrium at the
Unruh temperature given by
 
„ a
kB T D ; (2.44)
c 2
34 2 Horizons

where kB is the Boltzmann constant, „ is the reduced Planck constant, and the
speed of light c has been restored. The Unruh effect is analogous to the thermal
Hawking emission by black holes, with the uniform acceleration a playing a role
analogous to that of the surface gravity at a black hole horizon. Although there
is no gravity in Minkowski space, the Killing vector  a D x.@=@t/a C t.@=@x /a of
Minkowski spacetime defines a surface gravity equal to a [27], as we will see in
Sect. 2.12.
By defining the Rindler coordinates .T; X; Y; Z/ as

t D X sinh.aT/ ; (2.45)
x D X cosh.aT/ ; (2.46)
yDY; (2.47)
z D Z; (2.48)

or, inversely, as

1 t
TD tanh1 ; (2.49)
a x
p
X D x2  t 2 ; (2.50)
Y D y; (2.51)
Z D z; (2.52)

for jtj < jxj, the Minkowski line element ds2 D dt2 C dx2 C dy2 C dz2 is turned
into the Rindler line element

ds2 D a2 X 2 dT 2 C dX 2 C dY 2 C dZ 2 : (2.53)

The Rindler coordinate chart covers the region

1 < T < C1 ;
0 < X < C1 ;
1 < Y < C1 ;
1 < Z < C1 :

The Rindler metric is nothing but the Minkowski metric in a chart covering the
wedge jtj < jxj. Although the Rindler metric does not look flat, a calculation of
the Riemann tensor (which vanishes identically) shows that the geometry is indeed
flat. The Rindler metric has a coordinate singularity at X D 0, which corresponds
to the Rindler horizon t D ˙x. The metric can be continued analytically beyond the
horizon by going back to the original Minkowski coordinates .t; x; y; z/.
2.4 Event Horizons 35

2.4 Event Horizons

The traditional notion of horizon emerging from the study of static or stationary
black holes in General Relativity is that of event horizon. An event horizon is a
connected component of the boundary of the causal past of future null infinity [47,
48, 78, 95]. In the notations of Ref. [95], in which I C is future null infinity and
J  .I C / is its causal past, i.e.,
 the set of
 all events which can send lightlike signals
to I C , the event horizon is @ J  .I C / . This definition embodies the most peculiar
feature of a black hole, i.e., the horizon is a causal boundary which separates a region
from which nothing can come out to reach a distant observer from a region in which
signals can be sent out and eventually arrive to this observer. An event horizon is
generated by the null geodesics which fail to reach infinity and, therefore (provided
that it is smooth) is always a null hypersurface.
In black hole research and in astrophysics the concept of event horizon is
implicitly taken to define the concept of static or stationary black hole itself.
However, since to define and locate an event horizon one must know all the future
history of spacetime (one must know all the geodesics which do reach null infinity
and, tracing them back, the boundary of the region from which they originate),
an event horizon is a globally defined concept. To state that an event horizon has
formed (which traditionally is understood to mean that a black hole has formed or
is about to form) requires knowledge of the spacetime outside our future light cone,
which is impossible to achieve (unless, of course, the spacetime is stationary and
the black hole has existed forever—then nothing changes and by knowing the state
of the world now one knows it forever). It is often said that the event horizon has a
teleological nature. It has been shown [6, 19] that, inside a collapsing spherical shell
in Vaidya spacetime, an event horizon forms and grows, starting from the centre,
and an observer can cross it and be unaware of it even though his or her causal
past consists entirely of a portion of flat Minkowski space. In other words, the event
horizon “knows” about events belonging to a spacetime region very far away and
with no causal connection (a property called “clarvoyance” [18]).
Because of its global nature, an event horizon is not a practical notion to work
with, and it is nearly impossible to locate precisely an event horizon in a general
dynamical situation. In practice, astrophysical black holes did not exist forever but
formed in a highly dynamical process of gravitational collapse. Numerical relativity
codes are written to follow a gravitational collapse, the merger of a binary system,
or other dynamical situations ending in a black hole, and they crash at some point.
It is clearly impossible to follow the evolution of a system all the way to future null
infinity. Numerical relativists routinely use marginally trapped surfaces as proxies
for event horizons (see, e.g., [16, 29]).
Strictly speaking, an event horizon H is a tube in spacetime; it is a common
abuse of terminology to refer to the intersections of H with surfaces of constant
time (which produce 2-surfaces) as “event horizons”. Although improper, this
terminology is widespread and extends to the other notions of horizon that we define
below.
36 2 Horizons

Example 2.1. In the Schwarzschild geometry

 
2M dR2
ds2 D  1  dt2 C C R2 d˝.2/
2
(2.54)
R 1  2M
R

the surface R D 2M is a well known event horizon.

p
Example 2.2. The outer horizon RC D m C m2  Q2 of the Reissner-Nordström
p
metric (1.44) is an event horizon. The inner horizon R D m  m2  Q2 is not
because in the entire region R < RC there are outgoing radial null geodesics which
fail to reach future null infinity and the surface R D R is not a boundary of a region
with this property.

2.5 Killing Horizons

Remember that a Killing vector field ka is one that satisfies the Killing equation

ra kb C rb ka D 0 : (2.55)

A Killing vector describes a symmetry of spacetime in a geometric, coordinate-

invariant way [38, 48, 95].
A Killing horizon H of the spacetime .M; gab / is a null hypersurface which
is everywhere tangent to a Killing vector field ka which becomes null, kc kc D 0,
on H . This Killing vector field is timelike, kc kc < 0, in a spacetime region which
has H as boundary. Stationary event horizons in General Relativity are usually
Killing horizons for a suitably chosen Killing vector [28]. For example, in the
Schwarzschild geometry (1.3), the timelike Killing vector ka D .@=@t/a in the
R > 2M region outside the horizon becomes null at R D 2M and spacelike
inside (i.e., for R < 2M). The event horizon R D 2M is also a Killing horizon.
More generally, any event horizon in a locally static spacetime is also a Killing
horizon for the Killing vector ka D .@=@t/a associated with the time symmetry. If
the spacetime is stationary and asymptotically flat (but not necessarily static), it
must be axisymmetric and any event horizon is a Killing horizon for the Killing
vector

ka D .@=@t/a C ˝H .@=@'/a ; (2.56)

which is a linear combination of the vectors associated with time and rotational sym-
metries, and where ˝H is the angular velocity at the horizon (this statement requires
the assumption that the Einstein-Maxwell equations hold and some assumption on
the matter stress-energy tensor [48, 96]).
The concept of Killing horizon ceases to be useful in spacetimes, or spacetime
regions, which are not stationary and do not admit timelike Killing vectors. There
2.6 Apparent Horizons 37

have been attempts to use conformal Killing horizons in spacetimes which are
conformal to the Schwarzschild one ([37, 84], see also [61–63, 85]) but this approach
is a priori very restrictive and, in retrospect, it does not seem to have been very
productive. Instead, the introduction of the Kodama vector, which resembles in some
way a Killing field but is defined in spacetimes without Killing vectors, seems to be
much more useful in defining surface gravities and in the thermodynamics of time-
evolving horizons.
When present, a Killing horizon defines a notion of surface gravity Killing , as we
will see in Sect. 2.10.

2.6 Apparent Horizons

A future apparent horizon2 is the closure of a surface (usually a 3-surface) which

is foliated by marginal surfaces (l D 0). The future apparent horizon is a surface
defined by the conditions on the time slicings [50]

l D 0 ; (2.57)
n < 0 ; (2.58)

where l and n are the expansions of the future-directed outgoing and ingoing null
geodesic congruences, respectively. Equation (2.57) tells us that the future-pointing
outgoing null geodesics momentarily stop propagating outward and, presumably,3
turn around at the horizon, while the condition (2.58) distinguishes between black
holes and white holes.
Apparent horizons are defined quasi-locally4 and do not refer to the global causal
structure of spacetime—they don’t have the teleological nature of event horizons.
Apparent horizons (and also trapping horizons, see below) depend on the choice
of the foliation of the 3-surface with marginal surfaces [82, 97]. Also the ingoing
and outgoing null geodesics orthogonal to these surfaces and their expansions l
and n depend on the foliation [42]. The expansions l and n are scalars and are,

2
This is not the definition of apparent horizon originally introduced in the book by Hawking and
Ellis [48], which is not easy to work with in practice [21]. It is unfortunate that the term “apparent
horizon” corresponds to different precise definitions in the literature. The definition that we provide
here is more useful than that of [48] in practical (including numerical) applications and is the one
which is adopted in most of the recent literature.
3
The fact that these null rays “hesitate” (l D 0) does not necessarily imply that they are turning
around and will subsequently propagate inward. One could have, for example, a wormhole throat
at which outgoing null rays “hesitate” and then propagate outward again.
4
“Quasi-local” refers to a quantity which can be measured by an observer in a finite lifespan
experiment, as opposed to a global quantity which requires the observer to know the entire future
history or causal structure of spacetime which is, of course, physically impossible and would
require an infinite observation time in the non-stationary case.
38 2 Horizons

therefore, independent of the coordinate system chosen, but sometimes a choice

of coordinates helps in specifying locally the foliation (for example by choosing
spacelike surfaces of constant time coordinate—different time coordinates identify
different families of hypersurfaces of constant time), which is a geometric construct
and is coordinate-independent. However, congruences of outgoing and ingoing null
geodesics orthogonal to these surfaces will, of course, change by changing the
foliation while being coordinate-independent. The dependence of apparent horizons
on the spacetime slicing is made clear by the fact that non-symmetric slicings of the
Schwarzschild spacetime exist which do not admit any apparent horizon [82, 97].
Apparent horizons are, in general, distinct from event horizons: for example,
these horizons do not coincide in the Reissner-Nordström black hole (inner horizon)
and in the Vaidya spacetime [78]. Also in static black holes which are perturbed,
the apparent and the event horizons do not coincide. During the spherical collapse
of uncharged matter, an event horizon forms before the apparent horizon does and
these two horizons approach each other until they eventually coincide as the final
static state is reached [48].
Apparent horizons are always found inside the event horizon provided that
the null curvature condition Rab la lb  0 8 null vector la is satisfied [48]. This
requirement coincides with the null energy condition Tab ła lb  0 8 null vector la
if the Einstein equations are imposed, and in this case it is believed to be a
“reasonable” condition on physical matter. But the Hawking radiation produced
by horizons violates the weak and the null energy conditions [92]. Also a simple
scalar field non-minimally coupled to the curvature can violate the energy conditions
[17] and the null curvature condition is easily violated in alternative theories of
gravity (for example, Brans-Dicke [26] and scalar-tensor [20, 74, 94] theories)
and the apparent horizon is reported to lie outside of the event horizon during
spherical collapse in Brans-Dicke gravity, although it eventually settles inside the
event horizon when the static Schwarzschild configuration is achieved [81]. One
should not be too attached to the notion of event horizon and should perhaps look at
other notions of horizon as more realistic even though they depend on the spacetime
slicing (apparent, trapping, and dynamical horizons have this dependence).

2.7 Trapping Horizons

A future outer trapping horizon (FOTH) is the closure of a surface (usually a 3-

surface) foliated by marginal surfaces such that on its 2-dimensional “time slicings”
the conditions ([50], see also [73] and references therein)

l D 0 ; (2.59)
n < 0 ; (2.60)
Ln l D na ra l < 0 ; (2.61)
2.7 Trapping Horizons 39

are satisfied, where l and n are the expansions of the future-directed outgoing and
ingoing null geodesic congruences, respectively. The condition (2.61) is introduced
to distinguish between inner and outer horizons (e.g., in the non-extremal Reissner-
Nordström solution) and also distinguishes between apparent horizons and trapping
horizons (it is not imposed for apparent horizons but it is required for trapping ones).
One obtains the definition of past inner trapping horizon (PITH) by exchanging
la with na while reversing the signs of the inequalities,

n D 0 ; (2.62)
l > 0 ; (2.63)
Ll n D l ra n > 0 :
a
(2.64)

The past inner trapping horizon characterizes a white hole or a cosmological

horizon.
If we don’t exchange la and na , we can say that, in general, a trapping horizon
(TH) satisfies the definition requirements [21]

l D 0 ; (2.65)
n ¤ 0 ; (2.66)
Ln l ¤ 0 ; (2.67)

and is
• Future if n < 0,
• Past if n > 0,
• Outer if na ra l < 0,
• Inner if na ra l > 0.
For black holes, the trapping horizon has been associated with thermodynamics,
and it has even been claimed that it is the trapping horizon area and not the
area of the event horizon which should be associated with entropy in black hole
thermodynamics [31, 45, 56, 68]. This claim, however, is the subject of controversy
[32, 70, 83]. The Parikh-Wilczek “tunneling” approach [75] predicts Hawking
radiation also for apparent and for trapping horizons, not only for event horizons
[30, 35, 54, 58, 67, 71, 93] but also this aspect is not entirely free of controversy
[15].
Trapping horizons do not, in general, coincide with event horizons. Dramatic
examples are spacetimes which possess trapping horizons but not event horizons
[53, 80]. The difference between the areas of the trapping and the event horizon in
particular spacetimes have been studied in Ref. [69].
Example 2.3. In the Reissner-Nordström spacetime with the natural spherically
symmetric foliation, the event horizon r D rC is a future outer trapping horizon
(FOTH), the inner Cauchy horizon r D r is a future inner trapping horizon (FITH),
while the white hole horizons are past trapping horizons (PTHs).
40 2 Horizons

2.8 Isolated and Dynamical Horizons

In the realm of black holes, isolated horizons correspond to isolated systems in

thermal equilibrium not interacting with their surroundings, which could instead
be dynamical. The concept of isolated horizon has been introduced in the context
of loop quantum gravity [3, 7–11, 13, 41]. In general, this horizon construct is too
restrictive when one wants to allow mass-energy to cross the “horizon” in whatever
direction.
A weakly isolated horizon is a null surface H with nullnormal la such that
l D 0, Tab la is a future-oriented and causal vector, and Ll nb ra lb D 0. In this
context la is a Killing vector of the intrinsic geometry on H , without reference to
the surroundings, and can be used to define a “completely local Killing horizon” in
the absence of energy fluxes across H . The vector field la generates a congruence
of null geodesics on H , which can be employed to define a surface gravity  using
the non-affinely parametrized geodesic equation

l a ra l b D  l b ; (2.68)

which also yields

 D nb la ra lb ; (2.69)

where nb lb D 1. The surface gravity  defined in this way is constant on the weakly
isolated horizon H , corresponding to the zeroth law of thermodynamics. The vector
field na is not unique, hence this notion of surface gravity is not defined uniquely.
An Hamiltonian analysis of the phase space of isolated horizons identifies
boundary terms with energies of these boundaries and produces a first law of
thermodynamics for isolated horizons with rotational symmetry, i.e.,

ıHH D ıA C ˝H ıJ : (2.70)
8
Here J is the angular momentum, HH is the Hamiltonian, A is the area of the
2-dimensional cross-sections of H , and ˝H is the angular velocity of the horizon.
Next, a dynamical horizon [6] is a spacelike marginally trapped tube (MTT).
This definition allows energy flows to cross the dynamical horizon. A set of flux laws
describing the related changes of the area of the dynamical horizons are available
[6]. An apparent horizon which is everywhere spacelike is a dynamical horizon, but
an apparent horizon is not required to be spacelike.5 Being spacelike, dynamical
horizons can be crossed only in one direction by causal curves, while this is not the
case for apparent horizons which can be partially or entirely non-spacelike.

5
An apparent horizon which is everywhere timelike is called a timelike membrane [4, 5, 24].
2.9 Kodama Vector and Surface Gravity 41

To end this string of definitions, slowly evolving horizons have also been defined
and examined [21–23, 59]: they are “almost isolated” FOTHs and they are meant
to characterize black hole horizons which evolve very slowly in time. This slow
evolution is expected in many astrophysical processes but not, for example, in the
final stages of collapse or evaporation. Slowly evolving horizons are analogous to
thermodynamic systems in a regime near equilibrium.

2.9 Kodama Vector and Surface Gravity

Various notions of surface gravity associated with horizons have been introduced
in the literature. In static and stationary situations a timelike Killing vector field is
present outside the horizon and becomes null on it, and these notions of surface
gravity coincide and are well known from the study of the Kerr-Newman black
holes of General Relativity. In dynamical situations, however, there is no timelike
Killing vector and the various notions of surface gravity encountered in the literature
turn out to be inequivalent. In spherical symmetry, the Kodama vector mimics the
properties of a Killing vector and gives rise to a conserved current and a surface
gravity.
The Kodama vector [60] generalizes the notion of Killing vector field to
spacetimes which do not admit one and has been used as a substitute of a Killing
vector in the thermodynamics of time-dependent horizons. The Kodama vector is
defined only in spherical symmetry.6 Write the spacetime metric as

ds2 D hab dxa dxb C R2 d˝.2/

2
; (2.71)

where a; b D 0; 1 and R is the areal radius, and let ab be the volume form of the
2-metric hab [95]; then the Kodama vector is defined as [60]

K a   ab rb R (2.72)

with K  D K ' D 0. The Kodama vector lies in the 2-dimensional .t; R/ surface
(where t is the time coordinate) orthogonal to the 2-spheres of symmetry.7 In a
static spacetime the Kodama vector is parallel (in general, not equal) to the timelike
Killing vector. In the region in which it is timelike, p
the Kodama vector defines a
class of preferred observers with 4-velocity u  K = jK c Kc j.
a a

6
Reference [88] attempts to generalize the Kodama vector to non-spherically symmetric space-
times.
7
Moreover, K a ra R D  ab ra Rrb R D 0 because  ab is antisymmetric and ra Rrb R is symmetric.
42 2 Horizons

It can be proved ([60], see [1] for a simplified proof) that the Kodama vector has
zero divergence

ra K a D 0 I (2.73)

this property has the consequence that the Kodama energy current

J a  Gab Kb (2.74)

(where Gab is the Einstein tensor) is covariantly conserved, r a Ja D 0 even if there

is no timelike Killing vector. This surprising property is sometimes called the
“Kodama miracle” [1, 60].
If the spherically symmetric metric is written in Schwarzschild-like coordinates,

ds2 D A .t; R/ dt2 C B .t; R/ dR2 C R2 d˝.2/

2
; (2.75)

then the Kodama vector takes on the simple form (e.g., [60, 79])
 a
1 @
Ka D p : (2.76)
AB @t
p p
Proof. Denoting by h the determinant of the 2-metric hab , we have jhj D AB
and the volume form of hab is
p
ab D AB .ra trb R  ra Rrb t/

so that
p
 ab D gac gbd cd D gac gbd AB .ıc0 ıd1  ıd0 ıc1 /
p  
D AB ga0 gb1  gb0 ga1 :

Then the only non-vanishing components of  ab are  01 D  10 D .AB/1=2 and the

Kodama vector has components
 
 1
p 0 00 11 ı 0 1 @
K D  r R D  D AB ı g g Dp Dp :
AB AB @t

t
u
It is shown in [51] that the Noether charge associated with the Kodama conserved
current is the Misner-Sharp-Hernandez mass [55, 64]. This notion of mass energy,
too, is defined only in spherical symmetry.
2.10 Surface Gravities 43

The Kodama vector was used in Ref. [79] (see also [33]) to simplify the evolution
equations and the initial value problem for the coupled Einstein-Klein-Gordon
equations with spherical symmetry.

2.10 Surface Gravities

Surface gravity is defined classically in terms of geometric properties of the metric

tensor. However, the same quantity appears in black hole thermodynamics as the
proportionality factor between the variation of the black hole mass (acting as the
internal energy) dM and the variation of the event horizon area (proportional to the
entropy) dA. Since it is not established at present which definition of black hole mass
is appropriate in non-trivial backgrounds (different definitions are possible, see the
review [86]), it is obvious that also the definition of surface gravity will be subject to
the same ambiguities. Moreover, surface gravity appears also semiclassically since,
up to a numerical constant, it coincides with the Hawking temperature of a black
hole.
The traditional definition of surface gravity is given on a Killing horizon for
stationary spacetimes [95]. However, given that Killing horizons are not defined
in general non-stationary situations in which one considers quasi-local horizons
instead of event/Killing horizons, a different concept of surface gravity is necessary
in these cases. There are several definitions of surface gravity in the literature, and
they are inequivalent. The recurrent definitions are reviewed in Ref. [72] and are
briefly recalled here.

2.10.1 Killing Horizon Surface Gravity

As already remarked, a Killing horizon defines a notion of surface gravity Killing as

follows [95]: on the Killing horizon the Killing vector ka satisfies

ka ra kb  Killing kb ; (2.77)

so Killing measures the failure of the geodesic Killing vector ka to be affinely-

parametrized (the “inaffinity”) on the Killing horizon.
Proof. We need to prove [72] that, on the Killing horizon, ka ra kb is proportional
to kb . First, note that the equation defining the Killing horizon is kc kc D 0, hence
rb .kc kc / D 2kc rb kc is orthogonal to the horizon. But since ka is also normal to the
horizon (because ka is null there) and it generates this horizon, then kc rb kc / kb .
Using the Killing equation ra kb C rb ka D 0, one obtains

kc rb kc D kc rc kb / kb ;

so kc rc kb / kb , or kc rc kb D Killing kb . t
u
44 2 Horizons

Another property of the Killing surface gravity is (cf. Ref. [95], p. 332)

2 1  a b
Killing D r k .ra kb / : (2.78)
2
In static spacetimes (and only in those!) one can interpret the surface gravity Killing
as the limiting force required at spatial infinity to hold in place a unit test mass just
above the event horizon by means of an infinitely long massless string [95]. This
interpretation shows the non-local nature of the notion of Killing surface gravity.
Since the Killing equation ra kb C rb ka D 0 determines the Killing vector ka
only up to an overall normalization, there is freedom to rescale ka and the value
of the surface gravity depends on the non-affine parametrization chosen for ka .
However, in stationary situations one has the freedom of imposing that kc kc D 1
at spatial infinity.
Example 2.4. For the Kerr-Newman black hole (1.70)–(1.74), the Killing surface
gravity is [95]
p
M 2  a2  Q2
Killing D  p  : (2.79)
2M M C M 2  a2  Q2  Q2

The Killing surface gravity can be generalized to any event horizon that is not
a Killing horizon by replacing the Killing vector ka with the null generator of the
event horizon [72].

2.10.2 Surface Gravity of Marginally Trapped Surfaces

Consider the outgoing and ingoing null normals la and na to a marginally trapped
(spacelike compact 2-dimensional) surface, and assume that the expansion of la
vanishes, with la and na normalized so that lc nc D 1. la is not a horizon generator
in general, but it is still a non-affinely parametrized geodesic vector on the trapping
horizon. This fact allows one to introduce a surface gravity  by

l a ra l b   l b (2.80)

or

 D nb la ra lb : (2.81)

The value of this surface gravity depends on the parametrization of la , for which
various proposals have been advanced. By writing la as the tangent to a null curve
x . / with parameter , a parameter change (which depends on the spacetime point)
! 0 forces the components of la to change as
2.10 Surface Gravities 45

dx 0 dx d
l D !l D Dl  ˝.x/ l (2.82)
d d 0 d 0

so that
0 0 0
l  r 0 l D 0l ;
˝l r .˝l / D  0 ˝l ;
˝ l C .l r ˝/ l D  0 l ; (2.83)

and

 !  0 D ˝  C l  r ˝ : (2.84)

The Hayward proposal for spherical symmetry [52] uses the Kodama vector K a
which is always available in spherical symmetry [60]. This future-directed vector
satisfies
 
rb Ka T ab D 0 : (2.85)

The Hayward notion of surface gravity Kodama for a trapping horizon is given by

1 ab c
g K .rc Ka  ra Kc / D Kodama K b : (2.86)
2
This definition is unique because of the uniqueness of the Kodama vector. The
surface gravity Kodama agrees with the surface gravity on the horizon of a Reissner-
Nordström black hole but not with other dynamical surface gravity constructs. An
expression equivalent to (2.86) is [52]

1 1 p 
Kodama D .h/ R D p @ h h  @ R ; (2.87)
2 2 h

where h is the determinant of the metric hab in the 2-space orthogonal to the
2-spheres of symmetry. The Hamilton-Jacobi approach (a variant of the Parikh-
Wilczek method [75]) to study the Hawking radiation of time-dependent horizons,
leads naturally to the Kodama-Hayward definition of surface gravity [36] (see
Ref. [90] for a review of tunneling methods).

2.10.3 Fodor et al. Surface Gravity

The proposal by Fodor et al. [43] applies to spherically symmetric and asymptoti-
cally flat spacetimes and is based on the normalization la ta D 1 of the ingoing
46 2 Horizons

null normal na , where ta is the asymptotic time-translational Killing vector at

spatial infinity. The curve with tangent na is affinely-parametrized everywhere. By
requiring that lc nc D 1, this choice fixes the parametrization of la and one obtains
[43]

Fodor D nb la ra lb : (2.88)

2.10.4 Isolated Horizon Surface Gravity

This proposal, due to Ashtekar, Beetle, and Fairhurst [8] applies to an isolated
horizon. The null normal na is normalized in such a way that its expansion
agrees with that of the Reissner-Nordström black hole and with la na D 1.
This normalization identifies a unique surface gravity as a function of the horizon
parameters. This notion of surface gravity appears to be quite limited, for example
it cannot be extended to the Einstein-Yang-Mills case [12, 72].

2.10.5 Proposal for Slowly Evolving Horizons

The notion of slowly evolving horizons [22] extends the previous proposal for
surface gravity. On the isolated horizon the normal is  a D Bla C Cna , with B and C
scalar fields defined there, which weigh the contributions of la and na (for an isolated
horizon it is B D 1 ; C D 0). The slowly evolving horizon surface gravity is

SE  Bna lb rb la  Cla nb rb na : (2.89)

2.10.6 Other Proposals

There are other proposals for surface gravity in the literature, including Hayward’s
trapping gravity [50]

1p a
trapping  n ra l (2.90)
2
and the Mukohyama-Hayward proposal ([66], see also [72]).
All these definitions have been computed, for a general spherically symmetric
metric in Eddington-Finkelstein coordinates and in terms of the Misner-Sharp-
Hernandez mass [55, 64], in Ref. [72]. A critical comparison of these definitions
for black holes in spherical symmetry and using Painlevé-Gullstrand coordinates is
contained in Ref. [77].
2.11 Spherical Symmetry 47

2.11 Spherical Symmetry

The assumption of spherical symmetry greatly simplifies the study of horizons and
the solution of the field equations. While exact spherical symmetry is unrealistic
for astrophysical black holes (which rotate and may be distorted by other bodies
and by magnetic fields) and for realistic universes perturbed by non-spherical
inhomogeneities, it is an important assumption for the fundamental theory and it
may also be a realistic approximation in certain situations, especially in cosmology.
In spherical symmetry, the apparent horizons (existence, location, dynamics,
surface gravity, etc.) can be studied by using the Misner-Sharp-Hernandez mass
MMSH [55, 64], which coincides with the Hawking-Hayward quasi-local energy
[46, 49] for spherical spacetimes. In General Relativity the Misner-Sharp-Hernandez
mass is defined only for spherically symmetric spacetimes. In terms of the areal
radius R and angular coordinates .; '/, a spherically symmetric line element is
written as

ds2 D hab dxa dxb C R2 d˝.2/

2
; (2.91)

where a; b D 1; 2. The Misner-Sharp-Hernandez mass MMSH is defined by [55, 64]

2MMSH
1  r c R rc R (2.92)
R

or by8

R 
MMSH D 1  hab ra R rb R : (2.93)
2
Note that the Misner-Sharp-Hernandez mass is an invariant quantity of the 2-space
normal to the 2-spheres of symmetry.
Horizons in the presence of spherical symmetry are discussed in a clear
and elegant way using the Nielsen-Visser formalism ([73], see also Ref. [71]).
They consider the most general spherically symmetric metric with a spherically
symmetric spacetime slicing, which assumes the form

2M.t; R/ dR2
ds2 D e2.t;R/ 1  dt2 C C R2 d˝.2/
2
(2.94)
R 1  2M.t;R/
R

8
In .n C 1/ spacetime dimensions, the Misner-Sharp-Hernandez mass is
n.n  1/ n2  
MMSH D R Vn 1  hab ra Rrb R , where the line element is
16
n=2
ds2 D hab dxa dxb C R2 d˝.n1/
2
(a; b D 1; 2) and Vn D n  is the volume of the .n  1/-
 2
C1
dimensional unit ball [14].
48 2 Horizons

in Schwarzschild-like coordinates. A posteriori, M.t; R/ turns out to be the Misner-

Sharp-Hernandez mass. This form is ultimately inspired by the Morris-Thorne
wormhole metric [65], it compromises between the latter and the gauge (2.75), and
is particularly convenient in the study of both static and time-varying black holes
[73, 91]. It is not assumed that the spacetime is stationary or asymptotically flat.
The line element (2.94) can be rewritten in Painlevé-Gullstrand coordinates as
(see Appendix A.1)
  r
2 e2 2M 2 2e 2M
ds D  2
1 d C ddRCdR2 CR2 d˝.2/
2
; (2.95)
.@=@t/ R @=@t R

where .; R/ and M.; R/ are implicit functions of .; R/ and the hypersurfaces9
 D constant are flat (setting d D 0 gives ds2.3/ D dR2 CR2 d˝.2/
2
, the 3-dimensional
Euclidean metric in spherical coordinates). By defining the implicit functions of
.; R/ [73]

e.t;R/
c .; R/  ; (2.96)
.@=@t/
r r
2M.t; R/ e.t;R/ 2M
v .; R/  Dc ; (2.97)
R @=@t R

the line element is rewritten as


ds2 D  c2 .; R/  v 2 .; R/ d 2 C 2v .; R/ ddR C dR2 C R2 d˝.2/
2
: (2.98)

We now list a number of results obtained in [73] which are useful for practical
computations in spherical symmetry.
The outgoing radial null geodesic congruence has tangent field
!
1
l D 1; c.; R/  v.; R/; 0; 0 ; (2.99)
c.; R/

in Painlevé-Gullstrand coordinates .; R; ; '/, while the ingoing radial null

geodesics have tangent field
!
1
n D 1; c.; R/  v.; R/; 0; 0 ; (2.100)
c.; R/

9
The hypersurfaces  D constant are obviously spacelike since  is a timelike coordi-
nate. To be explicit, the normal Na D ra  to a surface  D constant has norm squared
Na N a D g00
H D e
2H
< 0 and N a is timelike.
2.11 Spherical Symmetry 49

with the normalization [73]

gab la nb D 2 : (2.101)

The expansions of these radial null geodesic congruences are computed as

r !
2 2M
l D 1 ; (2.102)
R R
r !
2 2M
n D  1C : (2.103)
R R

A 2-sphere of symmetry of radius R is [51, 73, 78]

• Trapped if R < 2M,
• Marginal if R D 2M,
• Untrapped if R > 2M.
The apparent horizon is the boundary between trapped and untrapped surfaces and
corresponds to l D 0 and n < 0 and is, therefore, given by

2M .; RAH /
D 1 ” r c Rrc R jAH D 0 ” gRR jAH D 0 ; (2.104)
RAH . /

where the last equation holds in both Painlevé-Gullstrand coordinates and in the
gauge (2.94) and is obtained by using the inverse of the metric (2.98)
0 1
1 v 0 0
B C
B C
B v c2  v 2 0 0 C
B
1 B C
 C
.g / D 2 B C: (2.105)
c B C
B 0 0 R2
0 C
B C
@ A
0 0 0 R2 sin2 

The condition gRR D 0 is a very convenient and practical recipe to locate the
apparent horizons in spherical symmetry when the areal radius R is used as a
coordinate. Sometimes it is convenient to perform a coordinate transformation to
this radial coordinate to rewrite the line element using R explicitly.
The gradient of the areal radius R and the normal na D ra R to the surfaces
R D constant become null at the apparent horizon. It is clear that this recipe
resembles the change in causal character of the Schwarzschild radial coordinate
(which is also an areal radius) on the Schwarzschild event horizon. In general,
however, the apparent horizon is not a null surface like the event horizon of the
Schwarzschild black hole (see the examples below).
50 2 Horizons

Nielsen and Visser provide also the derivative

" r !#   !
2 2M 0
2 1  2MAH RP AH
Ln l jAH D n ra
a
1 D 1C ;
R R R2AH 2cAH
AH
(2.106)
with a prime and an overdot denoting partial differentiation with respect to R and
 , respectively, and with the subscript AH denoting quantities evaluated on the
0
apparent horizon. It is noted in Refs. [71, 73] that 1  2MAH > 0 is required for
the horizon to be outer in a spacetime with regular asymptotic region, hence the
condition for the apparent horizon to be also a trapping horizon is

RP AH > 2cAH : (2.107)

Equation (2.103) tells us that along ingoing radial null geodesics with tangent na it
is RP D 2e , hence the apparent horizon at RAH D 2MAH is a trapping horizon if
0
it is outer (2MAH < 1) and does not move inward faster than the ingoing radial null
geodesics (in which case the latter would not be trapped).
If matter satisfies the null energy condition, and assuming the Einstein equations,
the area of the apparent horizon cannot decrease. Various energy fluxes across the
apparent horizon are also introduced and computed in [73]. The Nielsen-Visser
surface gravity at the horizon is computed using lb rb la D l la , obtaining [73]

1  2M 0 .; RH . //
l . / D : (2.108)
2RH . /

An extremal horizon is one which has vanishing surface gravity,

1  2M 0 .; RH . // D 0 : (2.109)

The fact that the Misner-Sharp-Hernandez mass (which coincides with the Haw-
king-Hayward quasi-local energy in spherical symmetry [46, 49]) can be employed
to locate apparent horizons in spherical symmetry makes it clear once again that the
apparent horizon is a quasi-local concept and is independent of the global causal
structure. However, it is not a completely local concept.
Using the Misner-Sharp-Hernandez mass it is easy to see when there are an
0
inner and an outer horizon. When this situation happens, it is MMSH > 1=2 at the
0
inner horizon and MMSH < 1=2 at the outer horizon. Graphically, this means that
(at a given time  , or at all times if the metric is stationary) at the intersections
between the graph of the function MMSH .R/ and the line MMSH D R=2 (which are
apparent horizons), the curve MMSH .R/ is steeper [respectively, less steep] than this
line. If MMSH .R/ rises faster than R=2, in an asymptotically flat spacetime in which
MMSH .R/ eventually asymptotes to a constant as R ! C1, a continuous MMSH will
have to cross the line R D MMSH =2 again, and there will be an outer horizon [71].
2.12 Rindler Horizons Revisited 51

2.12 Rindler Horizons Revisited

Armed with the notions of Killing horizon and Killing surface gravity, let us
revisit now the Rindler horizon of a uniformly accelerated observer in Minkowski
spacetime. The Killing vector field of this space associated with Lorentz boosts in
the x-direction has components
 ˛  ˛
˛ @ @
 Dx Ct ; (2.110)
@t @x

or  D .x; t; 0; 0/ in Cartesian coordinates .t; x; y; z/. Since  c c D x2 C t2 , this

vector is timelike if jxj > jtj, null on the light cone t D ˙x, and spacelike if jxj < jtj.
Therefore, there is a Killing horizon at t D ˙x.
Along the worldline (2.40) of a uniformly accelerated observer it is

1
 c c D x2 C t2 D  (2.111)
a2
and, therefore, we are led to normalize the Killing vector according to
 ˛  ˛
˛ ˛@ @
k  a D ax C at ; (2.112)
@t @x

or k D .ax; at; 0; 0/ in Cartesian coordinates with norm squared

 
ka ka D a2 x2 C t2 :

Along the worldline (2.40) of a uniformly accelerated observer, ka is normalized to

1.
The vanishing of the norm of the Killing vector ka describes a Killing horizon for
the uniformly accelerated observer; the entire Minkowski space can be threaded by
observers and their associated Rindler horizons.10 Now, we can associate a Killing
surface gravity Killing to this Rindler-Killing horizon as specified by Eq. (2.77). We
have

k˛ r˛ kˇ D k0 @t kˇ C k1 @x kˇ D ax @t kˇ C at @x kˇ

and
 
k r k D a2 t; a2 x; 0; 0 : (2.113)

10
This fact is essential in Jacobson’s thermodynamics of spacetime formalism [40, 57].
52 2 Horizons

 
On the Rindler-Killing horizon t D ˙x, it is k r k jRH D ˙a2 x; a2 x; 0; 0 and
Eqs. (2.77) and (2.113) give
 2 2 
˙a x; a x; 0; 0 D Killing .ax; ˙ax; 0; 0/ : (2.114)

In order for the Killing vector (2.112) to be future-oriented on the horizon t D ˙x,
one needs to choose the positive sign in k jRH D .˙at; at; 0; 0/ and in Eq. (2.114),
obtaining the Killing surface gravity of the Rindler horizon

Killing D a ; (2.115)

which coincides with the uniform acceleration of the Rindler observer. The Unruh
temperature (2.44) can then be written as
 
„ Killing
kB T D : (2.116)
c 2

2.13 Conclusions

We are now aware of the various notions of horizon and of surface gravity in the
literature. We have studied, in particular, the situation of spherical symmetry which
will accompany us for the rest of these lectures. Our simplified exposition is not
comprehensive and can certainly be made more rigorous: the reader can find more
detailed and technically more satisfactory treatments in the references provided, but
those would break the flow of our discussion. We are now going to apply the theory
of horizons discussed in this chapter to specific solutions of the Einstein equations
(and, later, of the field equations of alternative theories of gravity), in the presence
of spherical symmetry. The simplest situation that comes to mind is that of horizons
in spatially homogeneous and isotropic cosmologies, and its study is our next step.
In spite of its simplicity, this situation is not entirely trivial.

Problems

2.1. The tangent field la to a null geodesic congruence is rescaled according to

la ! ˛ la , where ˛ is a positive constant. How does the expansion l change? How
does it change if the null vector na used to define the 2-metric hab orthogonal to la is
simultaneously rescaled as na ! ˛ 1 na preserving the normalization lc nc D 1?
2.2. Find the outgoing and ingoing radial null geodesic congruences la and na of
the Reissner-Nordström metric and normalize them so that la na D 1.
2.3. Derive Eqs. (2.42) and (2.43).
References 53

2.4. Compute l ; n , and na ra l for the non-extremal Reissner-Nordström solution.

2.5. Verify that (2.110) is a Killing vector of Minkowski spacetime.
2.6. Use Eq. (2.78) to obtain directly Eq. (2.115).
2.7. Verify that the time and spatial translations, spatial rotations, and Lorentz
boosts given by
 a
@
.t/
a
D ;
@t
 a
@
.i/
a
D .i D 1; 2; 3/;
@xi
 a
k j @
a
L.i/ D ij x k .i; j; k D 1; 2; 3 cyclic/;
@x
 a  a
@ @
.i/
a
Dx i
Ct .i D 1; 2; 3/;
@t @xi

respectively, are Killing fields of the Minkowski metric (in fact they are all the
Killing vector fields of this metric).
2.8. Compute the Misner-Sharp-Hernandez mass of a sphere of constant areal
radius for the Reissner-Nordström metric (1.44). Does it reduce to a familiar result
in the Q ! 0 limit to Schwarzschild? Check11 that the extremal Reissner-Nordström
0
black hole corresponds to 1  2MMSH .RH / D 0.
2.9. Given the Rindler metric in spherical coordinates

ds2 D a2 R2 cos2  cos2 ' dT 2 C dR2 C R2 d˝.2/

2
;

a
compute the tangents l.˙/ to the congruences of ingoing and outgoing radial null
geodesics and their expansions .˙/ . Show that there are no apparent horizons in the
spacetime region covered by the Rindler chart.

References

1. Abreu, G., Visser, M.: Kodama time: geometrically preferred foliations of spherically symmet-
ric spacetimes. Phys. Rev. D 82, 044027 (2010)
2. Andersson, L., Mars, M., Simon, W.: Local existence of dynamical and trapping horizons.
Phys. Rev. Lett. 95, 11102 (2005)

11
Cf. Refs. [49, 73].
54 2 Horizons

3. Ashtekar, A., Corichi, A.: Laws governing isolated horizons: inclusion of dilaton couplings.
Class. Quantum Grav. 17, 1317 (2000)
4. Ashtekar, A., Galloway, G.J.: Some uniqueness results for dynamical horizons. Adv. Theor.
Math. Phys. 9, 1 (2005)
5. Ashtekar, A., Krishnan, B.: Dynamical horizons and their properties. Phys. Rev. D 68, 104030
(2003)
6. Ashtekar, A., Krishnan, B.: Isolated and dynamical horizons and their applications. Living Rev.
Relat. 7, 10 (2004)
7. Ashtekar, A., Beetle, C., Fairhurst, S.: Isolated horizons: a generalization of black hole
mechanics. Class. Quantum Grav. 16, L1 (1999)
8. Ashtekar, A., Beetle, C., Fairhurst, S.: Mechanics of isolated horizons. Class. Quantum Grav.
17, 253 (2000)
9. Ashtekar, A., Beetle, C., Lewandowski, J.: Geometry of generic isolated horizons. Class.
Quantum Gravity 19, 1195 (2002)
10. Ashtekar, A., Beetle, C., Lewandowski, J.: Mechanics of rotating isolated horizons. Phys. Rev.
D 64, 044016 (2002)
11. Ashtekar, A., Corichi, A., Krasnov, K.: Isolated horizons: the classical phase space. Adv. Theor.
Math. Phys. 3, 419 (2000)
12. Ashtekar, A., Fairhurst, S., Krishnan, B.: Isolated horizons: Hamiltonian evolution and the first
law. Phys. Rev. D 62, 104025 (2000)
13. Ashtekar, A., Beetle, C., Dreyer, O., Fairhurst, S., Krishnan, B., Lewandowski, J., Wiśnieski,
J.: Isolated horizons and their applications. Phys. Rev. Lett. 85, 3564 (2000)
14. Bak, D., Rey, S.-J.: Cosmic holography. Class. Quantum Grav. 17, L83 (2000)
15. Barceló, C., Liberati, S., Sonego, S., Visser, M.: Hawking-like radiation does not require a
trapped region. Phys. Rev. Lett. 97, 171301 (2006)
16. Baumgarte, T.W., Shapiro, S.L.: Numerical relativity and compact binaries. Phys. Rept. 376,
41 (2003)
17. Bellucci, S., Faraoni, V.: Energy conditions and classical scalar fields. Nucl. Phys. B 640, 453
(2002)
18. Bengtsson, I., Senovilla, J.M.M.: Region with trapped surfaces in spherical symmetry, its core,
and their boundaries. Phys. Rev. D 83, 044012 (2011)
19. Ben-Dov, I.: Outer trapped surfaces in Vaidya spacetimes. Phys. Rev. D 75, 064007 (2007)
20. Bergmann, P.G.: Comments on the scalar tensor theory. Int. J. Theor. Phys. 1, 25 (1968)
21. Booth, I.: Black hole boundaries. Can. J. Phys. 83, 1073 (2005)
22. Booth, I., Fairhurst, S.: The first law for slowly evolving horizons. Phys. Rev. Lett. 92, 011102
(2004)
23. Booth, I., Fairhurst, S.: Isolated, slowly evolving, and dynamical trapping horizons: Geometry
and mechanics from surface deformations. Phys. Rev. D 75, 084019 (2007)
24. Booth, I., Brits, L., Gonzalez, J.A., Van den Broeck, V.: Marginally trapped tubes and
dynamical horizons. Class. Quantum Grav. 23, 413 (2006)
25. Born, M.: Die theorie des starren elektrons in der kinematik des relativitätsprinzips. Ann.
Physik (Leipzig) 335, 1 (1909)
26. Brans, C., Dicke, R.H.: Mach’s principle and a relativistic theory of gravitation. Phys. Rev.
124, 925 (1961)
27. Carroll, S.M.: Spacetime and Geometry—An Introduction to General Relativity (Addison-
Wesley, San Francisco, 2004)
28. Chriusciel, P.T.: Uniqueness of stationary, electro-vacuum black holes revisited. Helv. Physica
Acta 69, 529 (1996)
29. Chu, T., Pfeiffer, H.P., Cohen, M.I.: Horizon dynamics of distorted rotating black holes. Phys.
Rev. D 83, 104018 (2011)
30. Clifton, T.: Properties of black hole radiation from tunnelling. Class. Quantum Grav. 25,
175022 (2008)
31. Collins, W.: Mechanics of apparent horizons. Phys. Rev. D 45, 495 (1992)
32. Corichi, A., Sudarsky, D.: When is S D A=4? Mod. Phys. Lett. A 17, 1431 (2002)
References 55

33. Csizmadia, P., Rácz, I.: Gravitational collapse and topology change in spherically symmetric
dynamical systems. Class. Quantum Grav. 27, 015001 (2010)
34. Davies, P.C.W.: Scalar production in Schwarzschild and Rindler metrics. J. Phys. A 8, 609
(1975)
35. Di Criscienzo, R., Nadalini, M., Vanzo, L., Zerbini, S., Zoccatelli, G.: On the Hawking
radiation as tunneling for a class of dynamical black holes. Phys. Lett. B 657, 107 (2007)
36. Di Criscienzo, R., Hayward, S.A., Nadalini, M., Vanzo, L., Zerbini, S.: Hamilton-Jacobi
method for dynamical horizons in different coordinate gauges. Class. Quantum Grav. 27,
015006 (2010)
37. Dyer, C.C., Honig, E.: Conformal Killing horizons. J. Math. Phys. 20, 409 (1979)
38. d’Inverno, R.: Introducing Einstein’s Relativity. Oxford University Press, Oxford (2002)
39. Eardley, D.: Black hole boundary conditions and coordinate conditions. Phys. Rev. D 57, 2299
(1998)
40. Eling, C., Guedens, R., Jacobson, T.: Nonequilibrium thermodynamics of spacetime. Phys.
Rev. Lett. 96, 121301 (2006)
41. Fairhurst, S., Krishnan, B.: Distorted black holes with charge. Int. J. Mod. Phys. D 10, 691
(2001)
42. Figueras, P., Hubeny, V.E., Rangamani, M., Ross, S.F.: Dynamical black holes and expanding
plasmas. J. High Energy Phys. 0904, 137 (2009)
43. Fodor, G., Nakamura, K., Oshiro, Y., Tomimatsu, A.: Surface gravity in dynamical spherically
symmetric space-times. Phys. Rev. D 54, 3882 (1996)
44. Fulling, S.A.: Nonuniqueness of canonical field quantization in Riemannian space-time. Phys.
Rev. D 7, 2850 (1973)
45. Haijcek, P.: Origin of Hawking radiation. Phys. Rev. D 36, 1065 (1987)
46. Hawking, S.W.: Gravitational radiation in an expanding universe. J. Math. Phys. 9, 598 (1968)
47. Hawking, S.W.: Black holes in general relativity. Commun. Math. Phys. 25, 152 (1972)
48. Hawking, S.W., G.Ellis, F.R.: The Large Scale Structure of Space-Time. Cambridge University
Press, Cambridge (1973)
49. Hayward, S.A.: Quasilocal gravitational energy. Phys. Rev. D 49, 831 (1994)
50. Hayward, S.A.: General laws of black hole dynamics. Phys. Rev. D 49, 6467 (1994)
51. Hayward, S.A.: Gravitational energy in spherical symmetry. Phys. Rev. D 53, 1938 (1996)
52. Hayward, S.A.: Unified first law of black-hole dynamics and relativistic thermodynamics.
Class. Quantum Grav. 15, 3147 (1998)
53. Hayward, S.A.: Formation and Evaporation of Nonsingular Black Holes. Phys. Rev. Lett. 96,
031103 (2006)
54. Hayward, S.A., Di Criscienzo, R., Nadalini, M., Vanzo, L., Zerbini, S.: Local Hawking
temperature for dynamical black holes. Class. Quantum Grav. 26, 062001 (2009)
55. Hernandez, W.C., Misner, C.W.: Observer time as a coordinate in relativistic spherical
hydrodynamics. Astrophys. J. 143, 452 (1966)
56. Hiscock, W.A.: Gravitational entropy of nonstationary black holes and spherical shells. Phys.
Rev. D 40, 1336 (1989)
57. Jacobson, T.: Thermodynamics of spacetime: the Einstein equation of state. Phys. Rev. Lett.
75, 1260 (1995)
58. Jang, K.-X., Feng, T., Peng, D.-T.: Hawking radiation of apparent horizon in a FRW universe
as tunneling beyond semiclassical approximation. Int. J. Theor. Phys. 48, 2112 (2009)
59. Kavanagh, W., Booth, I.: Spacetimes containing slowly evolving horizons. Phys. Rev. D 74,
044027 (2006)
60. Kodama, H.: Conserved energy flux from the spherically symmetric system and the back
reaction problem in the black hole evaporation. Progr. Theor. Phys. 63, 1217 (1980)
61. McClure, M.L., Dyer, C.C.: Asymptotically Einstein-de Sitter cosmological black holes and
the problem of energy conditions. Class. Quantum Grav. 23, 1971 (2006)
62. McClure, M.L., Anderson, K., Bardahl, K.: Cosmological versions of Vaidya’s radiating
stellar exterior, an accelerating reference frame, and Kinnersley’s photon rocket. Preprint
arXiv:0709.3288
56 2 Horizons

63. McClure, M.L., Anderson, K., Bardahl, K.: Nonisolated dynamical black holes and white
holes. Phys. Rev. D 77, 104008 (2008)
64. Misner, C.W., Sharp, D.H.: Relativistic equations for adiabatic, spherically symmetric gravita-
tional collapse. Phys. Rev. 136, 571 (1964)
65. Morris, M.S., Thorne, K.S.: Wormholes in space-time and their use for interstellar travel: a tool
for teaching general relativity. Am. J. Phys. 56, 395 (1988)
66. Mukohyama, S., Hayward, S.A.: Quasilocal first law of black hole dynamics. Class. Quantum
Grav. 17, 2153 (2000)
67. Nielsen, A.B.: Black holes without boundaries. Int. J. Mod. Phys. D 17, 2359 (2009)
68. Nielsen, A.B.: Black holes and black hole thermodynamics without event horizons. Gen. Rel.
Gravit. 41, 1539 (2009)
69. Nielsen, A.B.: The spatial relation between the event horizon and trapping horizon. Class.
Quantum Gravity 27, 245016 (2010)
70. Nielsen, A.B., Firouzjaee, J.T.: Conformally rescaled spacetimes and Hawking radiation. Gen.
Rel. Gravit. 45, 1815 (2013)
71. Nielsen, A.B., Yeom, D.-H.: Spherically symmetric trapping horizons, the Misner-Sharp mass
and black hole evaporation. Int. J. Mod. Phys. A 24, 5261 (2009)
72. Nielsen, A.B., Yoon, J.H.: Dynamical surface gravity. Class. Quantum Grav. 25, 085010 (2008)
73. Nielsen, A.B., Visser, M.: Production and decay of evolving horizons. Class. Quantum Grav.
23, 4637 (2006)
74. Nordtvedt, K.: PostNewtonian metric for a general class of scalar tensor gravitational theories
and observational consequences. Astrophys. J. 161, 1059 (1970)
75. Parikh, M.K., Wilczek, F.: Hawking radiation as tunneling. Phys. Rev. Lett. 85, 5042 (2000)
76. Penrose, R.: Gravitational collapse and spacetime singularities. Phys. Rev. Lett. 14, 57 (1965)
77. Pielahn, M., Kunstatter, G., Nielsen, A.B.: Critical analysis of dynamical surface gravity in
spherically symmetric black hole formation. Phys. Rev. D 84, 104008 (2011)
78. Poisson, E.: A Relativist’s Toolkit: The Mathematics of Black-Hole Mechanics. Cambridge
University Press, Cambridge (2004)
79. Rácz, I.: On the use of the Kodama vector field in spherically symmetric dynamical problems.
Class. Quantum Grav. 23, 115 (2006)
80. Roman, T.A., Bergmann, P.G.: Stellar collapse without singularities? Phys. Rev. D 28, 1265
(1983)
81. Scheel, M.A., Shapiro, S.L., Teukolsky, S.A.: Collapse to black holes in Brans-Dicke theory.
2. Comparison with general relativity. Phys. Rev. D 51, 4236 (1995)
82. Schnetter, E., Krishnan, B.: Non-symmetric trapped surfaces in the Schwarzschild and Vaidya
spacetimes. Phys. Rev. D 73, 021502 (2006)
83. Sorkin, R.D.: In: Wiltshire, D. (ed.) Proceedings of the First Australasian Conference on
General Relativity and Gravitation, February 1996, Adelaide, pp. 163–174. University of
Adelaide (1996). [Preprint arXiv:gr-qc/9701056]
84. Sultana, J., Dyer, C.C.: Conformal killing horizons. J. Math. Phys. 45, 4764 (2004)
85. Sultana, J., Dyer, C.C.: Cosmological black holes: A black hole in the Einstein-de Sitter
universe. Gen. Rel. Gravit. 37, 1349 (2005)
86. Szabados, L.: Quasi-local energy-momentum and angular momentum in GR: a review article.
Living Rev. Relat. 7, 4 (2004)
87. Tolman, R.C.: Non-Newtonian Mechanics. Some transformation equations. Philos. Mag. 25
(125), 150 (1912)
88. Tung, R.-S.: Stationary untrapped boundary conditions in general relativity. Class. Quantum
Grav. 25, 085005 (2008)
89. Unruh, W.G.: Notes on black-hole evaporation. Phys. Rev. D 14, 870 (1976)
90. Vanzo, L., Acquaviva, G., Di Criscienzo, R.: Tunnelling methods and Hawking’s radiation:
achievements and prospects. Class. Quantum Grav. 28, 183001 (2011)
91. Visser, M.: Dirty black holes: thermodynamics and horizon structure. Phys. Rev. D 46, 2445
(1992)
References 57

92. Visser, M.: Gravitational vacuum polarization. I. Energy conditions in the Hartle-Hawking
vacuum. Phys. Rev. D 54, 5103 (1996)
93. Visser, M.: Essential and inessential features of Hawking radiation. Int. J. Mod. Phys. D 12,
649 (2003)
94. Wagoner, R.V.: Scalar-tensor theory and gravitational waves. Phys. Rev. D 1, 3209 (1970)
95. Wald, R.M.: General Relativity. Chicago University Press, Chicago (1984)
96. Wald, R.M.: The thermodynamics of black holes. Living Rev. Relat. 4, 6 (2001)
97. Wald, R.M., Iyer, V.: Trapped surfaces in the Schwarzschild geometry and cosmic censorship.
Phys. Rev. D 44, R3719 (1991)
Chapter 3
Cosmological Horizons

Simplicity is the ultimate sophistication.

—Leonardo da Vinci

3.1 Introduction

Friedmann-Lemaître-Robertson-Walker (FLRW) spacetimes are spherically sym-

metric about every spatial point and are, therefore, trivially spherically symmetric,
but are nevertheless important. These cosmological spacetimes are much simpler
than black hole spacetimes but still contain horizons. Cosmological horizons have
been studied in inflationary scenarios of the early universe in relation to the so-
called horizon problem. In general, FLRW spaces contain time-dependent apparent
horizons expressed by particularly simple equations and are interesting from our
point of view. Due to their simplicity, it is convenient to discuss these apparent
horizons before approaching the more complicated horizons of time-dependent
black holes in the following chapters.
We begin by using an analogy between black hole and cosmological horizons
and examining various coordinate systems for FLRW spaces which mimic the
corresponding ones for Schwarzschild black holes. Several coordinate systems are
employed in the study of Hawking radiation from black hole horizons and different
coordinates, or various families of observers, provide potentially different notions
of surface gravity and horizon temperature. The underlying idea is to develop
coordinate systems useful to study Hawking radiation from cosmological horizons
and, later, from the time-dependent horizons of black holes embedded in time-
varying cosmological backgrounds. We also compute useful geometric quantities
in these various coordinate systems.

© Springer International Publishing Switzerland 2015 59

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6_3
60 3 Cosmological Horizons

3.1.1 FLRW Cosmologies

A FLRW space has line element

 
2 2 2dr2
ds D dt C a .t/ C r2 d˝.2/
2
(3.1)
1  kr2

in comoving (or “synchronous”) coordinates .t; r; ; '/, where k is the curvature

index and a.t/ is the scale factor. The comoving coordinate r is not an areal radius;
the latter is instead a function of both t and r,

R.t; r/ D a.t/r : (3.2)

It is also common to use the conformal time  defined by dt D ad, in terms of

which
 
dr2
ds2 D a2 ./ d2 C C r 2
d˝ 2
.2/ : (3.3)
1  kr2

In the spatially flat case k D 0 this line element is explicitly conformal to the
Minkowski one, however all FLRW spaces are conformally flat since their Weyl
tensor vanishes identically. The conformal flatness of the FLRW line element for
k ¤ 0 can be made explicit by transforming to suitable coordinates ([31, 48, 49] and
references therein).
In General Relativity, the Einstein equations reduce to ordinary differential
equations in a FLRW space. If the FLRW universe is sourced by a perfect fluid
with energy-momentum tensor

Tab D .P C / ua ub C Pgab ; (3.4)

where , P, and ua are the energy density, pressure, and 4-velocity field of the fluid,
respectively, one has
8 k
H2 D  ; (3.5)
3 a2
aR P C H 2 D  4 . C 3P/
DH (3.6)
a 3
(Einstein-Friedmann equations), where an overdot denotes differentiation with
respect to the comoving time t and H.t/  aP =a is the Hubble parameter. The
covariant conservation equation r b Tab D 0 then yields the energy conservation
equation

P C 3H .P C / D 0 : (3.7)

This equation is not independent of Eqs. (3.5) and (3.6) (in the same way that the
covariant conservation equation r b Tab D 0 is not independent of the Einstein
3.3 Kruskal-Szekeres Coordinates for de Sitter Space 61

equations) and can be derived from them. Another useful relation which follows
from Eqs. (3.5) and (3.6) is

P D 4 .P C / C k :
H (3.8)
a2
The expression of the Ricci curvature in terms of the Hubble parameter and its
derivative is useful:
 
P 2 k
R D 6 H C 2H C 2 : (3.9)
a

Let t D 0 denote the Big Bang spacetime singularity (in the cases in which
it is present). All comoving observers whose worldlines have ua as tangent are
physically equivalent and, therefore, the following considerations apply to any of
them, although we refer explicitly to a comoving observer located at r D 0.

3.2 Hyperspherical Coordinates for FLRW Space

The FLRW line element can be written using hyperspherical coordinates

h i
ds2 D dt2 C a2 .t/ d2 C f 2 ./d˝.2/
2
; (3.10)

where
8
< sinh  if k < 0 ;
f ./ D r D  if k D 0 ; (3.11)
:
sin  if k > 0 ;

and
Z
1 dr
Df .r/ D p : (3.12)
1  kr2

3.3 Kruskal-Szekeres Coordinates for de Sitter Space

Kruskal-Szekeres coordinates for de Sitter space were introduced by Gibbons and

Hawking [41] in their study of the thermodynamics of the de Sitter event horizon.
These .U; V/ coordinates are defined by [41]

Hr D 1  U 2 V 2 ; (3.13)
 
U
2Ht D ln  : (3.14)
V

Kruskal-Szekeres coordinates for FLRW spaces other than de Sitter are not known.
62 3 Cosmological Horizons

3.4 Painlevé-Gullstrand and Schwarzschild-Like

Coordinates for k D 0 FLRW Space

Consider the spatially flat (k D 0) FLRW metric in comoving coordinates

 
ds2 D dt2 C a2 .t/ dr2 C r2 d˝.2/
2
I

upon use of the areal radius R  a.t/r and of the relation between differentials
dR 1
dr D  Hr dt D .dR  HRdt/ ; (3.15)
a a
the metric is recast in the Painlevé-Gullstrand form1
 
ds2 D  1  H 2 R2 dt2  2HR dtdR C dR2 C R2 d˝.2/
2
: (3.16)

This form is useful for comparison with solutions of the field equations describing
black holes or central objects embedded in a spatially flat FLRW background. The
history and the advantages of Painlevé-Gullstrand coordinates in the study of black
hole horizons are discussed in Refs. [55, 59, 68] (see [32] for Painlevé-Gullstrand
coordinates in Kerr spacetime). We can eliminate the cross-term proportional to
dtdR by introducing the new time coordinate T defined by

1
dT D .dt C ˇdR/ ; (3.17)
F

where F.t; R/ is an integrating factor satisfying2

   
@ 1 @ ˇ
D (3.18)
@R F @t F

to guarantee that dT is a locally exact differential, and ˇ .t; r/ is a function to be

determined. Then, dt D FdT  ˇdR and, substituting into the line element (3.16),
one obtains
  
ds2 D  1  H 2 R2 F 2 dT 2 C ˇ 2 dR2  2FˇdTdR
 2HRdR .FdT  ˇdR/ C dR2 C R2 d˝.2/ 2

   
D  1  H 2 R2 F 2 dT 2 C 2F 1  H 2 R2 ˇ  HR dTdR
  
C 1  1  H 2 R2 ˇ 2 C 2ˇHR dR2 C R2 d˝.2/ 2
: (3.19)

1
Also called “r-gauge” (e.g., [30]).
2
As in most situations, the integrating factor is not unique.
3.5 Schwarzschild-Like Coordinates for General FLRW Spaces 63

By setting

HR
ˇ .t; R/ D (3.20)
1  H 2 R2

one obtains the FLRW metric in the Schwarzschild-like form3

  dR2
ds2 D  1  H 2 R2 F 2 dT 2 C C R2 d˝.2/
2
; (3.21)
1  H 2 R2
which is again reminiscent of the Schwarzschild line element, except for the
presence of the factor F and for the fact that the Hubble parameter H is not constant
(unless the FLRW space reduces to de Sitter space, in which case F  1). This line
element is of the form (2.94). By comparing Eqs. (3.21) and (2.94), or using directly
the definition (2.92), one obtains the Misner-Sharp-Hernandez mass

H 2 R3 4 R3
MMSH D D
P (3.22)
2 3
(which matches Eq. (3.56) below and constitutes a consistency check) and

e D F : (3.23)

Here an overdot on the equality sign denotes the fact that this equality holds in
General Relativity in a FLRW universe sourced by a perfect fluid. The cosmological
apparent horizon of a spatially flat FLRW space is now easily located by setting
gRR D 0, which yields

1
RAH D : (3.24)
H

3.5 Schwarzschild-Like Coordinates for General FLRW

Spaces

Painlevé-Gullstrand coordinates are useful in the discussion of black hole horizons

because they are regular across these horizons, while Schwarzschild-like coordi-
nates (the ones commonly used to introduce the Schwarzschild solution and black

3
The coordinate system in which the metric assumes the form (3.21) is sometimes referred to
as Nolan gauge in the literature. Nolan [69–71] studied this coordinate system in the case of
the McVittie metric [61] representing a central object embedded in a FLRW universe, eventually
restricting to k D 0 or 1. The McVittie metric reduces to that of Eq. (3.21) in the limit in which
the mass of the central object vanishes, hence (3.21) is a trivial case of the McVittie line element
in the Nolan gauge.
64 3 Cosmological Horizons

holes) are not. A similar situation occurs for cosmological horizons, although
Schwarzschild-like coordinates are not so obvious in this case (in fact, most people
are unfamiliar with them, except for the case of de Sitter and Schwarzschild-de
Sitter spaces) and usually cosmology is formulated using comoving or hyper-
spherical coordinates. In the following we derive pseudo-Painlevé-Gullstrand and
Schwarzschild-like coordinates for a general FLRW space.
Begin from the FLRW metric in comoving coordinates (3.1); using the areal
radius R.t; r/  a.t/r and Eq. (3.15), the FLRW line element (3.1) assumes the
pseudo-Painlevé-Gullstrand form
 
H 2 R2 2HR dR2
ds2 D  1  dt2  dtdR C C R2 d˝.2/
2
:
1  kR2 =a2 2
1  kR =a2 1  kR2 =a2
(3.25)
In the absence of a better nomenclature, we use the name “pseudo Painlevé-
Gullstrand coordinates” because the coefficient of dR2 is not unity, as required for
Painlevé-Gullstrand coordinates4 and the spacelike surfaces t D constant are not
flat (unless k D 0), which is instead the essential property of Painlevé-Gullstrand
coordinates [59]. By using the fact that

H 2 R2 1  R2 =R2AH
1 2
D
1  kr 1  kr2
with
1
RAH  p ; (3.26)
H2 C k=a2

one can write the line element (3.25) as [20, 51, 56]

1  R2 =R2AH 2 2HR dR2

ds2 D  dt  dtdR C C R2 d˝.2/
2
: (3.27)
1  kR2 =a2 1  kR2 =a2 1  kR2 =a2

Setting ds2 D 0 and d D d' D 0 for radial null rays with tangents pa yields
s
1 R2
p D HR ˙ H 2 R2 C 1  (3.28)
R2AH

with the choice p0 D 1 (e.g., [20]).

In Painlevé-Gullstrand coordinates, a common trick consisting of setting p1 D
dR=d D 0 to locate the apparent horizon is based on a rigorous argument and does
indeed provide the correct result (this argument holds true for general spherically

4
The literature contains ambiguous terminology for general FLRW spaces (e.g., [20]), while the
de Sitter case does not lend itself to these ambiguities [62, 72].
3.5 Schwarzschild-Like Coordinates for General FLRW Spaces 65

symmetric metrics in Painlevé-Gullstrand coordinates [68]). A posteriori, this trick

works also for quasi-Painlevé-Gullstrand coordinates in our specific situation.
To transform to the Schwarzschild-like form, we first introduce the new time
coordinate T defined by

1
dT D .dt C ˇdR/ ; (3.29)
F
where F is a (non-unique) integrating factor satisfying again
   
@ 1 @ ˇ
D (3.30)
@R F @t F

to guarantee that dT is a locally exact differential, while ˇ.t; R/ is a function to be

determined. Substituting dt D FdT  ˇdR into the line element, one obtains
 
H 2 R2
ds2 D  1 F 2 dT 2
1  kr2
 
H 2 R2 2HRˇ 1
C  1 2
ˇ2 C 2
C dR2
1  kr 1  kr 1  kr2
 
H 2 R2 2HRF
C 2 1 2
FˇdTdR  dTdR C R2 d˝.2/
2
: (3.31)
1  kr 1  kr2

By choosing

HR HR
ˇD  D 2 R2  kr 2
; (3.32)
1 H 2 R2 2
.1  kr / 1  H
1kr2

the cross-term proportional to dTdR is eliminated and we obtain the FLRW line
element in the Schwarzschild-like form
 
H 2 R2 dR2
ds2 D  1  F 2 dT 2 C C R2 d˝.2/
2
; (3.33)
1  kR2 =a2 1  kR2 =a2  H 2 R2

where F.T; R/; a, and H are implicit functions of T. By using the expression of the
Misner-Sharp-Hernandez mass in FLRW space (3.56) below, this line element can
be cast as
!
2
1  2M 2 2 dR2
ds D  R
F dT C C R2 d˝.2/
2
: (3.34)
1  2M
R
C H 2 R2 1  2M
R
66 3 Cosmological Horizons

A spherically symmetric metric can always be put in the form (2.94) [68]. By
comparing Eqs. (3.33) and (2.94), it follows that

F.T; R/
e D p
1  kR2 =a2

and

2MMSH kR2
1 D 1  2  H 2 R2 (3.35)
R a
(this equation is consistent with the expression (3.56) of the Misner-Sharp-
Hernandez mass in FLRW space that we are going to discuss soon).
It is now easy to locate the apparent horizon of a general FLRW space using
the prescription (2.104). Setting gRR jAH D 0 and reading gRR D g1
RR from Eq. (3.33)
yield the radius of the FLRW apparent horizon

1
RAH D p : (3.36)
H2 C k=a2

The components of the Kodama vector in Schwarzschild-like coordinates are given

by Eq. (2.76), which yields
p !
1  kR2 =a2
K D  ; 0; 0; 0 (3.37)
F

and its norm squared is

   
kR2 R2
Kc K c D  1  H 2 R2  2 D  1  2 : (3.38)
a RAH

The Kodama vector is timelike (Kc K c < 0) if R < RAH , null if R D RAH ,
and spacelike (Kc K c > 0) outside the apparent horizon R > RAH . The Kodama
vector produces the Misner-Sharp-Hernandez mass as a Noether charge [45].
Equations (3.35) and the Hamiltonian constraint (3.5) then imply that

4 R3
MMSH .R/ D : (3.39)
3
It is often useful to know the components of the Kodama vector in pseudo-
Painlevé-Gullstrand coordinates recurrent in the literature (e.g., [20]) and in comov-
ing coordinates, which we report here (see Appendix A.2 for details):
 p 
K D  1  kR2 =a2 ; 0; 0; 0 (3.40)
3.6 Painlevé-Gullstrand Coordinates for General FLRW Spaces 67

in pseudo-Painlevé-Gullstrand coordinates, and

 p p 
K D  1  kr2 ; Hr 1  kr2 ; 0; 0 (3.41)

in comoving coordinates .t; r; ; '/. The norm squared of the Kodama vector is
 
K c Kc D  1  kr2  aP 2 r2 D  .1  2M=R/ : (3.42)

3.6 Painlevé-Gullstrand Coordinates for General FLRW

Spaces

To find Painlevé-Gullstrand coordinates for general (i.e., not necessarily spatially

flat) FLRW spacetimes, begin from the FLRW line element in Schwarzschild-like
coordinates (3.33). We search for a new time coordinate TQ .T; R/, with

@TQ @TQ PQ C TQ 0 dR :
dTQ D dT C dR  TdT (3.43)
@T @R
By substituting

1 Q TQ 0
dT D dT  dR (3.44)
TPQ TPQ
one obtains
  !2
2 H 2 R2 2 dTQ TQ 0 dR2
ds D  1  F  dR C
1  kR2 =a2 TQP TQP 1  H R2  kR2 =a2
2

CR2 d˝.2/
2

   2
H 2 R2 F
D  1 dTQ 2
1  kR2 =a2 PTQ
2 !2 3
 2 2  Q0
H R F T 1
C 4 1  C 5 dR2
1  kR2 =a2 TPQ 1  H 2 R2  kR2 =a2
 
2F 2 TQ 0 H 2 R2 Q C R2 d˝.2/
2
C 1 dTdR : (3.45)
TQP 2 1  kR2 =a2

By imposing that the coefficient of dR2 be unity one obtains

r
PTR   
Q H 2 C ak2 1  kR 2
2
a
TQ 0 D ˙  2
 (3.46)
F 1  H 2 R2  kR
a2
68 3 Cosmological Horizons

which, substituted into the previous equation, yields the FLRW line element in
Painlevé-Gullstrand coordinates
   2 s
2 2
H R F Q2 ˙ 2FR H 2 C k=a2 Q
ds2 D  1  d T dTdR C dR2
1  kR2 =a2 TPQ TPQ 1  kR2 =a2

CR2 d˝.2/
2
: (3.47)

Clearly, slicings of constant time TQ are flat.

3.7 Congruences of Radial Null Geodesics in FLRW Space

Having given the general definitions and having calculated the needed formulae, let
us move to the study of congruences of radial null geodesics and of cosmological
horizons. In FLRW space, which is spherically symmetric about every point of
space, the outgoing and ingoing radial null geodesics have tangent fields with
comoving components
p ! p !
1  kr2 1  kr2
l D 1; ; 0; 0 ; n D 1;  ; 0; 0 ; (3.48)
a.t/ a.t/

respectively.
Proof. Setting pc pc D 0 for the 4-tangents pa yields

a2
0 D pc pc D .p0 /2 C .p1 /2
1  kr2
and
p
1 1  kr2 0
p D˙ p :
a

Of course, one can also set ds2 D 0 and d D d' D 0, obtaining directly
p
dr dr=d p1 1  kr2
D D D˙ ; (3.49)
dt dt=d p0 a

where is a parameter along the null rays. t

u
There is freedom to rescale a future-directed null vector by an arbitrary regular
function (which must be positive if we want to keep this vector future-oriented).
The choice of normalization in Eq. (3.48) implies that lc nc D 2. The
pmore common
normalization lc nc D 1 is obtained by dividing both la and na by 2.
3.7 Congruences of Radial Null Geodesics in FLRW Space 69

The expansions of these null geodesic congruences are computed using

Eq. (2.27). One first computes
  "
2 1=2
#
1 p  1  kr 3a2 2
P
a r
rc l c D p @ g l D C 2a2 r
g a3 r2 .1  kr2 /1=2
2 p
D 3H C 1  kr2 ; (3.50)
ar
 1=2 " #
1 p  1  kr2 3a2 aP r2 2
rc n D p
c
@ g n D  2a r
g a3 r2 .1  kr2 /1=2
2 p
D 3H  1  kr2 : (3.51)
ar
Then, using

p a3 r2 sin2 
g D p ;
1  kr2
krı c1 C aPa ı c0
00c D 0 ; 01c D 10c D Hı c1 ; 11c D ;
1  kr2
gcd lc nd D 2 ;

Eq. (2.27) yields5

 p  0 s 1
2 aP r C 1  kr2 1 kR2 A
l D D 2 @H C 1 2 ; (3.52)
ar R a
 p  0 s 1
2 aP r  1  kr2 1 kR2 A
n D D 2 @H  1 2 : (3.53)
ar R a

The cosmological apparent horizon is located where n D 0 and l > 0, i.e., at

1
rAH D p ; (3.54)
2
aP C k
or
1
RAH .t/ D p (3.55)
H2 C k=a2

5
See, e.g., Ref. [36]. The factor 2 in Eqs. (3.52) and (3.53) does not appear in Ref. [5] and in
other works because of different normalizations of la and na . The expansions of the congruences
definitely depend on the choice made for lc nc .
70 3 Cosmological Horizons

in terms of the proper radius R  ar. Note that the apparent horizon is defined using
only null geodesic congruences and their expansions and there is no reference to the
global causal structure.
Sometimes it is tempting to locate apparent horizons by simply guessing “where
the outgoing radial null rays stop”, that is, by setting lr D 0. Although this shortcut
may sometimes provide the correct result (it does work for spherically symmetric
metrics in Painlevé-Gullstrand coordinates [68]), in general, it is not to be adopted
in place of the proper procedure which consists of finding the surfaces on which
l D 0 and n ¤ 0. The radial null geodesic congruences in FLRW space offer a
counterexample: using Eq. (3.48) and setting n1 D 0 would lead to the incorrect
conclusion that there are no apparent or trapping horizons in k D 0; 1 FLRW
spaces, and to the incorrect value of RAH for k D C1 FLRW space. This is obviously
incorrect: except for the Minkowski case H D 0, apparent horizons always exist and
are given by Eq. (3.55).

3.8 Horizons in FLRW Space

Two horizons of FLRW space are familiar from standard cosmology textbooks:
they are the particle horizon and the event horizon [75]. In addition, apparent and
trapping horizons are relevant for our discussion. Consider a FLRW universe with
line element (3.1) in comoving coordinates. The proper, or areal, radius is R  a.t/r
and the Misner-Sharp-Hernandez mass of a sphere of radius R, defined by Eq. (2.92),
is easily found to be
 
k R3 4 R3
MMSH D H 2 C 2 D
P ; (3.56)
a 2 3

where, again, a dot on the last equality denotes the fact that it holds in General
Relativity in a universe sourced by a perfect fluid (note that Eq. (3.56) is valid for
any value of the curvature index k). In non-spatially flat FLRW spaces, k ¤ 0, the
quantity 4 R3 =3 is not the proper volume of a sphere of radius R, which is instead
Z 2 Z Z r q
Vproper D d' d dr0 g.3/ ; (3.57)
0 0 0

a6 r4 sin2 
where g.3/ D is the determinant of the restriction of the metric gab to the
1  kr2
3-surfaces t D constant. Therefore,
Z Z 
3
r
dr0 r02 3
Vproper D 4 a .t/ p D 4 a .t/ d0 f 2 .0 / ; (3.58)
0 1  kr02 0
3.8 Horizons in FLRW Space 71

where  is the hyperspherical coordinate and the function f ./ is given by

Eq. (3.11). However, it turns out that only the “areal volume”

4 R3
V (3.59)
3
will be needed for our purposes, as a consequence of the use of the Misner-Sharp-
Hernandez mass which is usually identified with the internal energy U that appears
in the thermodynamics of the apparent horizon.
Proof. The FLRW line element can be rewritten in Schwarzschild-like coordinates
as (Sect. 3.5)
 
H 2 R2 dR2
ds2 D  1  F 2 dt2 C C R2 d˝.2/
2
;
1  kR2 =a2 1  kR2 =a2  H 2 R2

where F.T; R/ is an integrating factor. Equation (2.92) then gives

2MMSH kR2
1 D gRR D 1  H 2 R2  2
R a
and
 
2 k R3 4 R3
MMSH D H C 2 D
P ; (3.60)
a 2 3

using the Hamiltonian constraint (3.5) valid in General Relativity with a perfect
fluid. t
u
The Misner-Sharp-Hernandez mass of a sphere of radius R does not depend
explicitly on the pressure P of the cosmic fluid. Its time derivative, instead, does:
consider a sphere of proper radius R D Rs .t/, then, using R  a.t/r and Eq. (3.7),
one has
 
P MSH  dMMSH D 4 3R2s RP s C R3s P
M
dt 3
4 ˚
D 3R2s RP s C R3s Œ3H.P C /
3

D 4 R2s RP s  HR3s .P C /
" #
3 Rs
P
D 4 Rs  H .P C / :
Rs

If the sphere is comoving with radius Rs / a.t/ then RP s =Rs D H and

P MSH D 4 HR3s P I
M (3.61)
72 3 Cosmological Horizons

in this case MP MSH depends explicitly on P but not on . By taking the ratio of
Eqs. (3.61) and (3.56) one also obtains that, in General Relativity,

P MSH C 3H P MMSH D 0
M .comoving sphere/: (3.62)

3.8.1 Particle Horizon

The particle horizon [75] at time t is a sphere centered on the comoving observer at
r D 0 and with radius
Z t
dt0
RPH .t/ D a.t/ 0
: (3.63)
0 a.t /

The particle horizon contains every particle signal that has reached the observer
between the time of the Big Bang t D 0 and the time t.6
For the fastest particles which travel radially to an observer at light speed, it is
ds2 D 0 and d˝.2/2
D 0. We write the line element using hyperspherical coordinates
as in Eq. (3.10). Then, along ingoing radial null geodesics, it is d D dt=a and
the infinitesimal proper radius is a.t/d. Integrating along an ingoing radial null
geodesic between the emission of the signal at e at time te and its detection at
 D 0 at time t, one obtains
Z 0 Z t
dt0
d D  (3.64)
e te a.t0 /

and the use of

Z e Z 0
e D d D  d
0 e

leads to
Z t
dt0
e D : (3.65)
te a.t0 /

6
More realistically, photons propagate freely in the universe only after the time of the last
scattering or recombination, before which the Compton scattering due to free electrons in the
cosmic
Z t plasma makes it opaque. Therefore, cosmologists introduce the optical horizon with radius
dt0
a.t/ 0
[66]. However, the optical horizon is irrelevant for our purposes and will not
trecombination a.t /
be used here.
3.8 Horizons in FLRW Space 73

To obtain a physical (proper) radius R one multiplies by the scale factor obtaining7
Z t
dt0
Re D a.t/ ; (3.66)
te a.t0 /

Now take the limit te ! 0C :

Z t
dt0
• If the integral 0
diverges, it is possible for the observer at r D 0 to receive
0 a.t /
all the light signals emitted at sufficiently early times from any point in the
universe. The maximal volume that can be causally connected to the observer
Z t
at time t is infinite.
dt0
• If the integral 0
is finite, the observer at r D 0 receives, at time t,
0 a.t / Z t
dt0
only the light signals started within the sphere r  0
, which is called the
0 a.t /
(comoving) radius of the particle horizon.
The physical (proper) radius of the particle horizon is therefore given by
Eq. (3.63). At a given time t the particle horizon is the boundary between the
worldlines that can be seen by the observer and those (“beyond the horizon”) which
cannot be seen. This boundary hides events which cannot be known by that observer
at time t. The particle horizon evolves with time (see Fig. 3.1).
The particle horizon is the horizon commonly studied in inflationary cosmology.
The horizon problem of standard Big Bang cosmology consists of the fact that
cosmic microwave photons coming from two opposite directions in the sky have the
same temperature T ' 2:73 K (apart from small fluctuations ıT=T  5  105 ) even
though they come from regions which have never been in causal contact and cannot
have thermalized. In the context of standard Big Bang theory this property can only
be explained by invoking miracolously fine-tuned initial conditions. Inflation in the
early universe provides a solution of the horizon problem by postulating that a small
region of the universe has inflated at speeds much larger than those possible in the
standard Big Bang model and that the cosmic microwave background photons that
we see today were in causal contact in this small region before inflation occurred
[42, 54, 57, 66].
Note that the particle and event horizons depend on the observer. Contrary to the
case of the event horizon associated with a Schwarzschild black hole (often called
an absolute horizon), different comoving observers in FLRW space will see event
horizons located at different places (e.g., [26]). Another difference with respect to
a black hole horizon is that the observer is located inside the cosmological particle
horizon and signals sent from the outside cannot reach him or her. In this sense, the

7
The notation for the proper radius R  a.t/r D a.t/f ./ is consistent with our previous use of
this symbol to denote an areal radius because a.t/r is in fact an areal radius, as is obvious from the
inspection of the FLRW line element (3.10). If k ¤ 0, the proper radius a.t/ and the areal radius
a.t/f ./ do not coincide.
74 3 Cosmological Horizons

Fig. 3.1 The particle horizon of a comoving FLRW observer at time t. As the comoving observer
moves along his or her (vertical) worldline, the particle horizon becomes larger and larger and
this observer is able to see more and more signals coming from further and further away. Other
observers (represented by the curved wordlines) will have different particle horizons

cosmological horizon of an expanding FLRW space resembles more a white hole

than a black hole event horizon.
The cosmological particle horizon is a null surface.
Proof. This statement should be obvious from the fact that the event horizon is a
causal boundary and is generated by the null geodesics which barely fail to reach
the observer, but let us check it explicitly anyway. Using hyperspherical coordinates
.t; /, the equation of the particle horizon is
Z t
dt0
F .t; /    D 0: (3.67)
0 a.t0 /

The normal to this surface has components

ı 0
N D r F jPH D ı 1  (3.68)
a
and the norm squared of this normal is
  
 ı 0 ı0
N a Na D gab Na Nb D g  Cı 1  C ı1
a a
g00 1 1
D 2
C g11 D  2 C 2 D 0 ;
a a a
hence the particle horizon is a null surface. t
u
3.8 Horizons in FLRW Space 75

The cosmological particle horizon evolves according to the equation [24, 35]

RP PH D HRPH C 1 : (3.69)

Proof. Equation (3.63) gives

Z Z
t
dt0 a aP t
dt0
RP PH D aP 0
C D a.t/ C 1 D HRPH C 1 :
0 a.t / a a 0 a.t0 /

t
u
In an expanding universe with a particle horizon it is RP PH > 0, which means
that more and more signals emitted between the Big Bang and the time t reach the
observer as the time t progresses. If RPH .t/ does not diverge as t ! tmax (where tmax
is possibly C1), then there will always be a region unaccessible to the comoving
observers.
The acceleration of the particle horizon is

 
RR PH D HP C H 2 RPH C H D aR RPH C H D
P 
4
. C 3P/ RPH C H ; (3.70)
a 3
as follows from Eqs. (3.69) and (3.6).

3.8.2 Event Horizon

Consider now all the events which can be seen by a comoving observer at r D 0
between time t and future infinity t D C1 (in a closed universe which recollapses,
or in a Big Rip universe which ends at a finite time, substitute C1 with the time
tmax corresponding to the maximal expansion or the Big Rip, respectively). The
comoving radius of the region which can be seen by this observer is
Z C1
dt0
EH D I (3.71)
t a.t0 /

if this integral diverges as the upper limit of integration goes to infinity or to tmax , it
is said that there is no event horizon in this FLRW space and events arbitrarily far
away can eventually be seen by the observer by waiting a sufficiently long time. If
the integral converges, there is an event horizon: events beyond rEH will never be
known to the observer [75]. The physical (proper) radius of the event horizon is
Z C1
dt0
REH .t/ D a.t/ : (3.72)
t a.t0 /
76 3 Cosmological Horizons

In short, the event horizon can be said to be the “complement”of the particle horizon
[66]; it is the (proper) distance to the most distant event that the comoving observer
will ever see. Clearly, in order to define the event horizon one must know the entire
future history of the universe from time t to infinity and the event horizon is defined
globally, not locally.

The cosmological event horizon is a null surface.

Proof. Again, the statement follows from the fact that the event horizon is a causal
boundary. To check explicitly, the equation of the particle horizon in hyperspherical
coordinates .t; / is
Z tmax
dt0
F .t; /    D 0: (3.73)
t a.t0 /

The normal to this surface has components

ı 0
N D r F jEH D ı 1  (3.74)
a
and the norm squared of this normal is
  
 ı 0 ı0
N a Na D gab Na Nb D g ı 1 ı1 
a a
g00 1 1
D g11 C D 2  2 D 0: (3.75)
a2 a a
t
u
The cosmological event horizon evolves according to the equation [1, 24, 35, 64]

RP EH D HREH  1 : (3.76)

Proof. Equation (3.72) gives

Z  Z t  Z tmax
tmax
dt0 d dt0 dt0 a
RP EH D aP C a.t/  D Ha.t/ 
t a.t0 / dt tmax a.t 0/
t a.t 0/ a
D HREH  1 :

t
u
The acceleration of the event horizon is also straightforward to derive,
 
RR EH D HP C H 2 REH  H : (3.77)

The event horizon does not exist in every FLRW space.

3.8 Horizons in FLRW Space 77

To wit, consider a spatially flat (k D 0) FLRW universe sourced by a perfect

fluid with equation of state P D w and w D constant > 1. If w > 1=3 (i.e., in
General Relativity, for a decelerating universe8 ), there is no event horizon because
2
a.t/ D a0 t 3.wC1/ (3.78)

and the event horizon radius would be

Z C1 C1
2 dt0 2 3.w C 1/ 0 3.wC1/
3wC1
REH D a0 t 3.wC1/
2
D t 3.wC1/ t :
t a0 .t0 / 3.wC1/ 3w C 1 t

3w C 1
If w > 1=3 then the exponent is positive and the integral diverges: there
3.w C 1/
is no event horizon in this case. Indeed, the existence of cosmological event horizons
seems to require the violation of the strong energy condition in at least some region
of spacetime [10]. We can state that

in General Relativity with a perfect fluid the FLRW event horizon exists only for
accelerated universes with P <  =3.

Since conformal time  is defined by dt D ad, it is

REH ./ D a./ .max  / :

For a spatially flat universe with perfect fluid, P D w and w D constant ¤ 1 in

General Relativity, using Eq. (3.78) one obtains

3.w C 1/ 3.wC1/
3wC1
D t : (3.79)
.3w C 1/a0

If 1 < w < 1=3, then  is negative for t > 0 while, if w < 1 (phantom universe)
or if w > 1=3 (decelerating universe), it is  > 0 for t > 0.
For w D 1 we have de Sitter space with scale factor a.t/ D a0 eHt (a0 ; H
constants) and dt  ad D a0 eHt d, which yields
Z Z
eHt 1 eHt
D d D dt D D < 0:
a0 aH a0 H

Then, t ! 1 corresponds to  ! 1, t ! C1 to  ! 0 , and we have t D

H 1 ln .a0 Hjj/.

8
The Einstein-Friedmann equation (3.6) gives aR < 0 in this case.
78 3 Cosmological Horizons

3.8.3 Hubble Horizon

For completeness, we mention that the literature refers often to a Hubble horizon
with radius
1
RH  : (3.80)
H
This is just language: this equation is not derived from any particular physical
consideration, other than giving an order of magnitude of the radius of curvature
of a FLRW space and being used as an estimate of the radius of the event horizon
during slow-roll inflation, when the universe is close to a de Sitter space [54]. The
Hubble horizon coincides with the apparent horizon for spatially flat universes and
with the horizon of de Sitter space. However, the concept of Hubble horizon does
not add to the discussion of the physics of the various types of FLRW horizons and
it will not be used in the following.

3.8.4 Apparent Horizon

According to the definition (2.57)–(2.58), the apparent horizon of FLRW space

is located by setting to zero the expansion n given by Eq. (3.53) while l (given
by Eq. (3.52)) is positive. Alternatively, note that the FLRW metric is obviously
spherically symmetric and one can apply the prescription (2.104) to locate the
apparent horizon deriving from the Misner-Sharp-Hernandez mass, as we have
already done. Using the areal radius R, the FLRW line element reads

a2 .t/
ds2 D dt2 C dr2 C R2 d˝.2/
2
(3.81)
1  kr2
where r.t; R/ D R=a.t/, which is of the form (2.91) with
 
a2 .t/
hab D diag 1; ; 0; 0 (3.82)
1  kr2

in comoving coordinates. The apparent horizon is located by the equation

r c Rrc R D 0, or
 
1  kr2
 R2t C R2r D 0 : (3.83)
a2

Since Rt D aP .t/r and Rr D a.t/, we obtain aP 2 r2 D 1  kr2 or H 2 R2 D 1  kR2 =a2 ,

and the FLRW apparent horizon of a comoving observer is a sphere of proper radius

1
RAH .t/ D p (3.84)
H2 C k=a2
3.8 Horizons in FLRW Space 79

centered on this observer. Looking at Eqs. (3.52) and (3.53) it is clear that when
R > RAH it is l > 0 and n > 0, while the region 0  R < RAH has l > 0
and n < 0 (radial null rays coming from the region outside the horizon will not
cross it and reach the observer). The cosmological apparent horizon depends on the
observer (here chosen at r D 0), much like horizons in flat space. It acts as a sphere
surrounding the observer and hiding information.
For a spatially flat universe, the radius of the apparent horizon RAH coincides
with the Hubble radius H 1 while, for a positively curved (k > 0) universe, RAH
is smaller than the Hubble radius, and it is larger for an open (k < 0) universe.
Note that, in General Relativity, the Hamiltonian constraint (3.5) guarantees that the
argument of the square root in Eq. (3.84) is positive for positive densities . The
apparent horizon always exists (except for the trivial case of Minkowski spacetime
without gravity) while, as seen above, the event horizon does not exist in all FLRW
spaces.

The apparent horizon, in general, is not a null surface, contrary to the event and
particle horizons.

To prove this statement, note that the equation of the apparent horizon in
comoving coordinates is

1
F .t; r/ D a.t/r  p D 0: (3.85)
H C k=a2
2

The normal to this surface has components

" #
Pı
HH 0  kPa
ı 0
a3
N D r F jAH D aP r ı 0 C aı 1 C 3=2
.H 2 C 2
k=a / AH
("   # )
H HP  k=a2
D aP r C ı 0 C aı 1
.H 2 C k=a2 /3=2 AH
"   #
H H P  k=a2
D HRAH C ı 0 C aı 1
.H 2 C k=a2 /3=2
 
D HRAH 1 C H P  k R2AH ı 0 C aı 1
a2
aR
D HR3AH ı 0 C aı 1 : (3.86)
a
In General Relativity with a perfect fluid with equation of state P D w ,
Eqs. (3.5), (3.6), and (3.84) yield

.3w C 1/
N D
P  HRAH ı 0 C aı 1 : (3.87)
2
80 3 Cosmological Horizons

The norm squared of the normal is

N a Na D gab Na Nb
"   ! #
H HP  k=a2

Dg HRAH C ı 0 C aı 1 
.H 2 C k=a2 /3=2
"   ! #
H H P  k=a2
 HRAH C ı0 C a ı1
.H 2 C k=a2 /3=2
"   #2
n H H P  k=a2

Dg HRAH C ı 0 ı0
.H 2 C k=a2 /3=2
o
C    ı 1 ı0 C    ı 0 ı1 C a2 ı 1 ı1
"   #2
HP  k=a2
00 2
Dg H RAH C C a2 g11
.H 2 C k=a2 /3=2
"   #2
HP  k=a2
2 2
D 1 krAH H RAH C
.H 2 C k=a2 /3=2
"   #2
1 HP  k=a2
2
D 1  krAH  H2 p C
H 2 C k=a2 .H 2 C k=a2 /3=2
 
H2 H P C H2 2 H 2 .Ra=a/2
2 2
D 1  krAH  D 1  kr   3
.H 2 C k=a2 /3
AH
H 6 1 C a2kH 2
"  2 #
2 2 aR 4
D H RAH 1  RAH
a

3H 2 R2AH
D
P . C P/ .  3P/
4 2
 
D H 2 R2AH 1  q2 H 4 R4AH ; (3.88)

where q  Raa=Pa2 is the deceleration parameter. The horizon is null if and only if
P D  or if P D =3.
In General Relativity with a perfect fluid the Hamiltonian constraint (3.5) yields
 1  1
k 8
H 2 R2AH D 1 C  ˝ 1 ;
c
D  (3.89)
a2 H 2 3H 2

3H 2
where c  is the critical density and ˝  = c is the density parameter, i.e.,
8
the energy density measured in units of the critical density, and one obtains
3.8 Horizons in FLRW Space 81

3 ˝ 2  q2
N a Na D
P  .w C 1/ .3w  1/ H 2 R2AH D
P : (3.90)
4 ˝3
We can use Eq. (3.90) to establish that [9, 12, 35, 77]:
• If 1 < w < 1=3 then N c Nc > 0 and the apparent horizon is timelike. For a k D 0
P < 0.
universe in Einstein theory this condition corresponds to H
• If w D 1 or w D 1=3 then N c Nc D 0 and the apparent horizon is null (de Sitter
space, which has HP D 0 and q D 1, falls into this category but it is not the only
space with this horizon property).
• If w < 1 or w > 1=3 then N c Nc < 0, the normal is timelike, and the apparent
horizon is spacelike. In Einstein theory with k D 0 and a perfect fluid as the
source, w < 1 corresponds to H P > 0 (“superacceleration”). This is the case
of Big Rip universes and of a phantom fluid which violates the weak energy
condition.
The black hole dynamical horizons discussed in the literature are usually required
to be spacelike [4]. However, cosmological horizons in the presence of non-exotic
matter are timelike.
In General Relativity, the radius of the apparent horizon can be written as

1
RAH D p (3.91)
˝ jHj

in terms of the density parameter ˝ by using Eq. (3.89).

The apparent horizon evolves according to the equation [2, 19, 21, 24, 35, 60]
 
PRAH D HR3AH k  H P D P 4 HR3AH .P C / : (3.92)
a2

Proof. Differentiate Eq. (3.84) with respect to t, obtaining

"   #
H HP  k=a2
RP AH D 
.H 2 C k=a2 /3=2 AH
 
P k
D H H C 2 R3AH
a
 
1 P
D HR3AH  H 2
 H
R2AH
  
D HRAH 1  H P C H 2 R2AH
   
P 2AH  H2 3 k P
D HRAH 1  HR D HR  H :
H 2 C k=a2 AH
a2

t
u
82 3 Cosmological Horizons

In General Relativity with a perfect fluid as a source, the only way to obtain
a stationary apparent horizon is when P D  . This equation of state yields de
Sitter space, for which Eq. (3.92) reduces to RP AH D 0, consistent with RH D H 1
and H D constant. For non-spatially flat universes, the equation of state P D 
produces other solutions.
Example 3.1. For k D 1 and a cosmological constant  > 0 as the only source of
gravity, the scale factor
r r !
3 
a.t/ D sinh t (3.93)
 3

is a solution of the Einstein-Friedmann equations, as can be checked easily. The

radius of the event horizon has the time dependence
r r !ˇ " r !#ˇ
3  ˇ  t ˇ
ˇ ˇ
REH .t/ D sinh t ˇln tanh ˇ: (3.94)
 3 ˇ 3 2 ˇ
p
The apparent horizon, instead, has constant radius RAH D 3=, according to the
fact that  C P D 0 in Eq. (3.92) [35].
Example 3.2. Consider, for k D C1 and positive cosmological constant , the scale
factor
r r !
3 
a.t/ D cosh t I (3.95)
 3

the event horizon has radius

r r !" r !!#
3  1 
REH .t/ D cosh t C n  tan sinh t ; (3.96)
 3 2 3

where n D 0; ˙1; ˙2; : : : The multiple possible values of n correspond to the

infinite possible branches which one can consider when inverting the tangent
function, and to the fact that in a closed universe light rays can travel multiple times
around the universe. In this situation it is problematic to regard the event horizon
p
as a true horizon [26]. The apparent horizon has constant radius RAH D 3=,
according to the fact that  C P D 0 in Eq. (3.92).9

9
These two examples, together with de Sitter space for k D 0, are presented in Ref. [26]. However,
contrary to what is stated in this reference, in both cases the event horizon is not constant: it is the
apparent horizon which is constant.
3.8 Horizons in FLRW Space 83

3.8.5 Trapping Horizon

When is the FLRW apparent horizon also a trapping horizon? When Ll n > 0,
which gives

R
Ll n D > 0; (3.97)
3
where R is the Ricci scalar of FLRW space [8, 9, 12, 35, 77].
Proof. Using Eqs. (3.48) and (3.53), we have

Ll n D la ra n D la @a n
p ! p !
1  kr2 1  kr2
D 2 @t C @r H
a ar
0 s 1
2 2
D 2 @HR P 2 C HR 1  kR C 1A :
R a2

At the apparent horizon R D RAH it is

2 q 3
  P H 1  a2 .H 2 Ck=a
k
k 6 H 2/
7
Ll n jAH D 2 H 2 C 4 2 C p C 15
a2 H C k
a2
2
H C k=a 2

 
D2 HP C 2H 2 C k D R :
a2 3

t
u
In General Relativity with a perfect fluid it is

8
Ll n jAH D
P .  3P/ (3.98)
3
and, therefore [8, 9, 12, 35, 77]

the apparent horizon is also a trapping horizon iff R > 0 (equivalent to P < =3
in General Relativity with a perfect fluid).

When trapping, the apparent horizon is a past inner trapping horizon according to
Hayward’s definition [45]: it is an inner horizon because the region 0 < R < RAH .t/
is not trapped (i.e., l n < 0), and a past horizon because geodesics which have
exited the hypersurface R D RAH cannot come back to it.
84 3 Cosmological Horizons

3.8.6 Examples

Let us consider a couple of examples.

Example 3.3. As a first example consider Minkowski space with line element

ds2 D dt2 C dR2 C R2 d˝.2/

2
W (3.99)

this is a trivial k D 0 FLRW universe with constant scale factor a  1. The

expression (3.84) of the radius of the apparent horizon gives RAH D 1, i.e., the
apparent horizon is located at spatial infinity.
Example 3.4. As a second example consider the Milne universe with line element
 
ds2 D d 2 C  2 d2 C sinh2  d˝.2/
2
; (3.100)

which describes a k D 1 FLRW universe in hyperspherical coordinates with

linear scale factor a. / D . The areal radius is R D  sinh  and its gradient has
components

r R D sinh  ı 0 C  cosh  ı 1 ;

which gives

r c Rrc R D 1 I (3.101)

since rc R cannot be null there is no apparent horizon. This fact is consistent with
Eq. (3.84) which yields RAH D C1. This fact is not surprising because the Milne
universe is nothing but a portion of empty Minkowski space. In fact, for a D  and
k D 1, the Einstein-Friedmann equations with a perfect fluid (3.5) and (3.6) give
D P D 0: this FLRW universe is empty and must therefore have zero spacetime
curvature and be Minkowski space, which is confirmed by a direct calculation of
the Riemann tensor. The Milne line element can be obtained from the Minkowski
line element (3.99) with the coordinate transformation .t; R; ; '/ ! .; ; ; '/
given by

t D  cosh  ; (3.102)
R D  sinh  ; (3.103)

as is straightforward to verify. The Milne universe is a foliation of flat Minkowski

spacetime with curved three-dimensional hypersurfaces. Although, in general, the
existence of apparent horizons is foliation-dependent, in this case changing the
foliation does not introduce an apparent horizon in Minkowski spacetime. The
dependence on the spacetime slicing does not always lead to trouble.
3.9 Dynamics of Cosmological Horizons 85

3.9 Dynamics of Cosmological Horizons

Let us now focus on the dynamical evolution of the horizons that we have defined
and compare their evolutionary laws, which we summarize here. The first question
to ask is whether these horizons are comoving: they almost never are. The difference
between the expansion rate R=RP of a horizon of radius R and the expansion rate H
of the cosmic matter is, for the particle, event, and apparent horizons

RP PH 1
H D ; (3.104)
RPH RPH
RP EH 1
H D ; (3.105)
REH REH
" k  #  
RP AH P
H aR .3w C 1/ H
a2
H DH  1 D  H R2AH D
P ; (3.106)
RAH 2
H C a2 k a 2

respectively. Taking into consideration only expanding FLRW universes (H > 0),
when it exists the particle horizon always expands faster than comoving. The event
horizon (which only exists for accelerated universes) always expands slower than
comoving. The apparent horizon expands faster than comoving for decelerated
universes (Ra < 0); slower than comoving for accelerated universes (Ra > 0); and
comoving for coasting universes (a.t/ / t).
An even simpler way of looking at the evolution is by using the comoving radius
of the horizon: if this radius is constant, then the horizon is comoving. We have,

1
rPPH D > 0; (3.107)
a
1
rPEH D  < 0 ; (3.108)
a
aP aR
rPAH D  ; (3.109)
.Pa2 C k/3=2

respectively. All these horizons are practically never comoving.

In a universe which changes its expansion from accelerated to decelerated, the
event horizon suddenly disappears when aR .t/ changes sign. Vice-versa, an event
horizon appears when the universe goes from decelerated to accelerated. If an event
horizon was to be assigned an entropy equal to one quarter of its area (which
is not an established fact), then this behaviour causes discontinuous jumps in the
horizon entropy. In a de Sitter universe the horizon is, instead, static and so is the
entropy.
86 3 Cosmological Horizons

3.10 Another Notation

The notation and terminology may become confusing when switching from black
hole to cosmological horizons and from observers outside an horizon to observers
surrounded by a horizon. It is perhaps more convenient to adopt another common
notation: instead of using la and na for the tangents to the outgoing and ingoing null
geodesic congruences in a black hole spacetime and having to invert these when
a
discussing white hole or expanding FLRW spacetimes, one can simply use l.˙/
where “C” and “” denote outgoing or ingoing congruences as appropriate. The
terminology is as follows.10
• For a black hole C denotes the outgoing null congruence (with tangent la in our
previous notation) and  denotes the ingoing null congruence (with tangent na in
the other notation). The observer is outside the apparent horizon.
– A surface is normal (untrapped) if C >0 and  < 0.
– A surface is trapped if C < 0 and  < 0.
– A surface is (future) marginally trapped if C D 0 and  < 0 and is outer if
@ C < 0.
• For a white hole C denotes the outgoing and  the ingoing null congruences.
The observer is outside the apparent horizon.
– A surface is normal (untrapped) if C > 0 and  < 0.
– A surface is trapped if C > 0 and  > 0.
– A surface is (past) marginally trapped if C > 0 and  D 0 and is outer if
@C  < 0.
• For an expanding FLRW space C denotes the outgoing and  the ingoing null
congruences. The observer is inside the apparent horizon.
– A surface is normal (untrapped) if C > 0 and  < 0.
– A surface is trapped if C > 0 and  > 0.
– A surface is (past) marginally trapped if C > 0 and  D 0 and is inner if
@C  > 0.
• For a contracting FLRW space C denotes the outgoing and  the ingoing null
congruences. The observer is inside the apparent horizon.
– A surface is normal (untrapped) if C > 0 and  < 0.
– A surface is trapped if C < 0 and  < 0.
– A surface is (future) marginally trapped if C D 0 and  < 0 and is inner if
@ C > 0.

10
Cf. Table 1 of Ref. [39].
3.11 de Sitter Space 87

3.11 de Sitter Space

de Sitter space (see Ref. [44] for an introduction and Refs. [6, 16, 78, 79] for reviews)
is a maximally symmetric constant curvature space and is a spatially flat FLRW
space with line element
 
ds2 D dt2 C a20 e2Ht dr2 C r2 d˝.2/
2
(3.110)

in spherical comoving coordinates. The scale factor is a.t/ D a0 eHt , with a0 and
H constants. The de Sitter solution is obtained from the Einstein-Friedmann equa-
tions (3.5) and (3.6) with a positive cosmological constant  (which can formally
 p
be treated as a perfect fluid with D P D ) and no matter, and H D ˙ =3.
8 p
Sometimes the literature refers to expanding de Sitter p spaces with a.t/ D a0 e =3 t
and to contracting de Sitter spaces
 with a.t/ D a0 e =3 t . Other times the scale
p
factor a.t/ D a0 cosh =3 t is used which is a combination of the previous two.
We will call “de Sitter space” only the one with line element (3.110) and scale
factor a.t/ D a0 eHt , with H a positive constant. It has curvature index k D 0 and
Ricci scalar R D 12H 2 .
The de Sitter metric can be recast in static Schwarzschild-like coordinates as
follows. First, introduce the areal radius R.t; r/  a.t/r, in terms of which

dR RPa
dr D  2 dt I (3.111)
a a

this change of radius reduces the line element to the Painlevé-Gullstrand form11
 
ds2 D  1  H 2 R2 dt2  2HR dtdR C dR2 C R2 d˝.2/
2
: (3.112)

The Painlevé-Gullstrand coordinates penetrate the horizon R D 1=H and constant

time slices are Euclidean (by setting dt D 0 one obtains the 3-dimensional Euclidean
line element ds2 j.3/ D dR2 C R2 d˝.2/2
), properties which we have already seen to
hold for the Painlevé-Gullstrand coordinates of the Schwarzschild metric.
The cross-term proportional to dtdR in the line element (3.112) is eliminated by
the use of the new time coordinate T defined by

HR
dT D dt C dR : (3.113)
1  H 2 R2

dT is a locally exact differential: in fact, thinking of .t; R/ as independent coordi-

nates, it is dT D T1 dt C T2 dR with T1 D 1, T2 D HR=.1  H 2 R2 /, and

11
The coordinates .t; R; ; '/ are called “Painlevé-de Sitter coordinates” [72].
88 3 Cosmological Horizons

@T1 @T2
D D 0:
@R @t
HR
By substituting dt D dT  dR into the line element (3.112), one obtains
1  H 2 R2
!
2
 2 2
 2 2HR H 2 R2 2
ds D  1  H R dT  dTdR C dR
1  H 2 R2 .1  H 2 R2 /2
 
HR
 2HR dT  dR dR C dR2 C R2 d˝.2/ 2
1  H 2 R2
 
D  1  H 2 R2 dT 2
" #
 2 2
 H 2 R2 2H 2 R2
C 1 1H R C dR2
.1  H 2 R2 /2 .1  H 2 R2 /

CR2 d˝.2/
2
(3.114)

and finally the de Sitter metric in static Schwarzschild-like coordinates is obtained

(e.g., [41, 65]),

  dR2
ds2 D  1  H 2 R2 dT 2 C C R2 d˝.2/
2
: (3.115)
1  H 2 R2
This line element bears some resemblance with the Schwarzschild metric in
Schwarzschild coordinates. This static chart covers only the region 0 < R < REH of
de Sitter space interior to the horizon and, therefore, de Sitter space is only locally
static.
For H D 0 de Sitter space degenerates into Minkowski space, the static coor-
dinate system becomes global, and there are no apparent and trapping horizons,
RAH ! 1.
There is no particle horizon in de Sitter space (defined as the k D 0 FLRW space
with scale factor a.t/ D a0 eHt and H > 0) because there is no Big Bang and the
integral

Z " 0
#t
t
dt0 eHt
D (3.116)
1 a.t0 / a0 H
1

diverges.
There is an event horizon in de Sitter space, with radius
r
1 3
REH D D ; (3.117)
H 

which is time-independent.
3.11 de Sitter Space 89

Proof.

Z " #C1
C1 Ht0
dt0 Ht e 1
REH  a.t/ De D :
t a.t0 / H H
t

t
u
The area of the event horizon is AEH D 4 R2EH D 12 =.

Apparent and event horizons coincide in de Sitter space and this horizon is null.

This surface is also a Killing horizon of the Killing vector field

 a
@
ka  : (3.118)
@T

To check that this is a Killing field, consider its components in static Schwarzschild-
like coordinates
! !
k D 1; 0; 0; 0 ; k D  .1  H 2 R2 /; 0; 0; 0 (3.119)

and the covariant derivatives

 
r k D @ k   ˛ k˛ D @ k C  0 1  H 2 R2 : (3.120)

The only non-vanishing Christoffel symbols  0 in these coordinates are

H2R
010 D 100 D  (3.121)
1  H 2 R2
or

H2R  
 0 D  2 2
ı 0 ı1 C ı 1 ı0 I (3.122)
1H R
this equation in turn gives
 
r k D H 2 R ı 1 ı0  ı 0 ı1 (3.123)

and the only component of the left hand side of the Killing equation
ra kb C rb ka D 0 which is not trivially zero is
 
r0 k1 C r1 k0 D 2 H 2 R C H 2 R ; (3.124)
90 3 Cosmological Horizons

which also vanishes. The Killing equation is satisfied and ka is a Killing vector field
with norm squared
 
ka ka D  1  H 2 R2 : (3.125)

Therefore, ka is timelike for R < H 1 , null for R D H 1 , and spacelike for R > H 1
and the hypersurface R D H 1 is a Killing horizon. This situation is analogous to
that of the Killing vector field .@=@t/a in the Schwarzschild spacetime (1.3), which
changes its causal character on the Schwarzschild event horizon R D 2M.
The components of the de Sitter space Killing vector ka in comoving coordinates
are

k D .1; 0; 0; 0/ ; k D .1; 0; 0; 0/ : (3.126)

In Painlevé-Gullstrand coordinates the components of the Killing field are

instead

k D .1; aP r; 0; 0/ D .1; HR; 0; 0/ : (3.127)

The surface gravity of the Killing horizon is

r
1 
dS D DHD : (3.128)
REH 3

Proof. The Killing surface gravity is given by Eq. (2.78) which yields, using static
coordinates and Eq. (3.123),

2 1  a b 1   
Killing D r k .ra kb / D  g ˛ gˇ r k r˛ kˇ
2 2
1 ˛ ˇ 2    
D  g g H R ı 1 ı0  ı 0 ı1 H 2 R ı˛1 ıˇ0  ı˛0 ıˇ1
2
H 4 R2 ˛ ˇ 
D g g ı˛0 ıˇ1 ı 0 ı1  ı˛0 ıˇ1 ı 1 ı0  ı˛1 ıˇ0 ı 0 ı1
2

Cı˛1 ıˇ0 ı 1 ı0
H 4 R2  00 11
D g g  2.g01 /2 C g11 g00 D H 4 R2 :
2

On the Killing horizon R D H 1 , it is Killing

2
D H 2 and Eq. (3.128) follows by
choosing the positive sign arising from the square root of this equation. u
t
3.13 Thermodynamics of de Sitter Space 91

3.12 Thermodynamics of Cosmological Horizons in General

Relativity

One of the most interesting developments of black hole physics is black hole
thermodynamics. Originally developed for stationary event horizons, this branch
of theoretical physics is being extended to other types of horizons such as apparent,
trapping, isolated, dynamical, and slowly evolving horizons. It was realized early on
that also cosmological horizons have thermodynamics associated to them, beginning
with the static event horizon of de Sitter space [41]. It is claimed in the literature
that this cosmological thermodynamics extends also to FLRW apparent horizons.

3.13 Thermodynamics of de Sitter Space

The thermodynamics of de Sitter space was studied by Gibbons and Hawking [41]
(see also [11, 25, 27, 62, 63, 72, 76] and the references therein). The de Sitter horizon
is endowed with temperature and entropy, like the Schwarzschild event horizon, as
was deduced using Euclidean field theory techniques [41]. Gibbons and Hawking
computed the thermal bath seen by a timelike geodesic observer in de Sitter space
carrying a (scalar) particle detector confined to a small tube around the observer’s
worldine [41]. The result is the de Sitter horizon temperature TH given by

„ H „ Killing
kB TH D D (3.129)
c 2 c 2
where we have restored the fundamental constants and Killing is the surface gravity
defined using the Killing vector .@=@T/a of de Sitter space. The entropy of the de
Sitter horizon, commonly referred to as Gibbons-Hawking entropy, is

kB c3 AH kB c3
SH D D (3.130)
„G 4 „G H 2
or

„c 1
SH D ; (3.131)
kB G 4 TH2

where AH D 4 R2EH D 4 H 2 is the area of the event horizon. It is common to

write

AH
SH D kB 2
; (3.132)
4lPl
92 3 Cosmological Horizons

p
2
where lPl D G„=c3 is the Planck length and lPl is interpreted as a “quantum of
area”. There is radiation with a thermal spectrum near the horizon with characteristic
wavelength  H 1 and T  1028 K today. The entropy is constant in time, in
agrement with the fact that de Sitter space is static in the region 0  R  RH .
There is no extremal horizon analogous to that of extremal black holes because
the metric contains a single parameter.
The Kodama vector in de Sitter space is immediately given by Eq. (2.76) which,
compared with the de Sitter line element in static coordinates (3.115), yields
 a
@
Ka D ; (3.133)
@T

i.e., the Kodama vector coincides with the timelike Killing vector (3.118) of de
Sitter space. The surface gravity generated by this Killing-Kodama field is  D
p 
H D =3 (Eq. (3.128)) and then the corresponding temperature is TH D in
2
geometrized units. The first law of thermodynamics is satisfied: taking the internal
4
energy U D MMSH D R3 , where MMSH is the Misner-Sharp-Hernandez mass
3
and VH D 4 R3AH =3 as the volume bounded by the event horizon, one has
   
4 4
dU D d R3 Dd D0
3 H 3H 3

because both H and are constant, while

 
4
dVH D d D 0;
3H 3

while SH is constant; dSH ; dU, and dVH all vanish making the first law of thermody-
namics rather trivial.

3.14 Thermodynamics of Apparent/Trapping Horizons

in FLRW Space

The thermodynamic formulae valid for the de Sitter event (and apparent) horizon
are taken and adapted by many authors to the non-static apparent horizon of FLRW
space, which does not coincide with the event horizon (the latter may not even
exist). The apparent horizon is often argued to be a causal horizon associated
with gravitational temperature, entropy and surface gravity in dynamical spacetimes
([5, 15, 22, 40, 46, 47, 67] and references therein) and, if these arguments hold, then
they would apply also to cosmological horizons. That the thermodynamics is ill-
defined for the event horizon of FLRW space (except for de Sitter space) was argued
3.14 Thermodynamics of Apparent/Trapping Horizons in FLRW Space 93

in [15, 26, 39, 40, 85]. The authors of [51, 86] attempted to compute the Hawking
radiation of the FLRW apparent horizon. References [20, 62] rederived it using
the Hamilton-Jacobi method [3, 68, 83] in the Parikh-Wilczek approach originally
developed for black hole horizons [73]. In this context the particle emission rate in
the WKB approximation is the tunneling probability for the classically forbidden
trajectories from inside to outside the horizon,
   
2 Im.I/ „!
  exp  ' exp  ; (3.134)
„ kB T

where I is the Euclideanized action with imaginary part Im.I/, ! is the angular
frequency of the radiated quanta (taken, for simplicity, to be those of a massless
scalar field, which is the simplest field to perform Hawking effect calculations),
and the Hawking temperature is read off the expression of the Boltzmann fac-
„2 !
tor, kB T D . The particle energy „! is defined in an invariant way as
2 Im.I/
! D K ra I, where K a is the Kodama vector, and the action I satisfies the
a

Hamilton-Jacobi equation

hab ra I rb I D 0 : (3.135)

Although the definition of energy is coordinate-invariant, it depends on the choice

of time, here defined as the Kodama time, and ! is a “Kodama (angular) frequency”.
It is not yet established beyond doubt, however, that this prescription gives a correct
and consistent thermodynamics.
There has been a rather large literature on the thermodynamics of FLRW spaces
(e.g., [2, 19, 30] and references therein). A review of the thermodynamical properties
of the FLRW apparent horizon, as well as the computation of the Kodama vector,
Kodama-Hayward surface gravity, and Hawking temperature in various coordinate
systems are given in Ref. [30]. The Kodama-Hayward temperature of the FLRW
apparent horizon is given by
 
  RAH H 2 C HP C k  
„ 2 2a2 „
kB T D  D RAH R
c 2 24 c
 
„G RAH
D
P .3P  / : (3.136)
c 3

For a spatially flat universe this expression reduces to

  P  
„G H C H=.2H/ „G .  3P/
kB T D D
P : (3.137)
c 2 c 3H

This temperature depends on the choice of surface gravity  since T D =2 in

geometrized units and, as we have seen, there are several inequivalent prescriptions
94 3 Cosmological Horizons

for this quantity. The choice of  giving the temperature reported here is the
Kodama-Hayward choice of Eq. (2.87) [30]. In fact, the Kodama-Hayward surface
gravity given by Eq. (2.87) is
 
RAH P C k RAH
Kodama D  2H 2 C H D R: (3.138)
2 a2 12

Proof. Using comoving coordinates and the decomposition

 of the metric
 ds2 D
2
2 a
hab dxa dxb C R2 d˝.2/ (a; b D 1; 2), where hab D diag 1; , Eq. (2.87)
1  kr2
gives
p
1  kr2 a
Kodama D @a p hab @b .a.t/r/
2a 1  kr2
p "   !#
1  kr2 a 1  kr2 a1
D @a p Parı Ca0
ı
2a 1  kr2 a
 2   
1 aP aR kr R 2 P k
D C ar C D 2H C H C 2 :
2 a2 a a 2 a

t
u
The Kodama-Hayward dynamical p surface gravity (3.138) vanishes if the scale
factor has the special form a.t/ D ˛t2 C ˇt C , where ˛; ˇ, and are constants.
On the apparent horizon, Eq. (3.138) can be written as [19]
!
1 RP AH
Kodama D  2H (3.139)
2HRAH RAH

using Eqs. (3.106) and (3.138). In the usual prescription of stationary spacetime
which assigns to a horizon of area A the entropy A =4 (in geometrized units), the
putative entropy of the FLRW apparent horizon would be
   
kB c3 AAH kB c3
SAH D D ; (3.140)
„G 4 „G H 2 C k=a2

where
4
AAH D 4 R2AH D (3.141)
H 2 C k=a2

is the area of the event horizon. The Hamiltonian constraint (3.5) gives

3 9H
SAH D
P ; SP AH D
P .P C / : (3.142)
8 8 2
3.14 Thermodynamics of Apparent/Trapping Horizons in FLRW Space 95

In an expanding universe the apparent horizon entropy increases if P C > 0, stays

constant if P D  , and decreases if the weak energy condition is violated, P <  .
The horizon temperature is positive only if the Ricci scalar is, which is equivalent
to equations of state satisfying P < =3 for a perfect fluid in Einstein theory [35].
This is the condition for the apparent horizon to be also a trapping horizon. A “cold
horizon” with T D 0 is obtained for vanishing Ricci scalar but the entropy is positive
for such an horizon. These properties induce caution in regarding Eqs. (3.136)
and (3.140) as established.
What about the first law of thermodynamics for the apparent horizon of FLRW
space? Several authors argue that the natural choice of surface gravity is the
Kodama-Hayward one, which produces the apparent horizon temperature (3.136).
AAH
The entropy usually attributed to the apparent horizon is SAH D D R2AH , and
4
the internal energy U should be identified with the Misner-Sharp-Hernandez mass
4 R3AH
MAH D contained inside the apparent horizon. As already remarked, the
3
4 R3AH
factor appearing in this espression is not the proper volume of a sphere of
3
areal radius RAH unless the universe has flat spatial sections, k D 0. This fact points
us to use the areal volume

4 R3AH
VAH  (3.143)
3
instead of the proper volume in thermodynamics (failing to do so would jeopardize
the possibility to write the first law). However, even with this caveat, the first law
does not assume the form

P AH C PVP AH
TAH SP AH D M (3.144)

that one might expect.

Let us proceed to illustrate what the Kodama-Hayward proposal for T entails.
0th law of thermodynamics. The 0th law states that the temperature (or, equiva-
lently, the surface gravity) is constant on the horizon. This law ensures that all points
of the horizon are at the same temperature, or that there is no temperature gradient
on it. This statement corresponds to thermal equilibrium on the horizon and is a
rather trivial consequence of spherical symmetry.
1st law of thermodynamics. The first law of thermodynamics for apparent horizons
is more complicated than (3.144) and was given in Refs. [46, 47] under the
name of “unified first law”. While using the Misner-Sharp-Hernandez mass MAH
enclosed by the apparent horizon as internal energy, the Kodama-Hayward horizon
temperature (3.136), the areal volume (3.143), one introduces further quantities as
follows. Decompose the metric as in Eq. (2.71); then the work density is

1
w0   Tab hab (3.145)
2
96 3 Cosmological Horizons

(proportional to the trace of the matter stress-energy tensor over the 2-space normal
to the 2-spheres of symmetry);

a  Ta b rb R C w0 ra R (3.146)

is the energy flux (localized Bondi flux) across the apparent horizon, when computed
on this hypersurface. The quantity AAH a is called the energy supply vector. If K a
denotes the Kodama vector,

ja  a C w0 Ka (3.147)

is a divergence-free energy-momentum vector which can be used in lieu of a. The

Einstein equations then give [46, 47]

MMSH D R2 C 4 R3 w0 ; (3.148)

ra MMSH D A ja : (3.149)

The last equation is rewritten as [46, 47]

A a D ra MMSH  w0 ra VAH (3.150)

(“unified first law”). The energy supply vector is then written as

   
 A MMSH
A a D ra C Rra : (3.151)
2 4 R

Along the apparent horizon, defined by r c Rrc R D 0, it is

R RAH
MAH D .1  r c Rrc R/ jAH D
2 2
and

AAH a D ra SAH D TAH ra SAH :
2
This equation is interpreted by saying that the energy supply across the apparent
horizon AAH a is the “heat” TAH ra SAH gained. Writing the energy supply explicitly
gives

TAH ra SAH D ra MAH  w0 ra VAH (3.152)

and w0 ra VAH is a work term. The “heat” entering the apparent horizon goes into
changing the internal energy MAH and performing work due to the change in size of
this horizon.
3.14 Thermodynamics of Apparent/Trapping Horizons in FLRW Space 97

Let us compute now the time component of Eq. (3.152) in comoving coordinates
for a FLRW  space 2sourced
 by a perfect fluid in General Relativity. We have
a
hab D diag 1; and
1  kr2

1
w0   Œ.P C / ua ub C Pgab  hab
2
1  
D  .P C / h00 .u0 /2 C P h00 h00 C h11 h11
2
1
D  Π.P C / C 2P
2
P
D : (3.153)
2
One computes also
 
P RP AH aR 2
VAH D 3VAH D 3HVAH 1  RAH
RAH a
3HVAH
D  3 .P C / (3.154)
2

and, using the last equation,

P AH D d .VAH / D VP AH C VAH P
M
dt
3HVAH
D .P C / : (3.155)
2
We also have

3 R2AH
SP AH D 2 RAH RP AH D
P H .P C / (3.156)

and, using this relation,

RAH R2
TAH SP AH D .3P  / 3 AH H .P C /
3
3HVAH
D .P C / .3P  / : (3.157)
4

Therefore we have, using Eqs. (3.157), (3.155), and (3.154),

.P  / P
P AH C
TAH SP AH D M VAH : (3.158)
2
98 3 Cosmological Horizons

In the infinitesimal interval of comoving time dt the changes in the thermodynamical

quantities are related by

.P  /
TAH dSAH D dMAH C dWAH ; dWAH D dVAH :
2

The coefficient of dVAH , i.e., w0 D .P  / =2 equals the pressure P (the naively
expected coefficient) only if P D  (the case of de Sitter space in which dMAH ,
dVAH , and dSAH all vanish). The fact that the coefficient appearing in the work term
is not simply P can be understood as a consequence of the fact that the apparent
horizon is not comoving [35]. For a comoving sphere of radius Rs it is RP s =Rs D H
and VP s D 3HVs , while

P s D d .Vs / D VP s C Vs P D 3HVs  3HVs .P C / D 3HVs P ;

M
dt

hence M P s C PVPs D 0. Indeed, the covariant conservation equation (3.7) is often

presented as the first law of thermodynamics for a comoving volume V. Because of
spatial homogeneity and isotropy there can be no preferred directions and physical
spatial vectors in FLRW space, therefore the heat flux through a comoving volume
must be zero. In fact, consider a comoving volume Vc (which, by definition, is
constant in time) and the corresponding proper volume at time t, V D a3 .t/Vc .
Multiplying Eq. (3.7) by V one obtains

a3 Vc P C 3a2 aP Vc .P C / D 0

or
d
V P C VP .P C / D . V/ C PVP D 0 :
dt
By interpreting U  V as the total internal energy of matter in the volume V one
obtains the relation between variations in the time dt

dU C PdV D 0 ; (3.159)

and the first law (with work term coefficient P) then gives TdS D 0, which is
consistent with the above-mentioned absence of entropy flux vectors and with the
well known fact that, in curved space, there is no entropy generation in a perfect
fluid (the entropy along fluid lines remains constant and there is no exchange of
entropy between neighbouring fluid lines [80]). Indeed, Eq. (3.62) for the evolution
of the Misner-Sharp-Hernandez mass contained in a comoving sphere reduces to
MP MSH C PVP D 0 or P C 3H .P C / D 0. However, for a non-comoving volume,
the work term is more complicated than PdV.
3.14 Thermodynamics of Apparent/Trapping Horizons in FLRW Space 99

Attempts to write the first law for the event, instead of the apparent, horizon lead
to inconsistencies [15, 26, 39, 85]. This fact supports the belief that it is the apparent
horizon which is the relevant quantity in the thermodynamics of cosmological
horizons.
(Generalized) 2nd law of thermodynamics.
A second law of thermodynamics for the de Sitter horizon was already given
in the original Gibbons-Hawking paper [41] and re-proposed in [65]. Davies [26]
has considered the event horizon of FLRW space and, for General Relativity with
a perfect fluid as the source, has proved the following theorem: if the cosmological
fluid satisfies P C  0 and a.t/ ! C1 as t ! C1, then the area of the
event horizon
 3 is non-decreasing. The entropy of the event horizon is taken to be
kB c AEH
SEH D , where AEH is the area of the event horizon. The validity of the
„G 4
generalized 2nd law for certain radiation-filled universes was discussed in [28, 29].
4 4
For radiation, the energy density is rad D T , where  is the Stefan-Boltzmann
c
12
constant, and the entropy density is

4 rad
srad D : (3.160)
3 T
The entropy of the radiation contained inside the volume VEH enclosed by the event
horizon is
p
4 2   1=4 3=4
Srad D srad VEH D rad VEH :
3 c
The total entropy of radiation contained in a comoving volume would stay constant
because rad  a4 and srad  a3 while a proper volume scales as V  a3 ; however,
the event horizon is not comoving and the radiation entropy within it decreases
as VEH expands slower than comoving (according to RP EH =REH D H  R1 EH ) and
radiation crosses outside the event horizon. For realistic universes the entropy of the
event horizon is much larger than the radiation entropy and, for universes departing
slightly from a de Sitter universe due to the presence of a cosmological constant in
addition to radiation, it can be shown analytically that the generalized 2nd law is
valid, SP rad C SP EH > 0 [28, 29]. However, a general proof is not available.
Another question raised in [29] is the following: if the universe contains a gas of
black holes, there is an entropy loss when these cross outside the horizon. For small
black holes, this loss in more than compensated by the increase in area of the event
horizon. However, for larger black holes, a preliminary study suggests a violation of
the generalized 2nd law in certain open universes. These results cannot be fully
relied upon because they assume relations valid for the Schwarzschild-de Sitter

PC
12
In general, the entropy density of a perfect fluid is s D [54].
T
100 3 Cosmological Horizons

black hole, which probably do not hold for large dynamical black holes embedded
in a FLRW background. However, this issue remains unsolved and deserves further
investigation with more reliable models.
Due to the difficulties with the event horizon one is led to consider the apparent
horizon instead. Then, Eq. (3.156) tells us that, in an expanding universe in Einstein
theory with perfect fluid, the apparent horizon area increases except for the quantum
vacuum equation of state P D  (for which SAH stays constant) and for phantom
fluids with P <  , in which case SAH decreases, adding another element of
weirdness to the behaviour of phantom matter.
The generalized 2nd law states that the total entropy of matter and of the horizon
Stotal D Smatter C SAH cannot decrease in any physical process,

ıS D ıSmatter C ıSAH  0 : (3.161)

(We refer here to the apparent horizon but several authors refer instead to the event
or particle horizons. It is clear that the apparent horizon is more appropriate since it
is a quantity defined quasi-locally and it always exists.) There is no definitive proof
that the apparent horizon thermodynamics is consistent.
Thermodynamics of spacetime and cosmic holography. In the spirit of the thermo-
dynamics of spacetime [34, 50], the Einstein-Friedmann equations of cosmology in
Einstein theory have been derived from the first law of thermodynamics (3.158),
first for spatially flat universes [15, 23] and then for general curvature index k [19].
An earlier work by Verlinde [82] derived the Einstein-Friedmann equations in a
radiation-dominated FLRW universe from the Cardy-Verlinde formula, which gives
the entropy for a conformal field theory, in the spirit of the holographic principle.
The Fischler-Susskind version of this principle can be formulated by saying that
the matter entropy contained in the volume enclosed by the particle horizon cannot
exceed the entropy of the particle horizon itself. This principle restricts the matter
content of the universe.
The cosmic holographic inequality is written as

APH
sVPH  ; (3.162)
4
where s is the entropy density of matter. Fischler and Susskind [37] find that
the cosmic holographic principle is violated for fluids with P D w if w < 1=3.
Bak and Rey instead apply the holographic inequality to the apparent horizon [5]
with the following results. For k D 0 or 1, the cosmic holographic inequality
sVAH  AAH =4 is satisfied for perfect fluids with jwj  1 (phantom fluids, in
particular, violate the holographic bound) and if the inequality is satisfied at the
Planck time [5]. Hence inflation, with w ' 1 violates also this version of the
cosmic holographic principle.
Various entropy bounds have been discussed for FLRW space [5, 7, 13, 14, 17,
18, 33, 38, 52, 53, 58, 74, 81, 84]. These discussions use the particle horizon or the
Problems 101

event horizon and seem to ignore the apparent horizon. Given the highly speculative
nature of the subject and the fact that results in this area do not seem to be settled,
we will not discuss it further and we refer the reader to these references.

3.15 Conclusions

The thermodynamics of apparent horizons is very intriguing, but is it correct? Is the

entire construction consistent? After all, the work term appearing in the first law is
somehow introduced a posteriori and, if one did not already know the first law, it is
unlikely that one would discover it this way.
The apparent horizon thermodynamics is formulated for horizons changing in
an arbitrary way. However, it would seem that equilibrium thermodynamics could
only be introduced for physical systems in equilibrium and near equilibrium, and
therefore the appropriate constructs should be slowly varying apparent horizons. In
the literature, this restriction appears only in the Hamilton-Jacobi approach to the
Wilczek-Parikh tunneling formalism, in which an high frequency approximation is
used which requires that the background is varying slowly (although usually this
requirement is not spelled out).
Apparent horizons seem to have implications also for the black hole information
loss paradox and are seen as an alternative to firewalls [43], but this viewpoint need
much work to be developed.

Problems

3.1. Compute the Misner-Sharp-Hernandez mass for a sphere in FLRW space using
comoving coordinates.
3.2. In a k D 0 FLRW universe with a perfect fluid and constant equation of state
P D w and 1 < w < 1=3 (accelerating but not superaccelerating universe),
show that the event horizon is always outside the apparent horizon and is, therefore,
unobservable.13
3.3. Compute the surface gravity of de Sitter space (3.128) using the prop-
erty (2.78).
3.4. Check that Eq. (3.144) is not satisfied in FLRW space.

13
Cf. Refs. [15, 85].
102

References

1. Akbar, M., Cai, R.-G.: Friedmann equations of FRW universe in scalar-tensor gravity, f(R)
gravity and first law of thermodynamics. Phys. Lett. B 635, 7 (2006)
2. Akbar, M., Cai, R.-G.: Thermodynamic behavior of the Friedmann equation at the apparent
horizon of the FRW universe. Phys. Rev. D 75, 084003 (2007)
3. Angheben, M., Nadalini, M., Vanzo, L., Zerbini, S.: Hawking radiation as tunneling for
extremal and rotating black holes. J. High Energy Phys. 05, 014 (2005)
4. Ashtekar, A., Krishnan, B.: Isolated and dynamical horizons and their applications. Living Rev.
Relat. 7, 10 (2004)
5. Bak, D., Rey, S.-J.: Cosmic holography. Class. Quantum Grav. 17, L83 (2000)
6. Balasubramanian, V., de Boer, J., Minic, D.: Mass, entropy and holography in asymptotically
de Sitter spaces. Phys. Rev. D 65, 123508 (2002)
7. Banks, T., Fischler, W.: An upper bound on the number of e-foldings. Preprint arXiv:astro-
ph/0307459
8. Ben-Dov, I.: Outer trapped surfaces in Vaidya spacetimes. Phys. Rev. D 75, 064007 (2007)
9. Bengtsson, I., Senovilla, J.M.M.: Region with trapped surfaces in spherical symmetry, its core,
and their boundaries. Phys. Rev. D 83, 044012 (2011)
10. Bhattacharya, S., Lahiri, A.: On the existence of cosmological event horizons. Class. Quantum
Grav. 27, 165015 (2010)
11. Birrell, N.D., Davies, P.C.W.: Quantum Fields in Curved Space. Cambridge University Press,
Cambridge (1982)
12. Booth, I., Brits, L., Gonzalez, J.A., Van den Broeck, V.: Black hole boundaries. Class. Quantum
Grav. 23, 413 (2006)
13. Bousso, R.: A covariant entropy conjecture. J. High Energy Phys. 07, 004 (1999)
14. Bousso, R.: Positive vacuum energy and the N bound. J. High Energy Phys. 11, 038 (2000)
15. Bousso, R.: Cosmology and the S-matrix. Phys. Rev. D 71, 064024 (2005)
16. Bousso, R.: Adventures in de Sitter space. Preprint arXiv:hep-th/0205177
17. Brustein, R., Veneziano, G.: Causal entropy bound for a spacelike region. Phys. Rev. Lett. 84,
5695 (2000)
18. Cai, R.-G.: Holography, the cosmological constant and the upper limit of the number of e-
foldings. J. Cosmol. Astropart. Phys. 02, 007 (2004)
19. Cai, R.-G., Kim, S.P.: First law of thermodynamics and Friedmann equations of Friedmann-
Robertson-Walker universe. J. High Energy Phys. 0502, 050 (2005)
20. Cai, R.-G., Cao, L.-M., Hu, Y.-P.: Hawking radiation of an apparent horizon in a FRW universe.
Class. Quantum Grav. 26, 155018 (2009)
21. Chakraborty, S., Mazumder, N., Biswas, R.: Cosmological evolution across phantom crossing
and the nature of the horizon. Astrophys. Sp. Sci. 334, 183 (2011)
22. Collins, W.: Mechanics of apparent horizons. Phys. Rev. D 45, 495 (1992)
23. Danielsson, U.H.: Transplanckian energy production and slow roll inflatioon. Phys. Rev. D 71,
023516 (2005)
24. Das, A., Chattopadhyay, S., Debnath, U.: Validity of generalized second law of thermodynam-
ics in the logamediate and intermediate scenarios of the universe. Found. Phys. 42, 266 (2011)
25. Davies, P.C.W.: Mining the universe. Phys. Rev. D 30, 737 (1984)
26. Davies, P.C.W.: Cosmological horizons and entropy. Class. Quantum Grav. 5, 1349 (1988)
27. Davies, P.C.W., Ford, L.H., Page, D.N.: Gravitational entropy: beyond the black hole. Phys.
Rev. D 34, 1700 (1986)
28. Davis, T.M., Davies, P.C.W.: How far can the generalized second law be generalized? Found.
Phys. 32, 1877 (2002)
29. Davis, T.M., Davies, P.C.W., Lineweaver, C.H.: Black hole versus cosmological horizon
entropy. Class. Quantum Grav. 20, 2753 (2003)
30. Di Criscienzo, R., Hayward, S.A., Nadalini, M., Vanzo, L., Zerbini, S.: On the Hawking
radiation as tunneling for a class of dynamical black holes. Class. Quantum Grav. 27, 015006
(2010)
References 103

31. d’Inverno, R.: Introducing Einstein’s Relativity. Oxford University Press, Oxford (2002)
32. Doran, C.: A new form of the Kerr solution. Phys. Rev. D 61, 06750 (2000)
33. Easther, R., Lowe, D.A.: Holography, cosmology, and the second law of thermodynamics.
Phys. Rev. Lett. 82, 4967 (1999)
34. Eling, C., Guedens, R., Jacobson, T.: Nonequilibrium thermodynamics of spacetime. Phys.
Rev. Lett. 96, 121301 (2006)
35. Faraoni, V.: Apparent and trapping cosmological horizons. Phys. Rev. D 84, 024003 (2011)
36. Faraoni, V., Nielsen, A.B.: Quasi-local horizons, horizon-entropy, and conformal field redefi-
nitions. Class. Quantum Grav. 28, 175008 (2011)
37. Fischler, W., Susskind, L.: Holography and cosmology. Preprint arXiv:hep-th/9806039
38. Fischler, W., Loewy, A., Paban, S.: The entropy of the microwave background and the
acceleration of the universe. J. High Energy Phys. 09, 024 (2003)
39. Frolov, A., Kofman, L.: Inflation and de Sitter thermodynamics. J. Cosmol. Astropart. Phys.
0305, 009 (2003)
40. Ghersi, J.T.G., Geshnizjani, G., Piazza, F., Shandera, S.: Eternal inflation and a thermodynamic
treatment of Einstein’s equations. J. Cosmol. Astropart. Phys. 1106, 005 (2011)
41. Gibbons, G.W., Hawking, S.W.: Cosmological event horizon, thermodynamics, and particle
creation. Phys. Rev. D 15, 2738 (1977)
42. Guth, A.H.: The inflationary universe: a possible solution to the horizon and flatness problems.
Phys. Rev. D 23, 347 (1981)
43. Hawking, S.W.: Information preservation and weather forecasting for black holes. Preprint
arXiv:1401.5761
44. Hawking, S.W., Ellis, G.F.R.: The Large Scale Structure of Space-Time. Cambridge University
Press, Cambridge (1973)
45. Hayward, S.A.: Gravitational energy in spherical symmetry. Phys. Rev. D 53, 1938 (1996)
46. Hayward, S.A.: Unified first law of black-hole dynamics and relativistic thermodynamics.
Class. Quantum Grav. 15, 3147 (1998)
47. Hayward, S.A., Mukohyama, S., Ashworth, M.C.: Dynamic black holes and entropy. Phys.
Lett. A 256, 347 (1999)
48. Ibison, M.: On the conformal forms of the Robertson-Walker metric. J. Math. Phys. 48, 122501
(2007)
49. Infeld, L., Schild, A.: A new approach to kinematic cosmology. Phys. Rev. 68, 250 (1945)
50. Jacobson, T.: Thermodynamics of spacetime: the Einstein equation of state. Phys. Rev. Lett.
75, 1260 (1995)
51. Jang, K.-X., Feng, T., Peng, D.-T.: Hawking radiation of apparent horizon in a FRW universe
as tunneling beyond semiclassical approximation. Int. J. Theor. Phys. 48, 2112 (2009)
52. Kaloper, N., Linde, A.D.: Cosmology versus holography. Phys. Rev. D 60, 103509 (1999)
53. Kaloper, N., Kleban, M., Sorbo, L.: Observational implications of cosmological event horizons.
Phys. Lett. B 600, 7 (2004)
54. Kolb, E.W., Turner, M.S.: The Early Universe. Addison-Wesley, Reading (1990)
55. Kraus, P., Wilczek, F.: Some applications of a simple stationary line element for the Schwarz-
schild geometry. Mod. Phys. Lett. A 9, 3713 (1995)
56. Li, R., Ren, J.-R., Shi, D.-F.: Fermions tunneling from apparent horizon of FRW universe.
Phys. Lett. B 670, 446 (2009)
57. Liddle, A.R., Lyth, D.H.: Cosmological Inflation and Large Scale Structure. Cambridge
University Press, Cambridge (2000)
58. Lowe, D.A., Marolf, D.: Holography and eternal inflation. Phys. Rev. D 70, 026001 (2004)
59. Martel, K., Poisson, E.: Regular coordinate systems for Schwarzschild and other spherical
spacetimes. Am. J. Phys. 69, 476 (2001)
60. Mazumder, N., Biswas, R., Chakraborty, S.: Interacting three fluid system and thermodynamics
of the universe bounded by the event horizon. Gen. Rel. Gravit. 43, 1337 (2011)
61. McVittie, G.C.: The mass-particle in an expanding universe. Mon. Not. R. Astr. Soc. 93, 325
(1933)
104

62. Medved, A.J.M.: Radiation via tunneling from a de Sitter cosmological horizon. Phys. Rev. D
66, 124009 (2002)
63. Medved, A.J.M.: A brief editorial on de Sitter radiation via tunneling. Preprint
arXiv:0802.3796
64. Mosheni Sadjadi, H.: Generalized second law in a phantom-dominated universe. Phys. Rev. D
73, 0635325 (2006)
65. Mottola, E.: Thermodynamic instability of de Sitter space. Phys. Rev. D 33, 1616 (1986)
66. Mukhanov, V.: Physical Foundations of Cosmology. Cambridge University Press, Cambridge
(2005)
67. Nielsen, A.B., Yeom, D.-H.: Black holes without boundaries. Int. J. Mod. Phys. A 24, 5261
(2009)
68. Nielsen, A.B., Visser, M.: Production and decay of evolving horizons. Class. Quantum Grav.
23, 4637 (2006)
69. Nolan, B.C.: Sources for McVittie’s mass particle in an expanding universe. Phys. Rev. D 58,
064006 (1998)
70. Nolan, B.C.: A Point mass in an isotropic universe. 2. Global properties. Class. Quantum Grav.
16, 1227 (1999)
71. Nolan, B.C.: A Point mass in an isotropic universe 3. The region R  2m. Class. Quantum
Grav. 16, 3183 (1999)
72. Parikh, M.K.: New coordinates for de Sitter space and de Sitter radiation. Phys. Lett. B 546,
189 (2002)
73. Parikh, M.K., Wilczek, F.: Hawking radiation as tunneling. Phys. Rev. Lett. 85, 5042 (2000)
74. Piao, Y.-S.: Entropy of the microwave background radiation in the observable universe. Phys.
Rev. D 74, 47301 (2006)
75. Rindler, W.: Visual horizons in world-models. Mon. Not. R. Astr. Soc. 116, 663 (1956)
(reprinted in Gen. Rel. Gravit. 34, 133 (2002))
76. Sekiwa, Y.: Decay of the cosmological constant by Hawking radiation as quantum tunneling.
Preprint arXiv:0802.3266
77. Senovilla, J.M.M.: Singularity theorems and their consequences. Gen. Rel. Grav. 30, 701
(1998)
78. Spradlin, M.A., Strominger, A., Volovich, A.: Les Houches lectures on de Sitter space. In Les
Houches 2001, Gravity, Gauge Theories and Strings, Proceedings of the Les Houches Summer
School 76, Les Houches, pp. 423–453 (2001). (Preprint arXiv:hep-th/011007)
79. Spradlin, M.A., Volovich, A.: Vacuum states and the S matrix in dS/CFT. Phys. Rev. D 65,
104037 (2002)
80. Stephani, H.: General Relativity. Cambridge University Press, Cambridge (1982)
81. Veneziano, G.: Pre-bangian origin of our entropy and time arrow. Phys. Lett. B 454, 22 (1999)
82. Verlinde, E.: On the holographic principle in a radiation-dominated universe. Preprint
arXiv:hep-th/0008140
83. Visser, M.: Essential and inessential features of Hawking radiation. Int. J. Mod. Phys. D 12,
649 (2003)
84. Wang, B., Abdalla, E.: Plausible upper limit on the number of e-foldings. Phys. Rev. D 69,
104014 (2004)
85. Wang, B., Gong, Y., Abdalla, E.: Thermodynamics of an accelerated expanding universe. Phys.
Rev. D 74, 083520 (2006)
86. Zhu, T., Ren, J.-R.: Corrections to Hawking-like radiation for a Friedmann-Robertson-Walker
universe. Eur. Phys. J. C 62, 413 (2009)
Chapter 4
Inhomogeneities in Cosmological
“Backgrounds” in Einstein Theory

One has no right to love or hate anything if one has not

acquired a thorough knowledge of its nature. Great love springs
from great knowledge of the beloved object, and if you know it
but little you will be able to love it only a little or not at all.
—Leonardo da Vinci

4.1 Introduction

There is much motivation for studying analytic solutions of General Relativity and
of alternative theories of gravity representing central inhomogeneities embedded
in cosmological “backgrounds”. Our main interest is understanding apparent and
trapping horizons and their dynamics. Another motivation comes from the fact that
the present acceleration of the cosmic expansion [8, 87, 136, 137, 142–145, 162]
requires, in the context of General Relativity, that approximately 73 % of the energy
content of the universe is in the form of a mysterious dark energy [90] (see Ref. [5]
for a discussion). Dark energy appears as an ad hoc explanation and an alternative
to it could be that gravity deviates from General Relativity at large scales, which
leads one to take more seriously alternative theories of gravity. Further motivation
for alternative gravity comes from the fact that virtually all theories attempting to
quantize gravity produce, in their low-energy limits, actions containing corrections
to Einstein theory such as nonminimally coupled dilatons and/or higher derivative
terms. These ideas have led to the introduction of f .R/ gravity in cosmology to
modify Einstein theory at large scales [26, 30, 152–154, 156, 157, 168] and explain
the cosmic acceleration without dark energy (see Refs. [40, 155] for reviews).
Given that the f .R/ theories of interest for cosmology (which are the most relevant
in today’s theoretical physics landscape) are designed to produce a time-varying
effective cosmological “constant”, spherically symmetric solutions representing
black holes or other inhomogeneities in these theories are expected to be dynamical
and to have FLRW asymptotics, not to be static and asymptotically flat. Relatively
speaking, very few such solutions are known. However, analytic solutions describ-
ing central objects in cosmological “backgrounds” are interesting also in General
Relativity and not only in alternative gravity. The first study of this kind of solution
by McVittie in 1933 [119] was motivated by the need to understand whether, and to

© Springer International Publishing Switzerland 2015 105

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6_4
106 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

what extent, the cosmic expansion affects the dynamics of local systems (Ref. [28]
reviews this subject). This question constitutes independent motivation for studying
black holes embedded in FLRW cosmologies. If objects which are strongly bound
gravitationally (and nothing is more strongly bound than a black hole) become
comoving and follow the expansion of the cosmic substratum, any weakly bound
object should do the same. This issue can only be studied by analytic solutions of
the Einstein equations if the central object is to be strongly bound.
The old McVittie solution [119] has been largely overlooked for decades and
recent studies show that its structure and details are not yet completely understood
[38, 55, 85, 94, 123, 124]. Relatively few other solutions describing central conden-
sations in otherwise spatially homogeneous universes have been reported over the
years (see Ref. [92] for an in-depth discussion of inhomogeneous cosmologies from
a more general point of view).
Cosmological condensations in General Relativity have received recent attention
also for other reasons, following other attempts to explain the present acceleration
of the universe without exotic dark energy and without modifying gravity. The first
idea consists of using the backreaction of inhomogeneities on the cosmic dynamics
to produce the observed acceleration [20–24, 88, 97, 98, 101, 102, 134, 139, 140,
173, 174]. This backreaction idea is implemented in a formalism plagued by formal
problems and it has not been demonstrated that it is able to explain the cosmic
acceleration. The sign of the backreaction terms (let alone their magnitude) in the
equation giving the averaged acceleration has not been shown to be the correct one
[17, 96, 163, 167]. Even more serious doubts are cast on this proposed solution to the
cosmic acceleration problem by the (admittingly more formal) work of Ref. [71].
A second idea to move beyond these riddles attributes the cosmic acceleration
to the possibility that we live inside a giant void, which involves the consideration
of analytic solutions of Einstein theory describing cosmological inhomogeneities
(Lemaître-Tolman-Bondi, Swiss-cheese, and other models) [89, 92, 112–114, 133,
135, 139, 140]. There has also been interest in evolving horizons in relation to the
accretion of dark energy [6, 32, 41, 69, 72, 78, 83, 108, 159]. Accretion onto a
primordial black hole in the early universe could have been so rapid to make its
growth very fast [27, 74–76]. The accretion of dark energy and, in particular, of
phantom energy (if this extreme form of dark energy exists at all) by a black hole,
and the consequent backreaction and mass change have been the subject of much
recent literature. Analytic solutions which accrete from their surroundings are useful
to elucidate questions in this area [66, 121].
Another major motivation for studying cosmological black holes is that explicit
examples of time-varying horizons would be very useful to understand Hawking
radiation and formulate black hole and horizon thermodynamics in fully dynamical
situations.
Due to the non-linearity of the Einstein equations, it is impossible to split a metric
into a cosmological “background” and a part describing a spherical inhomogeneity.
However, in the solutions described in this chapter, the spacetime reduces to a
FLRW universe when a parameter (related to the mass of the central inhomogeneity)
vanishes and, therefore, we will loosely use the word “background” to denote this
FLRW spacetime.
4.2 Schwarzschild-de Sitter-Kottler Spacetime 107

Let us proceed to analyze general-relativistic spacetimes which describe cosmo-

logical black holes and have time-evolving horizons at least part of the time. All
these solutions of the Einstein equations are spherically symmetric.

4.2 Schwarzschild-de Sitter-Kottler Spacetime

The line element of the Schwarzschild-de Sitter-Kottler spacetime [91] is

   1
2 2m 2 2 2 2m 2 2
ds D  1   H R dt C 1  H R dR2 C R2 d˝.2/
2
; (4.1)
R R
p
where R is obviously an areal radius, the constant H D =3 is the Hubble
parameter of the de Sitter “background”.  > 0 is the cosmological constant,
and the positive parameter m describes the mass of the central inhomogeneity (e.g.,
[14, 77]). The static coordinates .t; R; ; '/ cover the region

0 < t < C1 ;
R1 < R < R2 ;
0< < ;
0<'<2 ;

where R1;2 denote the horizon radii, see below.

The usual recipe gRR D 0 locates the apparent horizons, with radii given by the
positive roots of the cubic

H 2 R3  R C 2m D 0 : (4.2)

The solutions are

2
R1 D p sin ; (4.3)
3H
1 1
R2 D cos  p sin ; (4.4)
H 3H
1 1
R3 D  cos  p sin ; (4.5)
H 3H
p
where sin.3 / D 3 3 mH. In an expanding universe, both m and H are positive,
which implies that the root R3 is negative and unphysical, hence there can be at
most two apparent horizons. When R1 and R2 are real, R1 is a black hole apparent
horizon (it reduces to the Schwarzschild horizon R D 2m in the limit H ! 0), while
108 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

R2 is a cosmological apparent horizon (it becomes the static de Sitter horizon with
R D H 1 in the limit m ! 0). The Schwarzschild-de Sitter-Kottler line element is
static in the region covered by the coordinates .t; R; ; '/ and comprised between
the black hole and the cosmological horizons.
Both apparent horizons exist only if 0 < sin.3 / < 1. Then, since the metric
is static between the horizons, the black hole and cosmological apparent horizons
are also event horizons, hence they are null surfaces. When sin.3 / D 1 the
two horizons coincide: this extremal case corresponds to the Nariai black hole
[125, 126]. When instead sin.3 / > 1, the radii of both apparent horizons assume
complex values and are unphysical: the spacetime then contains a naked singularity.
To summarize:
p
• If mH < 1=.3 p3/ there are two horizons with radii R1 and R2 .
• If mH D 1=.3p 3/ it is R1 D R2 and the two horizons coincide.
• If mH > 1=.3 3/ there are no apparent horizons.
The last situation in this list is interpreted by noting that the black hole horizon
becomes larger than the cosmological horizon and any observer in the region
R1 < R < R2 cannot know about it. The spacetime region below the cosmological
horizon can only accommodate a black hole smaller than (or as large as) this
horizon.
The cosmological apparent horizon has a smaller radius than the radius that the
de Sitter apparent horizon would have were the black hole absent: R2 < H 1 . What
is more, the black hole apparent horizon is larger than that of a Schwarzschild
black hole of the same mass m: R1 > 2m. The physical interpretation is that the
cosmological “background” stretches the horizon of a black hole embedded in it,
while the black hole contracts the cosmological horizon.
The area of the black hole horizon A D 4 R21 is, of course, time-independent.
The central singularity is eternal and the black hole event horizon (when it exists)
surrounds it.
The Misner-Sharp-Hernandez mass (2.92) of a sphere of radius R is calculated
as

H 2 R3 4
MMSH D m C DmC R3 ; (4.6)
2 3

where D is the energy density of the de Sitter “background”.
8
The Schwarzschild-de Sitter-Kottler black hole is the subject of an extensive
literature devoted to the thermodynamical properties of its horizons. The thermo-
dynamics is particularly interesting because of the simultaneous presence of a black
hole and a cosmological horizon (e.g., [151]). There are two distinct temperatures
associated with the two horizons and it appears that thermal equilibrium is only
possible for an extremal (Nariai) black hole, in which these two temperatures
become equal. Black holes embedded in anti-de Sitter space have also been the
subject of much recent interest due to the AdS/CFT correspondence and to the
broader fluid-gravity duality (e.g., [59, 81]), but they will not be discussed here.
4.3 McVittie Solution 109

4.3 McVittie Solution

The McVittie spacetime discovered in 1933 [119] generalizes the Schwarzschild-de

Sitter-Kottler solution and is commonly interpreted as describing a central object
embedded in an FLRW space which, in general, is not de Sitter. Therefore, the
geometry in the region between the black hole and the cosmological horizon is time-
dependent. It was shown long ago that the McVittie spacetime is the only perfect
fluid solution of the Einstein equations which is spherically symmetric, shear free,
and asymptotically FLRW [141]. In spite of much work [1, 3, 38, 85, 92, 94, 123,
124, 130–132, 161], the McVittie spacetime still eludes our full understanding. A
simplifying assumption stated explicitly by McVittie in constructing his solution
is the no-accretion condition G10 D 0 (in spherical coordinates). According to the
Einstein equations this condition, equivalent to T01 D 0, excludes the radial flow
of cosmic fluid (although such flow would realistically occur when a spherical
local overdensity alters the “background”). Generalizations of the McVittie solution
allowing for the possibility of radial flow of energy are discussed later.
McVittie [119] intended to elucidate the extent of the effects of the cosmological
expansion on local gravitationally bound systems. Other approaches to this problem
generated various solutions of the Einstein equations, such as the Swiss-cheese
model of Einstein and Straus [44, 45]. The issue of cosmological expansion versus
local dynamics has been debated extensively but is not completely solved (see
the review in Ref. [28]). A complication (and an opportunity for us) is that,
unlike the Schwarzschild-de Sitter-Kottler spacetime, black holes in general FLRW
“backgrounds” are dynamical.
The McVittie line element is
 2
 
1  m.t/
2Nr m.t/ 4  2 
ds2 D   2 dt 2
C a2
.t/ 1 C dNr C rN 2 d˝.2/
2
(4.7)
2Nr
1 C m.t/
2Nr

in isotropic coordinates, where the McVittie no-accretion condition1 G10 D 0 [119]

dictates that the function m.t/ satisfy

P
m aP
C D 0; (4.8)
m a
hence it is
m0
m.t/ D ; (4.9)
a.t/

1
This is hypothesis e) of Ref. [119].
110 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

with m0 a non-negative constant. The physical mass of the central object is

the Misner-Sharp-Hernandez mass, while m.t/ is just a metric coefficient in the
particular coordinates adopted. We can then write the McVittie line element in
isotropic coordinates as
h i2
1  2Nrma.t/
0
m0 4  
ds2 D  h 2 2
i2 dt C a .t/ 1 C dNr2 C rN 2 d˝.2/
2
: (4.10)
2N
r a.t/
1 C 2Nrma.t/
0

The McVittie spacetime becomes the Schwarzschild one written in isotropic

coordinates if a  1, and it reduces to the FLRW metric when m vanishes. Apart
from the special case of a de Sitter “background”, the line element (4.7) is singular
on the 2-sphere rN D m=2 (which reduces to the Schwarzschild horizon if a  1)
[58, 130–132, 161]. This singularity is spacelike [130–132] and there is another
spacetime singularity at rN D 0. McVittie originally interpreted the metric (4.7) as
describing a point mass at rN D 0 but, in general, this point mass would be surrounded
by the rN D m=2 singularity, the interpretation of which is elusive [58, 130–132, 161].
According to Nolan [130], this singularity is weak in the sense that an object
falling across rN D m=2 is not crushed to zero volume and, therefore, the energy
density of the surrounding fluid must be finite. However, it is undeniable that the
pressure of this fluid
"  #
1 P 1C m
2H
2 2Nr
PD 3H C (4.11)
8 1  2N
m
r

diverges at rN D m=2 together with the Ricci scalar R D 8 .  3P/ [58, 117,
118, 130–132, 161], violating the weak and null energy conditions (but not the
positivity of the energy density) in a neighbourhood of the singularity. The de Sitter
“background” is an exception: in this case it is H P D 0 identically and the second
term on the right hand side of Eq. (4.11), which causes P to diverge, is absent
[51, 130–132].
Let us rewrite the McVittie line element (4.7) in terms of the areal radius
 m 2
R  a.t/Nr 1 C I (4.12)
2Nr
the differentials dNr and dR satisfy the relation
 m  m m P  m m
dR D 1 C aNr H 1 C C dt C a 1 C 1 dr :
2Nr 2Nr rN 2Nr 2Nr

Now Eq. (4.8) yields

 m m P  m
H 1C C DH 1 (4.13)
2Nr rN 2Nr
4.3 McVittie Solution 111

and
dR
dNr D     HNrdt : (4.14)
a 1C m
2Nr
1 m
2Nr

Upon use of the identity

 2
1  2N
m
2m0
r
D1 ; (4.15)
1 C 2N
m
r
R

substitution into Eq. (4.7) and manipulations yields2

 
2 2m0 dR2 2HR
ds D  1   H R dt2 C
2 2
2m0
q dtdR C R2 d˝.2/
2
;
R 1 R 1 R 2m0

(4.16)
where H  aP =a is the Hubble parameter of the FLRW “background”. In order to
remove the cross-term in dtdR from the line element we define a new time coordinate
T.t; R/ with the differential relation

1
dT D .dt C ˇdR/ ; (4.17)
F

where F.t; R/ is an integrating factor and the function ˇ.t; R/ must be determined
so that in the new coordinates the time-radius component of the metric vanishes. dT
is an exact differential if the 1-form (4.17) is closed,

@F @ˇ @F
D F Cˇ : (4.18)
@R @t @t

The substitution dt D FdT  ˇdR in Eq. (4.16) yields

 
2M
ds2 D  1   H 2 R2 F 2 dT 2
R
2 3
 
6 2M 1 2ˇHR 7 2
C 4 1   H 2 R2 ˇ 2 C Cq 5 dR
R 1  2M
R 1  2M
R
2 3
 
6 2M 2 2 HR 7 2 2
C2F 4 1  H R ˇ q 5 dTdR C R d˝.2/ : (4.19)
R 1  2M
R

2
Use m=Nr D ma=R D m0 =R, where ma is constant.
112 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

We now choose
HR
ˇ.t; R/ D q  ; (4.20)
2m0
1 R
1  2mR 0  H 2 R2

and the line element becomes

 
2 2m0 dR2
ds D  1   H R F 2 dT 2 C
2 2
2m0
C R2 d˝.2/
2
: (4.21)
R 1 R
 H 2 R2

For a de Sitter “background” with H D constant, the integrating factor F can be

set to unity, recovering the Schwarzschild-de Sitter-Kottler metric (4.1): clearly
the latter is a special case of the McVittie metric (4.21). In the new coordinates,
the spacetime singularity rN D m=2 corresponds to R D 2m a.t/ D 2m0 and does not
expand with the cosmic substratum.
Using Eqs. (4.21) and (2.92), the Misner-Sharp-Hernandez mass of a sphere of
symmetry with proper radius R is found to be

H 2 R3 4
MMSH D m0 C D m0 C R3 ; (4.22)
2 3
which is interpreted as a time-independent contribution m0 from the central object
plus the mass of the cosmic fluid contained in the sphere. Except for the static de
Sitter case, the Misner-Sharp-Hernandez mass of the sphere changes due to the fact
that the contribution from the cosmic fluid changes (because the radius R changes in
time, or because the density of the fluid itself varies with time, or both). In particular,
apparent horizons are not comoving and the mass enclosed by a black hole apparent
horizon (when this exists) changes even in the absence of accretion because of the
time evolution of both the areal radius RAH of the apparent horizon and of the time
evolution of the density .t/.

4.3.1 Apparent Horizons

We now restrict ourselves to a spatially flat FLRW “background” for simplicity

and we assume a perfect fluid stress-energy tensor asymptotically. The Einstein
equations then provide the energy density and the pressure P of the McVittie
fluid source. The density is the same as for the FLRW “background”,

3 2
.t/ D H .t/ : (4.23)
8
The “background” perfect fluid can be assigned any equation of state. For the sake
of illustration, however, we consider only a cosmic fluid which reduces to a timelike
4.3 McVittie Solution 113

dust at spatial infinity, with equation of state parameter w D 0. The pressure is then
[123, 124]
0 1
B 1 C
P .t; R/ D .t/ @ q  1A : (4.24)
2m0
1 R

The gRR D 0 recipe locates the apparent horizons of the McVittie metric and
gives

H 2 .t/ R3  R C 2m0 D 0 : (4.25)

This cubic equation is nothing but the Schwarzschild-de Sitter-Kottler horizon

condition (4.2) but now with a time-dependent Hubble parameter H.t/. The radii of
the time-dependent apparent horizons are labelled R1 .t/ and R2 .t/ and correspond
to the solutions R1;2 of Eq. (4.2) with the replacement H ! H.t/. Therefore, the
location of the apparent horizons of the McVittie spacetime varies with the cosmic
time.
As for the Schwarzschild-de Sitter-Kottler case, both p horizons exist if 0 <
sin.3 / < 1, which corresponds to m0 H.t/ < 1=.3 3/. However, contrary to
the Schwarschild-de Sitter-Kottler space with constant Hubble parameter, this
inequality is only satisfied at certain times,pbut not at other times, during the
pcosmic
history. There is a unique instant t D 2 3 m0 at which m0 H.t/ D 1=.3 3/ in a
dust-dominated “background” with H.t/ D 2=.3t/. Three possibilities occur:
1
• Early on, as t < t , it is m0 > p and both R1 .t/ and R2 .t/ are complex.
3 3 H.t/
There are no apparent horizons.
1
• At the time t it is m0 D p and two apparent horizons R1 .t/ and R2 .t/
3 3 H.t/
1
coincide at the real physical location R1 D R2 D p .
3 H.t/
1
• At late times t > t , we have m0 < p and both R1 .t/ and R2 .t/ are real.
3 3 H.t/
There are then two distinct apparent horizons.
Figure 4.1 illustrates the qualitative dynamics of the McVittie apparent horizons.
As there are no apparent horizons at early times t < t , a naked singularity
raises its head at R D 2m0 , where the Ricci scalar and the pressure diverge. In
fact, the Hubble parameter H.t/ diverges in the early universe and the mass m0
1
stays supercritical, with m0 > p . It seems appropriate to interpret this naked
3 3 H.t/
singularity on the lines of the one in the Schwarzschild-de Sitter-Kottler spacetime:
a black hole horizon cannot be accommodated in a “universe” which is too small to
contain it. At these early times the interpretation of the “background” as a “universe”
114 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

Black Hole Horizon

Cosmological Horizon

2m

0
t∗ t

Fig. 4.1 The radii of the McVittie black hole (continuous) and cosmological (dotted) apparent
horizons versus comoving time in a dust-dominated FLRW “background”. Time t and radius R are
in units of m and we arbitrarily fix m D 1

is not granted because the McVittie spacetime represents neither an isolated object
nor a FLRW universe; at early times it is an inhomogeneous spacetime which is
drastically different from both.
At the critical time t a black hole apparent horizon appears simultaneously with,
and coinciding with, a cosmological apparent horizon

1
R1 .t / D R2 .t / D p : (4.26)
3 H.t /

In the dust-dominated cosmological “background” it is easy to compute the radius of

this horizon as R1 D R2 D 3m0 . This extremal situation resembles the Nariai black
hole of the Schwarzschild-de Sitter-Kottler case, but it is instantaneous.
Later on, for t > t , the extremal horizon bifurcates into a black hole apparent
horizon surrounded by a cosmological horizon (both evolving in time). The
black hole apparent horizon contracts and asymptotes to the spacetime singularity
R D 2m0 from above as t ! C1. The cosmological apparent horizon expands
monotonically, its radius approaching 1=H.t/.
The singularity R D 2m0 has been discussed extensively [94, 123, 124, 130–132].
The surface of equation f .R/  R  2m0 D 0 has normal vector N D r f D ı1 ,
which has norm squared

Na N a D gab Na Nb jRD2m0 D 4m20 H 2 .t/ < 0 : (4.27)

As N c is timelike, the singularity R D 2m0 is spacelike. The curvature scalar

4.3 McVittie Solution 115

0 1
B 3 C
R D 8 Taa D 8 .  3P/ D 8 .t/ @4  q A (4.28)
2m0
1 R

diverges as R ! 2mC 0 . The two spacetime regions R < 2m0 and R > 2m0 are dis-
connected and separated by the R D 2m0 singularity [130–132], with the geometry
of the outer region described by the line element (4.21).p
At the critical time t , when R1 .t p
/ D R2 .t / D 1=. 3H.t //, the normal to the
surface of equation F .R/  R  1=. 3H.t // D 0 is M  r F D ı1 and
 
11 1 2 p 
M Mc D g
c
RD p D 1  3 m0 H.t / D 0 (4.29)
3 H.t / 3

and the instantaneous extremal apparent horizon is null.

One can compare the time rate of change of the apparent horizon radii with that
of the cosmic substratum by differentiating Eq. (4.25) and solving for RP AH , which
yields

2H HRP 3
RP AH D  AH
(4.30)
3H 2 R2AH  1

and
!
RP AH 2HRP 2
 H D H 1 C AH
: (4.31)
RAH 3H 2 R2AH  1

The apparent horizons are not comoving, except for trivial cases.3 Even in a pure
(spatially flat) FLRW universe (obtained in the limit m D 0), the cosmological
apparent horizon at RAH .t/  Rc .t/ D 1=H.t/ is not comoving, as discussed in
Chap. 3. If an entropy can be ascribed to apparent horizons in General Relativity
by the S D A =4 prescription, then a natural question would be whether the total
area of the McVittie apparent horizons is non-decreasing in time. The area A1 of the
black hole apparent horizon is decreasing, but it is bounded from below by 16 m20 ,
while this behaviour is more than compensated for by the increase of the area A2 of
the cosmological apparent horizon. The total area
 2  A A1 C A2
S D S1 C S2 D R1 C R22 D D (4.32)
4 4
is plotted in Fig. 4.2.
Since the apparent horizons emerge as a pair at t D t , the horizon entropy S
exhibits a discontinuous jump from zero value at this critical time.

3
Also the known analytic solutions describing wormholes embedded in cosmological “back-
grounds”, which are few, show that these wormholes evolve in time [10, 50, 109, 160].
116 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

k B c3
Fig. 4.2 The putative total horizon entropy S (in units , where kB is the Boltzmann constant)
„G
associated with the apparent horizons as a function of time

4.3.2 Phantom McVittie Spacetime

Another possibility is that of a “background” FLRW universe sourced by a phantom

fluid [55], defined by the equation of state parameter w  P= < 1 and violating
the weak energy condition. Phantom fluids have raised much interest in relation with
the present acceleration of the universe [5]. A notable feature is that they cause a
Big Rip singularity at a finite time in the future trip [25]. The scale factor solving the
Friedmann equation for a spatially flat FLRW universe with phantom scalar field is
[25]
a0
a.t/ D   2 ; (4.33)
trip  t 3jwC1j

where w < 1, a0 , and trip are constants. The Hubble parameter

2 1
H.t/ D ; (4.34)
3jw C 1j trip  t

is the time-reverse of that of a dust-dominated universe, H.t/ D 2=.3t/. While

the latter diverges at the Big Bang singularity and gradually decreases to zero,
the former is finite at t D 0 and increases monotonically, diverging at the Big
Rip. The McVittie apparent horizons describing black holes embedded in a FLRW
phantom universe also behave as the time-reversal of those embedded in a FLRW
“background” with w > 1, as shown in Fig. 4.3 [55].
4.3 McVittie Solution 117

Fig. 4.3 The proper radii of the McVittie black hole (continuous) and cosmological (dotted)
apparent horizons versus comoving time in a phantom-dominated universe. The parameter values
are w D 1:5 and trip D 0

The evolution of the apparent horizons in an expanding phantom universe

proceeds as follows. Early on, there exist a black hole and a cosmological apparent
horizon, which are approximately located at R D 2m0 and R D 1=H.t/, respectively.
As the universe grows older, the cosmological apparent horizon contracts while
the black hole one expands, until the critical time t at which the two apparent
horizons merge. At t > t they both disappear exposing a naked singularity. The
total apparent horizon area decreases and jumps discontinuously to zero value at the
time t . This phenomenology is pretty bizarre, but it reflects the weirdness of the
phantom fluid, which seems to violate the second law of thermodynamics in many
ways [15, 70, 80, 104, 122, 127, 128] and may turn out to be completely unphysical.

4.3.3 Nolan Interior Solution

The Nolan interior solution [129] describes a relativistic star of uniform density in
a FLRW “background” and provides, at least formally, a possible source for the
McVittie metric. It mimics in the FLRW context the Schwarzschild interior solution
with a Minkowski “background”. The Nolan line element in isotropic coordinates is
2   32  6
r2
1  rNm0 C mN 1  m
1 C 2Nmr0  
6 rN 3 4N
r0 7 2
ds2 D  4  0  5 dt C a2
.t/  2
dN
r 2
C N
r 2
d˝ 2
.2/
r2
1 C 2Nmr0 1 C mN 2Nr3 1 C mNr3
2
0 2Nr0
(4.35)
118 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

where rN0 is the star radius, m.t/ satisfies Eq. (4.8) (the condition forbidding accretion
onto the star surface), and 0  rN  rN0 . The interior metric is regular at the centre
and is matched to the exterior McVittie metric at rN D rN0 through the Darmois-
Israel junction conditions. The pressure is continuous at the surface rN D rN0 while
the otherwise uniform energy density has a discontinuity there; they are [129]
2 3
1 6 2 6m 7
.t/ D 43H C  6 5 ; (4.36)
8
a2 rN03 1 C 2Nmr0
2   
mNr2
1 6 1 C m
2Nr0
1 C 2Nr03
P .t; rN / D 2
43H  2H P   2
8 1  rN0 C 1  4Nr0 mN
m m r
rN3 0

  3
3m2 rN2
1
rN04 rN02 7
C  6 h   i5 : (4.37)
mNr2
a2 1 C m
2Nr0
1 m
rN0
C 1 m
4Nr0 rN03

The Nolan interior solution (a member of the Kustaanheimo-Qvist family of

shear-free solutions [93]) embeds the Schwarzschild interior solution with uniform
constant density [169] in a time-dependent FLRW “background”. The Schwarz-
schild interior solution is recovered if a  1. The energy density is positive-definite
and the condition P  0 imposed in Ref. [129] coincides with aR C 3Pa2 =2 < 0.
By continuity, if ˙0 .t/ D f.t; rN ; ; '/ W rN D rN0 g is the star surface at time t, the
metric on this 2-sphere must coincide with the restriction of the McVittie metric to
this sphere
 2
1 m.t/  
2Nr0 m.t/ 4 2 2
ds2 j˙0 D   2 dt 2
C a2
.t/ 1 C rN0 d˝.2/ : (4.38)
m.t/ 2Nr0
1C 2Nr0

The proper area of the star surface ˙0 is

Z Z  4
p 2 m.t/
A˙0 .t/ D dd' g˙0 D 4 a .t/ rN02 1C ; (4.39)
˙0 2Nr0

where gab j˙0 is the metric on ˙0 at time t and g˙0 is its determinant. Upon use of
 m 2
the proper radius R  a.t/Nr 1 C , one finds A˙0 .t/ D 4 R20 . The star surface
2Nr  
m 2
comoves with the cosmic fluid and has areal radius R0 .t/ D a.t/ rN0 1 C .
2Nr0
4.3 McVittie Solution 119

The Nolan interior solution provides an example of a local, strongly gravitationally

bound, relativistic object which is comoving: in this case the cosmic expansion
prevails over the local dynamics.
Following [51], let us compute the generalization of the Tolman-Oppenheimer-
Volkoff equation [169] for the Nolan model. The covariant conservation r b Tab D 0
of a perfect fluid with energy-momentum tensor Tab D .P C / ua ub C Pgab splits
into the two equations [169]

uc rc C .P C / r c uc D 0 ; (4.40)
h a @c P C .P C / u rc ua D 0 ;
c c
(4.41)

where ua is the fluid 4-velocity and hab  gab C ua ub defines the projection operator
ha c onto the 3-space orthogonal to ua . Since u / ı 0 in comoving coordinates and
uc uc D 1, we have
  
mNr2
1C m
2Nr0
1C 2Nr03
u D u ı0 D    .1; 0; 0; 0/ (4.42)
mNr2
1 m
rN0
C 1 m
4Nr0 rN03

(or u D jg00 j1=2 ). Then Eqs. (4.40) and (4.8) give

8 2 39
<  2 1 =
2
@ m 3 rN mNr
C 3H .P C / 1  4    2 1C 3 5 D 0: (4.43)
@t : rN0 2 1 C m 2Nr0 2Nr0 ;
2Nr0

Equation (4.43) is a more general form of the usual FLRW conservation equation
P C 3H .P C / D 0, which is reobtained if m vanishes identically. There is no
analogue of Eq. (4.43) for the Schwarzschild interior solution which has H D 0 and
static energy density.
We can put Eq. (4.41) in the form

@c P C uc ub @b P C .P C / ub rb uc D 0 I (4.44)

for c set to 1, the computation of the covariant derivative yields

 
@P mNr 1C m
2Nr0
C .P C / 3  h  i D 0 : (4.45)
@r rN0 1 C mNr2
1 m
C mNr2
1 m
2Nr03 rN0 rN03 4Nr0

In the Newtonian limit m=Nr; m=Nr0 1, P , r ' rN [51], this equation reduces to

@P d˚N
C D 0; (4.46)
@r dr
120 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

where
 1
4 3
Dm r (4.47)
3 0

and
mNr2
˚N D (4.48)
2Nr03

is the Newtonian potential. The density (4.47) can also be obtained by setting the
scale factor to unity in Eq. (4.36) and using the curvature radius. The equation of
hydrostatic equilibrium receives the first order correction

dP d˚N 3
C 1 Œ˚N .r/  ˚N .r0 / D 0 : (4.49)
dr dr 2

4.4 Charged McVittie Spacetime

A charged version of the original McVittie metric was found by Shah and Vaidya
[150] and later generalized by Mashhoon and Partovi [115]. Charged and uncharged
McVittie solutions are special cases of the Kustaanheimo-Qvist family [93]; related
solutions and relevant research are reviewed in Chap. 4 of Ref. [92]. The charged
McVittie solution was rediscovered by Gao and Zhang [64], who also generalized it
to higher dimensions [65], and was studied in [56, 116, 117]. Conformal diagrams
of the McVittie spacetime for various “backgrounds” were obtained in [38, 94, 95,
100].
Restricting again to a spatially flat FLRW “background”, the charged McVittie
line element and the only nonzero component of the Maxwell tensor are4
2
.m20 Q2 /
1 4a2 rN2
ds2 D  h  i2 dt
2
m0 2 Q2
1C 2aNr
 4a2 rN2

 m0 2 Q2
2  
2
Ca .t/ 1 C  2 2 dNr2 C rN 2 d˝.2/
2
; (4.50)
2aNr 4a rN
Q
F 01 D h i2 ; (4.51)
.m2 Q2 /  m0 2
 Q2
a2 rN 2 1  4a2 rN2 1C 2aNr
 4a2 rN2

4
A typographical error is present in the numerator of g00 in Ref. [64], but the line element appears
correctly in the later references [65, 116, 117].
4.4 Charged McVittie Spacetime 121

where rN is the isotropic radius and m0 > 0 and Q are the usual mass parameter and
an electric charge parameter, respectively. If a is set to unity the line element (4.50)
reduces to the Reissner-Nordström one in isotropic coordinates. As rN ! C1, it
reduces again to the spatially flat FLRW line element, and if both m0 and Q are set
to zero, the spacetime is exactly the spatially flat FLRW one.
The areal radius is now
" 2 #
m0 Q2
R.t; r/ D a.t/Nr 1 C  2
2a.t/Nr 4a .t/Nr2

m20  Q2
D m0 C a.t/Nr C ; (4.52)
4a.t/Nr
with R .t; rN /  m0 for jQj  m0 .q
When jQj  m0 , the function R .t; rN / decreases
from C1 in the range 0 < aNr < m20  Q2 =2, reaches an absolute minimum
q
Rmin D m0 C m20  Q2
q q
at aNr D m20  Q2 =2, and increases again to C1 for aNr > m20  Q2 =2 because
5
the isotropic
q radius corresponds to a double covering of the spacetime region
R > m0 C m20  Q2  m0 . When jQj  m0 , the areal radius R increases mono-
jQj  m0
tonically with rN and the physical region R  0 corresponds to rN   0.
2a.t/
In the following it will be useful to invert the relation between areal and isotropic
radii, which yields
 2 
.R  m0 / m0  Q2
rN 2  rN C D 0: (4.53)
a 4a2
The positive root obeys
p
2aNr D R  m0 C R2 C Q2  2m0 R  f .R/ : (4.54)

The Ricci scalar is [56]

6 aR  m0 2 Q2
RD 1C  2 2
.m20 Q2 / a 2aNr 4a rN
1 4a2 rN2
"   #)
2 m0 3 Q2  m20
CH 1 C (4.55)
aNr 4a2 rN 2

5
This fact is well known in the special case a  1; Q D 0 corresponding to the Schwarzschild
spacetime [19, 170].
122 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

and the non-vanishing components of the Einstein tensor are6

256 Q2 a4 rN 4 3Pa2
Gt t D  C ; (4.56)
Œm20  Q2 C 4m0 aNr C 4a2 rN 2 4 a2
256 Q2 a4 rN 4
GrN rN D
Œm20  Q2 C 4m0 aNr C 4a2 rN 2 4
aP 2 .5m20 C 5Q2  8m0 aNr C 4a2 rN 2 /
C
a2 .m20 C Q2 C 4a2 rN 2 /
2Ra.m20  Q2 C 4m0 aNr C 4a2 rN 2 /
C ; (4.57)
a.m20 C Q2 C 4a2 rN 2 /
256 Q2 a4 rN 4
G  D G' ' D 
Œm20  Q2 C 4m0 aNr C 4a2 rN 2 4
aP 2 .5m20 C 5Q2  8m0 aNr C 4a2 rN 2 /
C
a2 .m20 C Q2 C 4a2 rN 2 /
2Ra.m20  Q2 C 4m0 aNr C 4a2 rN 2 /
C : (4.58)
a.m20 C Q2 C 4a2 rN 2 /

As in the McVittie case, one has G01 D 0, which forbids the radial flow of cosmic
fluid. For m0 D Q D 0 Eq. (4.55)reduces to the Ricci scalar of the spatially flat
FLRW universe R D 6 H P C 2H 2 . For a  1 it reduces to zero (the Ricci scalar of
the Reissner-Nordström metric: then the only matter source is the electromagnetic
field with traceless energy-momentum tensor). If jQj  m0 there is a spacetime
singularity at
q
m20  Q2
aNr D ; (4.59)
2
corresponding to
q
R D m0 C m20  Q2 ; (4.60)

which divides again the spacetime into two disconnected parts. It occurs where
the outer event horizon of the Reissner-Nordström spacetime would be without
FLRW “background”. In the extremal case jQj D m0 the singularity occurs at rN D 0
or R D m0 , again coinciding with the location of the outer event horizon of the
Reissner-Nordström limit (the event horizon of the extremal Reissner-Nordström
black hole).

6
The Einstein tensor appearing in Refs. [117] and [116] misses a scale factor in the denominators.
4.4 Charged McVittie Spacetime 123

This new spherical singularity is not present if jQj > m0 , but then the invariant
of the Maxwell tensor

Q2
Fab F ab D  h  i2 (4.61)
m0 2 Q2
a2 rN 4 1 C 2aNr
 4a2 rN2

is singular at

jQj  m0
aNr D ; (4.62)
2
corresponding to R D 0, because the radial electric field (the only non-zero compo-
nent F 01 of F ab ) is singular. The Big Bang singularity a D 0 is also present.
The spacetime singularity (4.60) for jQj  m0 is spacelike, because it is described
by the equation D 0, where
q
m20  Q2
.t; rN /  a.t/Nr  I (4.63)
2
and
h  i2
m0 2 Q2
1C 2aNr
 4a2 rN2
rc r c D Pa2 rN 2 2
.m20 Q2 /
1 4a2 rN2

1
C h  i2 I (4.64)
m0 2 Q2
1C 2aNr
 4a2 rN2

q C
in the limit 2aNr ! m20  Q2 , this expression tends to 1. The norm of the
normal to the surfaces D constant is negative, and this surface and its limit are
spacelike [56, 115].
The location of the apparent horizons is given by r c Rrc R D 0, which reads
8 "  2  #2 9
<  m0 2 Q2
4
m0  Q2 =
aP 2 rN 2 1 C  2 2  1
: 2aNr 4a rN 4a2 rN 2 ;

 m0 2 Q2
2
 1C  2 2 D 0; (4.65)
2aNr 4a rN
124 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

which gives
" 2  #2
 m0 2 Q2
4
m0  Q2
aP 2 rN 2 1C  2 2 D 1 (4.66)
2aNr 4a rN 4a2 rN 2
or
 2
4f 2 R4 H 2  f 2  m20 C Q2 D 0 ; (4.67)

with f .R/ as in Eq. (4.54). We then have to solve the quartic

H 2 R4  R2 C 2m0 R  Q2 D 0 : (4.68)

For large R, this equation reduces to the radius of the spatially flat FLRW
“background” R H 1 . In any regime in which H ! 0 (for example for a power-
law scale factor a.t/ D a0 tp ), Eq. (4.67) becomes asymptotically

R2  2m0 R C Q2 ' 0 (4.69)

so that, in this limit

q
R D m0 ˙ m20  Q2 : (4.70)

In any region of a background for which H ! 0, a black hole apparent horizon

asymptotes to the singularity (4.60). The smaller root is always covered by the
spherical singularity and these two roots coincide with the locations of the two event
horizons of the Reissner-Nordström spacetime. The limit H ! 0 reproduces the
horizon structure of Reissner-Nordström (but with the spherical singularity added).
This spherical singularity persists when jQj D m0 and is time-independent (as long
as the scale factor a it is not exactly constant).
2
We restrict again to a dust-dominated FLRW substratum with H.t/ D and we
3t
begin with the case jQj < m0 . Then Eq. (4.68) can be solved explicitly at different
times t giving the location of the apparent horizons. The dynamics of the apparent
horizon radii in comoving time is qualitatively the same that we have seen for
the uncharged McVittie spacetime [55, 56, 85, 94, 95, 130, 131]. The innermost
apparent horizon is located in the inner disconnected region, which is separated
from the external geometry by the singularity (4.60) (which converged to R D 0
at the Big Bang). This innermost horizon asymptotes to the location of the inner
unstable Cauchy horizon of the Reissner-Nordström geometry.
In the extremal case jQj D m0 the quartic (4.68) is solved exactly, with roots
p p
1 1  4m0 H 1 1 C 4m0 H
Rextremal  ˙ ; ˙ : (4.71)
2H 2H 2H 2H
Explicit analytic expressions for the apparent horizon radii are rare to find in studies
of time-evolving black holes [49].
4.5 An Application to the Quantization of Black Hole Areas 125

The root
p
1 1 C 4m0 H
R D   (4.72)
2H 2H
is always negative and unphysical; the smallest positive root
p
1 1 C 4m0 H
Rinner   C (4.73)
2H 2H
always exists, while the other two roots
p
1 1  4m0 H
R˙  ˙ (4.74)
2H 2H
can merge (become complex) or appear simultaneously (become real) depending on
the evolution of H. In a dust-dominated FLRW “background”, H tends to zero at
late times and Rinner and R in fact converge to the same radius R D m0 , which is
the location of the single event horizon of the extremal Reissner-Nordström black
hole.
In the supercritical case jQj > m0 we have a naked singularity, as expected since
the limit a  1 reproduces the Reissner-Nordström spacetime. There is a cosmolog-
ical horizon, given by the only root of Eq. (4.68) which is real and positive [56].
In the Reissner-Nordström black hole (to which the charged McVittie spacetime
reduces if a  1) there are an outer event horizon and an inner (apparent) horizon.
However, for the relevant range of parameters jQj  m0 , there are only one
black hole apparent horizon and one cosmological apparent horizon. This fact is
interpreted as follows [56]. It is well known that the inner horizon of the Reissner-
Nordström black hole is unstable with respect to linear perturbations [138] and
the cosmological “background” perturbs the central inhomogeneity, only this is a
non-linear (or exact) “large perturbation”. It is conjectured in [56] that such an
horizon will not appear in all exact solutions of General Relativity which describe a
Reissner-Nordström black hole interacting with non-trivial environments.

4.5 An Application to the Quantization of Black Hole Areas

As an application, the charged McVittie spacetime was used7 to disprove the

universality of certain quantization laws for quantities constructed with the areas
of black hole apparent horizons and inspired by string theories [53].

7
The analysis of Ref. [53] makes use of the line element of [64] which contains an error but the
qualitative behaviour of the apparent horizons for jQj  m0 does not change and the argument of
Ref. [53], which is qualitative, is still valid.
126 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

The research community working on the holographic principle and stringy or

supergravity black holes was excited by the observation that products of Killing
horizon areas for certain multi-horizon black holes are independent of the black hole
mass—they depend only on quantized charges (supergravity and extra-dimensional
black holes with angular momentum and electric and magnetic charges were
considered) [31, 35–37, 63, 99, 120]. This literature, inspired by the holographic
principle and string theories (although the results are not strictly derived from
string theories), betrays the idea that quantized products of areas depending on
combinations of integers must carry the signature of some specific microphysics.
If the entropy S of an horizon is given by the S D A =4 prescription, statistical
mechanics based on microscopic models counts microstates determined by quantum
gravity and the horizon area should be quantized. When there are outer (C) and inner
() horizons, the quantization rules appearing in the literature are
p p 
2
A˙ D 8 lPl N1  N2 N1 ; N2 2 N ; (4.75)

or
 2 2
AC A D 8 lPl N; N 2 N; (4.76)

where lPl is the Planck length [31, 35–37, 63, 99, 120]. N1;2 and N are integers
for supersymmetric extremal black holes, and depend on the number of branes,
antibranes, and strings in more complicated situations [79]. A weaker rule states
that the product of horizon areas is independent of the black hole mass and depends
only on the quantized charges. These rules are often reported as universal ones
valid for all black holes with multiple horizons. However, a warning about uni-
versality was issued by Visser [165, 166]. He studied black holes in 4-dimensional
General Relativity and found that products of areas do not give mass-independent
quantities, and they are not related in any simple way to integers. Instead, specific
quadratic combinations of the various horizon radii (with the dimension of an area)
generate mass-independent quantities and are, presumably, the best candidates to
be quantized [165, 166]. It is essential to include in these algebraic combinations
both cosmological and virtual horizons, in addition to physical black hole horizons
[166]. (Virtual horizons are negative or imaginary roots of the equation locating the
horizons.)
The horizons considered in the literature are Killing (and event) horizons.
Realistic fundamental black holes cannot be stationary because, already at the
semiclassical level, they emit Hawking radiation and the backreaction changes
their masses, which become time-dependent, together with their horizon areas. For
astrophysical black holes the effect is completely negligible but this cannot be the
case for quantum black holes. Then, there will be no timelike Killing vector and
apparent horizons should be considered instead of Killing and event horizons.
Visser’s discussion of the Schwarzschild-de Sitter-Kottler black hole [165, 166]
can be repeated almost without changes: since the calculations performed in these
4.6 Generalized McVittie Spacetimes 127

works are algebraic, the only change is that the constant Hubble parameter of
the Schwarzschild-de Sitter-Kottler spacetime is replaced by the time-dependent
H.t/ of a FLRW “background”. Including the virtual horizon in the count, it is
straightforward to see that the quantities

1
RV .RBH C RC / C RBH RC D  (4.77)
H 2 .t/

and
1
.RBH C RC /2  RBH RC D (4.78)
H 2 .t/

are independent of the black hole mass m. This situation is a special case
[166]: mass-independent combinations of apparent horizon radii exist whenever
the Misner-Sharp-Hernandez mass is a Laurent polynomial of the areal radius R.
This is clearly the case of the uncharged and charged McVittie spacetimes. Taking
the McVittie spacetime as an example, the physical mass contained in a sphere is
the Misner-Sharp-Hernandez one and the cosmic fluid serves the only purpose of
generating a cosmological “background” to make the black hole dynamical. It seems
that the relevant mass to consider when mass-independent quantities such as (4.77)
and (4.78) are searched for is the black hole contribution m, not the total MMSH . In
any case, the radii of the apparent horizons identify different spheres and correspond
.i/ .i/
to different Misner-Sharp-Hernandez masses MMSH D 2RAH . Here we stick to m.
Following [166], we include the virtual horizon to obtain mass-independent
quantities. Now, when the apparent horizons are time-dependent, the combina-
tions (4.77) and (4.78) are also time-dependent. Even if they are expressed by
combinations of integers at an initial time, they will not be such immediately
afterward, and at all other times.
Although the cosmological black holes that we consider are just toy models for
non-stationary black hole horizons, the point is that realistic black holes are time-
dependent and far-reaching conclusions about quantizing black hole horizon areas,
or quantities quadratic in the radii of Killing horizons, are unwarranted. Generic
statements should be put on a firmer ground before being promoted to the role
of universal results. Our dynamical cosmological black hole examples reinforce
the argument of Refs. [165, 166] that the black holes of 4-dimensional General
Relativity do not reconcile with the quantization rules (4.75) and (4.76).

4.6 Generalized McVittie Spacetimes

The McVittie no-accretion condition can be relaxed to allow for generalized

McVittie solutions [51]. In the Synge approach, a metric can be prescribed and
forced to solve the Einstein equations by running the latter from left to right and by
128 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

computing the energy-momentum tensor corresponding to this solution. However,

this stress-energy tensor usually turns out to be completely unphysical and to violate
any reasonable energy condition, even the non-negativity of the energy density. It
comes as a surprise that generalized McVittie solutions with relatively well-behaved
matter sources exist.
Generalized McVittie solutions have line element

B2 .t; rN / 2  
ds2 D  dt C a2
.t/A4
.t; N
r / dN
r 2
C N
r 2
d˝ 2
.2/ (4.79)
A2 .t; rN /

in isotropic coordinates, where m.t/ is a positive function of time and

m.t/ m.t/
A .t; rN /  1 C ; B .t; rN /  1  : (4.80)
2Nr 2Nr
The Einstein tensor admits the only non-vanishing mixed components
 
3A2
aP mP 2
G00 D  C ; (4.81)
B2
a rN A
 
1 2m P
m aP
G0 D 2 2 5 C ; (4.82)
a rN A B m a
   
A2 d aP mP aP mP
G11 D G22 D G33 D 2 2 C C C
B dt a rN A a rN A
  
aP mP 2m P
 3 C C : (4.83)
a rN A rN AB

It is useful to consider the expression

aP mP PH m
m P B
C C D  (4.84)
a rN A mH m A

P H =mH , where
which appears in the Einstein equations and reduces to m

mH  m.t/a.t/ (4.85)

for the special subclass of solutions with m D m0 D constant referred to as

“comoving mass” solutions (this class will be of some importance later). On the
surface rN D m=2, C reduces to

aP P
m PH
m
C˙ D C D : (4.86)
a m mH
4.6 Generalized McVittie Spacetimes 129

McVittie solutions have C˙ D 0, while comoving mass solutions are characterized

by C D C˙ D H everywhere.
The Ricci scalar
 
3A2 P 2 2mCP
R D 2 2C C 4C C (4.87)
B rN AB

diverges as rN ! m=2 except in the case when m D constant. Distinct cosmic fluids
or effective fluids can be contemplated as matter sources of generalized McVittie
spacetimes, which we consider separately in what follows.

4.6.1 A Single Perfect Fluid

If the generalized McVittie metric is sourced by a single perfect fluid with stress
energy tensor

Tab D .P C / ua ub C Pgab ; (4.88)

 
a radial flow of cosmic fluid is described by the fluid 4-velocity u D u0 ; u; 0; 0 .
Then, the only possible solution in General Relativity is the Schwarzschild-de Sitter-
Kottler black hole [51, 66, 94]. In fact, using the normalization uc uc D 1, one finds

Ap
u0 D 1 C a2 A4 u2 (4.89)
B
and, using Eqs. (4.81)–(4.83),
p
P H D GB2 au .P C / A
m 1 C a2 A4 u2 ; (4.90)

where
Z Z
p
A D dd' g˙ D 4 a2 A4 rN 2 (4.91)

is the area of a sphere of isotropic radius rN and

 
AC 2 
3 D 8 .P C / a2 A4 u2 C ; (4.92)
B
 2  
A 2mCP 
2CP C 3C2 C D 8 .P C / a2 A4 u2 C P ; (4.93)
B rN AB
 2  
A P 2 2mCP
2C C 3C C D 8 P : (4.94)
B rN AB
130 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

Comparing Eqs. (4.93) and (4.94) gives P D  and the de Sitter equation of state
is the unique possibility; Eq. (4.90) then implies mP H D 0. A single perfect fluid
cannot source the generalized McVittie spacetime (with the trivial exception of
the non-accreting Schwarzschild-de Sitter-Kottler). A mixture of two perfect fluids
constitutes a potential source.8

4.6.2 Imperfect Fluid Without Radial Flow of Material

Another possibility is to allow an imperfect fluid with energy-momentum tensor

Tab D .P C / ua ub C Pgab C qa ub C qb ua (4.95)

to source the generalized McVittie spacetime. The vector qc is purely spatial9 and
describes a radial energy flow, but there is no flow of cosmic fluid:
 
A
u D ; 0; 0; 0 ; q˛ D .0; q; 0; 0/ ; qc uc D 0 : (4.96)
B

The .0; 1/ component of the Einstein equations is

P
m aP 4 2 2 4 2
C D rN a A B q : (4.97)
m a m
and
PH
m P
m aP
D C : (4.98)
mH m a
The area of a sphere ˙ of constant time and isotropic radius is
Z Z
p
A D dd' g˙ D 4 a2 A4 rN 2 ;

and the relation between energy flow, area A , and accretion rate

P H .t/ D aB2 A q
m (4.99)

holds true. An energy inflow is described by q < 0 and this condition can be written
on a sphere of radius rN m as

P H ' aA jqj :
m (4.100)

8
A mixture of two perfect fluids is the matter source for the Sultana-Dyer solution (Sect. 4.7),
which does not belong to the McVittie class.
9
In principle, one could take this vector to be spacelike instead of purely spatial.
4.6 Generalized McVittie Spacetimes 131

The .0; 0/ and .1; 1/ (or .2; 2/ or .3; 3/) components of the Einstein equations
provide the energy density and pressure10
 
3A2 aP mP 2
.t; rN / D C ; (4.101)
8 B2 a rN A
      
A2 d aP m P aP mP aP mP 2m P
P .t; rN / D 2 C C C 3 C C :
8 B2 dt a rN A a rN A a rN A rN AB
(4.102)

The energy density is always non-negative. The expansion scalar is 3C and

Eq. (4.102) provides a generalized Raychaudhuri equation,

3C2 P
m B2
CP D   C4 P; (4.103)
2 rN AB A2
2
P D  3H  4 P when
which reduces to the usual equation of FLRW cosmology H
2
m ! 0, and the Hamiltonian constraint H 2 D 8 =3 then yields

P D 4 .P C / :
H (4.104)

In the general case, Eq. (4.101) gives the more general equation

B2 P
mC
CP D 4 .P C /  : (4.105)
A2 rN AB

4.6.3 Imperfect Fluid with Radial Flow of Material

Yet another possibility is to have an imperfect fluid with energy-momentum

tensor (4.95) and to allow for both radial fluid flow and an energy current:
 p 
A 2 4 2
u D 1 C a A u ; u; 0; 0 ; q D .0; q; 0; 0/ : (4.106)
B

The Einstein components (4.81)–(4.83) then give

p
P H D aB2 A
m 1 C a2 A4 u2 Œ.P C / u C q ; (4.107)
 2
AC 8 
D .P C / a2 A4 u2 C ; (4.108)
B 3

10
Contrary to the McVittie spacetime, now depends also on the radial coordinate.
132 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

 2  
A 2mCP 
2CP C 3C2 C D 8 .P C / a2 A4 u2 C P C 2a2 A4 qu ;
B rN AB
(4.109)
 2  
A 2 P
mC
2CP C 3C2 C D 8 P : (4.110)
B rN AB

Subtracting the last two equations yields

u
q D  .P C / ; (4.111)
2
that is, an ingoing radial flow of mass is associated with an outgoing radial heat
current if P >  . The substitution of Eq. (4.111) into (4.107) provides the accretion
rate
aB2 p
PH D 
m 1 C a2 A4 u2 .P C / A u ; (4.112)
2
where .P C / A u is a flux of gravitating energy through the surface of area A .
Since u < 0 for inflow, mH increases if P C > 0, stays constant in a de Sitter
“background”, and decreases if phantom energy with P <  is accreted (which
lends support to the findings of Ref. [6] about the fate of a black hole in a phantom
universe11 ).
The energy density is
 
A2 2a2 A4 u2 P
mC
D 2
3C2 C CP C (4.113)
8 B 1 C a2 A4 u2 rN AB

and, in the special case m D m0 D constant, it reduces to

" #
A2 2 P
Ha 2 4 2
A u
D 3H 2 C : (4.114)
8 B2 1 C a2 A4 u2

P > 0. The velocity of the fluid is

It is positive in a superaccelerating universe with H
found to be
8r 91=2
ˆ
ˆ 4m20 H 2 a2 A4
 1>
>
< 1C B4 A 2 .PC /2
=
uD : (4.115)
ˆ 2a2 A4 >
>
:̂ ;

The fluid flow becomes superluminal as rN ! m0 =2, where B ! 0, an unphysical

feature probably due to the oversimplified hydrodynamical model. In principle, it

11
This study analyzes a test fluid in great detail and finds the same qualitative behaviour for the
mass of a black hole accreting cosmic fluid.
4.6 Generalized McVittie Spacetimes 133

does not make sense to study black holes in the presence of superluminal flows
because, then, it is obvious that an horizon will not confine energy. However we
restrict the flow to be inward, hence matter cannot flow superluminally outside of a
black hole apparent horizon.12 In reality, the matter/energy flow is subluminal and
becomes supersonic at a certain radius, a feature which can only be modeled in a
more realistic model of accretion.

4.6.4 “Comoving Mass” McVittie Solution

In the class of generalized McVittie solutions, the one corresponding to the choice
mH .t/ D m0 a.t/ is singled out because it is a late-time attractor within this class
(Sect. 4.6.6). The line element of the “comoving mass” McVittie solution is
 2 
2
1  m2Nr0 2 2 m0 4  2 2 2

ds D   2
dt C a .t/ 1 C dN
r C r d˝.2/ (4.116)
1 C m2Nr0 2Nr

 m0 2
in isotropic coordinates. Using the radial coordinate rQ  rN 1 C , the line
2Nr
element (4.116) becomes
   
2m0 2m0 1 2
ds2 D  1  dt2 C a2 1  dQr C a2 rQ 2 d˝.2/
2
: (4.117)
rQ rQ

We now introduce the areal radius R D aQr, obtaining

"   #  
2 2m0 a 2m0 a 1 2 2 2 2m0 a 1 2
ds D  1   1 H R dt C 1  dR
R R R
 
2m0 a 1
2HR 1  dt dR C R2 d˝.2/
2
(4.118)
R

and we eliminate the cross-term in dtdR by introducing the new time T as [66]
" #
1 HR
dT D dt C   dR ; (4.119)
F .t; R/ 2m0 a 2
1 R
 H 2 R2

where F .t.T; R/; R/ is an integrating factor. The line element (4.118) assumes the
form

12
In principle energy can still flow superluminally inward across the cosmological horizon. The
magnitude of the flux density qc decreases with the radial distance from the black hole.
134 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

"   #
2 2m0 a 2m0 a 1 2 2 2 2
ds D  1   1 H R F dT
R R

"   #1
2m0 a 2m0 a 1 2 2
C 1  1 H R dR2 C R2 d˝.2/
2
; (4.120)
R R

where H.t/ and F.t; R/ are now functions of T and R. The areal radii of the apparent
horizons, located by gRR D 0, satisfy [66]13

 HR2 C R  2m0 a D 0 : (4.121)

Discarding the negative radius, there are two positive roots

1  p 
RC D 1 C 1  8m0 aP ; (4.122)
2H
1  p 
RBH D 1  1  8m0 aP I (4.123)
2H
RC is a cosmological and RBH is a black hole apparent horizon. The singular surface
rN D m0 =2 corresponds to rQ D 2m0 and to R D 2m0 a D 2mH .t/ and RC,BH >
2m0 a D 2mH .
The Misner-Sharp-Hernandez mass of the black hole is expressed analytically as
 p 
MMSH .t/ D 2RBH D H 1 1  1  8m0 aP : (4.124)

For a small black hole with m0 aP D m

PH 1, it is

MMSH ' 4m0 aP .t/ D 4m

P H .t/ : (4.125)

4.6.5 More General Solutions

In the wider class of generalized McVittie solutions, the function m.t/ has arbitrary
time dependence. The generalized McVittie line element is
" #  
2 2m P2 
a2 m m 2 2 2 2m 1 2
ds D  1   1 C dt C a 1  dQr
rQ 1  2mQr
2Qr rQ
 
2 2 2
P 2 1C
2ma m
2r
Ca rQ d˝.2/  2m
dtdQr ; (4.126)
1 Qr

13
This expression appears also in Ref. [43] and it can be derived also from Eq. (4.116) by
expressing it in terms of R.
4.6 Generalized McVittie Spacetimes 135

 
M.t/ 2
where rQ D r 1 C and m.t/  M.t/=a.t/. Rewriting Eq. (4.126) with the
2ra.t/
areal radius14 R  aQr, one obtains
2  q 2 3
Qr
6 HR C ma
P 7
6 2M r
7 2 dR2
ds2 D  61   7 dt C C R2 d˝.2/
2
4 R 1  2M
R
5 1  2M
R

r !
2 rQ
 HR C ma
P dt dR : (4.127)
1  2M
R
r

We introduce
2M
A.t; R/  1  ; (4.128)
R
r
rQ
 .t; R/  HR C ma
P (4.129)
r

and we change from time t to the new time T which satisfies

 
1 C
dT D dt C 2 dR ; (4.130)
F A 2

with F .T; R/ an integrating factor, as usual. Then we have

2  q 2 3
Qr
  HR C ma
P
2M 66 r 7
7 2 2
ds2 D  1  61    7 F dT
R 4 2M 2 5
1 R

2  q 2 31
Qr
  HR C ma
P
2M 1 6
6 r 7
7
C 1 61    7 dR2
R 4 2M 2 5
1 R

CR2 d˝.2/
2
: (4.131)

14
R.t; r/  is an increasing
  function of r for r > m=2 since, in this range,
@R M M
Da 1C 1 is positive.
@r 2ar 2ar
136 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

The usual recipe gRR D 0 states that the apparent horizons are located at
 D ˙A, which means that
r  
rQ 2M
HR C ma
P D˙ 1 : (4.132)
r R

It is
!
 m M
P 2M
HR C M 1 C H D1 ; (4.133)
2r M R
 m
where the factor M 1 C quantifies the deviation of the radius from 2M
2r  
2m
(here r > m=2 is equivalent to R > 2M and to M 1 C > 2M). The quantity
  r
P
M
M
 H is nothing but the difference between the percent rate of change of M
and the corresponding rate of change of the “background” FLRW scale factor.
When this quantity is zero, we have an analogue of the condition for stationary
accretion, but now in a time-dependent “background” and under this condition the
special solution with M.t/ D m0 a.t/ describes stationary accretion relative to the
FRW “background”.
Equation (4.132) becomes (excluding the negative root)
" ! #
 m M
P
2
HR C M 1 C  H  1 R C 2M D 0 : (4.134)
2r M
 m
Since now M.t/ does not scale as a.t/, we have the coefficient 1 C , and
2r
Eq. (4.134) is not a second degree algebraic equation. However, one can decide to
manipulate it (blindly) as a quadratic equation with formal “roots”
8 s 9
1 <  m m
P  m m
P 2 =
RC,BH D 1M 1C ˙ 1M 1C  8mPa :
2H : 2r m 2r m ;
(4.135)
Because r D r.R/, Eq. (4.135) is an implicit equation for the radii RC,BH of the
cosmological and black hole apparent horizons. If the argument of the square root
in Eq. (4.135) is positive, there exist a cosmological apparent horizon (with proper
radius RC ) and a black hole apparent horizon (with radius RBH ). If the argument of
the square root is negative (which occurs near a Big Bang or a Big Rip singularity),
there are no apparent horizons and there is a naked singularity embedded in a FLRW
space. The critical situation corresponding to zero square
p root describes a moment
of time at which the two apparent horizons coincide at 2M=H.
4.6 Generalized McVittie Spacetimes 137

Conformal diagrams describing the global structure of the cosmological black

holes depend on the FLRW “backgrounds”: see [39, 66] for a discussion.

4.6.6 “Comoving Mass” Solution as an Attractor

The “comoving mass” solutions are generic under certain assumptions in the sense
that at late times they are approached by all generalized McVittie solutions [54]. To
prove this statement, one needs to assume that the universe always expands and that
the function m.t/ is non-negative and is continuous with its first derivative.
Begin by writing Eq. (4.121) in terms of rQ  R=a:

2m  m 1
HQr C D m
P 1C C : (4.136)
rQ a 2r a
Given that m  0, the left hand side of this equation cannot be negative and
 m 1
P 1C
m < : (4.137)
2r a
m
Since 1 C > 0 in an expanding universe in which a ! C1, it must be
2r

P 1  lim m.t/
m P  0:
t!C1

Now, if mP 1 D 0, the function m.t/ becomes asymptotically comoving: i.e. m.t/

m0 a.t/ for some positive constant m0 .
The other possibility is that m P 1 < 0; then there is a time Nt such that, for all times
N
t > t, we have m.t/P < 0 and only two possibilities are left: since m.t/  0, either
m.t/ reaches the value zero at a finite time t with derivative m P   m.t
P  / < 0, or
m.t/ ! m0 D constant with m.t/ P ! 0 (that is, m.t/ has a horizontal asymptote).
If the first situation occurs, then at t D t it is

HR D jm
P j a C 1

and the radius of the black hole apparent horizon at t is

 
1 1
r  rhorizon .t / D jm
P j C : (4.138)
H.t / a

Late in the history of the universe we have a black hole of zero mass M.t / D
a.t /m.t / but finite radius r . Evolving this situation in time generates a “black
hole” with negative mass M and finite apparent horizon, a situation which is clearly
unphysical and rules out the case m.t / D 0 with m.t > t / < 0 [54].
138 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

P
The remaining possibility is that m.t/ ! 0 at late times (which means t ! C1
in a universe which expands forever or t ! trip if there is a Big Rip singularity at
P ! 0 means that at late times the rate of increase of m.t/ is at
trip ). The condition m
most the Hubble rate and this function becomes comoving (this conclusion does not
hold at early times when 1=a in Eq. (4.136) does not tend to zero).
A trivial subcase of the second possibility, occurs if m0  0 and the solution
reduces to a FLRW universe. Perhaps there is some merit in interpreting this
situation as a black hole that evaporates completely, but the radial flow considered
in these solutions is not described by a null vector.
Further, one can speculate that if the assumptions are relaxed, the black hole
could avoid becoming comoving; for example, if m.t/ P is discontinuous, m.t/ could
tend to zero in a finite time t , but this spacetime would have discontinuous
connection coefficients and distributional curvature.

4.6.7 Recent Developments and Scalar Field Sources

The causal structure of the McVittie spacetime is rather complicated, and has
been the subject of various recent studies [38, 85, 94, 130–132]. There is finally
agreement that the McVittie metric describes what should be called a black hole
when the black hole apparent horizon is present [94], but the motion of timelike
and null test particles and the horizon structure depend heavily on the cosmological
“background” [38].
It came as a surprise that scalar field sources for McVittie geometries are possible
in various theories15 [1, 3]. It was realized that the McVittie spacetime is not only
a perfect fluid solution of the Einstein equations, but is also an analytic solution
of a special form of k-essence called cuscuton [1] (k-essence theories have been
originally formulated as dark energy models for cosmology, but since then they have
taken a life of their own as possible fundamental theories). The McVittie spacetime
admits constant mean curvature surfaces in its constant time foliation, and this fact
makes the McVittie metric also a solution of Hor̆ava-Lifschitz gravity (a theory
very popular in the search for quantum gravity because of its renormalizability
properties) with anisotropy. The McVittie solution is also an exact solution of shape
dynamics [68] (another approach to quantum gravity). The cuscuton theory is the
only form of k-essence which supports McVittie solutions [1].
In this optics, the generalized McVittie solution is also interesting, since it
turns out to be an exact solution of Horndeski theory (the most general scalar-
tensor theory admitting second order field equations) [3]. Rather than being a
generalization of an old solution of General Relativity, a curiosity from the past,
the McVittie and generalized McVittie solutions relate scalar fields and gravity in

15
See Ref. [13] for scalar field sources of Lemaître-Tolman-Bondi models and the rest of this
chapter for other scalar field solutions.
4.7 Sultana-Dyer Spacetime 139

modern theories and allow research into the basic physics of black holes with scalar
hair. This renaissance of the McVittie solution is still in its infancy and new results
are expected in the near future.

4.7 Sultana-Dyer Spacetime

Another inhomogeneous and time-dependent solution of the Einstein equations

which can be interpreted as a black hole embedded in a FLRW “background”
universe is the Sultana-Dyer spacetime [158]. This solution is of Petrov type D
and the “background” is a spatially flat FLRW universe evolving with the scale
factor a.t/ D a0 t2=3 of a dust fluid, however the matter source is a combination of a
timelike dust and a null dust.
The Sultana-Dyer solution is the extension of the geometry generated by
conformally transforming the Schwarzschild metric g(Schw)
ab :

g(Schw)
ab ! ˝ 2 g(Schw)
ab : (4.139)

The conformal factor is

˝ D a. / D  2 ; (4.140)

where a is the scale factor of a k D 0 FLRW universe and  is a time coordinate

(which is neither the comoving nor the conformal time). The Sultana-Dyer metric
is conformally static and possesses a conformal Killing vector  a satisfying the
conformal Killing equation [169]

1
L gab D ra b C rb a D gab r c c : (4.141)
2
The original intention of Ref. [158] was to transform the timelike Killing field
 c of the Schwarzschild spacetime into a conformal Killing field (defined for
 c rc ˝ ¤ 0), generating a conformal Killing horizon in the conformal cousin of
the Schwarzschild spacetime. Nowadays, conformal Killing horizons seem to have
little relevance in the study of evolving horizons, but the Sultana-Dyer spacetime
remains a useful example.
The Sultana-Dyer line element is
   
2m0 4m0 2m0
ds2 D a2 . /  1  d 2 C ddr C 1 C dr2 C r2 d˝.2/
2
r r r
(4.142)
or

2m0
ds2 D a2 . / d 2 C dr2 C r2 d˝.2/
2
C .d C dr/2 ; (4.143)
r
140 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

where m0 is a constant and a. / D  2 . The coordinate change

ˇ ˇ
ˇ r ˇ
ˇ
 .t; r/ D  C 2m0 ln ˇ  1ˇˇ (4.144)
2m0

(where  is the conformal time of the FLRW “background”) gives

2m0
d D d C   dr (4.145)
r 1  2mr 0

which, substituted into Eq. (4.142), turns the line element into
"   #
2 2 2m0 2 dr2 2 2
ds D a .; r/  1  d C C r d˝.2/ : (4.146)
r 1  2mr 0

In this form, the metric is explicitly conformal to the Schwarzschild one with
conformal factor
 ˇ ˇ2
ˇ r ˇ
2 ˇ
˝ D a.; r/ D  .; r/ D  C 2m0 ln ˇ  1ˇˇ (4.147)
2m 0

dependent on both  and r. The comoving time of the FLRW “background” is related
to the conformal time  by dt D ad. The Sultana-Dyer spacetime reduces to the
spatially flat FLRW universe if m ! 0 or for r ! C1.
Isotropic coordinates are also used for the Sultana-Dyer metric. The isotropic
radius rN is defined by
 m0 2
r D rN 1 C : (4.148)
2Nr
Using this coordinate and the fact that
 m0   m0 
dr D 1 C 1 dNr ; (4.149)
2Nr 2Nr
one obtains
"   #
1 m0 2  m0 4  2 
2 2 2Nr 2 2 2
ds D a .; r/    d C 1 C dNr C rN d˝.2/ : (4.150)
m0 2 2Nr
1C 2Nr

The matter source of the Sultana-Dyer spacetime is a combination of two non-

interacting perfect fluids. The total energy-momentum tensor is
(I) (II)
Tab D Tab C Tab  ua ub C n ka kb ; (4.151)
4.7 Sultana-Dyer Spacetime 141

(I)
where Tab D ua ub is associated with an ordinary dust with timelike 4-velocity uc ,
(II)
while Tab D n ka kb describes a null dust with density n and null vector kc [158].
A problem pointed out already in the original work [158] is that in the Sultana-
Dyer spacetime the cosmological fluid becomes tachyonic and the energy density
becomes negative at late times near the spacetime singularity rN D m0 =2 [158].
The Misner-Sharp-Hernandez mass of a sphere of radius r in the Sultana-Dyer
spacetime is easily computed
 
r3 a2; 2m0
MMSH D m0 a  2m0 ra; C 1C : (4.152)
2a r

Using the identity

@a
a; D a; D a  aPa ; (4.153)
@t

the definition of the Hubble parameter H  aP =a, it is easy to obtain [52]

H 2 R3
MMSH D m0 a .1  HR/2 C ; (4.154)
2
where R D ar is the areal radius of the Sultana-Dyer spacetime. The mass MMSH
consists of two contributions: the first one is the mass m0 of the Schwarzschild
“seed” metric rescaled by the conformal factor a but scaled down by the cosmic
expansion by .1  HR/2 . This factor vanishes at R D 1=H, which is the radius
of the cosmological horizon of the FLRW “background”. The factors a and
.1  HR/2 have competing effects which are not easy to interpret. The second
4 3H 2
contribution to MMSH can be written as R3 , where D is the density
3 8
of the “background” cosmological fluid. This second term is obviously the mass of
cosmic fluid contained in the sphere of areal radius R. Alternatively, one can write

H 2 R3 .1/ .2/ (int)

MMSH D m0 a C  m0 aHR.2  HR/  MMSH C MMSH C MMSH ; (4.155)
2
where the first term is purely local, the second one is purely cosmological, and the
third one is an interaction term which is small for an object of size R much smaller
than the Hubble radius H 1 of the “background”.
The quantity

M.Nt/  m0 a.Nt/ (4.156)

is not constant in the Sultana-Dyer solution. The locus r D 2m0 is not a spacetime
singularity; the conformal factor ˝ diverges there but the original metric (4.142) is
not singular and it can be considered as an extension of the conformally transformed
Schwarzschild metric (4.146).
142 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

The Ricci curvature is

 
12 2m0 2m0 
R D 6 1 C 2 ; (4.157)
 r r

and is not singular at r D 2m0 (where  ! 1). There are spacetime singularities
at r D 0 (central singularity) and at  D 0 (Big Bang).
To locate the apparent horizons16 [52, 149] one must solve the equation gRR D 0,
where R D ar, obtaining
 
r3 a2; 2m0
2m0 a C 1C  4m0 ra; D ar : (4.158)
a r

Since a. / D  2 , one obtains the cubic equation for r

4r3 C 8m0 r2  .8m0 C  /  r C 2m0  2 D 0 ; (4.159)

the real positive roots of which are [52, 149]

q
4m0   C  2 C 24m0  C 16m20
r1 D ; (4.160)
4
 .; r/
r2 D ; (4.161)
2
with r1 < r2 . The surface r D 2m0 , which is the null event horizon of the Schwarz-
schild seed metric, remains an event horizon of the Sultana-Dyer metric. The radii
of the apparent horizons expressed in terms of the areal radius R D ar are

3
R1 D ; R2 D REH D 2m0  2 ; (4.162)
2
and
q
4m0   C  2 C 24m0  C 16m20
R4 D 2 : (4.163)
4
These are implicit equations for the apparent horizons radii in terms of t and R.
The Sultana-Dyer spacetime was studied as an example of a time-dependent
black hole horizon for which the Hawking temperature can be derived explicitly
[149] to shed light on the Hawking effect and the thermodynamics in dynamical

16
Beware of an error at the beginning of Ref. [48] consisting of imposing a coordinate condition
which cannot be satisfied. This error was corrected in [29] and, later, in [159].
4.7 Sultana-Dyer Spacetime 143

situations. In this context, the bad behaviour of matter near the horizon is not impor-
tant. The authors of [149] studied the radiation of a massless, conformally coupled,
scalar field  in the Sultana-Dyer spacetime and computed the renormalized stress-
energy tensor hTab i of . What makes the calculation feasible is the simplification
due to the fact that the Sultana-Dyer geometry is conformal to the Schwarzschild
one. The conformal anomaly and particle creation by the FLRW “background” were
taken into account. Under the assumption that the Sultana-Dyer black hole evolves
slowly, its effective Hawking temperature can be computed neglecting non-adiabatic
terms. The result is [149]

1
TQ D C ::: (4.164)
8 m0 a

where the ellipsis denotes corrections which are small in the limit of a slowly
1
evolving black hole [149]. Since T D is the Hawking temperature of the
8 m0
“seed” Schwarzschild black hole generating the Sultana-Dyer geometry, one can
write
T
TQ D C ::: (4.165)
˝
This result is a special case of a more general relation

TSchw
TD (4.166)
˝
which is conjectured to hold [149] in spacetimes conformally related to the
Schwarzschild spacetime by a transformation with conformal factor ˝. There is
some independent support for this conjecture from naive dimensional considerations
[47].
According to Dicke [42] (who followed earlier ideas of Weyl [171, 172]), a
conformal transformation gab ! ˝ 2 gab is nothing but a rescaling of the lengths of
vectors and of the units used in a measurement, with the rescaling factor varying
with the spacetime point. An experiment measures the ratio between a certain
quantity x and its unit xu . The quantity x itself is not meaningful unless a unit xu
is fixed for that quantity, and only the ratio x=xu makes sense operationally. For
example, the length of a ruler divided by the unit of length lu is the same in the
Minkowski metric ab and in a conformally related metric gab D ˝ 2 ab if a new
unit of length Qlu D ˝lu is associated with the length Ql in a measurement. Two
metrics gab and gQ ab D ˝ 2 gab are physically equivalent, at least from the classical
point of view, when the units of the fundamental quantities length, time, and mass-
energy are scaled according to Qlu D ˝ lu , Qtu D ˝ tu , and m
Q u D ˝ 1 mu [42] (units
derived from the fundamental units are scaled according to their dimensions). There
is no difference between using the Schwarzschild metric g(Schw) ab and the Sultana-
(SD) Q
Dyer metric conformal to it, gQ ab D gab , provided that the units lu , Qtu , and m
Q u are
144 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

scaled appropriately (expanding for lengths and times, and redshifting for masses
and energies). Since kB T is an energy, the ratio between kB T and mu must be the
same when using g(Schw)
ab or g(SD)
ab :

kB TQ kB T (Schw)
D : (4.167)
mQu mu

Then the effective temperature of the Sultana-Dyer black hole follows,

T (Schw) 1
TQ D D ; (4.168)
˝ 8 m0 a

in agreement with Eq. (4.165).

In scalar-tensor gravity it is well-known that the transformation law of the stress-
energy tensor of matter under conformal transformations gab ! gQ ab D ˝ 2 gab is

TQ ab
(m)
D ˝ 2 Tab
(m)
; (4.169)

which agrees with a direct calculation of TQ ab

(m)
[46, 169]. By applying this rescaling
to the semiclassical energy-momentum tensor of a scalar field near the Sultana-Dyer
black hole, the renormalized hTQ ab i should be

hTab i
hTQ ab i D : (4.170)
a2
The explicit renormalization of Tab in Ref. [149] yields instead

hTab i 1
hTQ ab i D hTab
(SD)
iD  .Xab  Yab / ; (4.171)
a2 2880 2

where

RQ
Q RQ C
Xab D 2rQ a rQ b RQ  2Qgab  gQ ab  2RQ RQ ab ; (4.172)
2
2 1 RQ
Yab D RQ ca RQ bc C RQ RQ ab C RQ cd RQ cd gQ ab  gQ ab : (4.173)
3 2 2
The extra terms in (4.171) are interpreted as quantum particle creation by the
expanding FLRW “background” [149] (which cannot be predicted with Dicke’s
classical argument). When the black hole horizon evolves slowly, these terms can
be neglected and the rescaling argument agrees with the hTQ ab i of Eq. (4.170).
An independent argument in favour of Eq. (4.168) is the following [47]. The
first law of black hole thermodynamics for a Schwarzschild black hole of mass
m0 is TdS D dm0 . The expression of the Bekenstein-Hawking entropy S D A =4,
where A D 4 rH2 is the horizon area, together with the expression rH D 2m0 for the
4.8 Husain-Martinez-Nuñez Spacetime 145

horizon radius, yields the Hawking temperature T (Schw) D 1= .8 m0 /. For a Sultana-

Dyer expanding black hole, the purely local contribution to the Misner-Sharp-
.1/
Hernandez mass is MMSH D m0 a.t/ and the proper horizon radius is RH .t/ D a.t/rH .
For this expanding horizon, the first law of black hole thermodynamics includes a
work term dW:
.1/
TdS D dMMSH C dW : (4.174)

If the black hole entropy is S D A =4, then

.1/ .1/ .1/

8 TMMSH dMMSH D dMMSH C dW : (4.175)

In the adiabatic approximation, the black hole should be in a state of quasi-

equilibrium and the work terms should be negligible, which gives again

1 T (Schw)
T' D : (4.176)
8 MMSH .t/ a

The conformal factor of the Sultana-Dyer black hole does not depend on the
radial coordinate and, in the adiabatic approximation in which its time variation
is small, the Hawking temperature does have the scaling behaviour expected on
dimensional grounds. This scaling law, however, will break down as soon as the
conformal transformation is allowed to be radial-dependent, or the apparent horizon
is allowed to vary rapidly.
The temperature of the Sultana-Dyer apparent horizon was calculated in [110]
using the method of chiral anomalies. The result confirms the calculation of [149]
and the guess of [47]. Moreover, the temperature is related to the entropy and the
Misner-Sharp-Hernandez mass by the algebraic expression MMSH D 2ST, which is
the Smarr formula for stationary black holes.

4.8 Husain-Martinez-Nuñez Spacetime

The Husain-Martinez-Nuñez spacetime [82] is a solution of the Einstein equations

which shows us a dynamics of the apparent horizons which is different from what
we have encountered thus far. This spacetime is inhomogeneous, with a spatially
flat FLRW “background” and the matter source is not a fluid but a free scalar field
minimally coupled with gravity. The coupled Einstein-Klein-Gordon equations are

1 ./
Rab  gab R D 8 Tab ; (4.177)
2
 D 0 ; (4.178)
146 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

where

./ 1
Tab D ra rb   gab rc r c  : (4.179)
2
Equation (4.177) simplifies to

Rab D 8 ra rb  : (4.180)

The line element and scalar field are presented as [82]

"  
2 2C ˛ 2 dr2
ds D .A0  C B0 /  1  d C  ˛
r 1  2C
r
  #
2 2C 1˛ 2
Cr 1  d˝.2/ ; (4.181)
r
"   p #
1 2C ˛= 3 p
3
.; r/ D ˙ p ln D 1  .A0  C B0 / ; (4.182)
4 r
p
where A0 ; B0 ; C, and D are non-negative constants, ˛ D ˙ 3=2 is a parameter
which can assume only two possible values, and  > 0. The additive constant B0 can
be dropped if A0 ¤ 0 because, in this case, it is redundant. If A0 D 0 the Husain-
Martinez-Nuñez line element (4.181) reduces to that of the Fisher spacetime [61]

dr2
ds2 D V  .r/ d2 C C r2 V 1 .r/d˝.2/
2
; (4.183)
V  .r/

where V.r/ D 1  2 =r,  is a dimensionless parameter, is a mass, and the scalar

field is

.r/ D 0 ln V.r/ : (4.184)

The static Fisher solution of the Einstein-Klein-Gordon equations is better known

as the Janis-Newman-Winicour or the Wyman solution because it has been redis-
covered again and again over the years and has picked up several names [4,
9, 18, 84, 164, 175]. The Fisher spacetime is asymptotically flat and contains a
naked singularity at r D 2C. It is identified with the most general static spherically
symmetric solution of the Einstein equations with zero cosmological constant and
a massless minimally coupled scalar field as the source [146], but it is unstable
[2]. The Husain-Martinez-Nuñez line element is conformal to the Fisher one,
with the
p scale factor of the “background” FLRW space as the conformal factor,
˝ D A0  C B0 , and with only two possible values of the Fisher parameter . In
the following we set B0 D 0 and we denote the time of the Big Bang singularity
with  D 0. The sign appearing in Eq. (4.182) is independent of the sign of ˛. As
4.8 Husain-Martinez-Nuñez Spacetime 147

r ! C1 the metric (4.181) is asymptotically FLRW, and it is FLRW if C D 0 (then

the constant A0 can be eliminated by rescaling ).
It is straightforward to compute the Ricci scalar as
 ˛2
2˛ 2 C2 1  2C 3A20
R D 8 r rc  D
c
4
r
 3
 
2C ˛
: (4.185)
3r A0  2 .A0 / 1  r

Equation (4.185) shows that there is a spacetime singularity at r D 2C for both

values of ˛. The scalar field  also diverges there (in addition, a Big Bang singularity
occurs at  D 0). Only the range 2C < r < C1 is physically meaningful. The areal
radius is
  1˛
p 2C 2
R.; r/ D A0  r 1  (4.186)
r

and the value r D 2C of the radial coordinate corresponds to R D 0,

It is physically more rewarding to express the Husain-Martinez-Nuñez line
element in terms of the comoving
p time t of the “background” FLRW space defined
by dt D ad, where a./ D A0  is the FLRW scale factor. Since
Z p
2 A0 3=2
tD d a./ D  (4.187)
3

fixing  D 0 at t D 0, it is
 2=3
3
D p t (4.188)
2 A0

and
 1=3
p 1=3 3A0
a.t/ D A0  D a0 t ; a0 D : (4.189)
2

This scale factor is, of course, the one dictated by the stiff equation of state P D
of a free massless scalar field in a FLRW universe. The general FLRW solution17
for equation of state parameter w  P= is
2
a.t/ D const. t 3.wC1/ : (4.190)

17
In a FLRW universe there are no spatial scalar field gradients (which would identify a
P 2
preferred spatial direction) and the energy density and pressure are simply ./ D C V./,
2
P 2
P./ D  V./. If V./ D 0, then it is P./ D ./ .
2
148 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

The Husain-Martinez-Nuñez line element in comoving time is

 ˛ "   #
2 2C 2 2 2C 1˛
dr2 22
ds D  1  dt Ca .t/  ˛ C r 1  d˝.2/ (4.191)
r 1  2C
r
r

and the scalar field sourcing it is

"   p #
1 2C ˛= 3 2p3
.t; r/ D ˙ p ln D 1  a .t/ : (4.192)
4 r

The areal radius (4.186) increases with r for r > 2C. We now recast the Husain-
Martinez-Nuñez metric using the areal radius R. Using the notation

2C .˛ C 1/C
A.r/  1  ; B.r/  1  ; (4.193)
r r
1˛
one has R.r/ D a.t/rA 2 .r/ and

˛C1 dR 1
dr D A 2  AH rdt ; (4.194)
a B.r/

which leads to

H 2 R2 A2.1˛/ H 2 R2 A2˛ .r/ 2

ds2 D A˛ 1  2
dt2 C dR
B .r/ B2 .r/
3˛
2HRA 2
 dt dR C R2 d˝.2/
2
: (4.195)
B2 .r/

The time-radius cross-term is eliminated by using a new time T given by

1
dT D .dt C ˇdR/ ; (4.196)
F

where ˇ.t; R/ is a function to be determined and F.t; R/ is an integrating factor

obeying the usual equation
   
@ 1 @ ˇ
D : (4.197)
@R F @t F

Using dt D FdT  ˇdR in Eq. (4.195) and choosing

3.1˛/
HRA 2
ˇ.t; R/ D 2 ; (4.198)
B .r/  H 2 R2 A2.1˛/
4.8 Husain-Martinez-Nuñez Spacetime 149

the line element becomes

H 2 R2 A2.1˛/ .r/ 2 2
ds2 D A˛ .r/ 1  F dt
B2 .r/
H 2 R2 A2˛ .r/ A1˛ .r/
C 1 C dR2 C R2 d˝.2/
2
:
B2 .r/ B2 .r/  H 2 R2 A2.1˛/ .r/
(4.199)

The apparent horizons, located by gRR D 0, solve

B.r/ D H.t/RA1˛ .r/ ; (4.200)

where r D r.t; R/. In terms of the original coordinates  and r one has [82]

1 2h i 
2C ˛1
D 2 r  .˛ C 1/C 1  : (4.201)
 r r

For r ! C1 (or R ! C1), Eq. (4.201) reduces to R ' H 1 , which is the radius
of the cosmological apparent horizon of spatially flat FLRW space. Equation (4.200)
can only be solved numerically. Let x  C=r, then the equation locating the apparent
horizons is
 
.˛ C 1/C 2C ˛1
HR D 1  1 : (4.202)
r r

The left hand side can be written as

a0 2C 1˛
HR D .1  2x/ 2
3 t2=3 x

expressing the radius of the apparent horizons in units of H 1 ; this is the radius
of the cosmological apparent horizon of the FLRW “background”. The right hand
side of Eq. (4.202) is Œ1  .˛ C 1/x .1  2x/˛1 , hence Eq. (4.202) and the equation
defining the areal radius give
 3=2
2Ca0 .1  2x/3.1˛/
t.x/ D ; (4.203)
3 x Œ1  .˛ C 1/x
2C 1˛
R.x/ D a0 t1=3 .x/ .1  2x/ 2 ; (4.204)
x

respectively, which is a parametric representation of the function R.t/ andpis then

used to plot this function in Figs. 4.4 and 4.5. If the parameter value ˛ D 3=2 is
adopted, in the time interval between the Big Bang and a critical time t there is
150 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

Horizon
production

Merger

0
0 t*
t

Fig. 4.4
p The radii of the Husain-Martinez-Nuñez apparent horizons versus comoving time for
˛ D 3=2. Time t and radius R are both measured in arbitrary units of length and the parameters
C and a0 are chosen so that .Ca0 /3=2 D 103 in Eq. (4.203)

Fig. 4.5 The areal radius

p of the single Husain-Martinez-Nuñez apparent horizon present for the
parameter value ˛ D  3=2 versus comoving time. This cosmological apparent horizon expands
and there is always a naked singularity at R D 0
4.8 Husain-Martinez-Nuñez Spacetime 151

only one apparent horizon which expands. At the time t , two additional apparent
horizons appear; one is a cosmological apparent horizon which expands forever,
while the other is a black hole horizon which contracts until it merges with the first
apparent horizon which has been growing in the meantime. When they encounter
each other, these two apparent horizons merge and disappear, causing a naked
singularity at R D 0 to be present in a FLRW universe for the rest of the time.
This apparent horizons dynamics differs from that of the McVittie and generalized
McVittie solutions. The “S-curve” of Fig. 4.4 is recurrent in analytic solutions of
Brans-Dicke and f .R/ gravity (Chap. 5). The scalar field is regular on the apparent
horizons. p
For the parameter value ˛ D  3=2 only one cosmological apparent horizon
exists at all times, and it expands. A naked singularity is present at R D 0 (Fig. 4.5),
in addition to the usual spacelike Big Bang singularity at t D 0.
For the Husain-Martinez-Nuñez spacetime, it is possible to establish that the
apparent horizons are spacelike by studing the normal vector to these surfaces and
seeing that this vector always lies inside the light cone in an .; r/ diagram (we
follow Ref. [82] here). According to Eq. (4.201), along the apparent horizons it is
 1˛
r2 1  2C
D r
: (4.205)
2 Œr  C.1 C ˛/

Differentiating with respect to r, we obtain

ˇ   (   )
ˇ 2C ˛ r2 1  2C
;r ˇ D 1 1 r
: (4.206)
AH r 2 Œr  C.1 C ˛/2

By comparison, along radial null geodesics, we have

ˇ  
ˇ 2C ˛
;r ˇ D˙ 1 (4.207)
light cone r

(which can be obtained by setting ds2 D 0 together with d D d' D 0), so that [82]
ˇ ˇ ˇ
ˇ ˇ
ˇ ;r ˇˇ ˇ 
1  2C

ˇ ˇ AH ˇ
ˇ ˇ ˇ D 1  h
r
i2  1 : (4.208)
ˇ ˇ ˇ .˛C1/C
ˇ ;r light cone ˇ 2 1 r

The normal to the apparent horizon is always enclosed by the light cone, hence it
is timelike, except where this vector becomes tangent to the light cone itself and
it is null (which occurs when a pair of apparent horizons is created or disappears)
[82].
The nature of the singularity at r D 2C (or R D 0) is easily assessed. All surfaces
described by f .R/  R  const: D 0 have gradient N  r f D ı 1 in .t; R; ; '/
152 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

coordinates. The norm squared of this gradient is

B2 1
Nc N c D g11 D : (4.209)
H 2 R2 A2˛ 1C A1˛
B2 H 2 R2 A2.1˛/

1˛
Now, B.r/ ! and A.r/ ! 0C as r ! 2CC , therefore Nc N c > 0 and Nc N c !
2
C1 as r ! 2CC . The R D 0 singularity is timelike for both values of the parameter
˛.
The Husain-Martinez-Nuñez spacetime is quoted as describing p scalar field
collapse, but a better description (for the parameter value ˛ D C 3=2) is that it
exhibits the creation and annihilation of pairs of black hole apparent horizons. The
R D 0 singularity (for both values of ˛) is created with the universe in the Big Bang
and is not the product of gravitational collapse.
According to what already seen in this chapter,
p the physical interpretation of
the apparent horizon dynamics for ˛ D C 3 is that a black hole larger than
the cosmological horizon cannot fit in the early “universe”. When this “universe”
becomes sufficiently large, a black hole appears with inner and outer apparent
horizons. These black hole horizons then merge into an extremal (null) black hole
horizon and disappear.

4.9 Fonarev Solutions

The Fonarev spacetime of General Relativity has as the matter source a minimally
coupled scalar field with an exponential self-interaction potential [62]. It describes
a central inhomogeneity in an otherwise FLRW universe. The theory is described
by the action
Z
p 1
SD d4 x g R  ra r a   V ./ ; (4.210)
2

where

V ./ D V0 e 
; (4.211)

with and V0 positive constants. This potential has been investigated at length
in FLRW cosmology [34, 73, 86, 103, 111]. The coupled Einstein-Klein-Gordon
equations are

1
Gab D 8 ra rb   gab rc r c   gab V./ ; (4.212)
2
4.9 Fonarev Solutions 153

dV
  D 0: (4.213)
d

Equation (4.212) can be simplified to

Rab D 8 .ra rb  C gab V/ : (4.214)

The spherically symmetric Fonarev line element and scalar field are

dr2
ds2 D a2 ./ f 2 .r/ d2 C C S2 .r/ d˝.2/
2
; (4.215)
f 2 .r/
  "  2 #
1 2w 1 V0 2  2
 .; r/ D p ln 1  C ln a C ln ;
2C2 r 2A20 .6  2 /
(4.216)

where
 ˛
2w 2
f .r/ D 1  ; ˛Dp ; (4.217)
r 2C2

  1˛
2w 2 2
S.r/ D r 1  ; a./ D A0 jj 2 2 ; (4.218)
r

with w and A0 constants. For simplicity we choose A0 D 1.  is the conformal

time of the FLRW “background”. When w D 0 the metric (4.215) reduces to
a spatially flat FLRW one while, when a  1 and ˛ D 1, it degenerates into
the Schwarzschild solution (however, the value ˛ D 1 cannot be obtained if the
condition ˛ D p holds). The line element becomes asymptotically that of a
2C2
spatially flat FLRW space as r ! C1 (see the end of this section for a discussion
of the apparent horizons).
The Fonarev metric is clearly a generalization of the Husain-Martinez-Nuñez
class of solutions (4.181) to exponential
p potentials. The Husain-Martinez-Nuñez
metric is recovered by setting D ˙ 6 and V0 D 0.
A phantom Fonarev solution corresponding to a dynamical phantom scalar field
solution of General Relativity was introduced in Ref. [66]. It is obtained from the
Fonarev solution using the transformation

 ! i ; ! i : (4.219)
154 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

The corresponding action is

Z  
p 1
SD d4 x g R C ra r a   V0 e  (4.220)
2

with a phantom field endowed with the “wrong” sign of the kinetic term. The
coupled Einstein-scalar field equations are now
1
Gab D ra rb  C gab rc r c   gab V ; (4.221)
2
dV
 C D 0: (4.222)
d
The generalized Fonarev metric and phantom scalar are

dr2
ds2 D a2 ./ f 2 .r/ d2 C C S2 .r/ d˝.2/
2
; (4.223)
f .r/2
"  2 #  
2
1 V0 C2 1 2w
 .; r/ D ln 2 C 6/
 ln a  p ln 1  ;
2. 22 r
(4.224)

where
 
2w ˛=2
f .r/ D 1  ; ˛ D p ; (4.225)
r 22

  1˛
2w 2  2
S.r/ D r 1  ; a./ D  2 C2 : (4.226)
r
p
Assuming that > 2,pit is of interest to understand the physical meaning of the
constant w. When 2 it is a 1 and ˛ 1 and the metric becomes [66]
 1    
2 2w 2 2w 2 2 2w 2 2
ds  1 d C 1  dr C r 1  d˝.2/ : (4.227)
r r r

Using the coordinate transformation [66]

 
2w
yDr 1 ; (4.228)
r

the line element (4.227) is rewritten as

   
2 2w 2 2w 1 2
ds D  1 C d C 1 C dy C y2 d˝.2/
2
; (4.229)
y y
4.9 Fonarev Solutions 155

which is the Schwarzschild solution with mass w. Therefore, the parameter w
corresponds to the negative mass in this limit and we will we use M instead of w.
Let us locate the apparent horizons of the generalized Fonarev metric as the
parameters M and ˛ vary. By writing the line element as
   
2 1 2M ˛ 2 2M ˛ 2
ds D 2˛2 2  1 C d C 1 C dr
r r
 2˛2 1
  #
2 2M 1C˛ 2
Cr 1 C d˝.2/ (4.230)
r

and replacing the conformal time  with the comoving time t, one obtains
 
2M ˛ 2
ds2 D  1 C dt
r
"    #
2 2M ˛ 2 2 2M 1C˛ 2
Ca .t/ 1 C dr C r 1 C d˝.2/ ; (4.231)
r r
. /
2 ˛ 2 1
a .t/ D .t0  t/ ˛2 ; (4.232)
p
where the integration constant t0 marks the Big Rip and ˛ < 1 since > 2.
The exponent ˛ is determined by the slope of the potential according to Eq. (4.225).
When M D 0 the metric (4.231) reduces to a phantom-dominated FLRW one. By
setting, for the sake of illustration, ˛ D 3 or D 3=2, the line element (4.231)
reduces to
 
2 2M 3 2
ds D  1 C dt
r
"    #
2 2M 3 2 2 2M 2 2
Ca .t/ 1 C dr C r 1 C d˝.2/ ;
r r

a .t/ D .t0  t/16=9 : (4.233)

In terms of the areal radius R D ar .1 C 2M=r/1 , the equation locating the apparent
horizons is
r !1 r !5
8Ma 8Ma HR 8Ma
1C 1C 1C  1C 1C D 0; (4.234)
R R 32 R

 aP =a is the Hubble parameter of the FLRW “background”. Further setting

where H r
8Ma
x1C 1C yields
R

aMHx4  4x2 C 12x  8 D 0 : (4.235)

156 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

This quartic equation has only two real positive roots corresponding to a cosmologi-
cal apparent horizon RC and a black hole apparent horizon RBH [66]. The qualitative
behaviour of the apparent horizons is the same as for the McVittie and generalized
McVittie solutions: a black hole apparent horizon inflates while a cosmological
apparent horizon shrinks. At a critical time these two apparent horizons coincide and
disappear leaving behind a naked singularity [66]. The time reverse of this picture
gives the apparent horizons of the Fonarev geometry with canonical scalar.

4.10 Other Analytic Cosmological Black Hole Solutions

of the Einstein Equations

Apparently unaware of McVittie’s 1933 work, in 1946 Einstein and Straus [44]
derived the solution of General Relativity now called Einstein-Straus vacuole or
Swiss-cheese model by pasting a Schwarzschild-like region of spacetime onto a
dust-dominated FLRW universe across a timelike hypersurface. There is a black
hole event horizon in this spacetime and the usual energy conditions are satisfied.
The Einstein-Straus model is discussed at length in the literature (see, e.g.,
Ref. [92] and references therein) and we will not repeat such discussions here.
The Einstein-Straus vacuole was later generalized to include a cosmological con-
stant, obtaining a Schwarschild-(anti-)de Sitter instead of Schwarzschild interior [7],
or to include a fluid with pressure in the interior region [11]. Also the generalization
obtained by matching a Schwarschild interior with an inhomogeneous Lemaître-
Tolman-Bondi exterior has been studied [12]. The Hawking radiation emitted by
the Einstein-Straus black hole has been studied in [148, 149]. It is found that a black
hole in an expanding universe is excited to a non-equilibrium state and emits with
stronger intensity than a thermal one.
Apparent and trapping horizons were studied also in Oppenheimer-Schneider,
Vaidya, and Lemaître-Tolman-Bondi spacetimes [13, 67]. In this last class of
models, in particular, the use of different coordinates determines different foliations
and, potentially, different apparent horizons. Multiple “S-curve” phenomenology is
reported and interpreted as a single apparent horizon tube which goes back and forth
in time and, when sliced with hypersurfaces of constant time, produces multiple
apparent horizons which appear and disappear in pairs [13].
McClure and Dyer [117] found a spherical solution of the Einstein equations
presumably describing a central inhomogeneity in a radiation-dominated universe
with a radial heat current, which satisfies the energy conditions everywhere and
is perfectly comoving. There is a spacetime singularity at rN D m=2. Similarly,
another analytic solution of General Relativity with a dust-dominated “background”
universe exhibits singular energy density and a spacetime singularity [117]. Charged
versions of the Vaidya, Sultana-Dyer, and Thakurta solutions are reported in
Ref. [147].
The Kerr-de Sitter black hole is well known in the literature, but a Kerr-FLRW
(or Kerr-McVittie) solution is not reported. Solutions describing spherical shells in
4.11 Conclusions 157

FLRW space were found by studying inflation and phase transitions in the early
universe: the most well known are the Coleman-de Luccia [33] and the Farhi-Guth
[57] spacetimes.18
Common techniques used to generate cosmological black holes consist of:
1. Performing a conformal transformation of a static black hole solution (for
example, Schwarzschild in some coordinate system) using a time-dependent
conformal factor (usually given by the scale factor of the FLRW “background”
universe):

gab ! gQ ab D ˝ 2 gab D a2 gab : (4.236)

This technique generates, for example, the Sultana-Dyer black hole and various
solutions studied in [116–118].
2. Performing a Kerr-Schild transformation of a static black hole metric gab :

gab ! gN ab D gab C ka kb ; (4.237)

where is a function and kc is a null and geodesic vector field with respect to
both gab and gN ab (see Sect. 1.3.6).
In general, by conformally transforming or Kerr-Schild transforming a “seed”
solution of the Einstein equations with standard matter source (including vacuum),
it is not guaranteed that the product of this transformation will satisfy the Einstein
equations with the same form of matter, or with any reasonable matter source at
all. Indeed, one can use the Synge approach consisting of running the Einstein
equations from the left to the right, i.e., prescribing a metric motivated in some
way and computing the corresponding energy-momentum tensor. But the latter
will in general violate the energy conditions and will be physically unreasonable
because it is built in a completely artificial way and is devoid of physical content.
Indeed, this is the problem of most solutions obtained by conformally or Kerr-Schild
transforming a known black hole metric. Moreover, the conformal transformation of
a static black hole metric does not always generate a black hole: often it generates a
naked singularity instead.

4.11 Conclusions

It is rare to find explicit analytic expressions of the apparent horizons for solutions
of the Einstein equations representing cosmological (or other dynamical) black
holes. To the best of our knowledge such an expression is available only for the

18
Sometimes one encounters in the literature also dynamical black hole spacetimes which are
constructed by hand and are not known to be solutions of the Einstein equations or of the field
equations of other theories of gravity (e.g., [16, 60, 105–107]).
158 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

extremal charged McVittie black hole and for the “comoving” generalized McVittie
spacetime. For other exact solutions of General Relativity the apparent horizon can
only be located numerically or given by implicit analytic expressions (for example,
for the Schwarzschild-de Sitter-Kottler and the McVittie black holes).
When they represent black holes, the various inhomogeneous solutions consid-
ered in this chapter describe:
• Eternal black holes which have not been created in a collapse process but are
created together with the universe in the Big Bang, or have existed forever (for
example, in a de Sitter background); or
• Black holes that appear when a naked singularity is suddenly covered by an
apparent horizon which is created simultaneously with another (cosmological)
apparent horizon.
In any case, when a timelike naked singularity is present, the initial value problem
[169] is not well posed and the spacetime cannot be obtained as the development of
regular Cauchy data.
The subject of dynamical and cosmological black holes is still too young to
classify all the possibilities allowed by the Einstein equations in a physically
meaningful way, and it is even debatable whether apparent and trapping horizons
provide a truly satisfactory notion of dynamical black hole. Nevertheless, one can
tentatively group the known solutions of the Einstein equations with these features
in two ways:
1. On the basis of the type of matter sourcing the “background” FLRW universe
(dust, perfect fluid, imperfect fluid, scalar field, etc.);
2. On the basis of the dynamics and phenomenology of the apparent horizons.
Solutions of General Relativity with a perfect fluid and an electric field include
the McVittie and charged McVittie solution (and its special case, the Schwarzschild-
de Sitter and Schwarzschild-anti de Sitter black holes); solutions sourced by a
canonical scalar field include the Husain-Martinez-Nuñez, the Fonarev, and the
phantom Fonarev solutions. At the moment of writing, the phenomenology of
apparent horizons distinguishes between the McVittie type with two appearing
or disappearing (one black hole and one cosmological) apparent horizons, and
the Husain-Martinez-Nuñez-type phenomenology with three apparent horizons.
Multiple “S-curve” phenomenology is reported in Lemaître-Tolman-Bondi models
[13]. It is not clear whether completely different horizon phenomenologies are
possible in Einstein theory.
The rather bizarre phenomenology of the apparent horizons in the solutions
examined begs the question of whether they are, after all, physically significant.
Naked singularities form during simulations of gravitational collapse but, generally
speaking, they are not “typical”. General choices of the initial data result in
black holes rather than naked singularities. It could well be that the solutions
examined here are non-generic or even very special. The “comoving mass solution”
References 159

is a late-time attractor in the generalized McVittie class but it is not a generic

cosmological black hole solution with spherical symmetry [28]. No definitive
statement can be made at present.

References

1. Abdalla, E., Afshordi, N., Fontanini, M., Guariento, D.C., Papantonopoulos, E.: Cosmological
black holes from self-gravitating fields. Phys. Rev. D 89, 104018 (2014)
2. Abe, S.: Stability of a collapsed scalar field and cosmic censorship. Phys. Rev. D 38, 1053
(1988)
3. Afshordi, N., Fontanini, M., Guariento, D.C.: Horndeski meets McVittie: a scalar field theory
for accretion onto cosmological black holes. Phys. Rev. D 90, 084012 (2014)
4. Agnese, A.G., La Camera, M.: Gravitation without black holes. Phys. Rev. D 31, 1280 (1985)
5. Amendola, L., Tsujikawa, S.: Dark Energy, Theory and Observations. Cambridge University
Press, Cambridge (2010)
6. Babichev, E., Dokuchaev, V., Eroshenko, Yu.: Black hole mass decreasing due to phantom
energy accretion. Phys. Rev. Lett. 93, 021102 (2004)
7. Balbinot, R., Bergamini, R., Comastri, A.: Phys. Rev. D 38, 2415 (1988)
8. Barris, B., et al.: Twenty-three high-redshift supernovae from the Institute for Astronomy
Deep Survey: doubling the supernova sample at z > 0:7. Astrophys. J. 602, 571 (2004)
9. Bergman, O., Leipnik, R.: Phys. Rev. 107, 1157 (1957)
10. Bochicchio, I., Faraoni, V.: A Lemaître-Tolman-Bondi cosmological wormhole. Phys. Rev. D
82, 044040 (2010)
11. Bona, C., Stela, J.: “Swiss cheese” models with pressure. Phys. Rev. D 36, 2915 (1987)
12. Bonnor, W.B.: A generalization of the Einstein-Straus vacuole. Class. Quantum Grav. 17,
2739 (2000)
13. Booth, I., Brits, L., Gonzalez, J.A., Van den Broeck, V.: Marginally trapped tubes and
dynamical horizons. Class. Quantum Grav. 23, 413 (2006)
14. Bousso, R.: Adventures in de Sitter space. Preprint arXiv:hep-th/0205177
15. Brevik, I., Nojiri, S., Odintsov, S.D., Vanzo, L.: Entropy and universality of the Cardy-
Verlinde formula in a dark energy universe. Phys. Rev. D 70, 043520 (2004)
16. Brown, B.A., Lindesay, J.: Class. Quantum Grav. 26, 045010 (2009)
17. Brown, I., Behrend, J., Malik, K.: Gauges and cosmological backreaction. J. Cosmol.
Astropart. Phys. 11, 027 (2009)
18. Buchdahl, H.A.: Static solutions of the brans-dicke equations. Int. J. Theor. Phys. 6, 407
(1972)
19. Buchdahl, H.A.: Isotropic coordinates and Schwarzschild metric. Int. J. Theor. Phys. 24, 731
(1985)
20. Buchert, T.: On average properties of inhomogeneous fluids in general relativity: dust
cosmologies. Gen. Rel. Gravit. 32, 105 (2000)
21. Buchert, T.: Backreaction issues in relativistic cosmology and the dark energy debate. AIP
Conf. Proc. 910, 361 (2007)
22. Buchert, T.: Gen. Rel. Gravit. 40, 467 (2008)
23. Buchert, T., Carfora, M.: Regional averaging and scaling in relativistic cosmology. Class.
Quantum Grav. 19, 6109 (2002)
24. Buchert, T., Carfora, M.: On the curvature of the present-day universe. Class. Quantum Grav.
25, 195001 (2008)
25. Caldwell, R.R., Kamionkowski, M., Weinberg, N.N.: Phantom energy and cosmic doomsday.
Phys. Rev. Lett. 91, 071301 (2003)
160 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

26. Capozziello, S., Carloni, S., Troisi, A.: Quintessence without scalar fields. Recent Res. Dev.
Astron. Astrophys. 1, 625 (2003)
27. Carr, B.J.: Primordial black holes: do they exist and are they useful? Preprint astro-
ph/0511743
28. Carrera, M., Giulini, D.: Influence of global cosmological expansion on local dynamics and
kinematics. Rev. Mod. Phys. 82, 169 (2010)
29. Carrera, M., Giulini, D.: On the generalization of McVittie’s model for an inhomogeneity in
a cosmological spacetime. Phys. Rev. D 81, 043521 (2010)
30. Carroll, S.M., Duvvuri, V., Trodden, M., Turner, M.S.: Is cosmic speed-up due to new
gravitational physics? Phys. Rev. D 70, 043528 (2004)
31. Castro, A., Rodriguez, M.J.: Universal properties and the first law of black hole inner
mechanics. Phys. Rev. D 86, 024008 (2012)
32. Chen, S., Jing, J.: Quasinormal modes of a black hole surrounded by quintessence. Class.
Quantum Grav. 22, 4651 (2005)
33. Coleman, S.R., De Luccia, F.: Gravitational effects on and of vacuum decay. Phys. Rev. D 21,
3305 (1980)
34. Coley, A.A., van den Hoogen, R.J.: Dynamics of multi-scalar-field cosmological models and
assisted inflation. Phys. Rev. D 62, 023517 (2000)
35. Cvetic, M., Larsen, F.: General rotating black holes in string theory: greybody factors and
event horizons. Phys. Rev. D 56, 4994 (1997)
36. Cvetic, M., Larsen, F.: Greybody factors and charges in Kerr/CFT. J. High Energy Phys. 0909,
088 (2009)
37. Cvetic, M., Gibbons, G.W., Pope, C.N.: Universal area product formulae for rotating and
charged black holes in four and higher dimensions. Phys. Rev. Lett. 106, 121301 (2011)
38. da Silva, A., Fontanini, M., Guariento, D.C.: How the expansion of the universe determines
the causal structure of McVittie spacetimes. Phys. Rev. D 87, 064030 (2013)
39. da Silva, A., Guariento, D.C., Molina, C.: Cosmological black holes and white holes with
time-dependent mass. Phys. Rev. D 91, 084043 (2015)
40. De Felice, A., Tsujikawa, S.: f .R/ theories. Living Rev. Relat. 13, 3 (2010)
41. de Freitas Pacheco, J.A., Horvath, J.E.: Generalized second law and phantom cosmology.
Class. Quantum Grav. 24, 5427 (2007)
42. Dicke, R.H.: Mach’s principle and invariance under transformation of units. Phys. Rev. 125,
2163 (1962)
43. Di Criscienzo, R., Nadalini, M., Vanzo, L., Zerbini, S., Zoccatelli, G.: On the Hawking
radiation as tunneling for a class of dynamical black holes. Phys. Lett. B 657, 107 (2007)
44. Einstein, A., Straus, E.G.: The influence of the expansion of space on the gravitation fields
surrounding the individual stars. Rev. Mod. Phys. 17, 120 (1945)
45. Einstein, A., Straus, E.G.: Corrections and additional remarks to our paper: the influence of
the expansion of space on the gravitation fields surrounding the individual stars. Rev. Mod.
Phys. 18, 148 (1946)
46. Faraoni, V.: Cosmology in Scalar-Tensor Gravity. Kluwer Academic, Dordrecht (2004)
47. Faraoni, V.: Hawking temperature of expanding cosmological black holes. Phys. Rev. D 76,
104042 (2007)
48. Faraoni, V.: Analysis of the Sultana-Dyer cosmological black hole solution of the Einstein
equations. Phys. Rev. D 80, 044013 (2009)
49. Faraoni, V.: Evolving black hole horizons in general relativity and alternative gravity. Galaxies
1, 114 (2013)
50. Faraoni, V., Israel, W.: Dark energy, wormholes, and the big rip. Phys. Rev. D 71, 064017
(2005)
51. Faraoni, V., Jacques, A.: Cosmological expansion and local physics. Phys. Rev. D 76, 063510
(2007)
52. Faraoni, V., Vitagliano, V.: Horizon thermodynamics and spacetime mappings. Phys. Rev. D
89, 064015 (2014)
References 161

53. Faraoni, V., Zambrano Moreno, A.F.: Are quantization rules for horizon areas universal? Phys.
Rev. D 88, 044011 (2013)
54. Faraoni, V., Gao, C., Chen, X., Shen, Y.-G.: What is the fate of a black hole embedded in an
expanding universe? Phys. Lett. B 671, 7 (2009)
55. Faraoni, V., Zambrano Moreno, A.F., Nandra, R.: Making sense of the bizarre behavior of
horizons in the McVittie spacetime. Phys. Rev. D 85, 083526 (2012)
56. Faraoni, V., Zambrano Moreno, A.F., Prain, A.: Charged McVittie spacetime. Phys. Rev. D
89, 103514 (2013)
57. Farhi, E., Guth, A.H., Guven, J.: Is it possible to create a universe in the laboratory by quantum
tunneling? Nucl. Phys. B 339, 417 (1990)
58. Ferraris, M., Francaviglia, M., Spallicci, A.: Associated radius, energy and pressure of
McVittie’s metric in its astrophysical application. Nuovo Cimento 111B, 1031 (1996)
59. Figueras, P., Hubeny, V.E., Rangamani, M., Ross, S.F.: Dynamical black holes and expanding
plasmas. J. High Energy Phys. 0904, 137 (2009)
60. Finch, T.K., Lindesay, J.: Global causal structure of a transient black object. Preprint
arXiv:1110.6928
61. Fisher, I.Z.: Scalar mesostatic field with regard for gravitational effects. Zh. Eksp. Teor. Fiz.
18, 636 (1948) (translated in arXiv:gr-qc/9911008)
62. Fonarev, O.A.: Exact Einstein scalar field solutions for formation of black holes in a
cosmological setting. Class. Quantum Grav. 12, 1739 (1995)
63. Galli, P., Ortin, T., Perz, J., Shahbazi, C.S.: Non-extremal black holes of N D 2; d D 4
supergravity. J. High Energy Phys. 1107, 041 (2011)
64. Gao, C.J., Zhang, S.N.: Reissner-Nordstrom metric in the Friedman-Robertson-Walker
universe. Phys. Lett. B 595, 28 (2004)
65. Gao, C.J., Zhang, S.N.: Higher dimensional Reissner-Nordstrom-FRW metric. Gen. Rel.
Gravit. 38, 23 (2006)
66. Gao, C., Chen, X., Faraoni, V., Shen, Y.-G.: Does the mass of a black hole decrease due to
accretion of phantom energy? Phys. Rev. D 78, 024008 (2008)
67. Gao, C., Chen, X., Shen, Y.-G., Faraoni, V.: Black holes in the universe: generalized Lemaître-
Tolman-Bondi solutions. Phys. Rev. D 84, 104047 (2011)
68. Gomes, H., Gryb, S., Koslowski, T.: Einstein gravity as a 3D conformally invariant theory.
Class. Quantum Grav. 28, 045005 (2011)
69. Gonzalez, J.A., Guzman, F.S.: Accretion of phantom scalar field into a black hole. Phys. Rev.
D 79, 121501 (2009)
70. Gonzalez-Diaz, P.F., Siguenza, C.L.: Phantom thermodynamics. Nucl. Phys. B 697, 363
(2004)
71. Green, S.R., Wald, R.M.: New framework for analyzing the effects of small scale inhomo-
geneities in cosmology. Phys. Rev. D 83, 084020 (2011)
72. Guariento, D.C., Horvath, J.E., Custodio, P.S., de Freitas Pacheco, J.A.: Evolution of
primordial black holes in a radiation and phantom energy environment. Gen. Rel. Gravit.
40, 1593 (2008)
73. Guo, Z.-K., Piao, Y.-S., Cai, R.-G., Zhang, Y.-Z.: Cosmological scaling solutions and cross
coupling exponential potential. Phys. Lett. B 576, 12 (2003)
74. Harada, T., Carr, B.J.: Upper limits on the size of a primordial black hole. Phys. Rev. D 71,
104009 (2005)
75. Harada, T., Carr, B.J.: Growth of primordial black holes in a universe containing a massless
scalar field. Phys. Rev. D 71, 104010 (2005)
76. Harada, T., Maeda, H., Carr, B.J.: Nonexistence of self-similar solutions containing a black
hole in a universe with a stiff fluid or scalar field or quintessence. Phys. Rev. D 74, 024024
(2006)
77. Hawking, S.W., Ellis, G.F.R.: The Large Scale Structure of Space-Time. Cambridge Univer-
sity Press, Cambridge (1973)
78. He, X., Wang, B., Wu, S.-F., Lin, C.-Y.: Quasinormal modes of black holes absorbing dark
energy. Phys. Lett. B 673, 156 (2009)
162 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

79. Horowitz, G.T., Maldacena, J.M., Strominger, A.: Nonextremal black hole microstates and U
duality. Phys. Lett. B 383, 151 (1996)
80. Hsu, D.H., Jenskins, A., Wise, M.B.: Gradient instability for w < 1. Phys. Lett. B 597, 270
(2004)
81. Hubeny, V.: The fluid/gravity correspondence: a new perspective on the membrane paradigm.
Class. Quantum Grav. 28, 114007 (2011)
82. Husain, V., Martinez, E.A., Nuñez, D.: Exact solution for scalar field collapse. Phys. Rev. D
50, 3783 (1994)
83. Izquierdo, G., Pavon, D.: The Generalized second law in phantom dominated universes in the
presence of black holes. Phys. Lett. B 639, 1 (2006)
84. Janis, A.I., Newman, E.T., Winicour, J.: Reality of the Schwarzschild singularity. Phys. Rev.
Lett. 20, 878 (1968)
85. Kaloper, N., Kleban, M., Martin, D.: McVittie’s legacy: black holes in an expanding universe.
Phys. Rev. D 81, 104044 (2010)
86. Kitada, Y., Maeda, K.: Cosmic no hair theorem in homogeneous space-times. 1. Bianchi
models. Class. Quantum Grav. 10, 703 (1993)
87. Knop, R., et al.: New constraints on ˝M ; ˝ , and w from an independent set of 11 high-
redshift supernovae observed with the Hubble Space Telescope. Astrophys. J. 598, 102 (2003)
88. Kolb, E.W., Matarrese, S., Riotto, A.: On cosmic acceleration without dark energy. New J.
Phys. 8, 322 (2006)
89. Kolb, E., Marra, V., Matarrese, S.: Cosmological background solutions and cosmological
backreactions. Gen. Rel. Gravit. 42, 1399 (2010)
90. Komatsu, E., et al.: Seven-year Wilkinson Microwave Anisotropy Probe (WMAP ) observa-
tions: cosmological interpretation. Astrophys. J. (Suppl.) 192, 18 (2011)
91. Kottler, F.: Über die physikalischen ndlagen der Einsteinschen gravitationstheorie. Ann. Phys.
(Leipzig) 361, 401 (1918)
92. Krasiński, A.: Inhomogeneous Cosmological Models. Cambridge University Press, Cam-
bridge (1997)
93. Kustaanheimo, P., Qvist, B.: A note on some general solutions of the Einstein field equations
in a spherically symmetric world. Comm. Phys. Math. Soc. Sci. Fennica 13(16), 1 (1948)
(reprinted in Gen. Rel. Gravit. 30, 659 (1998))
94. Lake, K., Abdelqader, M.: More on McVittie’s legacy: a Schwarzschild-de Sitter black and
white hole embedded in an asymptotically CDM cosmology. Phys. Rev. D 84, 044045
(2011)
95. Landry, P., Abdelqader, M., Lake, K.: McVittie solution with a negative cosmological
constant. Phys. Rev. D 86, 084002 (2012)
96. Larena, J.: Spatially averaged cosmology in an arbitrary coordinate system. Phys. Rev. D 79,
084006 (2009)
97. Larena, J., Buchert, T., Alimi, J.-M.: Correspondence between kinematical backreaction and
scalar field cosmologies: the ‘morphon field’. Class. Quantum Grav. 23, 6379 (2006)
98. Larena, J., Alimi, J.-M., Buchert, T., Kunz, M., Corasaniti, P.: Testing backreaction effects
with observations. Phys. Rev. D 79, 083011 (2009)
99. Larsen, F.: String model of black hole microstates. Phys. Rev. D 56, 1005 (1997)
100. Le Delliou, M., Mimoso, J.P., Mena, F.C., Fontanini, M., Guariento, D.C., Abdalla, E.:
Separating expansion and collapse in general fluid models with heat flux. Phys. Rev. D 88,
027301 (2013)
101. Li, N., Schwarz, D.J.: Onset of cosmological backreaction. Phys. Rev. D 76, 083011 (2007)
102. Li, N., Schwarz, D.J.: Scale dependence of cosmological backreaction. Phys. Rev. D 78,
083531 (2008)
103. Liddle, A.R., Mazumdar, A., Schunck, F.E.: Assisted inflation. Phys. Rev. D 58, 061301
(1998)
104. Lima, J.A.S., Alcaniz, J.S.: Thermodynamics and spectral distribution of dark energy. Phys.
Lett. B 600, 191 (2004)
105. Lindesay, J.: Found. Phys. 37, 1181 (2007)
References 163

106. Lindesay, J.: Foundations of Quantum Gravity, p. 282. Cambridge University Press, Cam-
bridge (2013)
107. Lindesay, J., Sheldon, P.: Class. Quantum Grav. 27, 215015 (2010)
108. Maeda, H., Harada, T., Carr, B.J.: Self-similar cosmological solutions with dark energy. II.
Black holes, naked singularities, and wormholes. Phys. Rev. D 77, 024023 (2008)
109. Maeda, H., Harada, T., Carr, B.J.: Cosmological wormholes. Phys. Rev. D 79, 044034 (2009)
110. Majhi, B.R.: Thermodynamics of Sultana-Dyer black hole. J. Cosmol. Astropart. Phys. 1405,
014 (2014)
111. Malik, K.A., Wands, D.: Dynamics of assisted inflation. Phys. Rev. D 59, 123501 (1999)
112. Marra, V.: A back-reaction approach to dark energy. Preprint arXiv:0803.3152
113. Marra, V., Kolb, E., Matarrese, S., Riotto, A.: Cosmological observables in a Swiss-cheese
universe. Phys. Rev. D 76, 123004 (2007)
114. Marra, V., Kolb, E., Matarrese, S.: Light-cone averages in a Swiss-cheese universe. Phys. Rev.
D 77, 023003 (2008)
115. Mashhoon, B., Partovi, M.H.: Gravitational collapse of a charged fluid sphere. Phys. Rev. D
20, 2455 (1979)
116. McClure, M.L.: Cosmological black holes as models of cosmological inhomogeneities. PhD
thesis, University of Toronto (2006)
117. McClure, M.L., Dyer, C.C.: Asymptotically Einstein-de Sitter cosmological black holes and
the problem of energy conditions. Class. Quantum Grav. 23, 1971 (2006)
118. McClure, M.L., Dyer, C.C.: Matching radiation-dominated and matter-dominated Einstein-
de Sitter universes and an application for primordial black holes in evolving cosmological
backgrounds. Gen. Rel. Gravit. 38, 1347 (2006)
119. McVittie, G.C.: The mass-particle in an expanding universe. Mon. Not. R. Astr. Soc. 93, 325
(1933)
120. Meessen, P., Ortin, T., Perz, J., Shahbazi, C.S.: Black holes and black strings of N D 2; d D 5
supergravity in the H-FGK formalism. J. High Energy Phys. 1209, 001 (2012)
121. Miller, J.C., Musco, I.: Causal horizons and topics in structure formation. Preprint
arXiv:1412.8660
122. Mosheni Sadjadi, H.: Generalized second law in a phantom-dominated universe. Phys. Rev.
D 73, 0635325 (2006)
123. Nandra, R., Lasenby, A.N., Hobson, M.P.: The effect of a massive object on an expanding
universe. Mon. Not. R. Astr. Soc. 422, 2931 (2012)
124. Nandra, R., Lasenby, A.N., Hobson, M.P.: The effect of an expanding universe on massive
objects. Mon. Not. R. Astr. Soc. 422, 2945 (2012)
125. Nariai, H.: On some static solutions of Einstein’s gravitational field equations in a spherically
symmetric case. Sci. Rep. Tohoku Univ. 34, 160 (1950)
126. Nariai, H.: On a new cosmological solution of Einstein’s field equations of gravitation. Sci.
Rep. Tohoku Univ. 35, 62 (1951)
127. Nojiri, S., Odintsov, S.D.: Final state and thermodynamics of a dark energy universe. Phys.
Rev. D 70, 103522 (2004)
128. Nojiri, S., Odintsov, S.D.: Quantum escape of sudden future singularity. Phys. Lett. B 595, 1
(2004)
129. Nolan, B.C.: Sources for McVittie’s mass particle in an expanding universe. J. Math. Phys.
34, 1 (1993)
130. Nolan, B.C.: A Point mass in an isotropic universe: existence, uniqueness and basic properties.
Phys. Rev. D 58, 064006 (1998)
131. Nolan, B.C.: A Point mass in an isotropic universe. 2. Global properties. Class. Quantum
Grav. 16, 1227 (1999)
132. Nolan, B.C.: A Point mass in an isotropic universe 3. The region R  2m. Class. Quantum
Grav. 16, 3183 (1999)
133. Paranjape, A., Singh, T.P.: The possibility of cosmic acceleration via spatial averaging in
Lemaitre-Tolman-Bondi models. Class. Quantum Grav. 23, 6955 (2006)
164 4 Inhomogeneities in Cosmological “Backgrounds” in Einstein Theory

134. Paranjape, A., Singh, T.P.: The spatial averaging limit of covariant macroscopic gravity: scalar
corrections to the cosmological equations. Phys. Rev. D 76, 044006 (2007)
135. Paranjape, A., Singh, T.P.: Explicit cosmological coarse graining via spatial averaging. Gen.
Rel. Gravit. 40, 139 (2008)
136. Perlmutter, S., et al.: Discovery of a supernova explosion at half the age of the Universe.
Nature 391, 51 (1998)
137. Perlmutter, S., et al.: Measurements of ˝ and  from 42 high-redshift supernovae. Astrophys.
J. 517, 565 (1999)
138. Poisson, E., Israel, W.: The internal structure of black holes. Phys. Rev. D 41, 1796 (1990)
139. Ras̈änen, S.: Dark energy from backreaction. J. Cosmol. Astropart. Phys. 02, 003 (2004)
140. Räsänen, S.: Backreaction in the Lemaitre-Tolman-Bondi model. J. Cosmol. Astropart. Phys.
11, 010 (2004)
141. Raychaudhuri, A.K.: Theoretical Cosmology. Clarendon Press, Oxford (1979)
142. Riess, A.G., et al.: Observational evidence from supernovae for an accelerating universe and
a cosmological constant. Astron. J. 116, 1009 (1998)
143. Riess, A.G., et al.: An indication of evolution of type Ia supernovae from their risetimes.
Astron. J. 118, 2668 (1999)
144. Riess, A.G., et al.: The farthest known supernova: support for an accelerating universe and a
glimpse of the epoch of deceleration. Astrophys. J. 560, 49 (2001)
145. Riess, A.G., et al.: Type Ia supernova discoveries at z > 1 from the Hubble Space Telescope:
evidence for past deceleration and constraints on dark energy evolution. Astron. J. 607, 665
(2004)
146. Roberts, M.D.: Massless scalar static spheres. Astrophys. Space Sci. 200, 331 (1993)
147. Rodrigues, M.G., Zanchin, V.T.: Charged black holes in expanding Einstein-de Sitter
universes. Class. Quantum Grav. 32, 115004 (2015)
148. Saida, H.: Hawking radiation in the Swiss-cheese universe. Class. Quantum Grav. 19, 3179
(2002)
149. Saida, H., Harada, T., Maeda, H.: Black hole evaporation in an expanding universe. Class.
Quantum Grav. 24, 4711 (2007)
150. Shah, Y.P., Vaidya, P.C.: Gravitational field of a charged particle embedded in a homogeneous
universe. Tensor 19, 191 (1968)
151. Shankaranarayanan, S.: Temperature and entropy of Schwarzschild.de Sitter space-time.
Phys. Rev. D 67, 084026 (2003)
152. Sotiriou, T.P.: f .R/ gravity and scalar-tensor theory. Class. Quantum Grav. 23, 5117 (2006)
153. Sotiriou, T.P.: PhD thesis, International School for Advanced Studies, Trieste (2007) (preprint
arXiv:0710.4438)
154. Sotiriou, T.P.: In: Kleinert, H., Jantzen, R.T., Ruffini, R. (eds.) Proceedings of the Eleventh
Marcel Grossmann Meeting on General Relativity, pp. 1223–1226. World Scientific, Singa-
pore (2008) (preprint arXiv:gr-qc/0611158)
155. Sotiriou, T.P., Faraoni, V.: f .R/ theories of gravity. Rev. Mod. Phys. 82, 451 (2010)
156. Sotiriou, T.P., Liberati, S.: Metric-affine f .R/ theories of gravity. Ann. Phys. (N.Y.) 322, 935
(2007)
157. Sotiriou, T.P., Liberati, S.: The metric-affine formalism of f .R/ gravity. J. Phys. Conf. Ser. 68,
012022 (2007)
158. Sultana, J., Dyer, C.C.: Cosmological black holes: a black hole in the Einstein-de Sitter
universe. Gen. Rel. Gravit. 37, 1349 (2005)
159. Sun, C.-Y.: Phantom energy accretion onto black holes in a cyclic universe. Phys. Rev. D 78,
064060 (2008)
160. Sushkov, S.V., Kim, S.-W.: Cosmological evolution of a ghost scalar field. Gen. Rel. Gravit.
36, 1671 (2004)
161. Sussman, R.: Conformal structure of a Schwarzschild black hole immersed in a Friedman
universe. Gen. Rel. Gravit. 17, 251 (1985)
162. Tonry, J.L., et al.: Cosmological results from high-z supernovae. Astrophys. J. 594, 1 (2003)
References 165

163. Tsagas, C.G., Challinor, A., Maartens, R.: Relativistic cosmology and large-scale structure.
Phys. Rep. 465, 61 (2008)
164. Virbhadra, K.S.: Janis-Newman-Winicour and Wyman solutions are the same. Int. J. Mod.
Phys. A 12, 4831 (1997)
165. Visser, M.: Quantization of area for event and Cauchy horizons of the Kerr-Newman black
hole. J. High Energy Phys. 1206, 023 (2012)
166. Visser, M.: Area products for stationary black hole horizons. Phys. Rev. D 88, 044014 (2013)
167. Vitagliano, V., Liberati, S., Faraoni, V.: Averaging inhomogeneities in scalar-tensor cosmol-
ogy. Class. Quantum Grav. 26, 215005 (2009)
168. Vollick, D.N.: Phys. Rev. D 68, 063510 (2003)
169. Wald, R.M.: General Relativity. Chicago University Press, Chicago (1984)
170. Weyl, H.: Zur Gravitationstheorie. Ann. Phys. (Leipzig) 54, 117 (1917)
171. Weyl, H.: Gravitation und Elektrizit. Stz. Preuss. Akad. Wiss. 1, 465 (1918)
172. Weyl, H.: Space, Time, Matter. Dover, New York (1950)
173. Wiltshire, D.L.: Cosmic clocks, cosmic variance and cosmic averages. New J. Phys. 9, 377
(2007)
174. Wiltshire, D.L.: Exact solution to the averaging problem in cosmology. Phys. Rev. Lett. 99,
251101 (2007)
175. Wyman, M.: Static spherically symmetric scalar fields in general relativity. Phys. Rev. D 24,
839 (1981)
Chapter 5
Cosmological Inhomogeneities in Alternative
Gravity

To such an extent does nature delight and abound in variety that

among her trees there is not one plant to be found which is
exactly like another; and not only among the plants, but among
the boughs, the leaves and the fruits, you will not find one which
is exactly similar to another.
—Leonardo da Vinci

5.1 Introduction

After studying inhomogeneous spacetimes describing central condensations embed-

ded in FLRW spaces in Einstein’s theory, we now turn to qualitatively similar
spacetimes in the context of alternative theories of gravity. In addition to the reasons
already mentioned in the previous chapter for the study of cosmological black holes,
there is motivation to extend the analysis to alternative theories of gravity. We
have already mentioned that the McVittie spacetime is a solution of the cuscuton
theory and that the generalized McVittie solution solves the field equations of
Horndeski gravity. Let us consider now, in particular, Brans-Dicke [5] and scalar-
tensor theories of gravity [4, 40, 56]: these theories are the prototypical alternatives
to General Relativity and allow for a varying gravitational strength. While usually
only the cosmological variation of the gravitational “constant” Geff with time is
studied, inhomogeneous spacetimes describing central objects embedded in FLRW
“backgrounds” constitute toy models to study the spatial variation of Geff [13, 48].
Another motivation is that explicit analytic examples of evolving apparent horizons
would be useful to study Hawking radiation and black hole thermodynamics in
a fully dynamical situation in alternative gravity, and bring into light possible
differences with General Relativity. Although we restrict ourselves to 4-dimensional
spacetime manifolds, once one begins to study alternative theories the door is open
for further generalization. For example, exact cosmological and time-dependent
black holes are of interest in higher-dimensional Gauss-Bonnet gravity [41], and
other examples arise from intersecting branes in supergravity [36].
The known solutions of field equations alternative to the Einstein equations with
the desired properties are usually spherically symmetric. It is unknown whether, in
general, the concepts of Misner-Sharp-Hernandez mass and Kodama vector can be

© Springer International Publishing Switzerland 2015 167

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6_5
168 5 Cosmological Inhomogeneities in Alternative Gravity

extended beyond General Relativity,1 and some of the quantities needed to write
down the first law of thermodynamics become obscure. Since the Misner-Sharp-
Hernandez mass is a special case of the Hawking-Hayward quasi-local energy [31],
it seems rather natural to use this construct as the internal energy when abandoning
spherical symmetry. The Kodama vector, however, is not defined once this special
symmetry is given up.
The concepts of apparent and trapping horizon do not depend on the field
equations of the theory. Moreover, there have been several studies of horizon entropy
for event horizons in alternative gravities (see Ref. [22] for a review).

5.2 Brans-Dicke Cosmological Black Holes

In string theories [30] a dilaton field coupled non-minimally to the Ricci curvature
mimics a Brans-Dicke scalar field (in its low-energy limit, bosonic string theory
reduces precisely to an !0 D 1 Brans-Dicke theory [6, 28]). This fact has added
motivation for the study of scalar-tensor theories in general. Brans-Dicke gravity
corresponds to the action [5]
Z
p !0 ab
SBD D d4 x g R  g ra rb  C 2 L (m) ; (5.1)


where   8 , L (m) is the matter Lagrangian density, and the Brans-Dicke scalar
field  effectively plays the role of the inverse of the gravitational coupling. !0 is
the “Brans-Dicke parameter”.
Contrary to General Relativity, even static, asymptotically flat, spherically
symmetric black holes in scalar-tensor gravity are not forced to be Schwarzschild:
the Jebsen-Birkhoff theorem is peculiar to Einstein theory and breaks down in more
general contexts, even in the simplest Brans-Dicke case (5.1). In this theory, what
can be rescued is only a very weak form of the theorem: if the Brans-Dicke scalar
field is required to be time-independent in electrovacuo, then the metric is static (but
not necessarily the Schwarzschild or Reissner-Nordström one) [23]. In this form,
however, the theorem is not very useful and it allows for substantial departures from
the Schwarzschild geometry.2 To the extent that astrophysical black holes can be
considered as isolated, however, all physically reasonable (that is, stable and not
fine-tuned) black holes of scalar-tensor gravity reduce to general-relativistic black
holes [51]. Let us consider now asymptotically FLRW solutions of Brans-Dicke
theory which do not fall into this category. The best known solutions are the analytic
ones found by Clifton, Mota and Barrow [13] (discussed in Ref. [27], which we fol-
low here) and the numerical ones of Sakai and Barrow [48]. We focus on the former.

1
See Ref. [14] for a proposal.
2
The Jebsen-Birkhoff theorem, however, holds in Gauss-Bonnet gravity [58].
5.2 Brans-Dicke Cosmological Black Holes 169

5.2.1 Clifton-Mota-Barrow Black Holes

The matter source is assumed to be a perfect fluid with energy density (m) ,
pressure P(m) , and equation of state P(m) D .  1/ (m) , with D constant [13].
The spherically symmetric and dynamical Clifton-Mota-Barrow line element is [13]

ds2 D e.Nr/ dt2 C a2 .t/e .Nr/

.dNr2 C rN 2 d˝.2/
2
/; (5.2)

where
 2˛
.Nr/
1  2˛N
m
e D r
 A2˛ ; (5.3)
1 C 2˛N
m
r
 m 4 2 .˛1/.˛C2/
.Nr/
e D 1C A˛ ; (5.4)
2˛Nr
  3!2!0 .2 /C2
t 0 .2 /C4
a.t/ D a0  a tˇ ; (5.5)
t0
  3! 2.43 /
t 0 .2 /C4 2 2
.t; rN / D 0 A ˛ .˛ 1/ ; (5.6)
t0
s
2.!0 C 2/
˛D ; (5.7)
2!0 C 3
 
a0 3 2˛
(m)
.t; rN / D 0(m) A ; (5.8)
a.t/

!0 is the Brans-Dicke parameter, m is a mass parameter, ˛; 0 ; a0 , 0(m) and t0 are

positive constants (0 , 0(m) , and t0 are not independent). rN is the isotropic radius
related to the Schwarzschild radial coordinate r by
 m 2
r  rN 1 C ; (5.9)
2˛Nr
so that
 
m2
dr D 1  2 2 dNr : (5.10)
4˛ rN

The constant ˛ is real for !0 < 2 and for !0 > 3=2. As customary in Brans-
Dicke theory [5, 17, 29], we assume that !0 > 3=2 and ˇ  0. The metric (5.2) is
separable and reduces to the spatially flat FLRW one if m is set to zero. If ¤ 2, set-
ting !0 D .  2/1 yields ˇ D 0 and the geometry becomes static (interestingly,
the scalar field remains time-dependent). If instead D 2 or D 4=3, then ˇ D 1=2
170 5 Cosmological Inhomogeneities in Alternative Gravity

p
and the scale factor a.t/  t irrespective of the value of the parameter !0 . These
special cases will be discussed separately.
The areal radius is
 m 2 1 .˛1/.˛C2/
R D a.t/Nr 1 C A˛
2˛Nr
1
D a.t/rA ˛ .˛1/.˛C2/ (5.11)

and the line element is written as

2 2 2/
ds2 D A2˛ dt2 C a2 .t/A ˛ .˛ dr2 C R2 d˝.2/
2
: (5.12)

Using the relation between differentials

1
dR  aP .t/rA ˛ .˛1/.˛C2/ dt
dr D 1  ; (5.13)
a.t/A ˛ .˛1/.˛C2/2 A2 C ˛m2 r .˛  1/.˛ C 2/

the line element becomes

aP 2 .t/r2 2 .˛2 C2˛2/

ds2 D  A2˛  A˛ dt2
B2 .Nr/
aP .t/r ˛2 C3˛2 A2 .Nr/
2 2
A ˛ dRdt C 2 dR2 C R2 d˝.2/
2
; (5.14)
B .Nr/ B .Nr/

where

.˛  1/.˛ C 2/ m
B.Nr/  A2 .Nr/ C (5.15)
˛2 r
s
2.!0 C 2/
is positive because ˛ D  1.
2!0 C 3
The dtdR cross-term is eliminated by introducing the new time Nt defined by [27]

1
dNt D Œdt C .t; R/dr ; (5.16)
F.t; R/

where .t; R/ is a function to be determined and F.t; R/ is an integrating factor, as

usual. With this coordinate, the line element becomes

aP 2 .t/r2 2 .˛2 C2˛2/ 2 2

ds2 D  A2˛  A˛ F dNt
B.Nr/2

2˛ aP 2 .t/r2 2 .˛2 C2˛2/ F aP .t/r ˛2 C3˛2
C 2 F A  A˛ 2 A ˛ dRdNt
B.Nr/2 B.Nr/2
5.2 Brans-Dicke Cosmological Black Holes 171

A2 2 aP 2 .t/r2 2 .˛2 C2˛2/

C  A2˛  A˛
B.Nr/2 B.Nr/2

aP .t/r ˛2 C3˛2
C2 A ˛ dR2 C R2 d˝.2/
2
: (5.17)
B.Nr/2

By setting

˛ 2 C3˛2
aP .t/r A ˛
D (5.18)
B2 D.t; rN /

with

aP 2 .t/r2 4 .˛1/
D.t; rN /  1  A˛ ; (5.19)
B2
the line element assumes the form [27]
 
2 2˛ 2 H 2 2 2.2˛/ A2
ds D A DF dt C N2 RA C 2 dR2 C R2 d˝.2/
2
; (5.20)
B4 D B

where H  aP .t/=a.t/. The apparent horizons, roots of gRR D 0, solve

B4 D
D 0; (5.21)
H 2 R2 A2.2˛/ C A2 B2 D
which reduces to D D 0, or

B2 A2.˛1/ D H 2 R2 : (5.22)

Then it must be
.˛  1/.˛ C 2/ ma.t/ .˛1/.˛C2/
A˛1 A2 C A ˛ D ˙HR : (5.23)
˛2 R

In an expanding universe, the square bracket is positive and the positive sign is the
only one appropriate; Eq. (5.23) becomes

.˛  1/.˛ C 2/ 2.˛1/.˛C1/
HR2  2
m a.t/A ˛  A˛C1 R D 0 : (5.24)
˛
If m > 0, the Ricci scalar diverges at R D 0, identifying a spacetime singularity, at
which also (m) is singular.3

3
The expression of the Ricci scalar is long and cumbersome and it is not reported here, see
Ref. [27].
172 5 Cosmological Inhomogeneities in Alternative Gravity

Let us discuss special limiting cases [27]. When m vanishes (no central object),
Eq. (5.24) gives R D H 1 , the radius of the FLRW Hubble horizon. This value of R
is also obtained if rN ! C1; then R becomes a comoving radius and the geometry
that of spatially flat FLRW space. In this limit Eq. (5.22) shows that A; B ! 1 (the
limit is not so straightforward in Eq. (5.24) as R ! 1 and rN ! 1). Based on these
features, the horizon at larger radii should be a cosmological one.
Consider now the static limit: when ˇ D 0, it is a.t/  a0 (Eq. (5.5)). This value
of ˇ follows from the choice !0 D .  2/1 with ¤ 2. For each value of the
Brans-Dicke parameter !0 there is at most one static Clifton-Mota-Barrow solution
corresponding to a specific choice of the equation of state of the cosmic fluid.
In order for ˛ to be real, it must be !0 < 2 or !0 > 3=2, which translates to
> 3=2 or < 4=3 when ˇ D 0.
Equations (5.6) and (5.8) yield
 2
t 2.˛ 2 1/
.t; R/ D 0 A ˛ ; (5.25)
t0
(m) 2˛
(m)
D 0 A : (5.26)

Even though the metric gab and the matter density (m)
are static, the scalar 
depends on time. By writing the line element as

A2 2
ds2 D A2˛ dt2 C dR C R2 d˝.2/
2
; (5.27)
B2

the apparent horizon equation gRR D 0 is B D 0, equivalent to the quadratic

m  2  m2
rN 2 C ˛  2 N
r C D 0: (5.28)
˛2 4˛ 2

m2 h 2 i
The discriminant .˛ 2 / D 2 ˛ 2  2  ˛ 2 is non-negative if ˛  1 or ˛  2
˛
(using the fact that ˛  0 according to Eq. (5.7)). In the parameter range 1 < ˛ < 2
the equation gRR D 0 has no real roots and there are no apparent horizons. If ˛  1
or ˛  2 there are the real roots
q
m  
rN˙ D 2
 ˛2  2 ˙ .˛ 2  2/2  ˛ 2 ; (5.29)
˛

but they are both negative and no apparent horizon exists: the static spacetime
always contains a naked singularity.
Let us discuss now the General Relativity limit !0 ! 1. If ¤ 0 and ¤ 2,
this limit implies ˛ ! 1,  ! 0 , and4

p
4
In the case D 2 the scale factor a.t/ / t, the scalar  / t1 and the density (m) / t3 are
independent of !0 . However, the limit !0 ! 1 still yields ˛ D 1 and the various functions of
5.2 Brans-Dicke Cosmological Black Holes 173

 m 2 
1  2N m 4  2 
ds2 D  r
dt 2
C a2
.t/ 1 C dN
r C N
r 2
d˝ 2
.2/ ; (5.30)
1 C 2N
m
r
2Nr
  32
t
a.t/ D a0 ; (5.31)
t0
 t 2
0
(m)
.t/ D 0(m) A2 : (5.32)
t
This line element is recognized as that of a generalized McVittie metric, which
becomes
!2  
M.t/ 4  2 
M.t/
1 2Nra.t/
2
ds D  dt2 C a2 .t/ 1 C dNr C rN 2 d˝.2/
2
(5.33)
1C M.t/ 2Nra.t/
2Nra.t/

in isotropic coordinates, with M.t/  0 an arbitrary function of time and G10 ¤ 0,

corresponding to a radial energy flow. As already seen, the solution with “comoving
mass function” M.t/ D M0 a.t/ (where M0 is a constant) is a late-time attractor in the
class of generalized McVittie solutions (5.33). This attractor solution is precisely the
!0 ! 1 limit of the Clifton-Mota-Barrow solutions (5.2)–(5.8) (which are indeed
accreting).5
For large values of !0 , the solution (5.2)–(5.8) asymptotes to the attractor of the
generalized McVittie family of solutions.
The D 0 case describes a cosmological constant “fluid”; in this case the
exponent ˇ of the scale factor a.t/ diverges as !0 ! 1 and this scale factor
becomes a power law (the General Relativity limit of the solution is expected to
be the Schwarzschild-de Sitter-Kottler spacetime). p
If D 2, the !0 ! 1 limit yields ˛ ! 1,  / t1 , a.t/ / t, (m) / t3 A2 ,
and the line element is as in Eq. (5.30).
Leaving behind all these special cases, let us inspect now the structure and
dynamics of the apparent horizons of the generic Clifton-Mota-Barrow spacetime
m
[27]. It is convenient to introduce the variable x  , in terms of which we have
2˛Nr
1x
AD ; (5.34)
1Cx

rN have the same functional dependence as in the ¤ 2 case. The line element is a generalized
McVittie one.
5
The generalized McVittie solutions of General Relativity were introduced 2 years after the
Clifton-Mota-Barrow paper, and the attractor solution with M D M0 a.t/ could, in principle, have
been found earlier as the limit to General Relativity, but its attractor role follows from different
considerations.
174 5 Cosmological Inhomogeneities in Alternative Gravity

Fig. 5.1 Apparent horizon radii versus time (both in units of .ma /1=.1ˇ/ ) for the value
!0 D 17=12 of the Brans-Dicke parameter. The dashed curve corresponds to D 1 (dust) and
the solid curve to both D 4=3 (radiation) and D 2 (stiff matter). For dust, there is only
one apparent horizon which expands to a maximum size and then shrinks. Universes filled with
radiation or stiff matter, instead, contain naked singularities

while H D ˇ=t. The areal radius R of the apparent horizon(s) and the time t can be
expressed parametrically as functions of x, obtaining

  .˛1/.˛C2/
ˇm .1 C x/2 1  x ˛
R.x/ D a t ; (5.35)
2˛ x 1Cx
(
2˛ x 
t.x/ D 2
.1  x/2=˛
m a ˇ .1 C x/ ˛ .˛C1/
 ˇ1
1
.˛  1/.˛ C 2/
C2x .1  x/2.˛1/=˛ : (5.36)
˛

Figures 5.1–5.5 show the areal radii of the apparent horizons versus time for the
Brans-Dicke parameter values !0 D 17=12, 1=3, 1, and for various large values
of !0 (of the order of 105 ), respectively, and for various choices of the equation of
state parameter . In these figures (which follow those of Ref. [27]), R and t are
reported in units of
5.2 Brans-Dicke Cosmological Black Holes 175

Fig. 5.2 The apparent horizon radii for the Brans-Dicke parameter value !0 D 1=3. The dotted
curve corresponds to D 0 (cosmological constant). In all cases there is a single expanding
horizon and the spacetime contains a naked singularity

  1
1 m 1ˇ
.ma / 1ˇ D a0 t0 (5.37)
t0

(this normalization absorbs completely the parameters m, a0 , and t0 ).

The dotted curves describe D 0 (cosmological constant); the dashed curves
correspond to D 1 (dust), while the solid curves correspond to both D 4=3
(radiation) and D 2 (stiff matter). For both values D 4=3; 2, ˇ assumes the
value 1=2 and is independent of !0 . For the value !0 D 17=12 of the Brans-Dicke
parameter (Fig. 5.1), a cosmological constant ( D 0) gives a contracting universe
and this case is not plotted.
For !0 D 17=12 and !0 D 1=3 there is only one apparent horizon for all
of the values of explored. In most cases, this horizon is expanding forever and
the spacetime contains a naked singularity. For !0 D 17=12 and for dust, the
apparent horizon expands to a maximum size, stops, and then contracts to zero size
asymptotically [27].
When !0 D 1 (Figs. 5.3 and 5.4), for dust, radiation, and stiff matter there is
initially a single expanding apparent horizon (Fig. 5.3), then two additional apparent
horizons appear; the outer horizon expands while the inner horizon eventually
approaches the initial one, at which point they merge and disappear, reproducing
the “S-curve” phenomenology of the Husain-Martinez-Nuñez solution [27].
176 5 Cosmological Inhomogeneities in Alternative Gravity

Fig. 5.3 Apparent horizon radii for !0 D 1. For all three values of , there is a single horizon
at early times. As time progresses, two more apparent horizons appear, covering the central
singularity. Two of these horizons eventually merge and disappear; then there remains a naked
singularity in a FLRW universe, which has its own cosmological horizon. The third curve, flattened
along the time axis, is zoomed on in Fig. 5.4

If !0 D 1 and D 0 (cosmological constant, Fig. 5.4), one has similar dynamics

of the apparent horizons but the new pair of horizons forms inside the original one.
Figure 5.5 corresponds to large values of the parameter !0 . The apparent horizons
exhibit an S-curve behaviour very similar to phenomenology already encountered
in solutions of General Relativity [27].

5.2.2 Conformally Transformed Husain-Martinez-Nuñez

Spacetime

In addition to the solutions just considered, Clifton, Mota, and Barrow [13]
generated another dynamical and spherically symmetric solution of the Brans-Dicke
field equations by performing a conformal transformation of the Husain-Martinez-
Nuñez metric:

g(HMN)
ab ! ˝ 2 g(HMN)
ab D  g(HMN)
ab ; (5.38)
r
2! C 3
 ! Q D ln  : (5.39)
16
5.2 Brans-Dicke Cosmological Black Holes 177

Fig. 5.4 The same as Fig. 5.3, zoomed in

While, in studying scalar-tensor gravity, it is customary to perform the inverse

conformal transformation (from the Jordan to the Einstein frame) to end up with an
Einstein frame formulation in which the Ricci scalar couples minimally to a scalar
field with canonical kinetic energy (but the latter couples non-minimally to all forms
of matter which are not trace-free), Ref. [13] transforms a General Relativity metric
to a Jordan frame to generate a solution of the Brans-Dicke field equations. The
two-parameter geometry thus obtained is described by the line element and scalar
field [13]
 
˛ 1 p1
ds2 D A 3ˇ
.r/ dt2
  p
˛ 1C p1 .
2 ˇ 3
p
/h i
CA 3ˇ
.r/ t 3ˇ 3 dr2 C r2 A.r/d˝.2/
2
; (5.40)
˙1 p 2
.t; r/ D A 2ˇ .r/ t 3 ˇ1 ; (5.41)

where
2C p p
A.r/ D 1  ; ˇD 2! C 3 ; ! > 3=2 ; ˛ D ˙ 3=2 : (5.42)
r
178 5 Cosmological Inhomogeneities in Alternative Gravity

Increasing w
Horizon
production
R

Horizon mergers

0
0 t

Fig. 5.5 Apparent horizon radii for various (increasing) values of !0 (remember that !0 ! C1
reproduces General Relativity). Here one finds the S-curve familiar from the Husain-Martinez-
Nuñez solution of the previous chapter. The lower bend in the S-curve gets pushed at infinity as
! ! C1

Spacetime singularities are present at r D 2C and at t D 0. The physical coordinate

range is 2C < r < C1 and t > 0. The result is an inhomogeneous spacetime with
a spatially flat FLRW “background” which has scale factor
p
ˇ p3
a.t/ D t 3ˇ 3 t : (5.43)

Following [26], let us rewrite the line element using the notation and the areal radius

ds2 D A .r/ dt2 C A .r/ a2 .t/dr2 C R2 .t; r/d˝.2/

2
; (5.44)

C1
R.t; r/ D A 2 .r/ a.t/ r ; (5.45)

where
 
1
 D ˛ 1 p ; (5.46)
3ˇ
 
1
 D ˛ 1 C p : (5.47)
3ˇ

Equation (5.45) gives

5.2 Brans-Dicke Cosmological Black Holes 179

C1
dR  A 2 .r/Pa.t/rdt
dr D 1 C1
(5.48)
A 2 .r/a.t/ C.C1/
r
CA 2 .r/a.t/

which, substituted into the line element, produces

1 n h C1
i
ds2 D  D1 A  A 2 aP 2 r2 dt2
D1 .r/
C1
o
2A 2 aP r dtdR C dR2 C R2 d˝.2/2
; (5.49)

with

C. C 1/
D1 .r/ D A.r/ 1 C : (5.50)
rA.r/

Once again, the time-radius cross-term is removed by introducing a new time

coordinate T defined by

dt D dT  .t; r/dR (5.51)

which, substituted into the line element, yields

1   
ds2 D  D1 A  AC1 aP 2 r2 dT 2
D1
 C1

C 2D1 A   2AC1 aP 2 r2  2A 2 aP r dTdR
 C1 C1
 i
C 1 C 2A 2 aP r C 2A 2 aP 2 r2  D1 A  2 dR2

CR2 d˝.2/
2
:

Let us choose
C1
A 2 aP r H 2 R2
.t; r/ D D I (5.52)
D1 A  A
C1
2 aP 2 r2 D1 A  H 2 R2

then the line element reduces to

1 ˚  
ds2 D  D1 A  H 2 R2 dT 2
D1

  H 2 R2
C 1 C 2HR C H 2 R2  D1 A dR2
D1 A  H 2 R2
CR2 d˝.2/
2
: (5.53)
180 5 Cosmological Inhomogeneities in Alternative Gravity

There are now two possibilities: either D1 A  H 2 R2 vanishes, or it is different from

zero. We consider these two cases separately.
Consider first the situations in which D1 A ¤ H 2 R2 ; then the line element is
simply

1     
ds2 D  D1 A  H 2 R2 dT 2 C 1 C 2HR C H 2 R2 dR2 CR2 d˝.2/
2
: (5.54)
D1

The usual recipe gRR D 0 locating the apparent horizons is equivalent to D1 .r/ D 0,
which is satisfied by

r D .1  / C  rAH : (5.55)

Expressed using the areal radius, this equation is

 2
C1
RAH .t/ D .1  / Ca.t/ : (5.56)
1

If  < 1, this formal root lies in the physical region rAH > 2C, in which the areal
radius can be written as

a.t/r
R.t; r/ D : (5.57)
.1  2C=r/j j
C1
2

We see that R ! C1 as r ! 2CC and, since

@R 1 2C.1  /
D a.t/A 2 .r/ 1  ; (5.58)
@r 2r

1
for  < 1 it is > 1 and @R=@r > 0 if
2
 
1
r > r0  2C > 2C ; (5.59)
2

while @R=@r D 0 at r D r0 , and @R=@r < 0 for r < r0 . Because

p  
3 1
D 1C p p (5.60)
2 3 2! C 3
p
for ˛ D ˙ 3=2, the condition p < 1 for the apparent horizon to exist in the
physical region requires ˛ D C 3=2. The sufficient condition  < 1 restricts
the Brans-Dicke parameter to
5.2 Brans-Dicke Cosmological Black Holes 181

2 3
16 1 7
!< 4 p 2  35  !0 ; (5.61)
2
2 3

hence the apparent horizon exists at r0 > 2C for the parameter range

3
 < ! < !0 (5.62)
2
and it has areal radius
ˇ ˇ jC1j
ˇ  C 1 ˇ 2
ˇ
RAH .t/ D ˇ ˇ j  1j a.t/C : (5.63)
 1ˇ

This apparent horizon is comoving with the cosmic fluid and disappears in the limit
C ! 0 in which there is nopcentral inhomogeneity.
If !  !0 or if ˛ D  3=2, there is no apparent horizon and the conformal
relative of the Husain-Martinez-Nuñez spacetime
p  contains a naked singularity. In

p 3 1
particular, when ˛ D  3=2 it is  D 1C p > 0 and the areal radius
2 3ˇ

  jC1j
2C 2
R.t; r/ D 1  a.t/r (5.64)
r

is a monotonically increasing function of r in the range 2C < r < C1.

Let us consider now the case D1 A D H 2 R2 ; the line element appropriate to this
situation is
A  
ds2 D 2 2
2HR dtdR C dR2 C R2 d˝.2/
2
: (5.65)
H R
The inverse metric has components
0 1
A  HR 0 0
B  HR 0A 0 0 C
B  C
.g  / D B A C (5.66)
@ 0 0 R12 0 A
1
0 0 0 R2 sin2 

with gRR identically vanishing, hence the condition D1 A  H 2 R2 D 0 imposed

from the outset is actually the equation locating the apparent horizons. One can
write it as
p C.  1/  1
HR D D1 A D 1 C A 2 (5.67)
r
182 5 Cosmological Inhomogeneities in Alternative Gravity

(choosing the positive sign for the square root), where r D r.T; R/. In the limit
r ! C1 (and also in the parameter limit C ! 0) it is R ' H 1 , the radius of the
FLRW apparent horizon. The apparent horizons can be located numerically by using
a parametric representation similar to that already seen for the Husain-Martinez-
Nuñez spacetime. The result is an “S-curve” completely analogous to that of the
Husain-Martinez-Nuñez geometry [26].

5.3 f .R/ Cosmological Black Holes

In General Relativity, the current acceleration of the universe [2, 32, 42–46, 54]
requires that approximately 73 % of the energy content of the universe be in an
exotic form, “dark energy” with pressure P(m)   (m) [1, 33, 34]. As an alternative
to this mysterious and ad hoc dark energy, it has been reasoned that perhaps gravity
deviates from General Relativity at large scales. A simple model of modified gravity
replacing Einstein theory at large scales is f .R/ gravity [8, 9, 49, 52, 53, 55], so
called from the form of the action
Z
p 
S D d4 x g f .R/ C L (m) (5.68)

which reduces to the Einstein-Hilbert action for a linear function f of the Ricci
scalar R. This modified gravity can in principle explain the cosmic acceleration
and be theoretically consistent [7, 15, 50]. Since f .R/ theories interesting for
cosmology contain a time-varying effective cosmological “constant”, black holes
or local objects are dynamical and asymptotically FLRW, not asymptotically flat.
Very few analytic solutions of this kind are known.
A rare spherically symmetric dynamical solution in vacuum f .R/ D R 1Cı
gravity was found by Clifton [10] and studied in [21]. The parameter ı of R 1Cı
gravity is severely constrained by Solar System experiments to be in the range
ı D .1:1 ˙ 1:2/  105 [3, 11, 12, 59]. The Clifton line element is
 
ds2 D A2 .Nr/dt2 C a2 .t/B2 .Nr/ dNr2 C rN 2 d˝.2/
2
; (5.69)

where
 
1  C2 =Nr 2=q
A2 .Nr/ D ; (5.70)
1 C C2 =Nr
 
C2 4
B2 .Nr/ D 1 C A2 .Nr/ qC2ı1 ; (5.71)
rN
ı.1C2ı/
a.t/ D t 1ı ; (5.72)
q2 D 1  2ı C 4ı 2 ; (5.73)
5.3 f .R/ Cosmological Black Holes 183

in isotropic coordinates. Equation (5.69) gives back the FLRW line element when
the mass parameter C2 vanishes. This modified gravity reduces to General Relativity
when ı ! 0, in which the metric (5.69) reproduces the Schwarzschild solution in
isotropic coordinates. We assume that C2 is positive and we assume that ı > 0 for
stability (in fact, local stability of the theory requires f 00 .R/  0 [18, 19]).
Clifton’s solution (5.69)–(5.73) is conformal to the Fonarev solution of General
Relativity already seen and, since the latter is conformally static [35], also the
Clifton solution is.
A first transformation to the radial coordinate
 
C2 2
r  rN 1 C ; (5.74)
rN
 1
C2
in terms of which dNr D 1  22 dr, followed by another transformation to the
rN
areal radius
p
a.t/ B2 .Nr/ r qC2ı1
R   D a.t/ r A2 .Nr/ 2 : (5.75)
C2 2
1 C rN

brings the line element (5.69) to the form [21]

ds2 D A2 dt2 C a2 A2ı1

2 dr2 C R2 d˝.2/
2
: (5.76)

Now we have
qC2ı1 qC2ı1
dR  A2 2
aP r dt dR  A2 2
aP r dt
dr D qC2ı1 2ı1q
 qC2ı1
(5.77)
2.qC2ı1/ C2
a A2 2
C q
A
r 2
2
aA2 2
C.Nr/

(where an overdot denotes differentiation with respect to the comoving time t of the
FLRW “background”) and

2.q C 2ı  1/ C2 q 2.q C 2ı  1/ C2 a 2ı1q

C.Nr/ D 1 C A2 D 1 C A 2 : (5.78)
q r q R 2

The metric is recast as

" 2.ı1/
# qC2ı1

2 A 2A2 2
ds D A2 1 2 2 2 2
aP r 2
dt  aP r dtdR
C C2

dR2
C q C R2 d˝.2/
2
: (5.79)
A2 C2
184 5 Cosmological Inhomogeneities in Alternative Gravity

The cross-term in dtdR is eliminated by replacing t with another time coordinate Nt

which satisfies
1
dNt D Œdt C ˇ.t; R/dR ; (5.80)
F.t; R/

where F.t; R/ is an integrating factor chosen to make dNt exact. In terms of this time
coordinate, the line element assumes the form [21]
"2.ı1/
#
2 A2
ds D A2 1  aP r F 2 dNt2
2 2
C2
8 " # qC2ı1
9
< A2
2.ı1/
A 2 =
2 2 2
C2F A2 ˇ 1  P
a r  P
a r dNtdR
: C2 C2 ;
8 " # qC2ı1
9
< A
2.ı1/
2A2 2
1 = 2
C A2 1 2 2 aP 2 r2 ˇ 2 C aP rˇ C q 2 dR
: C C2 A2 C ;

CR2 d˝.2/
2
: (5.81)

The choice of ˇ
qC2ı3
A2 2
aP r
ˇD (5.82)
C2 2.ı1/
A2
1 C2
aP 2 r2

removes the unwanted cross-term and leaves the geometry in the final form
" q1 2 2
#
1 A H R
ds2 D A2 DF 2 dNt2 C q 2 1 C 2 2 dR2 C R2 d˝.2/
2
; (5.83)
A2 C C D

where H  aP =a and

2.ı1/ q1
A2 A2
D1 aP 2 r2 D 1  H 2 R2 : (5.84)
C2 C2
Further manipulation yields [21]

dR2
ds2 D A2 DF 2 dNt2 C q C R2 d˝.2/
2
: (5.85)
A2 C2 D

The equation gRR D 0 locating the apparent horizons holds if A2 C2 D D 0, giving

q
5.3 f .R/ Cosmological Black Holes 185

 
q1
A2 C2  H 2 R2 A2
q
D 0; (5.86)

so the apparent horizons are located by A2 D 0 or H 2 R2 D C2 A2 . The expression

qC1

A2 goes to zero for rN D C2 , which describes the Schwarzschild event horizon in the
limit to General Relativity ı ! 0. Now A2 D 0 identifies a singularity because the
Ricci scalar
 
6 H P C 2H 2
RD (5.87)
A2 .Nr/

diverges as rN ! C2 .
The second possibility to satisfy gRR D 0 gives H 2 R2 D C2 A2 , which trans-
qC1

lates to

2.q C 2ı  1/ C2 a 2ı1q qC1

HR D ˙ 1 C A2 2 A2 2 ; (5.88)
q R

discarding the negative sign in an expanding universe. In the limit ı ! 0 Eq. (5.88)
reduces to

2ıC2 a .1 3ı2 / 1ı

HR D 1 C A2 A2 : (5.89)
R

It is instructive to discuss two limits. In the first limit, C2 ! 0 and the central object
disappears, leaving behind FLRW space; then the coordinate rN D r approaches the
comoving radius of FLRW space and R approaches the proper radius of this space.
Equation (5.88) then degenerates into Rc D 1=H (the FLRW cosmological horizon).
In the General Relativity limit ı ! 0, Eq. (5.88) reduces to rN D C2 with H  0.
Using x  C2 =Nr, Eqs. (5.72) and (5.75) allow one to write the left hand side
of (5.88) as
qC2ı1
ı .1 C 2ı/ 2ı2 C2ı1 C2 .1  x/ q
HR D t 1ı ; (5.90)
1ı x .1 C x/ qC2ı1
q

while the right hand side of Eq. (5.88) is

qC1
.1  x/ q 2 .q C 2ı  1/ x
1C (5.91)
.1  x/2
qC1
.1 C x/ q q

and Eq. (5.88) is written as6

p
1  2ı  2ı 2 3 1
6
Here > 0 for 0 < ı < ' 0:366.
1ı 2
186 5 Cosmological Inhomogeneities in Alternative Gravity

2qC2ı2
1 .1  ı/ x .1 C x/ q
12ı2ı 2
D
t 1ı ı .1 C 2ı/ C2 .1  x/ 2.ı1/
q

2 .q C 2ı  1/ x
 1C : (5.92)
q .1  x/2

The left hand side of Eq. (5.92) vanishes at late times, and x ' 0; then there is a
single root of gRR D 0, identified with the radius of a cosmological apparent horizon
(Nr ! 1 as x D C2 =Nr ! 0 and the limit x ! 0 can also be obtained as the parameter
1ı
C2 ! 0, in which case HR ! 1 and rN ' R ' H 1 D t). At late times
ı .1 C 2ı/
there is only one cosmological apparent horizon and the Clifton spacetime contains
a naked singularity at R D 0.
A parametric representation of the apparent horizon radius R and time t is

ı.1C2ı/ C2 qC2ı1 q2ıC1

R.x/ D t.x/ 1ı .1  x/ q .1 C x/ q ; (5.93)
x
8 9 2 1ı
< .1  ı/ x .1 C x/
2.qCı1/
q 2 .q C 2ı  1/ x =
2ı C2ı1

t.x/ D 1C :
(5.94)
: ı .1 C 2ı/ C2 2.ı1/
.1  x/ q q.1  x/2 ;

The qualitative behaviour of the apparent horizon radii is described by an “S-curve”

similar to that of the Husain-Martinez-Nuñez spacetime [21].

5.4 Conclusions

To end these lectures, we are now aware of several explicit examples of time-varying
apparent horizons in General Relativity and in alternative theories of gravity. These
examples will be useful to study Hawking radiation and black hole thermodynamics
in fully dynamical situations (and in part, they are already beginning to be used for
this purpose [16, 20, 24, 25, 37–39, 47]). Many aspects of fundamental gravitational
physics are touched upon in the study of apparent horizons: perhaps the most
obvious is the long-standing issue of cosmological expansion versus local dynamics,
which prompted McVittie to produce his solution of the Einstein equations in
1933. To turn things around, one could as well say that another aspect is present,
namely the effect of a central inhomogeneity on the cosmological expansion of a
“background” universe (although it is the first effect that is usually emphasized).
Although the McVittie geometry has been largely overlooked and does not make
it to the relativity textbooks, it has seen a resurgence of interest in recent years and
it has been shown to be a solution also of very interesting modern theories, such as
Hor̆ava-Lifschitz gravity, Horndeski theory, and shape dynamics.
Analytic solutions of the field equations describing cosmological black holes
allow us to study the effect of “normal” versus phantom backgrounds on black holes,
5.4 Conclusions 187

especially the accretion of dark energy (and of its extreme form, phantom energy)
by black holes, which has been the subject of a significant amount of literature.
There is, however, an intrinsic limitation in the test fluid approximation used in the
literature to address this problem, and only exact solutions can provide some answer
to the various questions which go unanswered.
The scope enlarges when one attempts to go beyond General Relativity. Alterna-
tive theories of gravity may be required already to explain the current acceleration
of the universe without dark energy. A huge amount of literature has been
devoted to f .R/ gravity (which is a special class of scalar-tensor theories with
a complicated scalar field potential) for this purpose.7 In f .R/ and scalar-tensor
gravity designed for cosmology, the theory contains an effective time-dependent
cosmological “constant” and black holes in these theories are asymptotically FLRW,
not asymptotically flat, and they are dynamical (except for the special case of a
de Sitter “background”). Only a few analytic solutions of these theories describing
cosmological black holes are known, and they exhibit various phenomenologies of
apparent horizons.
In the context of scalar-tensor gravity, cosmological black holes provide toy
models to study the spatial variation of fundamental constants (for example, the
variation of the gravitational coupling Geff ). Containing only a scalar extra degree
of freedom, scalar-tensor gravity is a minimal modification of Einstein’s theory. As
such, it is justly regarded as the prototypical alternative to Einstein’s theory [57]
and it was quite interesting to discuss cosmological black hole solutions of the
relevant field equations. The fact that such a variety of behaviours (cosmological
black holes, naked singularities, appearing/bifurcating and merging/disappearing
pairs of apparent horizons) is contained in this relatively simple theory of gravity
induces the suspicion that more complicated theories of gravity will disclose higher
degrees of richness and complication of non-stationary horizons, which have not yet
been unveiled.
As a general consideration, while it is necessary to find new solutions of the field
equations containing apparent and trapping time-evolving horizons, and it is good
to extend the catalogue, it is more important to understand the known solutions. At
least, this is the lesson that one draws from the history of the McVittie solution.
For certain analytic dynamical solutions of well known field equations, it is not
even known if they represent black holes or naked singularities, and probably new
apparent horizon phenomenology is waiting to be discovered and fully understood.
The types of questions that should be investigated, and that we tried to address,
in the general theory and in the various specific spacetimes discussed, include:
• Do these geometries contain spacetime singularities? In relation with Cosmic
Censorship, are singularities naked, or are they dressed by some kind of horizon?
Are singularities spacelike, timelike, null, and do they change their causal

7
At the moment of writing, a crude estimate is 2000 theoretical and observational journal articles
devoted to f .R/ gravity since 2003.
188 5 Cosmological Inhomogeneities in Alternative Gravity

character? The implication of timelike naked singularities is that the spacetime

cannot be derived by regular Cauchy data and the world becomes unpredictable.
• Are there apparent/trapping horizons and, if so, where are they? Are these
apparent horizons timelike, spacelike, or null?
• Do apparent horizons in a FLRW “background” expand? How? Are they
comoving? For spherically symmetric spacetimes in General Relativity, does the
Misner-Sharp-Hernandez mass enclosed by the horizon increase? If possible, the
contributions from the mass-energy of the local object, the accretion of cosmic
fluid, the expansion/contraction of the apparent horizon, and the time evolution
of the energy density of the “background” should be identified and separated.
• What are the dynamics and phenomenology of the apparent horizons? A single
apparent horizon may originate from a spacetime singularity (as in the Husain-
Martinez-Nuñez solution of General Relativity); apparent horizons are known
to appear and disappear in pairs (as in the McVittie and in Lemaître-Tolman-
Bondi models); and they appear to jump. This variation with time of the number
of apparent horizons is sometimes interpreted as a single trapping horizon tube
which goes back and forth in time, generating what looks like the appearance,
disappearance, or bifurcation of apparent horizons when the tube is sliced with a
hypersurface of constant time. The behaviour of apparent horizons can be much
richer than the simple McVittie or “S-curve” phenomenologies, as shown by
the Clifton-Mota-Barrow solutions of Brans-Dicke theory which, when zoomed
at closely, reveal unintuitive behaviour of their apparent horizons and a richer
phenomenology as a wider parameter space is spanned.
• What are the conformal diagrams for these spacetimes? Their construction, and
the detailed analysis of the causal structure of spacetime is not, in general, an easy
task and has taken almost 80 years to be analyzed in detail for the McVittie space.
Since there is freedom in fixing the FLRW “background” for cosmological black
holes, there is also a substantial range of possibilities for the causal structure.
• What are the matter sources for these solutions of the relevant field equations?
In these lectures we have encountered a single perfect fluid, a mixture of two
non-interacting perfect fluids, canonical and phantom scalar fields (free and with
exponential potentials), cuscuton fields in Hor̆ava-Lifschitz theory, scalars from
Horndeski theory, a mixture of a perfect fluid and a Brans-Dicke field, and pure
vacuum geometry in f .R/ gravity.
Two questions stand out and remain largely unanswered:
• Are apparent and trapping horizons the right constructs to use in the study of
dynamical black holes, as theoretical laboratories in general and for astrophysical
purposes? de facto, they are the objects used in numerical relativity to predict
accurate waveforms for gravitational wave detection experiments. However, it is
possible that better concepts of “horizon” will be discovered in the future.
• It is not established beyond doubt that a thermodynamics of apparent horizons is
meaningful. It is not clear why there should be a thermodynamics of equilibrium
at all in situations in which these horizons are changing fast (in comparison with
some given time scale, for example of a “background”). However, in slowly
References 189

evolving situations, it is certainly plausible that a thermodynamics of quasi-

equilibrium can be formulated in an adiabatic approximation. More work is
needed to establish once and for all the correct concept of horizon temperature
for apparent horizons. While, for stationary black holes, various independent
methods provide the same result (the Hawking temperature), we do not have
the luxury of comparing results computed with different methods for apparent
horizons.
Due to the many questions still open, these lectures end without giving a
complete view of the field of apparent and trapping horizons because there isn’t
one yet. This subject is an active area of research and it is hoped that, although the
research is sometimes difficult and there are not many clues to follow at the moment
of writing, these notes will soon be outdated by new advances in this area.

References

1. Amendola, L., Tsujikawa, S.: Dark Energy, Theory and Observations. Cambridge University
Press, Cambridge (2010)
2. Barris, B., et al.: Twenty-three high-redshift supernovae from the Institute for Astronomy Deep
Survey: doubling the supernova sample at z > 0:7. Astrophys. J. 602, 571 (2004)
3. Barrow, J.D., Clifton, T.: Exact cosmological solutions of scale-invariant gravity theories.
Class. Quantum Grav. 23, L1 (2006)
4. Bergmann, P.G.: Comments on the scalar tensor theory. Int. J. Theor. Phys. 1, 25 (1968)
5. Brans, C., Dicke, R.H.: Mach’s principle and a relativistic theory of gravitation. Phys. Rev.
124, 925 (1961)
6. Callan, C.G., Friedan, D., Martinez, E.J., Perry, M.J.: Strings in background fields. Nucl. Phys.
B 262, 593 (1985)
7. Capozziello, S., Faraoni, V.: Beyond Einstein Gravity, a Survey of Gravitational Theories for
Cosmology and Astrophysics. Springer, New York (2010)
8. Capozziello, S., Carloni, S., Troisi, A.: Quintessence without scalar fields. Recent Res. Dev.
Astron. Astrophys. 1, 625 (2003)
9. Carroll, S.M., Duvvuri, V., Trodden, M., Turner, M.S.: Is cosmic speed-up due to new
gravitational physics? Phys. Rev. D 70, 043528 (2004)
10. Clifton, T.: Spherically symmetric solutions to fourth-order theories of gravity. Class. Quantum
Grav. 23, 7445 (2006)
11. Clifton, T., Barrow, J.D.: The power of general relativity. Phys. Rev. D 72, 103005 (2005)
12. Clifton, T., Barrow, J.D.: Class. Quantum Grav. 23, 2951 (2005)
13. Clifton, T., Mota, D.F., Barrow, J.D.: Inhomogeneous gravity. Mon. Not. R. Astr. Soc. 358, 601
(2005)
14. Cognola, G., Gorbunova, O., Sebastiani, L., Zerbini, S.: On the energy issue for a class of
modified higher order gravity black hole solutions. Phys. Rev. D 84, 023515 (2011)
15. De Felice, A., Tsujikawa, S.: f .R/ theories. Living Rev. Relat. 13, 3 (2010)
16. Deruelle, N., Sasaki, M.: In: Odintsov, S.D., Sáez-Gómez, D., Xambó-Descamps, S. (eds.)
Proceedings, Cosmology, the Quantum Vacuum, and Zeta Functions, Barcelona, 8–10 Mar
2010. Springer Proceedings in Physics, vol. 137, p. 247. Springer, Berlin/New York (2011)
17. Faraoni, V.: Cosmology in Scalar-Tensor Gravity. Kluwer Academic, Dordrecht (2004)
18. Faraoni, V.: Matter instability in modified gravity. Phys. Rev. D 74, 104017 (2006)
19. Faraoni, V.: Phys. Rev. D 75, 067302 (2007)
190 5 Cosmological Inhomogeneities in Alternative Gravity

20. Faraoni, V.: Hawking temperature of expanding cosmological black holes. Phys. Rev. D 76,
104042 (2007)
21. Faraoni, V.: Clifton’s spherical solution in f(R) vacuum harbours a naked singularity. Class.
Quantum Grav. 26, 195013 (2009)
22. Faraoni, V.: Black hole entropy in scalar-tensor and f .R/ gravity: an overview. Entropy 12,
1246 (2010)
23. Faraoni, V.: Jebsen-Birkhoff theorem in alternative gravity. Phys. Rev. D 81, 044002 (2010)
24. Faraoni, V., Nielsen, A.B.: Quasi-local horizons, horizon-entropy, and conformal field redefi-
nitions. Class. Quantum Grav. 28, 175008 (2011)
25. Faraoni, V., Vitagliano, V.: Horizon thermodynamics and spacetime mappings. Phys. Rev. D
89, 064015 (2014)
26. Faraoni, V., Zambrano Moreno, A.F.: Interpreting the conformal cousin of the Husain-
Martinez-Nuñez solution. Phys. Rev. D 86, 084044 (2012)
27. Faraoni, V., Vitagliano, V., Sotiriou, T.P., Liberati, S.: Dynamical apparent horizons in
inhomogeneous Brans-Dicke universes. Phys. Rev. D 86, 064040 (2012)
28. Fradkin, E.S., Tseytlin, A.A.: Quantum string theory effective action. Nucl. Phys. B 261, 1
(1985)
29. Fujii, Y., Maeda, K.: The Scalar-Tensor Theory of Gravitation. Cambridge University Press,
Cambridge (2003)
30. Green, M.B., Schwarz, G.H., Witten, E.: Superstring Theory. Cambridge University Press,
Cambridge (1987)
31. Hayward, S.A.: Quasilocal gravitational energy. Phys. Rev. D 49, 831 (1994)
32. Knop, R., et al.: New constraints on ˝M ; ˝ , and w from an independent set of 11 high-redshift
supernovae observed with the Hubble Space Telescope. Astrophys. J. 598, 102 (2003)
33. Komatsu, E., et al.: Seven-year Wilkinson Microwave Anisotropy Probe (WMAP ) observa-
tions: cosmological interpretation. Astrophys. J. (Suppl.) 192, 18 (2011)
34. Linder, E.V.: Resource Letter DEAU-1: Dark energy and the accelerating universe. Am. J.
Phys. 76, 197 (2008)
35. Maeda, H.: Global structure and physical interpretation of the Fonarev solution for a scalar
field with exponential potential. Preprint arXiv:0704.2731
36. Maeda, K., Ohta, N., Uzawa, K.: Dynamics of intersecting brane systems-Classification and
their applications. J. High Energy Phys. 0906, 051 (2009)
37. Majhi, B.R.: Thermodynamics of Sultana-Dyer black hole. J. Cosmol. Astropart. Phys. 1405,
014 (2014)
38. Majhi, B.R.: Conformal transformation, near horizon symmetry, Virasoro algebra, and entropy.
Phys. Rev. D 90, 044020 (2014)
39. Nielsen, A.B., Firouzjaee, J.T.: Conformally rescaled spacetimes and Hawking radiation. Gen.
Rel. Gravit. 45, 1815 (2013)
40. Nordtvedt, K.: PostNewtonian metric for a general class of scalar tensor gravitational theories
and observational consequences. Astrophys. J. 161, 1059 (1970)
41. Nozawa, M., Maeda, H.: Dynamical black holes with symmetry in Einstein-Gauss-Bonnet
gravity. Class. Quantum Grav. 25, 055009 (2008)
42. Perlmutter, S., et al.: Discovery of a supernova explosion at half the age of the Universe. Nature
391, 51 (1998)
43. Riess, A.G., et al.: Observational evidence from supernovae for an accelerating universe and a
cosmological constant. Astron. J. 116, 1009 (1998)
44. Riess, A.G., et al.: An indication of evolution of type Ia supernovae from their risetimes.
Astron. J. 118, 2668 (1999)
45. Riess, A.G., et al.: The farthest known supernova: support for an accelerating universe and a
glimpse of the epoch of deceleration. Astrophys. J. 560, 49 (2001)
46. Riess, A.G., et al.: Type Ia supernova discoveries at z > 1 from the Hubble Space Telescope:
evidence for past deceleration and constraints on dark energy evolution. Astron. J. 607, 665
(2004)
References 191

47. Saida, H., Harada, T., Maeda, H.: Black hole evaporation in an expanding universe. Class.
Quantum Grav. 24, 4711 (2007)
48. Sakai, N., Barrow, J.D.: Cosmological evolution of black holes in Brans-Dicke gravity. Class.
Quantum Grav. 18, 4717 (2001)
49. Sotiriou, T.P.: f .R/ gravity and scalar-tensor theory. Class. Quantum Grav. 23, 5117 (2006)
50. Sotiriou, T.P., Faraoni, V.: f .R/ theories of gravity. Rev. Mod. Phys. 82, 451 (2010)
51. Sotiriou, T.P., Faraoni, V.: Black holes in scalar-tensor gravity. Phys. Rev. Lett. 108, 081103
(2012)
52. Sotiriou, T.P., Liberati, S.: Metric-affine f .R/ theories of gravity. Ann. Phys. (N.Y.) 322, 935
(2007)
53. Sotiriou, T.P., Liberati, S.: The metric-affine formalism of f .R/ gravity. J. Phys. Conf. Ser. 68,
012022 (2007)
54. Tonry, J.L., et al.: Cosmological results from high-z supernovae. Astrophys. J. 594, 1 (2003)
55. Vollick, D.N.: Phys. Rev. D 68, 063510 (2003)
56. Wagoner, R.V.: Scalar-tensor theory and gravitational waves. Phys. Rev. D 1, 3209 (1970)
57. Will, C.M.: Theory and Experiment in Gravitational Physics, 2nd edn. Cambridge University
Press, Cambridge (1993)
58. Wiltshire, D.L.: Spherically symmetric solutions of Einstein-Maxwell theory with a Gauss-
Bonnet term. Phys. Lett. B 169, 36 (1986)
59. Zakharov, A.F., Nucita, A.A., De Paolis, F., Ingrosso, G.: Solar system constraints on Rn
gravity. Phys. Rev. D 74, 107101 (2006)
Appendix

I have been impressed with the urgency of doing. Knowing is not

enough; we must apply. Being willing is not enough; we must
do.
—Leonardo da Vinci

A.1 Painlevé-Gullstrand Coordinates for General

Spherically Symmetric Metrics

Beginning from the metric given by Eq. (2.94) and following the notations of
Ref. [1], we search for a new time coordinate  (Painlevé-Gullstrand time). The
transformation t !  .t; R/ yields

@ @
d D dt C dR
@t @R
and
 
1 @
dt D d  dR ;
@=@t @R

which transforms the line element (2.94) into

 
e2 .1  2M=R/ 2M @=@R
ds2 D  2
d 2
C 2e 2
1  ddR
.@=@t/ R .@=@t/2
"    #
2 2M @=@R 2 1
C e 1 C dR2 C R2 d˝.2/
2
:
R @=@t 1  2M=R
(A.1)

We now impose that the new time coordinate  is such that g11 D 1, which implies
that

© Springer International Publishing Switzerland 2015 193

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6
194 Appendix

r
@ e 2M @
D˙ : (A.2)
@R 1  2M=R R @t

Then, the metric component g01 in the new coordinates is

r
e 2M
g01 D˙ (A.3)
@=@t R

and the line element assumes the form (2.95).

A.2 Kodama Vector in FLRW Space

Here we compute the components of the Kodama vector in FLRW space in pseudo-
Painlevé-Gullstrand and in comoving coordinates.

A.2.1 Pseudo-Painlevé-Gullstrand Coordinates

In these coordinates the 2-metric hab of Eq. (2.71) and its inverse are given by
0 .1H 2 R2 kR2 =a2 / 1
HR
B 1kR2 =a2 1kR2 =a2 C
.hab / D @ A; (A.4)
HR 1
1kR2 =a2 1kR2 =a2
0 1
  1 HR
hab D @ A (A.5)
 2 2 2 2

HR 1  H R  kR =a

by decomposing the metric (3.25). The volume form on the normal 2-space is
p 1
ab D jhj .dt/a ^ .dR/b D p .ıa0 ıb1  ıa1 ıb0 / ; (A.6)
1  kR2 =a2

while

.ıc0 ıd1  ıc1 ıd0 /

 ab D gac gbd p
1  kR2 =a2
 a0 b1 
h h  ha1 hb0
D p :
1  kR2 =a2
A.2 Kodama Vector in FLRW Space 195

The Kodama vector is

 a0 b1 
h h  ha1 hb0
K   rb R D p
a ab
ıb1
1  kR2 =a2
 a0 11 
h h  ha1 h10
D p
1  kR2 =a2

and, therefore,

1  
K0 D p 1  H 2 R2  kR2 =a2 C H 2 R2
1 kR2 =a2
 
 1  kR2 =a2 p
D p D  1  kR2 =a2 ;
1  kR2 =a2
 10 11 
1 h h  h11 h10
K D p D 0:
1  kR2 =a2

To conclude, we have
 p 
K D  1  kR2 =a2 ; 0; 0; 0 (pseudo-Painlevé-Gullstrand coordinates):
(A.7)

A.2.2 Comoving Coordinates

In comoving coordinates the FLRW line element is

a2 .t/
ds2 D dt2 C dr2 C R2 d˝.2/
2
D hab dxa dxb C R2 d˝.2/
2
; (A.8)
1  kr2

where R D a.t/r is the areal radius. The volume form on the 2-space .t; r/ has
components
p a  
˛ˇ D jhj .dt/˛ ^ .dr/ˇ D p ı˛0 ıˇ1  ı˛1 ıˇ0
1  kr 2

while
a  
 ˛ˇ D g˛ gˇı  ı Dp g˛ gˇı ı 0 ıı1  ı 1 ıı0
1 kr2
a  
D p g˛0 gˇ1  g˛1 gˇ0 :
1 kr2
196 Appendix

The components of the Kodama vector in comoving coordinates are

 
K ˛   ˛ˇ rˇ R D  ˛ˇ aP rıˇ0 C aıˇ1
a  ˛0 01 
D p aP rh h C ah˛0 h11  aP rh˛1 h00  ah˛1 h01
1  kr2
a  ˛0 11 
D p ah h  aP rh˛1 h00 :
1  kr2

Now,
 
a 1  kr2 a
a p
0 00 11
K D p ah h D p 2
D  1  kr2 ;
1  kr2 1  kr2 a
 
a aP ar 1  kr2 p
1 11 00
K D p aP rh h D p 2
D Hr 1  kr2 ;
1  kr2 1  kr2 a

and the components of the Kodama vector are

 p p 
K D  1  kr2 ; Hr 1  kr2 ; 0; 0 (comoving coordinates): (A.9)

The norm squared of K a is

a2
K a Ka D g00 .K 0 /2 C g11 .K 1 /2 D .1  kr2 / C H 2 r2 .1  kr2 /
1  kr2
   
D  1  aP 2 r2  kr2 DP  1  r2 =rAH
2
I (A.10)

it vanishes at the apparent horizon

1
rAH D p : (A.11)
aP 2 C k

Reference

1. Nielsen, A.B., Visser, M.: Production and decay of evolving horizons. Class. Quantum Grav. 23,
4637 (2006)
Index

Symbols conformal Killing horizon, 37

f .R/ gravity, 105, 151, 182, 183, 187, 188 conformal Killing vector, 139
conformal time, 140, 153, 155
conformal transformation, 139, 141, 143–145,
A 157, 176, 177
advanced time, 7, 8
anti-de Sitter, 108
anti-trapped surface, 30 D
apparent horizon, 2, 14, 37–40, 47, 49, 50, 53, de Sitter space, 1, 2, 61, 63, 77, 78, 81, 82,
59, 63, 64, 66, 69–71, 78, 79, 81–86, 87–92, 98, 101
92–96, 98–101, 107, 108, 112–117, deviation vector, 26, 27
123–127, 133, 134, 136–138, 142, dominant energy condition, 19
145, 149–153, 155–158, 167, 171–176, dragging of inertial frames, 17
180–182, 184, 186–189, 196 dynamical horizon, 1, 2, 38, 40, 81
apparent horizon tube, 156
areal radius, 5–7, 41, 47, 49, 53, 60, 62, 64, 73, E
78, 84, 87, 95, 107, 110, 112, 118, 121, Eddington-Finkelstein coordinates, 8, 11, 46
127, 133, 135, 141, 142, 147–150, 155, effective action, 3, 105
170, 174, 178, 180, 181, 183, 195 energy supply vector, 96
areal volume, 71, 95 ergosphere, 17
event horizon, 1, 2, 5, 6, 8, 9, 12, 17, 33,
35–39, 43, 44, 49, 61, 70, 73–79, 82, 85,
B 88–92, 94, 99–101, 108, 122, 124–126,
binary system, 2, 35 142, 156, 168, 185
Boyer-Linquist coordinates, 16, 18 extremal horizon, 1, 14, 17, 50, 92, 115, 122,
Brans-Dicke parameter, 168, 169, 172, 174, 124–126
175, 180
Brans-Dicke theory, 38, 168, 169, 188
F
Fisher spacetime, 146
C fluid-gravity duality, 108
Cauchy horizon, 1, 14, 39, 124 future apparent horizon, 37
conformal anomaly, 143 future inner trapping horizon, 39
conformal factor, 139–141, 143, 145, 146, 157 future null infinity, 2, 35, 36
conformal Killing equation, 139 future outer trapping horizon, 39

© Springer International Publishing Switzerland 2015 197

V. Faraoni, Cosmological and Black Hole Apparent Horizons,
Lecture Notes in Physics 907, DOI 10.1007/978-3-319-19240-6
198 Index

G M
Gauss-Bonnet gravity, 167, 168 marginal surface, 30
generalized Raychaudhuri equation, 131 marginally outer trapped tube, 30
geodesic deviation equation, 26 marginally trapped surface, 35, 44
geodesic equation, 25, 26, 28, 40 marginally trapped tube, 40
Gibbons-Hawking entropy, 91 Misner-Sharp-Hernandez mass, 1, 46, 47, 50,
53, 63, 65, 66, 70, 71, 78, 92, 95, 98,
101, 108, 110, 112, 127, 134, 141, 145,
H 167, 168, 188
Hawking radiation, 1, 2, 34, 39, 45, 59, 93,
106, 126, 156, 167, 186
Hawking temperature, 43, 93, 142, 143, N
145, 189 naked singularity, 14, 17, 108, 113, 117, 125,
Hawking-Hayward quasi-local energy, 1, 47, 136, 146, 150, 151, 156–158, 172, 175,
50, 168 176, 181, 186
Hor̆ava-Lifschitz gravity, 138, 186 Nolan gauge, 63
Horndeski theory, 167, 186, 188 Nolan interior solution, 117–119
Hubble horizon, 78, 172 normal surface, 29
Hubble parameter, 19, 60, 61, 63, 107, 111, null curvature condition, 38
113, 116, 127, 141, 155 null dominant energy condition, 19
hyperspherical coordinates, 61, 64, 72, 74, null energy condition, 19, 30, 38, 50, 110
76, 84
P
Painlevé-Gullstrand coordinates, 10–12, 46,
I
48, 49, 62–65, 67, 68, 70, 87, 90, 193
inflation, 2, 20, 59, 73, 78, 100, 157
particle creation, 143, 144
isolated horizon, 40, 46
particle horizon, 70, 72–76, 79, 85, 88, 100
isotropic radius, 6, 109, 110, 117, 121,
past inner trapping horizon, 83
128–130, 133, 140, 169, 173, 183
phantom energy, 19, 106, 132, 187
phantom field, 154
phantom fluid, 81, 100, 116, 117
J
phantom universe, 77, 116, 117, 132
Jebsen-Birkhoff theorem, 168
positive curvature condition, 28

K Q
Kerr-Schild coordinates, 12, 13 quantum gravity, 40, 126, 138
Kerr-Schild metric, 12, 13
Kerr-Schild transformation, 157
Killing equation, 36, 43, 44, 89, 90 R
Killing horizon, 36, 37, 40, 43, 44, 51, 89, 90, Raychaudhuri equation, 28, 29
126, 127, 139 retarded time, 7, 8
Killing vector, 10, 17, 34, 36, 37, 40–44, 46, Ricci tensor, 4
51–53, 89–92, 126 Riemann tensor, 4
Kodama vector, 37, 41, 66, 67, 92, 93, 96, 167, Rindler horizon, 1, 3, 30, 32–34, 51, 52, 79
168, 194–196 Rindler observer, 32, 52
Kretschmann scalar, 17
Kruskal-Szekeres coordinates, 7, 8, 12, 14,
15, 61 S
S-curve, 151, 156, 158, 175, 176, 178, 182,
186, 188
L slowly evolving horizon, 2, 41
Lemaître-Tolman-Bondi model, 106, 138, spacetime singularity, 5, 16, 61, 110, 112, 114,
156, 158 122, 123, 141, 147, 156, 171, 188
Index 199

static limit, 17 U
strong energy condition, 19, 77 uniform acceleration, 30, 31, 34, 52
supergravity, 126, 167 Unruh effect, 33, 34
supernovae, 3 Unruh temperature, 33, 52
Synge approach, 127 untrapped surface, 30, 49

T V
thermodynamics of spacetime, 3, 51, 100 Vaidya spacetime, 1, 35, 38, 156
timelike membrane, 40
Tolman-Oppenheimer-Volkoff equation, 119
tortoise coordinate, 7, 15 W
trapped surface, 2, 25, 30 weak energy condition, 19, 81, 95,
trapping horizon, 1, 2, 37–39, 44, 45, 50, 70, 110, 116
83, 88, 95, 105, 156, 158, 168, 188, 189 white hole, 8, 9, 12, 37, 39, 74, 86
trapping horizon tube, 188 wormhole, 19, 37, 48, 115